Skip to contentSkip to conversation

WireGuard: Secure Network Tunnel

PATCH Superseded Jason A. Donenfeld 3 participants 39 messages Oct 18, 2018 – Oct 26, 2018

Subsystems (30)

v8 of 9 Superseded Latest is v9 →

lore
Share & tools
lore

Jason A. Donenfeld

author maintainer
#0 lore raw

v8 cover WireGuard: Secure Network Tunnel

Changes v7->v8, along with who suggested it.
--------------------------------------------
- Implementations that fail the selftests are now disabled, after a warning is printed. This way users don't make wrong calculations, even in the face of a rather grave bug.
- [Sultan Alsawaf] When assigning to a boolean, prefer "BIT(i) & a" to "(a >>
i) & 1".
- [Andrew Lunn] Avoid control statements inside macros.
- [Jiri Pirko] Prefix functions used in callbacks with wg_.
- [Jiri Pirko] Rename struct wireguard_peer and struct wireguard_device to struct wg_peer and struct wg_device.
- [Eugene Syromiatnikov] Do not use nla type field as an index, and actually don't use an index at all, because it has no meaning or relevance at all.
- [Joe Perches] Do not place a space between for_each iterators and parentheses.
- Enumerable style cleanups and nits.
- [Arnd Bergmann] Swap endianness in allowedips early on in case optimizer is able to look a bit further in but not too far, resulting in a warning from -Wmaybe-uninitialized.
- [Jiri Pirko] Use textual error labels instead of numerical ones.
- [Jiri Pirko] Better module description string.
- [Eric Biggers] In poly1305 port to crypto API, account for short inputs in final function, in which case -ENOKEY should be returned.

-----------------------------------------------------------

This patchset is available on git.kernel.org in this branch, where it may be pulled directly for inclusion into net-next:

* https://git.kernel.org/pub/scm/linux/kernel/git/zx2c4/linux.git/log/?h=jd/wireguard

-----------------------------------------------------------

WireGuard is a secure network tunnel written especially for Linux, which has faced around three years of serious development, deployment, and scrutiny. It delivers excellent performance and is extremely easy to use and configure. It has been designed with the primary goal of being both easy to audit by virtue of being small and highly secure from a cryptography and systems security perspective. WireGuard is used by some massive companies pushing enormous amounts of traffic, and likely already today you've consumed bytes that at some point transited through a WireGuard tunnel. Even as an out-of-tree module, WireGuard has been integrated into various userspace tools, Linux distributions, mobile phones, and data centers. There are ports in several languages to several operating systems, and even commercial hardware and services sold integrating WireGuard. It is time, therefore, for WireGuard to be properly integrated into Linux.

Ample information, including documentation, installation instructions, and project details, is available at:

* https://www.wireguard.com/
* https://www.wireguard.com/papers/wireguard.pdf

As it is currently an out-of-tree module, it lives in its own git repo and has its own mailing list, and every commit for the module is tested against every stable kernel since 3.10 on a variety of architectures using an extensive test suite:

* https://git.zx2c4.com/WireGuard
https://git.kernel.org/pub/scm/linux/kernel/git/zx2c4/WireGuard.git/
* https://lists.zx2c4.com/mailman/listinfo/wireguard
* https://www.wireguard.com/build-status/

The project has been broadly discussed at conferences, and was presented to the Netdev developers in Seoul last November, where a paper was released detailing some interesting aspects of the project. Dave asked me after the talk if I would consider sending in a v1 "sooner rather than later", hence this patchset. Zinc was presented at Kernel Recipes in September, and a video is available online. Both Zinc and WireGuard will be presented at the conference in Vancouver in November.

* https://www.wireguard.com/presentations/
* https://www.wireguard.com/papers/wireguard-netdev22.pdf
* Zinc talk: https://www.youtube.com/watch?v=bFhdln8aJ_U
* Netdev talk: https://www.youtube.com/watch?v=54orFwtQ1XY

The cryptography in the protocol itself has been formally verified by several independent academic teams with positive results, and I know of two additional efforts on their way to further corroborate those findings. The version 1 protocol is "complete", and so the purpose of this review is to assess the implementation of the protocol. However, it still may be of interest to know that the thing you're reviewing uses a protocol with various nice security properties:

* https://www.wireguard.com/formal-verification/

This patchset is divided into four segments. The first introduces a very simple helper for working with the FPU state for the purposes of amortizing SIMD operations. The second segment is a small collection of cryptographic primitives, split up into several commits by primitive and by hardware. The third shows usage of Zinc within the existing crypto API and as a replacement to the existing crypto API. The last is WireGuard itself, presented as an unintrusive and self-contained virtual network driver.

It is intended that this entire patch series enter the kernel through DaveM's net-next tree. Subsequently, WireGuard patches will go through DaveM's net-next tree, while Zinc patches will go through Greg KH's tree in cases when an entire development cycle has no relationships with existing code in crypto/; however, if there are any relationships with code in crypto/, then pull requests will be sent to Herbert instead in case there are merge conflicts.

Enjoy,
Jason

author maintainer
#1 lore raw

PATCH v8 1/28 ARM: makefile: use ARMv3M mode for RiscPC

The purpose of CONFIG_CPU_32v3 is to avoid ldrh/strh on the RiscPC, which is pretty much an ARMv4 device, except its bus will choke on the half-words. The way to make the C compiler not output ldrh/strh is with -march=armv3, which doesn't support them in the ISA. However, this prevents certain cryptography code from working that uses instructions like umull. Fortunately there's also -march=armv3m that does support those, making it possible to continue assembling optimized cryptography routines for our beloved RiscPC.

---

Notes:
This commit has been submitted to the proper ARM tree and is working its way upstream. It's included in this series here so that kbuild 0-day bot doesn't get too nervous about RiscPC, but is already entering the tree through arm-next.

 arch/arm/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/Makefile +1 -1
index d1516f85f25d..7fd4bcaf0721 100644
@@ -74,7 +74,7 @@ endif
 arch-$(CONFIG_CPU_32v5)		=-D__LINUX_ARM_ARCH__=5 $(call cc-option,-march=armv5te,-march=armv4t)
 arch-$(CONFIG_CPU_32v4T)	=-D__LINUX_ARM_ARCH__=4 -march=armv4t
 arch-$(CONFIG_CPU_32v4)		=-D__LINUX_ARM_ARCH__=4 -march=armv4
-arch-$(CONFIG_CPU_32v3)		=-D__LINUX_ARM_ARCH__=3 -march=armv3
+arch-$(CONFIG_CPU_32v3)		=-D__LINUX_ARM_ARCH__=3 -march=armv3m
 
 # Evaluate arch cc-option calls now
 arch-y := $(arch-y)
Signature · 1 line
2.19.1
author maintainer
#2 lore raw

PATCH v8 2/28 asm: simd context helper API

In reply to Jason A. Donenfeld

Sometimes it's useful to amortize calls to XSAVE/XRSTOR and the related FPU/SIMD functions over a number of calls, because FPU restoration is quite expensive. This adds a simple header for carrying out this pattern:

    simd_context_t simd_context;

    simd_get(&simd_context);
    while ((item = get_item_from_queue()) != NULL) {
        encrypt_item(item, &simd_context);
        simd_relax(&simd_context);
    }
    simd_put(&simd_context);

The relaxation step ensures that we don't trample over preemption, and the get/put API should be a familiar paradigm in the kernel.

On the other end, code that actually wants to use SIMD instructions can accept this as a parameter and check it via:

   void encrypt_item(struct item *item, simd_context_t *simd_context)
   {
       if (item->len > LARGE_FOR_SIMD && simd_use(simd_context))
           wild_simd_code(item);
       else
           boring_scalar_code(item);
   }

The actual XSAVE happens during simd_use (and only on the first time), so that if the context is never actually used, no performance penalty is hit.

Cc Samuel Neves <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Greg KH <[email protected]>
---
 arch/alpha/include/asm/Kbuild      |  1 +
 arch/arc/include/asm/Kbuild        |  1 +
 arch/arm/include/asm/Kbuild        |  1 -
 arch/arm/include/asm/simd.h        | 63 ++++++++++++++++++++++++++++++
 arch/arm64/include/asm/simd.h      | 51 +++++++++++++++++++++---
 arch/c6x/include/asm/Kbuild        |  1 +
 arch/h8300/include/asm/Kbuild      |  1 +
 arch/hexagon/include/asm/Kbuild    |  1 +
 arch/ia64/include/asm/Kbuild       |  1 +
 arch/m68k/include/asm/Kbuild       |  1 +
 arch/microblaze/include/asm/Kbuild |  1 +
 arch/mips/include/asm/Kbuild       |  1 +
 arch/nds32/include/asm/Kbuild      |  1 +
 arch/nios2/include/asm/Kbuild      |  1 +
 arch/openrisc/include/asm/Kbuild   |  1 +
 arch/parisc/include/asm/Kbuild     |  1 +
 arch/powerpc/include/asm/Kbuild    |  1 +
 arch/riscv/include/asm/Kbuild      |  1 +
 arch/s390/include/asm/Kbuild       |  1 +
 arch/sh/include/asm/Kbuild         |  1 +
 arch/sparc/include/asm/Kbuild      |  1 +
 arch/um/include/asm/Kbuild         |  1 +
 arch/unicore32/include/asm/Kbuild  |  1 +
 arch/x86/include/asm/simd.h        | 44 ++++++++++++++++++++-
 arch/xtensa/include/asm/Kbuild     |  1 +
 include/asm-generic/simd.h         | 20 ++++++++++
 include/linux/simd.h               | 32 +++++++++++++++
 27 files changed, 224 insertions(+), 8 deletions(-)
 create mode 100644 arch/arm/include/asm/simd.h
 create mode 100644 include/linux/simd.h
arch/alpha/include/asm/Kbuild +1 -0
index 0580cb8c84b2..220dfd170d45 100644
@@ -13,3 +13,4 @@ generic-y += sections.h
 generic-y += trace_clock.h
 generic-y += current.h
 generic-y += kprobes.h
+generic-y += simd.h
arch/arc/include/asm/Kbuild +1 -0
index feed50ce89fa..a7f4255f1649 100644
@@ -22,6 +22,7 @@ generic-y += parport.h
 generic-y += pci.h
 generic-y += percpu.h
 generic-y += preempt.h
+generic-y += simd.h
 generic-y += topology.h
 generic-y += trace_clock.h
 generic-y += user.h
arch/arm/include/asm/Kbuild +0 -1
index 1d66db9c9db5..ebdc9eeb8d39 100644
@@ -16,7 +16,6 @@ generic-y += rwsem.h
 generic-y += seccomp.h
 generic-y += segment.h
 generic-y += serial.h
-generic-y += simd.h
 generic-y += sizes.h
 generic-y += timex.h
 generic-y += trace_clock.h
arch/arm/include/asm/simd.h new file +63 -0

63 lines not on this page · raw message

arch/arm64/include/asm/simd.h +45 -6

71 lines not on this page · raw message

arch/c6x/include/asm/Kbuild +1 -0
index 33a2c94fed0d..7543c38f7ade 100644
@@ -30,6 +30,7 @@ generic-y += pgalloc.h
 generic-y += preempt.h
 generic-y += segment.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += tlbflush.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/h8300/include/asm/Kbuild +1 -0
index a5d0b2991f47..1fcef25ee19d 100644
@@ -39,6 +39,7 @@ generic-y += preempt.h
 generic-y += scatterlist.h
 generic-y += sections.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += spinlock.h
 generic-y += timex.h
arch/hexagon/include/asm/Kbuild +1 -0
index dd2fd9c0d292..217d4695fd8a 100644
@@ -29,6 +29,7 @@ generic-y += rwsem.h
 generic-y += sections.h
 generic-y += segment.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/ia64/include/asm/Kbuild +1 -0
index 557bbc8ba9f5..41c5ebdf79e5 100644
@@ -4,6 +4,7 @@ generic-y += irq_work.h
 generic-y += mcs_spinlock.h
 generic-y += mm-arch-hooks.h
 generic-y += preempt.h
+generic-y += simd.h
 generic-y += trace_clock.h
 generic-y += vtime.h
 generic-y += word-at-a-time.h
arch/m68k/include/asm/Kbuild +1 -0
index a4b8d3331a9e..73898dd1a4d0 100644
@@ -19,6 +19,7 @@ generic-y += mm-arch-hooks.h
 generic-y += percpu.h
 generic-y += preempt.h
 generic-y += sections.h
+generic-y += simd.h
 generic-y += spinlock.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/microblaze/include/asm/Kbuild +1 -0
index 569ba9e670c1..7a877eea99d3 100644
@@ -25,6 +25,7 @@ generic-y += parport.h
 generic-y += percpu.h
 generic-y += preempt.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += syscalls.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/mips/include/asm/Kbuild +1 -0
index 58351e48421e..e8868e0fb2c3 100644
@@ -16,6 +16,7 @@ generic-y += qrwlock.h
 generic-y += qspinlock.h
 generic-y += sections.h
 generic-y += segment.h
+generic-y += simd.h
 generic-y += trace_clock.h
 generic-y += unaligned.h
 generic-y += user.h
arch/nds32/include/asm/Kbuild +1 -0
index dbc4e5422550..fb2f113716ce 100644
@@ -46,6 +46,7 @@ generic-y += sections.h
 generic-y += segment.h
 generic-y += serial.h
 generic-y += shmbuf.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += stat.h
 generic-y += switch_to.h
arch/nios2/include/asm/Kbuild +1 -0
index 8fde4fa2c34f..571a9d9ad107 100644
@@ -33,6 +33,7 @@ generic-y += preempt.h
 generic-y += sections.h
 generic-y += segment.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += spinlock.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/openrisc/include/asm/Kbuild +1 -0
index eb87cd8327c8..b6231211bbad 100644
@@ -34,6 +34,7 @@ generic-y += qrwlock_types.h
 generic-y += qrwlock.h
 generic-y += sections.h
 generic-y += segment.h
+generic-y += simd.h
 generic-y += string.h
 generic-y += switch_to.h
 generic-y += topology.h
arch/parisc/include/asm/Kbuild +1 -0
index 2013d639e735..97970b4d05ab 100644
@@ -17,6 +17,7 @@ generic-y += percpu.h
 generic-y += preempt.h
 generic-y += seccomp.h
 generic-y += segment.h
+generic-y += simd.h
 generic-y += topology.h
 generic-y += trace_clock.h
 generic-y += user.h
arch/powerpc/include/asm/Kbuild +1 -0
index 3196d227e351..2337190aaf69 100644
@@ -8,3 +8,4 @@ generic-y += preempt.h
 generic-y += rwsem.h
 generic-y += vtime.h
 generic-y += msi.h
+generic-y += simd.h
arch/riscv/include/asm/Kbuild +1 -0
index efdbe311e936..438a11d9c47a 100644
@@ -46,6 +46,7 @@ generic-y += setup.h
 generic-y += shmbuf.h
 generic-y += shmparam.h
 generic-y += signal.h
+generic-y += simd.h
 generic-y += socket.h
 generic-y += sockios.h
 generic-y += stat.h
arch/s390/include/asm/Kbuild +1 -0
index e3239772887a..3744c4c61fb5 100644
@@ -22,6 +22,7 @@ generic-y += mcs_spinlock.h
 generic-y += mm-arch-hooks.h
 generic-y += preempt.h
 generic-y += rwsem.h
+generic-y += simd.h
 generic-y += trace_clock.h
 generic-y += unaligned.h
 generic-y += word-at-a-time.h
arch/sh/include/asm/Kbuild +1 -0
index 6a5609a55965..8e64ff35a933 100644
@@ -16,6 +16,7 @@ generic-y += percpu.h
 generic-y += preempt.h
 generic-y += rwsem.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += trace_clock.h
 generic-y += xor.h
arch/sparc/include/asm/Kbuild +1 -0
index 410b263ef5c8..72b9e08fb350 100644
@@ -17,5 +17,6 @@ generic-y += msi.h
 generic-y += preempt.h
 generic-y += rwsem.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += trace_clock.h
 generic-y += word-at-a-time.h
arch/um/include/asm/Kbuild +1 -0
index b10dde6cb793..8c2bfa6e0494 100644
@@ -22,6 +22,7 @@ generic-y += param.h
 generic-y += pci.h
 generic-y += percpu.h
 generic-y += preempt.h
+generic-y += simd.h
 generic-y += switch_to.h
 generic-y += topology.h
 generic-y += trace_clock.h
arch/unicore32/include/asm/Kbuild +1 -0
index bfc7abe77905..98a908720bbd 100644
@@ -27,6 +27,7 @@ generic-y += preempt.h
 generic-y += sections.h
 generic-y += segment.h
 generic-y += serial.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += syscalls.h
 generic-y += topology.h
arch/x86/include/asm/simd.h +43 -1

50 lines not on this page · raw message

arch/xtensa/include/asm/Kbuild +1 -0
index 82c756431b49..7950f359649d 100644
@@ -24,6 +24,7 @@ generic-y += percpu.h
 generic-y += preempt.h
 generic-y += rwsem.h
 generic-y += sections.h
+generic-y += simd.h
 generic-y += topology.h
 generic-y += trace_clock.h
 generic-y += word-at-a-time.h
include/asm-generic/simd.h +20 -0
index d0343d58a74a..b3dd61ac010e 100644
@@ -1,5 +1,9 @@
 /* SPDX-License-Identifier: GPL-2.0 */
 
+#include <linux/simd.h>
+#ifndef _ASM_SIMD_H
+#define _ASM_SIMD_H
+
 #include <linux/hardirq.h>
 
 /*
@@ -13,3 +17,19 @@ static __must_check inline bool may_use_simd(void)
 {
 	return !in_interrupt();
 }
+
+static inline void simd_get(simd_context_t *ctx)
+{
+	*ctx = HAVE_NO_SIMD;
+}
+
+static inline void simd_put(simd_context_t *ctx)
+{
+}
+
+static __must_check inline bool simd_use(simd_context_t *ctx)
+{
+	return false;
+}
+
+#endif /* _ASM_SIMD_H */
include/linux/simd.h new file +32 -0
new file mode 100644
index 000000000000..4e0b8a9bdc14
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0
+ *
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _SIMD_H
+#define _SIMD_H
+
+typedef enum {
+	HAVE_NO_SIMD = 1 << 0,
+	HAVE_FULL_SIMD = 1 << 1,
+	HAVE_SIMD_IN_USE = 1 << 31
+} simd_context_t;
+
+#define DONT_USE_SIMD ((simd_context_t []){ HAVE_NO_SIMD })
+
+#include <linux/sched.h>
+#include <asm/simd.h>
+
+static inline bool simd_relax(simd_context_t *ctx)
+{
+#ifdef CONFIG_PREEMPT
+	if ((*ctx & HAVE_SIMD_IN_USE) && need_resched()) {
+		simd_put(ctx);
+		simd_get(ctx);
+		return true;
+	}
+#endif
+	return false;
+}
+
+#endif /* _SIMD_H */
Signature · 1 line
2.19.1
author maintainer
#3 lore raw

PATCH v8 3/28 zinc: introduce minimal cryptography library

In reply to Jason A. Donenfeld

Zinc stands for "Zinc Is Neat Crypto" or "Zinc as IN Crypto". It's also short, easy to type, and plays nicely with the recent trend of naming crypto libraries after elements. The guiding principle is "don't overdo it". It's less of a library and more of a directory tree for organizing well-curated direct implementations of cryptography primitives.

Zinc is a new cryptography API that is much more minimal and lower-level than the current one. It intends to complement it and provide a basis upon which the current crypto API might build, as the provider of software implementations of cryptographic primitives. It is motivated by three primary observations in crypto API design:

* Highly composable "cipher modes" and related abstractions from the 90s did not turn out to be as terrific an idea as hoped, leading to a host of API misuse problems.

* Most programmers are afraid of crypto code, and so prefer to integrate it into libraries in a highly abstracted manner, so as to shield themselves from implementation details. Cryptographers, on the other hand, prefer simple direct implementations, which they're able to verify for high assurance and optimize in accordance with their expertise.

* Overly abstracted and flexible cryptography APIs lead to a host of dangerous problems and performance issues. The kernel is in the business usually not of coming up with new uses of crypto, but rather implementing various constructions, which means it essentially needs a library of primitives, not a highly abstracted enterprise-ready pluggable system, with a few particular exceptions.

This last observation has seen itself play out several times over and over again within the kernel:

* The perennial move of actual primitives away from crypto/ and into lib/, so that users can actually call these functions directly with no overhead and without lots of allocations, function pointers, string specifier parsing, and general clunkiness. For example:
sha256, chacha20, siphash, sha1, and so forth live in lib/ rather than in crypto/. Zinc intends to stop the cluttering of lib/ and introduce these direct primitives into their proper place, lib/zinc/.

* An abundance of misuse bugs with the present crypto API that have been very unpleasant to clean up.

* A hesitance to even use cryptography, because of the overhead and headaches involved in accessing the routines.

Zinc goes in a rather different direction. Rather than providing a thoroughly designed and abstracted API, Zinc gives you simple functions, which implement some primitive, or some particular and specific construction of primitives. It is not dynamic in the least, though one could imagine implementing a complex dynamic dispatch mechanism (such as the current crypto API) on top of these basic functions. After all, dynamic dispatch is usually needed for applications with cipher agility, such as IPsec, dm-crypt, AF_ALG, and so forth, and the existing crypto API will continue to play that role. However, Zinc will provide a non- haphazard way of directly utilizing crypto routines in applications that do have neither the need nor desire for abstraction and dynamic dispatch.

It also organizes the implementations in a simple, straight-forward, and direct manner, making it enjoyable and intuitive to work on. Rather than moving optimized assembly implementations into arch/, it keeps them all together in lib/zinc/, making it simple and obvious to compare and contrast what's happening. This is, notably, exactly what the lib/raid6/ tree does, and that seems to work out rather well. It's also the pattern of most successful crypto libraries. The architecture- specific glue-code is made a part of each translation unit, rather than being in a separate one, so that generic and architecture-optimized code are combined at compile-time, and incompatibility branches compiled out by the optimizer.

All implementations have been extensively tested and fuzzed, and are selected for their quality, trustworthiness, and performance. Wherever possible and performant, formally verified implementations are used, such as those from HACL* [1] and Fiat-Crypto [2]. The routines also take special care to zero out secrets using memzero_explicit (and future work is planned to have gcc do this more reliably and performantly with compiler plugins). The performance of the selected implementations is state-of-the-art and unrivaled on a broad array of hardware, though of course we will continue to fine tune these to the hardware demands needed by kernel contributors. Each implementation also comes with extensive self-tests and crafted test vectors, pulled from various places such as Wycheproof [9].

Regularity of function signatures is important, so that users can easily "guess" the name of the function they want. Though, individual primitives are oftentimes not trivially interchangeable, having been designed for different things and requiring different parameters and semantics, and so the function signatures they provide will directly reflect the realities of the primitives' usages, rather than hiding it behind (inevitably leaky) abstractions. Also, in contrast to the current crypto API, Zinc functions can work on stack buffers, and can be called with different keys, without requiring allocations or locking.

SIMD is used automatically when available, though some routines may benefit from either having their SIMD disabled for particular invocations, or to have the SIMD initialization calls amortized over several invocations of the function, and so Zinc utilizes function signatures enabling that in conjunction with the recently introduced simd_context_t.

More generally, Zinc provides function signatures that allow just what is required by the various callers. This isn't to say that users of the functions will be permitted to pollute the function semantics with weird particular needs, but we are trying very hard not to overdo it, and that means looking carefully at what's actually necessary, and doing just that, and not much more than that. Remember: practicality and cleanliness rather than over-zealous infrastructure.

Zinc provides also an opening for the best implementers in academia to contribute their time and effort to the kernel, by being sufficiently simple and inviting. In discussing this commit with some of the best and brightest over the last few years, there are many who are eager to devote rare talent and energy to this effort.

Following the merging of this, I expect for the primitives that currently exist in lib/ to work their way into lib/zinc/, after intense scrutiny of each implementation, potentially replacing them with either formally-verified implementations, or better studied and faster state-of-the-art implementations.

Also following the merging of this, I expect for the old crypto API implementations to be ported over to use Zinc for their software-based implementations.

As Zinc is simply library code, its config options are un-menued, with the exception of CONFIG_ZINC_SELFTEST and CONFIG_ZINC_DEBUG, which enables various selftests and debugging conditions.

[1] https://github.com/project-everest/hacl-star
[2] https://github.com/mit-plv/fiat-crypto
[3] https://cr.yp.to/ecdh.html
[4] https://cr.yp.to/chacha.html
[5] https://cr.yp.to/snuffle/xsalsa-20081128.pdf
[6] https://cr.yp.to/mac.html
[7] https://blake2.net/
[8] https://tools.ietf.org/html/rfc8439
[9] https://github.com/google/wycheproof

Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 MAINTAINERS       |  8 ++++++++
 lib/Kconfig       |  2 ++
 lib/Makefile      |  2 ++
 lib/zinc/Kconfig  | 41 +++++++++++++++++++++++++++++++++++++++++
 lib/zinc/Makefile |  3 +++
 5 files changed, 56 insertions(+)
 create mode 100644 lib/zinc/Kconfig
 create mode 100644 lib/zinc/Makefile
MAINTAINERS +8 -0
index 7f1399ac028e..c0d078d4dbd8 100644
@@ -16230,6 +16230,14 @@ Q:	https://patchwork.linuxtv.org/project/linux-media/list/
 S:	Maintained
 F:	drivers/media/dvb-frontends/zd1301_demod*
 
+ZINC CRYPTOGRAPHY LIBRARY
+M:	Jason A. Donenfeld <[email protected]>
+M:	Samuel Neves <[email protected]>
+S:	Maintained
+F:	lib/zinc/
+F:	include/zinc/
+L:	[email protected]
+
 ZPOOL COMPRESSED PAGE STORAGE API
 M:	Dan Streetman <[email protected]>
 L:	[email protected]
lib/Kconfig +2 -0
index a3928d4438b5..3e6848269c66 100644
@@ -485,6 +485,8 @@ config GLOB_SELFTEST
 	  module load) by a small amount, so you're welcome to play with
 	  it, but you probably don't need it.
 
+source "lib/zinc/Kconfig"
+
 #
 # Netlink attribute parsing support is select'ed if needed
 #
lib/Makefile +2 -0
index 423876446810..7a04d7ce95a6 100644
@@ -213,6 +213,8 @@ obj-$(CONFIG_PERCPU_TEST) += percpu_test.o
 
 obj-$(CONFIG_ASN1) += asn1_decoder.o
 
+obj-y += zinc/
+
 obj-$(CONFIG_FONT_SUPPORT) += fonts/
 
 obj-$(CONFIG_PRIME_NUMBERS) += prime_numbers.o
lib/zinc/Kconfig new file +41 -0

41 lines not on this page · raw message

lib/zinc/Makefile new file +3 -0
new file mode 100644
index 000000000000..a61c80d676cb
@@ -0,0 +1,3 @@
+ccflags-y := -O2
+ccflags-y += -D'pr_fmt(fmt)="zinc: " fmt'
+ccflags-$(CONFIG_ZINC_DEBUG) += -DDEBUG
Signature · 1 line
2.19.1
author maintainer
#4 lore raw

PATCH v8 4/28 zinc: ChaCha20 generic C implementation and selftest

In reply to Jason A. Donenfeld

This implements the ChaCha20 permutation as a single C statement, by way of the comma operator, which the compiler is able to simplify terrifically.

Information: https://cr.yp.to/chacha.html

Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 include/zinc/chacha20.h      |   70 +
 lib/zinc/Kconfig             |    4 +
 lib/zinc/Makefile            |    3 +
 lib/zinc/chacha20/chacha20.c |  179 +++
 lib/zinc/selftest/chacha20.c | 2698 ++++++++++++++++++++++++++++++++++
 lib/zinc/selftest/run.h      |   48 +
 6 files changed, 3002 insertions(+)
 create mode 100644 include/zinc/chacha20.h
 create mode 100644 lib/zinc/chacha20/chacha20.c
 create mode 100644 lib/zinc/selftest/chacha20.c
 create mode 100644 lib/zinc/selftest/run.h
include/zinc/chacha20.h new file +70 -0
new file mode 100644
index 000000000000..0b98bd6946ae
@@ -0,0 +1,70 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _ZINC_CHACHA20_H
+#define _ZINC_CHACHA20_H
+
+#include <asm/unaligned.h>
+#include <linux/simd.h>
+#include <linux/kernel.h>
+#include <linux/types.h>
+
+enum {
+	CHACHA20_NONCE_SIZE = 16,
+	CHACHA20_KEY_SIZE = 32,
+	CHACHA20_KEY_WORDS = CHACHA20_KEY_SIZE / sizeof(u32),
+	CHACHA20_BLOCK_SIZE = 64,
+	CHACHA20_BLOCK_WORDS = CHACHA20_BLOCK_SIZE / sizeof(u32),
+	HCHACHA20_NONCE_SIZE = CHACHA20_NONCE_SIZE,
+	HCHACHA20_KEY_SIZE = CHACHA20_KEY_SIZE
+};
+
+enum { /* expand 32-byte k */
+	CHACHA20_CONSTANT_EXPA = 0x61707865U,
+	CHACHA20_CONSTANT_ND_3 = 0x3320646eU,
+	CHACHA20_CONSTANT_2_BY = 0x79622d32U,
+	CHACHA20_CONSTANT_TE_K = 0x6b206574U
+};
+
+struct chacha20_ctx {
+	union {
+		u32 state[16];
+		struct {
+			u32 constant[4];
+			u32 key[8];
+			u32 counter[4];
+		};
+	};
+};
+
+static inline void chacha20_init(struct chacha20_ctx *ctx,
+				 const u8 key[CHACHA20_KEY_SIZE],
+				 const u64 nonce)
+{
+	ctx->constant[0] = CHACHA20_CONSTANT_EXPA;
+	ctx->constant[1] = CHACHA20_CONSTANT_ND_3;
+	ctx->constant[2] = CHACHA20_CONSTANT_2_BY;
+	ctx->constant[3] = CHACHA20_CONSTANT_TE_K;
+	ctx->key[0] = get_unaligned_le32(key + 0);
+	ctx->key[1] = get_unaligned_le32(key + 4);
+	ctx->key[2] = get_unaligned_le32(key + 8);
+	ctx->key[3] = get_unaligned_le32(key + 12);
+	ctx->key[4] = get_unaligned_le32(key + 16);
+	ctx->key[5] = get_unaligned_le32(key + 20);
+	ctx->key[6] = get_unaligned_le32(key + 24);
+	ctx->key[7] = get_unaligned_le32(key + 28);
+	ctx->counter[0] = 0;
+	ctx->counter[1] = 0;
+	ctx->counter[2] = nonce & U32_MAX;
+	ctx->counter[3] = nonce >> 32;
+}
+void chacha20(struct chacha20_ctx *ctx, u8 *dst, const u8 *src, u32 len,
+	      simd_context_t *simd_context);
+
+void hchacha20(u32 derived_key[CHACHA20_KEY_WORDS],
+	       const u8 nonce[HCHACHA20_NONCE_SIZE],
+	       const u8 key[HCHACHA20_KEY_SIZE], simd_context_t *simd_context);
+
+#endif /* _ZINC_CHACHA20_H */
lib/zinc/Kconfig +4 -0
index 90e066ea93a0..d271be37cecb 100644
@@ -1,3 +1,7 @@
+config ZINC_CHACHA20
+	tristate
+	select CRYPTO_ALGAPI
+
 config ZINC_SELFTEST
 	bool "Zinc cryptography library self-tests"
 	help
lib/zinc/Makefile +3 -0
index a61c80d676cb..3d80144d55a6 100644
@@ -1,3 +1,6 @@
 ccflags-y := -O2
 ccflags-y += -D'pr_fmt(fmt)="zinc: " fmt'
 ccflags-$(CONFIG_ZINC_DEBUG) += -DDEBUG
+
+zinc_chacha20-y := chacha20/chacha20.o
+obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
lib/zinc/chacha20/chacha20.c new file +179 -0

179 lines not on this page · raw message

lib/zinc/selftest/chacha20.c new file +2698 -0

2,698 lines not on this page · raw message

lib/zinc/selftest/run.h new file +48 -0

40 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#5 lore raw

PATCH v8 5/28 zinc: import Andy Polyakov's ChaCha20 x86_64 implementation

In reply to Jason A. Donenfeld

These x86_64 vectorized implementations come from Andy Polyakov's implementation, and are included here in raw form without modification, so that subsequent commits that fix these up for the kernel can see how it has changed.

While this is CRYPTOGAMS code, the originating code for this happens to be the same as OpenSSL's commit cded951378069a478391843f5f8653c1eb5128da

Based-on-code-from: Andy Polyakov <[email protected]>
Cc: Andy Polyakov <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: [email protected]
Cc: Samuel Neves <[email protected]>
Cc: Jean-Philippe Aumasson <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 .../chacha20/chacha20-x86_64-cryptogams.S     | 3433 +++++++++++++++++
 1 file changed, 3433 insertions(+)
 create mode 100644 lib/zinc/chacha20/chacha20-x86_64-cryptogams.S
lib/zinc/chacha20/chacha20-x86_64-cryptogams.S new file +3433 -0
new file mode 100644
index 000000000000..2bfc76f7e01f
@@ -0,0 +1,3433 @@
+/* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
+/*
+ * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ */
+
+.text	
+
+
+
+.align	64
+.Lzero:
+.long	0,0,0,0
+.Lone:
+.long	1,0,0,0
+.Linc:
+.long	0,1,2,3
+.Lfour:
+.long	4,4,4,4
+.Lincy:
+.long	0,2,4,6,1,3,5,7
+.Leight:
+.long	8,8,8,8,8,8,8,8
+.Lrot16:
+.byte	0x2,0x3,0x0,0x1, 0x6,0x7,0x4,0x5, 0xa,0xb,0x8,0x9, 0xe,0xf,0xc,0xd
+.Lrot24:
+.byte	0x3,0x0,0x1,0x2, 0x7,0x4,0x5,0x6, 0xb,0x8,0x9,0xa, 0xf,0xc,0xd,0xe
+.Ltwoy:
+.long	2,0,0,0, 2,0,0,0
+.align	64
+.Lzeroz:
+.long	0,0,0,0, 1,0,0,0, 2,0,0,0, 3,0,0,0
+.Lfourz:
+.long	4,0,0,0, 4,0,0,0, 4,0,0,0, 4,0,0,0
+.Lincz:
+.long	0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
+.Lsixteen:
+.long	16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16
+.Lsigma:
+.byte	101,120,112,97,110,100,32,51,50,45,98,121,116,101,32,107,0
+.byte	67,104,97,67,104,97,50,48,32,102,111,114,32,120,56,54,95,54,52,44,32,67,82,89,80,84,79,71,65,77,83,32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115,115,108,46,111,114,103,62,0
+.globl	ChaCha20_ctr32
+.type	ChaCha20_ctr32,@function
+.align	64
+ChaCha20_ctr32:
+.cfi_startproc	
+	cmpq	$0,%rdx
+	je	.Lno_data
+	movq	OPENSSL_ia32cap_P+4(%rip),%r10
+	btq	$48,%r10
+	jc	.LChaCha20_avx512
+	testq	%r10,%r10
+	js	.LChaCha20_avx512vl
+	testl	$512,%r10d
+	jnz	.LChaCha20_ssse3
+
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+	subq	$64+24,%rsp
+.cfi_adjust_cfa_offset	64+24
+.Lctr32_body:
+
+
+	movdqu	(%rcx),%xmm1
+	movdqu	16(%rcx),%xmm2
+	movdqu	(%r8),%xmm3
+	movdqa	.Lone(%rip),%xmm4
+
+
+	movdqa	%xmm1,16(%rsp)
+	movdqa	%xmm2,32(%rsp)
+	movdqa	%xmm3,48(%rsp)
+	movq	%rdx,%rbp
+	jmp	.Loop_outer
+
+.align	32
+.Loop_outer:
+	movl	$0x61707865,%eax
+	movl	$0x3320646e,%ebx
+	movl	$0x79622d32,%ecx
+	movl	$0x6b206574,%edx
+	movl	16(%rsp),%r8d
+	movl	20(%rsp),%r9d
+	movl	24(%rsp),%r10d
+	movl	28(%rsp),%r11d
+	movd	%xmm3,%r12d
+	movl	52(%rsp),%r13d
+	movl	56(%rsp),%r14d
+	movl	60(%rsp),%r15d
+
+	movq	%rbp,64+0(%rsp)
+	movl	$10,%ebp
+	movq	%rsi,64+8(%rsp)
+.byte	102,72,15,126,214
+	movq	%rdi,64+16(%rsp)
+	movq	%rsi,%rdi
+	shrq	$32,%rdi
+	jmp	.Loop
+
+.align	32
+.Loop:
+	addl	%r8d,%eax
+	xorl	%eax,%r12d
+	roll	$16,%r12d
+	addl	%r9d,%ebx
+	xorl	%ebx,%r13d
+	roll	$16,%r13d
+	addl	%r12d,%esi
+	xorl	%esi,%r8d
+	roll	$12,%r8d
+	addl	%r13d,%edi
+	xorl	%edi,%r9d
+	roll	$12,%r9d
+	addl	%r8d,%eax
+	xorl	%eax,%r12d
+	roll	$8,%r12d
+	addl	%r9d,%ebx
+	xorl	%ebx,%r13d
+	roll	$8,%r13d
+	addl	%r12d,%esi
+	xorl	%esi,%r8d
+	roll	$7,%r8d
+	addl	%r13d,%edi
+	xorl	%edi,%r9d
+	roll	$7,%r9d
+	movl	%esi,32(%rsp)
+	movl	%edi,36(%rsp)
+	movl	40(%rsp),%esi
+	movl	44(%rsp),%edi
+	addl	%r10d,%ecx
+	xorl	%ecx,%r14d
+	roll	$16,%r14d
+	addl	%r11d,%edx
+	xorl	%edx,%r15d
+	roll	$16,%r15d
+	addl	%r14d,%esi
+	xorl	%esi,%r10d
+	roll	$12,%r10d
+	addl	%r15d,%edi
+	xorl	%edi,%r11d
+	roll	$12,%r11d
+	addl	%r10d,%ecx
+	xorl	%ecx,%r14d
+	roll	$8,%r14d
+	addl	%r11d,%edx
+	xorl	%edx,%r15d
+	roll	$8,%r15d
+	addl	%r14d,%esi
+	xorl	%esi,%r10d
+	roll	$7,%r10d
+	addl	%r15d,%edi
+	xorl	%edi,%r11d
+	roll	$7,%r11d
+	addl	%r9d,%eax
+	xorl	%eax,%r15d
+	roll	$16,%r15d
+	addl	%r10d,%ebx
+	xorl	%ebx,%r12d
+	roll	$16,%r12d
+	addl	%r15d,%esi
+	xorl	%esi,%r9d
+	roll	$12,%r9d
+	addl	%r12d,%edi
+	xorl	%edi,%r10d
+	roll	$12,%r10d
+	addl	%r9d,%eax
+	xorl	%eax,%r15d
+	roll	$8,%r15d
+	addl	%r10d,%ebx
+	xorl	%ebx,%r12d
+	roll	$8,%r12d
+	addl	%r15d,%esi
+	xorl	%esi,%r9d
+	roll	$7,%r9d
+	addl	%r12d,%edi
+	xorl	%edi,%r10d
+	roll	$7,%r10d
+	movl	%esi,40(%rsp)
+	movl	%edi,44(%rsp)
+	movl	32(%rsp),%esi
+	movl	36(%rsp),%edi
+	addl	%r11d,%ecx
+	xorl	%ecx,%r13d
+	roll	$16,%r13d
+	addl	%r8d,%edx
+	xorl	%edx,%r14d
+	roll	$16,%r14d
+	addl	%r13d,%esi
+	xorl	%esi,%r11d
+	roll	$12,%r11d
+	addl	%r14d,%edi
+	xorl	%edi,%r8d
+	roll	$12,%r8d
+	addl	%r11d,%ecx
+	xorl	%ecx,%r13d
+	roll	$8,%r13d
+	addl	%r8d,%edx
+	xorl	%edx,%r14d
+	roll	$8,%r14d
+	addl	%r13d,%esi
+	xorl	%esi,%r11d
+	roll	$7,%r11d
+	addl	%r14d,%edi
+	xorl	%edi,%r8d
+	roll	$7,%r8d
+	decl	%ebp
+	jnz	.Loop
+	movl	%edi,36(%rsp)
+	movl	%esi,32(%rsp)
+	movq	64(%rsp),%rbp
+	movdqa	%xmm2,%xmm1
+	movq	64+8(%rsp),%rsi
+	paddd	%xmm4,%xmm3
+	movq	64+16(%rsp),%rdi
+
+	addl	$0x61707865,%eax
+	addl	$0x3320646e,%ebx
+	addl	$0x79622d32,%ecx
+	addl	$0x6b206574,%edx
+	addl	16(%rsp),%r8d
+	addl	20(%rsp),%r9d
+	addl	24(%rsp),%r10d
+	addl	28(%rsp),%r11d
+	addl	48(%rsp),%r12d
+	addl	52(%rsp),%r13d
+	addl	56(%rsp),%r14d
+	addl	60(%rsp),%r15d
+	paddd	32(%rsp),%xmm1
+
+	cmpq	$64,%rbp
+	jb	.Ltail
+
+	xorl	0(%rsi),%eax
+	xorl	4(%rsi),%ebx
+	xorl	8(%rsi),%ecx
+	xorl	12(%rsi),%edx
+	xorl	16(%rsi),%r8d
+	xorl	20(%rsi),%r9d
+	xorl	24(%rsi),%r10d
+	xorl	28(%rsi),%r11d
+	movdqu	32(%rsi),%xmm0
+	xorl	48(%rsi),%r12d
+	xorl	52(%rsi),%r13d
+	xorl	56(%rsi),%r14d
+	xorl	60(%rsi),%r15d
+	leaq	64(%rsi),%rsi
+	pxor	%xmm1,%xmm0
+
+	movdqa	%xmm2,32(%rsp)
+	movd	%xmm3,48(%rsp)
+
+	movl	%eax,0(%rdi)
+	movl	%ebx,4(%rdi)
+	movl	%ecx,8(%rdi)
+	movl	%edx,12(%rdi)
+	movl	%r8d,16(%rdi)
+	movl	%r9d,20(%rdi)
+	movl	%r10d,24(%rdi)
+	movl	%r11d,28(%rdi)
+	movdqu	%xmm0,32(%rdi)
+	movl	%r12d,48(%rdi)
+	movl	%r13d,52(%rdi)
+	movl	%r14d,56(%rdi)
+	movl	%r15d,60(%rdi)
+	leaq	64(%rdi),%rdi
+
+	subq	$64,%rbp
+	jnz	.Loop_outer
+
+	jmp	.Ldone
+
+.align	16
+.Ltail:
+	movl	%eax,0(%rsp)
+	movl	%ebx,4(%rsp)
+	xorq	%rbx,%rbx
+	movl	%ecx,8(%rsp)
+	movl	%edx,12(%rsp)
+	movl	%r8d,16(%rsp)
+	movl	%r9d,20(%rsp)
+	movl	%r10d,24(%rsp)
+	movl	%r11d,28(%rsp)
+	movdqa	%xmm1,32(%rsp)
+	movl	%r12d,48(%rsp)
+	movl	%r13d,52(%rsp)
+	movl	%r14d,56(%rsp)
+	movl	%r15d,60(%rsp)
+
+.Loop_tail:
+	movzbl	(%rsi,%rbx,1),%eax
+	movzbl	(%rsp,%rbx,1),%edx
+	leaq	1(%rbx),%rbx
+	xorl	%edx,%eax
+	movb	%al,-1(%rdi,%rbx,1)
+	decq	%rbp
+	jnz	.Loop_tail
+
+.Ldone:
+	leaq	64+24+48(%rsp),%rsi
+.cfi_def_cfa	%rsi,8
+	movq	-48(%rsi),%r15
+.cfi_restore	%r15
+	movq	-40(%rsi),%r14
+.cfi_restore	%r14
+	movq	-32(%rsi),%r13
+.cfi_restore	%r13
+	movq	-24(%rsi),%r12
+.cfi_restore	%r12
+	movq	-16(%rsi),%rbp
+.cfi_restore	%rbp
+	movq	-8(%rsi),%rbx
+.cfi_restore	%rbx
+	leaq	(%rsi),%rsp
+.cfi_def_cfa_register	%rsp
+.Lno_data:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_ctr32,.-ChaCha20_ctr32
+.type	ChaCha20_ssse3,@function
+.align	32
+ChaCha20_ssse3:
+.cfi_startproc	
+.LChaCha20_ssse3:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	testl	$2048,%r10d
+	jnz	.LChaCha20_4xop
+	cmpq	$128,%rdx
+	je	.LChaCha20_128
+	ja	.LChaCha20_4x
+
+.Ldo_sse3_after_all:
+	subq	$64+8,%rsp
+	movdqa	.Lsigma(%rip),%xmm0
+	movdqu	(%rcx),%xmm1
+	movdqu	16(%rcx),%xmm2
+	movdqu	(%r8),%xmm3
+	movdqa	.Lrot16(%rip),%xmm6
+	movdqa	.Lrot24(%rip),%xmm7
+
+	movdqa	%xmm0,0(%rsp)
+	movdqa	%xmm1,16(%rsp)
+	movdqa	%xmm2,32(%rsp)
+	movdqa	%xmm3,48(%rsp)
+	movq	$10,%r8
+	jmp	.Loop_ssse3
+
+.align	32
+.Loop_outer_ssse3:
+	movdqa	.Lone(%rip),%xmm3
+	movdqa	0(%rsp),%xmm0
+	movdqa	16(%rsp),%xmm1
+	movdqa	32(%rsp),%xmm2
+	paddd	48(%rsp),%xmm3
+	movq	$10,%r8
+	movdqa	%xmm3,48(%rsp)
+	jmp	.Loop_ssse3
+
+.align	32
+.Loop_ssse3:
+	paddd	%xmm1,%xmm0
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,222
+	paddd	%xmm3,%xmm2
+	pxor	%xmm2,%xmm1
+	movdqa	%xmm1,%xmm4
+	psrld	$20,%xmm1
+	pslld	$12,%xmm4
+	por	%xmm4,%xmm1
+	paddd	%xmm1,%xmm0
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,223
+	paddd	%xmm3,%xmm2
+	pxor	%xmm2,%xmm1
+	movdqa	%xmm1,%xmm4
+	psrld	$25,%xmm1
+	pslld	$7,%xmm4
+	por	%xmm4,%xmm1
+	pshufd	$78,%xmm2,%xmm2
+	pshufd	$57,%xmm1,%xmm1
+	pshufd	$147,%xmm3,%xmm3
+	nop
+	paddd	%xmm1,%xmm0
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,222
+	paddd	%xmm3,%xmm2
+	pxor	%xmm2,%xmm1
+	movdqa	%xmm1,%xmm4
+	psrld	$20,%xmm1
+	pslld	$12,%xmm4
+	por	%xmm4,%xmm1
+	paddd	%xmm1,%xmm0
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,223
+	paddd	%xmm3,%xmm2
+	pxor	%xmm2,%xmm1
+	movdqa	%xmm1,%xmm4
+	psrld	$25,%xmm1
+	pslld	$7,%xmm4
+	por	%xmm4,%xmm1
+	pshufd	$78,%xmm2,%xmm2
+	pshufd	$147,%xmm1,%xmm1
+	pshufd	$57,%xmm3,%xmm3
+	decq	%r8
+	jnz	.Loop_ssse3
+	paddd	0(%rsp),%xmm0
+	paddd	16(%rsp),%xmm1
+	paddd	32(%rsp),%xmm2
+	paddd	48(%rsp),%xmm3
+
+	cmpq	$64,%rdx
+	jb	.Ltail_ssse3
+
+	movdqu	0(%rsi),%xmm4
+	movdqu	16(%rsi),%xmm5
+	pxor	%xmm4,%xmm0
+	movdqu	32(%rsi),%xmm4
+	pxor	%xmm5,%xmm1
+	movdqu	48(%rsi),%xmm5
+	leaq	64(%rsi),%rsi
+	pxor	%xmm4,%xmm2
+	pxor	%xmm5,%xmm3
+
+	movdqu	%xmm0,0(%rdi)
+	movdqu	%xmm1,16(%rdi)
+	movdqu	%xmm2,32(%rdi)
+	movdqu	%xmm3,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	subq	$64,%rdx
+	jnz	.Loop_outer_ssse3
+
+	jmp	.Ldone_ssse3
+
+.align	16
+.Ltail_ssse3:
+	movdqa	%xmm0,0(%rsp)
+	movdqa	%xmm1,16(%rsp)
+	movdqa	%xmm2,32(%rsp)
+	movdqa	%xmm3,48(%rsp)
+	xorq	%r8,%r8
+
+.Loop_tail_ssse3:
+	movzbl	(%rsi,%r8,1),%eax
+	movzbl	(%rsp,%r8,1),%ecx
+	leaq	1(%r8),%r8
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r8,1)
+	decq	%rdx
+	jnz	.Loop_tail_ssse3
+
+.Ldone_ssse3:
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.Lssse3_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_ssse3,.-ChaCha20_ssse3
+.type	ChaCha20_128,@function
+.align	32
+ChaCha20_128:
+.cfi_startproc	
+.LChaCha20_128:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	subq	$64+8,%rsp
+	movdqa	.Lsigma(%rip),%xmm8
+	movdqu	(%rcx),%xmm9
+	movdqu	16(%rcx),%xmm2
+	movdqu	(%r8),%xmm3
+	movdqa	.Lone(%rip),%xmm1
+	movdqa	.Lrot16(%rip),%xmm6
+	movdqa	.Lrot24(%rip),%xmm7
+
+	movdqa	%xmm8,%xmm10
+	movdqa	%xmm8,0(%rsp)
+	movdqa	%xmm9,%xmm11
+	movdqa	%xmm9,16(%rsp)
+	movdqa	%xmm2,%xmm0
+	movdqa	%xmm2,32(%rsp)
+	paddd	%xmm3,%xmm1
+	movdqa	%xmm3,48(%rsp)
+	movq	$10,%r8
+	jmp	.Loop_128
+
+.align	32
+.Loop_128:
+	paddd	%xmm9,%xmm8
+	pxor	%xmm8,%xmm3
+	paddd	%xmm11,%xmm10
+	pxor	%xmm10,%xmm1
+.byte	102,15,56,0,222
+.byte	102,15,56,0,206
+	paddd	%xmm3,%xmm2
+	paddd	%xmm1,%xmm0
+	pxor	%xmm2,%xmm9
+	pxor	%xmm0,%xmm11
+	movdqa	%xmm9,%xmm4
+	psrld	$20,%xmm9
+	movdqa	%xmm11,%xmm5
+	pslld	$12,%xmm4
+	psrld	$20,%xmm11
+	por	%xmm4,%xmm9
+	pslld	$12,%xmm5
+	por	%xmm5,%xmm11
+	paddd	%xmm9,%xmm8
+	pxor	%xmm8,%xmm3
+	paddd	%xmm11,%xmm10
+	pxor	%xmm10,%xmm1
+.byte	102,15,56,0,223
+.byte	102,15,56,0,207
+	paddd	%xmm3,%xmm2
+	paddd	%xmm1,%xmm0
+	pxor	%xmm2,%xmm9
+	pxor	%xmm0,%xmm11
+	movdqa	%xmm9,%xmm4
+	psrld	$25,%xmm9
+	movdqa	%xmm11,%xmm5
+	pslld	$7,%xmm4
+	psrld	$25,%xmm11
+	por	%xmm4,%xmm9
+	pslld	$7,%xmm5
+	por	%xmm5,%xmm11
+	pshufd	$78,%xmm2,%xmm2
+	pshufd	$57,%xmm9,%xmm9
+	pshufd	$147,%xmm3,%xmm3
+	pshufd	$78,%xmm0,%xmm0
+	pshufd	$57,%xmm11,%xmm11
+	pshufd	$147,%xmm1,%xmm1
+	paddd	%xmm9,%xmm8
+	pxor	%xmm8,%xmm3
+	paddd	%xmm11,%xmm10
+	pxor	%xmm10,%xmm1
+.byte	102,15,56,0,222
+.byte	102,15,56,0,206
+	paddd	%xmm3,%xmm2
+	paddd	%xmm1,%xmm0
+	pxor	%xmm2,%xmm9
+	pxor	%xmm0,%xmm11
+	movdqa	%xmm9,%xmm4
+	psrld	$20,%xmm9
+	movdqa	%xmm11,%xmm5
+	pslld	$12,%xmm4
+	psrld	$20,%xmm11
+	por	%xmm4,%xmm9
+	pslld	$12,%xmm5
+	por	%xmm5,%xmm11
+	paddd	%xmm9,%xmm8
+	pxor	%xmm8,%xmm3
+	paddd	%xmm11,%xmm10
+	pxor	%xmm10,%xmm1
+.byte	102,15,56,0,223
+.byte	102,15,56,0,207
+	paddd	%xmm3,%xmm2
+	paddd	%xmm1,%xmm0
+	pxor	%xmm2,%xmm9
+	pxor	%xmm0,%xmm11
+	movdqa	%xmm9,%xmm4
+	psrld	$25,%xmm9
+	movdqa	%xmm11,%xmm5
+	pslld	$7,%xmm4
+	psrld	$25,%xmm11
+	por	%xmm4,%xmm9
+	pslld	$7,%xmm5
+	por	%xmm5,%xmm11
+	pshufd	$78,%xmm2,%xmm2
+	pshufd	$147,%xmm9,%xmm9
+	pshufd	$57,%xmm3,%xmm3
+	pshufd	$78,%xmm0,%xmm0
+	pshufd	$147,%xmm11,%xmm11
+	pshufd	$57,%xmm1,%xmm1
+	decq	%r8
+	jnz	.Loop_128
+	paddd	0(%rsp),%xmm8
+	paddd	16(%rsp),%xmm9
+	paddd	32(%rsp),%xmm2
+	paddd	48(%rsp),%xmm3
+	paddd	.Lone(%rip),%xmm1
+	paddd	0(%rsp),%xmm10
+	paddd	16(%rsp),%xmm11
+	paddd	32(%rsp),%xmm0
+	paddd	48(%rsp),%xmm1
+
+	movdqu	0(%rsi),%xmm4
+	movdqu	16(%rsi),%xmm5
+	pxor	%xmm4,%xmm8
+	movdqu	32(%rsi),%xmm4
+	pxor	%xmm5,%xmm9
+	movdqu	48(%rsi),%xmm5
+	pxor	%xmm4,%xmm2
+	movdqu	64(%rsi),%xmm4
+	pxor	%xmm5,%xmm3
+	movdqu	80(%rsi),%xmm5
+	pxor	%xmm4,%xmm10
+	movdqu	96(%rsi),%xmm4
+	pxor	%xmm5,%xmm11
+	movdqu	112(%rsi),%xmm5
+	pxor	%xmm4,%xmm0
+	pxor	%xmm5,%xmm1
+
+	movdqu	%xmm8,0(%rdi)
+	movdqu	%xmm9,16(%rdi)
+	movdqu	%xmm2,32(%rdi)
+	movdqu	%xmm3,48(%rdi)
+	movdqu	%xmm10,64(%rdi)
+	movdqu	%xmm11,80(%rdi)
+	movdqu	%xmm0,96(%rdi)
+	movdqu	%xmm1,112(%rdi)
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.L128_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_128,.-ChaCha20_128
+.type	ChaCha20_4x,@function
+.align	32
+ChaCha20_4x:
+.cfi_startproc	
+.LChaCha20_4x:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	movq	%r10,%r11
+	shrq	$32,%r10
+	testq	$32,%r10
+	jnz	.LChaCha20_8x
+	cmpq	$192,%rdx
+	ja	.Lproceed4x
+
+	andq	$71303168,%r11
+	cmpq	$4194304,%r11
+	je	.Ldo_sse3_after_all
+
+.Lproceed4x:
+	subq	$0x140+8,%rsp
+	movdqa	.Lsigma(%rip),%xmm11
+	movdqu	(%rcx),%xmm15
+	movdqu	16(%rcx),%xmm7
+	movdqu	(%r8),%xmm3
+	leaq	256(%rsp),%rcx
+	leaq	.Lrot16(%rip),%r10
+	leaq	.Lrot24(%rip),%r11
+
+	pshufd	$0x00,%xmm11,%xmm8
+	pshufd	$0x55,%xmm11,%xmm9
+	movdqa	%xmm8,64(%rsp)
+	pshufd	$0xaa,%xmm11,%xmm10
+	movdqa	%xmm9,80(%rsp)
+	pshufd	$0xff,%xmm11,%xmm11
+	movdqa	%xmm10,96(%rsp)
+	movdqa	%xmm11,112(%rsp)
+
+	pshufd	$0x00,%xmm15,%xmm12
+	pshufd	$0x55,%xmm15,%xmm13
+	movdqa	%xmm12,128-256(%rcx)
+	pshufd	$0xaa,%xmm15,%xmm14
+	movdqa	%xmm13,144-256(%rcx)
+	pshufd	$0xff,%xmm15,%xmm15
+	movdqa	%xmm14,160-256(%rcx)
+	movdqa	%xmm15,176-256(%rcx)
+
+	pshufd	$0x00,%xmm7,%xmm4
+	pshufd	$0x55,%xmm7,%xmm5
+	movdqa	%xmm4,192-256(%rcx)
+	pshufd	$0xaa,%xmm7,%xmm6
+	movdqa	%xmm5,208-256(%rcx)
+	pshufd	$0xff,%xmm7,%xmm7
+	movdqa	%xmm6,224-256(%rcx)
+	movdqa	%xmm7,240-256(%rcx)
+
+	pshufd	$0x00,%xmm3,%xmm0
+	pshufd	$0x55,%xmm3,%xmm1
+	paddd	.Linc(%rip),%xmm0
+	pshufd	$0xaa,%xmm3,%xmm2
+	movdqa	%xmm1,272-256(%rcx)
+	pshufd	$0xff,%xmm3,%xmm3
+	movdqa	%xmm2,288-256(%rcx)
+	movdqa	%xmm3,304-256(%rcx)
+
+	jmp	.Loop_enter4x
+
+.align	32
+.Loop_outer4x:
+	movdqa	64(%rsp),%xmm8
+	movdqa	80(%rsp),%xmm9
+	movdqa	96(%rsp),%xmm10
+	movdqa	112(%rsp),%xmm11
+	movdqa	128-256(%rcx),%xmm12
+	movdqa	144-256(%rcx),%xmm13
+	movdqa	160-256(%rcx),%xmm14
+	movdqa	176-256(%rcx),%xmm15
+	movdqa	192-256(%rcx),%xmm4
+	movdqa	208-256(%rcx),%xmm5
+	movdqa	224-256(%rcx),%xmm6
+	movdqa	240-256(%rcx),%xmm7
+	movdqa	256-256(%rcx),%xmm0
+	movdqa	272-256(%rcx),%xmm1
+	movdqa	288-256(%rcx),%xmm2
+	movdqa	304-256(%rcx),%xmm3
+	paddd	.Lfour(%rip),%xmm0
+
+.Loop_enter4x:
+	movdqa	%xmm6,32(%rsp)
+	movdqa	%xmm7,48(%rsp)
+	movdqa	(%r10),%xmm7
+	movl	$10,%eax
+	movdqa	%xmm0,256-256(%rcx)
+	jmp	.Loop4x
+
+.align	32
+.Loop4x:
+	paddd	%xmm12,%xmm8
+	paddd	%xmm13,%xmm9
+	pxor	%xmm8,%xmm0
+	pxor	%xmm9,%xmm1
+.byte	102,15,56,0,199
+.byte	102,15,56,0,207
+	paddd	%xmm0,%xmm4
+	paddd	%xmm1,%xmm5
+	pxor	%xmm4,%xmm12
+	pxor	%xmm5,%xmm13
+	movdqa	%xmm12,%xmm6
+	pslld	$12,%xmm12
+	psrld	$20,%xmm6
+	movdqa	%xmm13,%xmm7
+	pslld	$12,%xmm13
+	por	%xmm6,%xmm12
+	psrld	$20,%xmm7
+	movdqa	(%r11),%xmm6
+	por	%xmm7,%xmm13
+	paddd	%xmm12,%xmm8
+	paddd	%xmm13,%xmm9
+	pxor	%xmm8,%xmm0
+	pxor	%xmm9,%xmm1
+.byte	102,15,56,0,198
+.byte	102,15,56,0,206
+	paddd	%xmm0,%xmm4
+	paddd	%xmm1,%xmm5
+	pxor	%xmm4,%xmm12
+	pxor	%xmm5,%xmm13
+	movdqa	%xmm12,%xmm7
+	pslld	$7,%xmm12
+	psrld	$25,%xmm7
+	movdqa	%xmm13,%xmm6
+	pslld	$7,%xmm13
+	por	%xmm7,%xmm12
+	psrld	$25,%xmm6
+	movdqa	(%r10),%xmm7
+	por	%xmm6,%xmm13
+	movdqa	%xmm4,0(%rsp)
+	movdqa	%xmm5,16(%rsp)
+	movdqa	32(%rsp),%xmm4
+	movdqa	48(%rsp),%xmm5
+	paddd	%xmm14,%xmm10
+	paddd	%xmm15,%xmm11
+	pxor	%xmm10,%xmm2
+	pxor	%xmm11,%xmm3
+.byte	102,15,56,0,215
+.byte	102,15,56,0,223
+	paddd	%xmm2,%xmm4
+	paddd	%xmm3,%xmm5
+	pxor	%xmm4,%xmm14
+	pxor	%xmm5,%xmm15
+	movdqa	%xmm14,%xmm6
+	pslld	$12,%xmm14
+	psrld	$20,%xmm6
+	movdqa	%xmm15,%xmm7
+	pslld	$12,%xmm15
+	por	%xmm6,%xmm14
+	psrld	$20,%xmm7
+	movdqa	(%r11),%xmm6
+	por	%xmm7,%xmm15
+	paddd	%xmm14,%xmm10
+	paddd	%xmm15,%xmm11
+	pxor	%xmm10,%xmm2
+	pxor	%xmm11,%xmm3
+.byte	102,15,56,0,214
+.byte	102,15,56,0,222
+	paddd	%xmm2,%xmm4
+	paddd	%xmm3,%xmm5
+	pxor	%xmm4,%xmm14
+	pxor	%xmm5,%xmm15
+	movdqa	%xmm14,%xmm7
+	pslld	$7,%xmm14
+	psrld	$25,%xmm7
+	movdqa	%xmm15,%xmm6
+	pslld	$7,%xmm15
+	por	%xmm7,%xmm14
+	psrld	$25,%xmm6
+	movdqa	(%r10),%xmm7
+	por	%xmm6,%xmm15
+	paddd	%xmm13,%xmm8
+	paddd	%xmm14,%xmm9
+	pxor	%xmm8,%xmm3
+	pxor	%xmm9,%xmm0
+.byte	102,15,56,0,223
+.byte	102,15,56,0,199
+	paddd	%xmm3,%xmm4
+	paddd	%xmm0,%xmm5
+	pxor	%xmm4,%xmm13
+	pxor	%xmm5,%xmm14
+	movdqa	%xmm13,%xmm6
+	pslld	$12,%xmm13
+	psrld	$20,%xmm6
+	movdqa	%xmm14,%xmm7
+	pslld	$12,%xmm14
+	por	%xmm6,%xmm13
+	psrld	$20,%xmm7
+	movdqa	(%r11),%xmm6
+	por	%xmm7,%xmm14
+	paddd	%xmm13,%xmm8
+	paddd	%xmm14,%xmm9
+	pxor	%xmm8,%xmm3
+	pxor	%xmm9,%xmm0
+.byte	102,15,56,0,222
+.byte	102,15,56,0,198
+	paddd	%xmm3,%xmm4
+	paddd	%xmm0,%xmm5
+	pxor	%xmm4,%xmm13
+	pxor	%xmm5,%xmm14
+	movdqa	%xmm13,%xmm7
+	pslld	$7,%xmm13
+	psrld	$25,%xmm7
+	movdqa	%xmm14,%xmm6
+	pslld	$7,%xmm14
+	por	%xmm7,%xmm13
+	psrld	$25,%xmm6
+	movdqa	(%r10),%xmm7
+	por	%xmm6,%xmm14
+	movdqa	%xmm4,32(%rsp)
+	movdqa	%xmm5,48(%rsp)
+	movdqa	0(%rsp),%xmm4
+	movdqa	16(%rsp),%xmm5
+	paddd	%xmm15,%xmm10
+	paddd	%xmm12,%xmm11
+	pxor	%xmm10,%xmm1
+	pxor	%xmm11,%xmm2
+.byte	102,15,56,0,207
+.byte	102,15,56,0,215
+	paddd	%xmm1,%xmm4
+	paddd	%xmm2,%xmm5
+	pxor	%xmm4,%xmm15
+	pxor	%xmm5,%xmm12
+	movdqa	%xmm15,%xmm6
+	pslld	$12,%xmm15
+	psrld	$20,%xmm6
+	movdqa	%xmm12,%xmm7
+	pslld	$12,%xmm12
+	por	%xmm6,%xmm15
+	psrld	$20,%xmm7
+	movdqa	(%r11),%xmm6
+	por	%xmm7,%xmm12
+	paddd	%xmm15,%xmm10
+	paddd	%xmm12,%xmm11
+	pxor	%xmm10,%xmm1
+	pxor	%xmm11,%xmm2
+.byte	102,15,56,0,206
+.byte	102,15,56,0,214
+	paddd	%xmm1,%xmm4
+	paddd	%xmm2,%xmm5
+	pxor	%xmm4,%xmm15
+	pxor	%xmm5,%xmm12
+	movdqa	%xmm15,%xmm7
+	pslld	$7,%xmm15
+	psrld	$25,%xmm7
+	movdqa	%xmm12,%xmm6
+	pslld	$7,%xmm12
+	por	%xmm7,%xmm15
+	psrld	$25,%xmm6
+	movdqa	(%r10),%xmm7
+	por	%xmm6,%xmm12
+	decl	%eax
+	jnz	.Loop4x
+
+	paddd	64(%rsp),%xmm8
+	paddd	80(%rsp),%xmm9
+	paddd	96(%rsp),%xmm10
+	paddd	112(%rsp),%xmm11
+
+	movdqa	%xmm8,%xmm6
+	punpckldq	%xmm9,%xmm8
+	movdqa	%xmm10,%xmm7
+	punpckldq	%xmm11,%xmm10
+	punpckhdq	%xmm9,%xmm6
+	punpckhdq	%xmm11,%xmm7
+	movdqa	%xmm8,%xmm9
+	punpcklqdq	%xmm10,%xmm8
+	movdqa	%xmm6,%xmm11
+	punpcklqdq	%xmm7,%xmm6
+	punpckhqdq	%xmm10,%xmm9
+	punpckhqdq	%xmm7,%xmm11
+	paddd	128-256(%rcx),%xmm12
+	paddd	144-256(%rcx),%xmm13
+	paddd	160-256(%rcx),%xmm14
+	paddd	176-256(%rcx),%xmm15
+
+	movdqa	%xmm8,0(%rsp)
+	movdqa	%xmm9,16(%rsp)
+	movdqa	32(%rsp),%xmm8
+	movdqa	48(%rsp),%xmm9
+
+	movdqa	%xmm12,%xmm10
+	punpckldq	%xmm13,%xmm12
+	movdqa	%xmm14,%xmm7
+	punpckldq	%xmm15,%xmm14
+	punpckhdq	%xmm13,%xmm10
+	punpckhdq	%xmm15,%xmm7
+	movdqa	%xmm12,%xmm13
+	punpcklqdq	%xmm14,%xmm12
+	movdqa	%xmm10,%xmm15
+	punpcklqdq	%xmm7,%xmm10
+	punpckhqdq	%xmm14,%xmm13
+	punpckhqdq	%xmm7,%xmm15
+	paddd	192-256(%rcx),%xmm4
+	paddd	208-256(%rcx),%xmm5
+	paddd	224-256(%rcx),%xmm8
+	paddd	240-256(%rcx),%xmm9
+
+	movdqa	%xmm6,32(%rsp)
+	movdqa	%xmm11,48(%rsp)
+
+	movdqa	%xmm4,%xmm14
+	punpckldq	%xmm5,%xmm4
+	movdqa	%xmm8,%xmm7
+	punpckldq	%xmm9,%xmm8
+	punpckhdq	%xmm5,%xmm14
+	punpckhdq	%xmm9,%xmm7
+	movdqa	%xmm4,%xmm5
+	punpcklqdq	%xmm8,%xmm4
+	movdqa	%xmm14,%xmm9
+	punpcklqdq	%xmm7,%xmm14
+	punpckhqdq	%xmm8,%xmm5
+	punpckhqdq	%xmm7,%xmm9
+	paddd	256-256(%rcx),%xmm0
+	paddd	272-256(%rcx),%xmm1
+	paddd	288-256(%rcx),%xmm2
+	paddd	304-256(%rcx),%xmm3
+
+	movdqa	%xmm0,%xmm8
+	punpckldq	%xmm1,%xmm0
+	movdqa	%xmm2,%xmm7
+	punpckldq	%xmm3,%xmm2
+	punpckhdq	%xmm1,%xmm8
+	punpckhdq	%xmm3,%xmm7
+	movdqa	%xmm0,%xmm1
+	punpcklqdq	%xmm2,%xmm0
+	movdqa	%xmm8,%xmm3
+	punpcklqdq	%xmm7,%xmm8
+	punpckhqdq	%xmm2,%xmm1
+	punpckhqdq	%xmm7,%xmm3
+	cmpq	$256,%rdx
+	jb	.Ltail4x
+
+	movdqu	0(%rsi),%xmm6
+	movdqu	16(%rsi),%xmm11
+	movdqu	32(%rsi),%xmm2
+	movdqu	48(%rsi),%xmm7
+	pxor	0(%rsp),%xmm6
+	pxor	%xmm12,%xmm11
+	pxor	%xmm4,%xmm2
+	pxor	%xmm0,%xmm7
+
+	movdqu	%xmm6,0(%rdi)
+	movdqu	64(%rsi),%xmm6
+	movdqu	%xmm11,16(%rdi)
+	movdqu	80(%rsi),%xmm11
+	movdqu	%xmm2,32(%rdi)
+	movdqu	96(%rsi),%xmm2
+	movdqu	%xmm7,48(%rdi)
+	movdqu	112(%rsi),%xmm7
+	leaq	128(%rsi),%rsi
+	pxor	16(%rsp),%xmm6
+	pxor	%xmm13,%xmm11
+	pxor	%xmm5,%xmm2
+	pxor	%xmm1,%xmm7
+
+	movdqu	%xmm6,64(%rdi)
+	movdqu	0(%rsi),%xmm6
+	movdqu	%xmm11,80(%rdi)
+	movdqu	16(%rsi),%xmm11
+	movdqu	%xmm2,96(%rdi)
+	movdqu	32(%rsi),%xmm2
+	movdqu	%xmm7,112(%rdi)
+	leaq	128(%rdi),%rdi
+	movdqu	48(%rsi),%xmm7
+	pxor	32(%rsp),%xmm6
+	pxor	%xmm10,%xmm11
+	pxor	%xmm14,%xmm2
+	pxor	%xmm8,%xmm7
+
+	movdqu	%xmm6,0(%rdi)
+	movdqu	64(%rsi),%xmm6
+	movdqu	%xmm11,16(%rdi)
+	movdqu	80(%rsi),%xmm11
+	movdqu	%xmm2,32(%rdi)
+	movdqu	96(%rsi),%xmm2
+	movdqu	%xmm7,48(%rdi)
+	movdqu	112(%rsi),%xmm7
+	leaq	128(%rsi),%rsi
+	pxor	48(%rsp),%xmm6
+	pxor	%xmm15,%xmm11
+	pxor	%xmm9,%xmm2
+	pxor	%xmm3,%xmm7
+	movdqu	%xmm6,64(%rdi)
+	movdqu	%xmm11,80(%rdi)
+	movdqu	%xmm2,96(%rdi)
+	movdqu	%xmm7,112(%rdi)
+	leaq	128(%rdi),%rdi
+
+	subq	$256,%rdx
+	jnz	.Loop_outer4x
+
+	jmp	.Ldone4x
+
+.Ltail4x:
+	cmpq	$192,%rdx
+	jae	.L192_or_more4x
+	cmpq	$128,%rdx
+	jae	.L128_or_more4x
+	cmpq	$64,%rdx
+	jae	.L64_or_more4x
+
+
+	xorq	%r10,%r10
+
+	movdqa	%xmm12,16(%rsp)
+	movdqa	%xmm4,32(%rsp)
+	movdqa	%xmm0,48(%rsp)
+	jmp	.Loop_tail4x
+
+.align	32
+.L64_or_more4x:
+	movdqu	0(%rsi),%xmm6
+	movdqu	16(%rsi),%xmm11
+	movdqu	32(%rsi),%xmm2
+	movdqu	48(%rsi),%xmm7
+	pxor	0(%rsp),%xmm6
+	pxor	%xmm12,%xmm11
+	pxor	%xmm4,%xmm2
+	pxor	%xmm0,%xmm7
+	movdqu	%xmm6,0(%rdi)
+	movdqu	%xmm11,16(%rdi)
+	movdqu	%xmm2,32(%rdi)
+	movdqu	%xmm7,48(%rdi)
+	je	.Ldone4x
+
+	movdqa	16(%rsp),%xmm6
+	leaq	64(%rsi),%rsi
+	xorq	%r10,%r10
+	movdqa	%xmm6,0(%rsp)
+	movdqa	%xmm13,16(%rsp)
+	leaq	64(%rdi),%rdi
+	movdqa	%xmm5,32(%rsp)
+	subq	$64,%rdx
+	movdqa	%xmm1,48(%rsp)
+	jmp	.Loop_tail4x
+
+.align	32
+.L128_or_more4x:
+	movdqu	0(%rsi),%xmm6
+	movdqu	16(%rsi),%xmm11
+	movdqu	32(%rsi),%xmm2
+	movdqu	48(%rsi),%xmm7
+	pxor	0(%rsp),%xmm6
+	pxor	%xmm12,%xmm11
+	pxor	%xmm4,%xmm2
+	pxor	%xmm0,%xmm7
+
+	movdqu	%xmm6,0(%rdi)
+	movdqu	64(%rsi),%xmm6
+	movdqu	%xmm11,16(%rdi)
+	movdqu	80(%rsi),%xmm11
+	movdqu	%xmm2,32(%rdi)
+	movdqu	96(%rsi),%xmm2
+	movdqu	%xmm7,48(%rdi)
+	movdqu	112(%rsi),%xmm7
+	pxor	16(%rsp),%xmm6
+	pxor	%xmm13,%xmm11
+	pxor	%xmm5,%xmm2
+	pxor	%xmm1,%xmm7
+	movdqu	%xmm6,64(%rdi)
+	movdqu	%xmm11,80(%rdi)
+	movdqu	%xmm2,96(%rdi)
+	movdqu	%xmm7,112(%rdi)
+	je	.Ldone4x
+
+	movdqa	32(%rsp),%xmm6
+	leaq	128(%rsi),%rsi
+	xorq	%r10,%r10
+	movdqa	%xmm6,0(%rsp)
+	movdqa	%xmm10,16(%rsp)
+	leaq	128(%rdi),%rdi
+	movdqa	%xmm14,32(%rsp)
+	subq	$128,%rdx
+	movdqa	%xmm8,48(%rsp)
+	jmp	.Loop_tail4x
+
+.align	32
+.L192_or_more4x:
+	movdqu	0(%rsi),%xmm6
+	movdqu	16(%rsi),%xmm11
+	movdqu	32(%rsi),%xmm2
+	movdqu	48(%rsi),%xmm7
+	pxor	0(%rsp),%xmm6
+	pxor	%xmm12,%xmm11
+	pxor	%xmm4,%xmm2
+	pxor	%xmm0,%xmm7
+
+	movdqu	%xmm6,0(%rdi)
+	movdqu	64(%rsi),%xmm6
+	movdqu	%xmm11,16(%rdi)
+	movdqu	80(%rsi),%xmm11
+	movdqu	%xmm2,32(%rdi)
+	movdqu	96(%rsi),%xmm2
+	movdqu	%xmm7,48(%rdi)
+	movdqu	112(%rsi),%xmm7
+	leaq	128(%rsi),%rsi
+	pxor	16(%rsp),%xmm6
+	pxor	%xmm13,%xmm11
+	pxor	%xmm5,%xmm2
+	pxor	%xmm1,%xmm7
+
+	movdqu	%xmm6,64(%rdi)
+	movdqu	0(%rsi),%xmm6
+	movdqu	%xmm11,80(%rdi)
+	movdqu	16(%rsi),%xmm11
+	movdqu	%xmm2,96(%rdi)
+	movdqu	32(%rsi),%xmm2
+	movdqu	%xmm7,112(%rdi)
+	leaq	128(%rdi),%rdi
+	movdqu	48(%rsi),%xmm7
+	pxor	32(%rsp),%xmm6
+	pxor	%xmm10,%xmm11
+	pxor	%xmm14,%xmm2
+	pxor	%xmm8,%xmm7
+	movdqu	%xmm6,0(%rdi)
+	movdqu	%xmm11,16(%rdi)
+	movdqu	%xmm2,32(%rdi)
+	movdqu	%xmm7,48(%rdi)
+	je	.Ldone4x
+
+	movdqa	48(%rsp),%xmm6
+	leaq	64(%rsi),%rsi
+	xorq	%r10,%r10
+	movdqa	%xmm6,0(%rsp)
+	movdqa	%xmm15,16(%rsp)
+	leaq	64(%rdi),%rdi
+	movdqa	%xmm9,32(%rsp)
+	subq	$192,%rdx
+	movdqa	%xmm3,48(%rsp)
+
+.Loop_tail4x:
+	movzbl	(%rsi,%r10,1),%eax
+	movzbl	(%rsp,%r10,1),%ecx
+	leaq	1(%r10),%r10
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r10,1)
+	decq	%rdx
+	jnz	.Loop_tail4x
+
+.Ldone4x:
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.L4x_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_4x,.-ChaCha20_4x
+.type	ChaCha20_4xop,@function
+.align	32
+ChaCha20_4xop:
+.cfi_startproc	
+.LChaCha20_4xop:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	subq	$0x140+8,%rsp
+	vzeroupper
+
+	vmovdqa	.Lsigma(%rip),%xmm11
+	vmovdqu	(%rcx),%xmm3
+	vmovdqu	16(%rcx),%xmm15
+	vmovdqu	(%r8),%xmm7
+	leaq	256(%rsp),%rcx
+
+	vpshufd	$0x00,%xmm11,%xmm8
+	vpshufd	$0x55,%xmm11,%xmm9
+	vmovdqa	%xmm8,64(%rsp)
+	vpshufd	$0xaa,%xmm11,%xmm10
+	vmovdqa	%xmm9,80(%rsp)
+	vpshufd	$0xff,%xmm11,%xmm11
+	vmovdqa	%xmm10,96(%rsp)
+	vmovdqa	%xmm11,112(%rsp)
+
+	vpshufd	$0x00,%xmm3,%xmm0
+	vpshufd	$0x55,%xmm3,%xmm1
+	vmovdqa	%xmm0,128-256(%rcx)
+	vpshufd	$0xaa,%xmm3,%xmm2
+	vmovdqa	%xmm1,144-256(%rcx)
+	vpshufd	$0xff,%xmm3,%xmm3
+	vmovdqa	%xmm2,160-256(%rcx)
+	vmovdqa	%xmm3,176-256(%rcx)
+
+	vpshufd	$0x00,%xmm15,%xmm12
+	vpshufd	$0x55,%xmm15,%xmm13
+	vmovdqa	%xmm12,192-256(%rcx)
+	vpshufd	$0xaa,%xmm15,%xmm14
+	vmovdqa	%xmm13,208-256(%rcx)
+	vpshufd	$0xff,%xmm15,%xmm15
+	vmovdqa	%xmm14,224-256(%rcx)
+	vmovdqa	%xmm15,240-256(%rcx)
+
+	vpshufd	$0x00,%xmm7,%xmm4
+	vpshufd	$0x55,%xmm7,%xmm5
+	vpaddd	.Linc(%rip),%xmm4,%xmm4
+	vpshufd	$0xaa,%xmm7,%xmm6
+	vmovdqa	%xmm5,272-256(%rcx)
+	vpshufd	$0xff,%xmm7,%xmm7
+	vmovdqa	%xmm6,288-256(%rcx)
+	vmovdqa	%xmm7,304-256(%rcx)
+
+	jmp	.Loop_enter4xop
+
+.align	32
+.Loop_outer4xop:
+	vmovdqa	64(%rsp),%xmm8
+	vmovdqa	80(%rsp),%xmm9
+	vmovdqa	96(%rsp),%xmm10
+	vmovdqa	112(%rsp),%xmm11
+	vmovdqa	128-256(%rcx),%xmm0
+	vmovdqa	144-256(%rcx),%xmm1
+	vmovdqa	160-256(%rcx),%xmm2
+	vmovdqa	176-256(%rcx),%xmm3
+	vmovdqa	192-256(%rcx),%xmm12
+	vmovdqa	208-256(%rcx),%xmm13
+	vmovdqa	224-256(%rcx),%xmm14
+	vmovdqa	240-256(%rcx),%xmm15
+	vmovdqa	256-256(%rcx),%xmm4
+	vmovdqa	272-256(%rcx),%xmm5
+	vmovdqa	288-256(%rcx),%xmm6
+	vmovdqa	304-256(%rcx),%xmm7
+	vpaddd	.Lfour(%rip),%xmm4,%xmm4
+
+.Loop_enter4xop:
+	movl	$10,%eax
+	vmovdqa	%xmm4,256-256(%rcx)
+	jmp	.Loop4xop
+
+.align	32
+.Loop4xop:
+	vpaddd	%xmm0,%xmm8,%xmm8
+	vpaddd	%xmm1,%xmm9,%xmm9
+	vpaddd	%xmm2,%xmm10,%xmm10
+	vpaddd	%xmm3,%xmm11,%xmm11
+	vpxor	%xmm4,%xmm8,%xmm4
+	vpxor	%xmm5,%xmm9,%xmm5
+	vpxor	%xmm6,%xmm10,%xmm6
+	vpxor	%xmm7,%xmm11,%xmm7
+.byte	143,232,120,194,228,16
+.byte	143,232,120,194,237,16
+.byte	143,232,120,194,246,16
+.byte	143,232,120,194,255,16
+	vpaddd	%xmm4,%xmm12,%xmm12
+	vpaddd	%xmm5,%xmm13,%xmm13
+	vpaddd	%xmm6,%xmm14,%xmm14
+	vpaddd	%xmm7,%xmm15,%xmm15
+	vpxor	%xmm0,%xmm12,%xmm0
+	vpxor	%xmm1,%xmm13,%xmm1
+	vpxor	%xmm14,%xmm2,%xmm2
+	vpxor	%xmm15,%xmm3,%xmm3
+.byte	143,232,120,194,192,12
+.byte	143,232,120,194,201,12
+.byte	143,232,120,194,210,12
+.byte	143,232,120,194,219,12
+	vpaddd	%xmm8,%xmm0,%xmm8
+	vpaddd	%xmm9,%xmm1,%xmm9
+	vpaddd	%xmm2,%xmm10,%xmm10
+	vpaddd	%xmm3,%xmm11,%xmm11
+	vpxor	%xmm4,%xmm8,%xmm4
+	vpxor	%xmm5,%xmm9,%xmm5
+	vpxor	%xmm6,%xmm10,%xmm6
+	vpxor	%xmm7,%xmm11,%xmm7
+.byte	143,232,120,194,228,8
+.byte	143,232,120,194,237,8
+.byte	143,232,120,194,246,8
+.byte	143,232,120,194,255,8
+	vpaddd	%xmm4,%xmm12,%xmm12
+	vpaddd	%xmm5,%xmm13,%xmm13
+	vpaddd	%xmm6,%xmm14,%xmm14
+	vpaddd	%xmm7,%xmm15,%xmm15
+	vpxor	%xmm0,%xmm12,%xmm0
+	vpxor	%xmm1,%xmm13,%xmm1
+	vpxor	%xmm14,%xmm2,%xmm2
+	vpxor	%xmm15,%xmm3,%xmm3
+.byte	143,232,120,194,192,7
+.byte	143,232,120,194,201,7
+.byte	143,232,120,194,210,7
+.byte	143,232,120,194,219,7
+	vpaddd	%xmm1,%xmm8,%xmm8
+	vpaddd	%xmm2,%xmm9,%xmm9
+	vpaddd	%xmm3,%xmm10,%xmm10
+	vpaddd	%xmm0,%xmm11,%xmm11
+	vpxor	%xmm7,%xmm8,%xmm7
+	vpxor	%xmm4,%xmm9,%xmm4
+	vpxor	%xmm5,%xmm10,%xmm5
+	vpxor	%xmm6,%xmm11,%xmm6
+.byte	143,232,120,194,255,16
+.byte	143,232,120,194,228,16
+.byte	143,232,120,194,237,16
+.byte	143,232,120,194,246,16
+	vpaddd	%xmm7,%xmm14,%xmm14
+	vpaddd	%xmm4,%xmm15,%xmm15
+	vpaddd	%xmm5,%xmm12,%xmm12
+	vpaddd	%xmm6,%xmm13,%xmm13
+	vpxor	%xmm1,%xmm14,%xmm1
+	vpxor	%xmm2,%xmm15,%xmm2
+	vpxor	%xmm12,%xmm3,%xmm3
+	vpxor	%xmm13,%xmm0,%xmm0
+.byte	143,232,120,194,201,12
+.byte	143,232,120,194,210,12
+.byte	143,232,120,194,219,12
+.byte	143,232,120,194,192,12
+	vpaddd	%xmm8,%xmm1,%xmm8
+	vpaddd	%xmm9,%xmm2,%xmm9
+	vpaddd	%xmm3,%xmm10,%xmm10
+	vpaddd	%xmm0,%xmm11,%xmm11
+	vpxor	%xmm7,%xmm8,%xmm7
+	vpxor	%xmm4,%xmm9,%xmm4
+	vpxor	%xmm5,%xmm10,%xmm5
+	vpxor	%xmm6,%xmm11,%xmm6
+.byte	143,232,120,194,255,8
+.byte	143,232,120,194,228,8
+.byte	143,232,120,194,237,8
+.byte	143,232,120,194,246,8
+	vpaddd	%xmm7,%xmm14,%xmm14
+	vpaddd	%xmm4,%xmm15,%xmm15
+	vpaddd	%xmm5,%xmm12,%xmm12
+	vpaddd	%xmm6,%xmm13,%xmm13
+	vpxor	%xmm1,%xmm14,%xmm1
+	vpxor	%xmm2,%xmm15,%xmm2
+	vpxor	%xmm12,%xmm3,%xmm3
+	vpxor	%xmm13,%xmm0,%xmm0
+.byte	143,232,120,194,201,7
+.byte	143,232,120,194,210,7
+.byte	143,232,120,194,219,7
+.byte	143,232,120,194,192,7
+	decl	%eax
+	jnz	.Loop4xop
+
+	vpaddd	64(%rsp),%xmm8,%xmm8
+	vpaddd	80(%rsp),%xmm9,%xmm9
+	vpaddd	96(%rsp),%xmm10,%xmm10
+	vpaddd	112(%rsp),%xmm11,%xmm11
+
+	vmovdqa	%xmm14,32(%rsp)
+	vmovdqa	%xmm15,48(%rsp)
+
+	vpunpckldq	%xmm9,%xmm8,%xmm14
+	vpunpckldq	%xmm11,%xmm10,%xmm15
+	vpunpckhdq	%xmm9,%xmm8,%xmm8
+	vpunpckhdq	%xmm11,%xmm10,%xmm10
+	vpunpcklqdq	%xmm15,%xmm14,%xmm9
+	vpunpckhqdq	%xmm15,%xmm14,%xmm14
+	vpunpcklqdq	%xmm10,%xmm8,%xmm11
+	vpunpckhqdq	%xmm10,%xmm8,%xmm8
+	vpaddd	128-256(%rcx),%xmm0,%xmm0
+	vpaddd	144-256(%rcx),%xmm1,%xmm1
+	vpaddd	160-256(%rcx),%xmm2,%xmm2
+	vpaddd	176-256(%rcx),%xmm3,%xmm3
+
+	vmovdqa	%xmm9,0(%rsp)
+	vmovdqa	%xmm14,16(%rsp)
+	vmovdqa	32(%rsp),%xmm9
+	vmovdqa	48(%rsp),%xmm14
+
+	vpunpckldq	%xmm1,%xmm0,%xmm10
+	vpunpckldq	%xmm3,%xmm2,%xmm15
+	vpunpckhdq	%xmm1,%xmm0,%xmm0
+	vpunpckhdq	%xmm3,%xmm2,%xmm2
+	vpunpcklqdq	%xmm15,%xmm10,%xmm1
+	vpunpckhqdq	%xmm15,%xmm10,%xmm10
+	vpunpcklqdq	%xmm2,%xmm0,%xmm3
+	vpunpckhqdq	%xmm2,%xmm0,%xmm0
+	vpaddd	192-256(%rcx),%xmm12,%xmm12
+	vpaddd	208-256(%rcx),%xmm13,%xmm13
+	vpaddd	224-256(%rcx),%xmm9,%xmm9
+	vpaddd	240-256(%rcx),%xmm14,%xmm14
+
+	vpunpckldq	%xmm13,%xmm12,%xmm2
+	vpunpckldq	%xmm14,%xmm9,%xmm15
+	vpunpckhdq	%xmm13,%xmm12,%xmm12
+	vpunpckhdq	%xmm14,%xmm9,%xmm9
+	vpunpcklqdq	%xmm15,%xmm2,%xmm13
+	vpunpckhqdq	%xmm15,%xmm2,%xmm2
+	vpunpcklqdq	%xmm9,%xmm12,%xmm14
+	vpunpckhqdq	%xmm9,%xmm12,%xmm12
+	vpaddd	256-256(%rcx),%xmm4,%xmm4
+	vpaddd	272-256(%rcx),%xmm5,%xmm5
+	vpaddd	288-256(%rcx),%xmm6,%xmm6
+	vpaddd	304-256(%rcx),%xmm7,%xmm7
+
+	vpunpckldq	%xmm5,%xmm4,%xmm9
+	vpunpckldq	%xmm7,%xmm6,%xmm15
+	vpunpckhdq	%xmm5,%xmm4,%xmm4
+	vpunpckhdq	%xmm7,%xmm6,%xmm6
+	vpunpcklqdq	%xmm15,%xmm9,%xmm5
+	vpunpckhqdq	%xmm15,%xmm9,%xmm9
+	vpunpcklqdq	%xmm6,%xmm4,%xmm7
+	vpunpckhqdq	%xmm6,%xmm4,%xmm4
+	vmovdqa	0(%rsp),%xmm6
+	vmovdqa	16(%rsp),%xmm15
+
+	cmpq	$256,%rdx
+	jb	.Ltail4xop
+
+	vpxor	0(%rsi),%xmm6,%xmm6
+	vpxor	16(%rsi),%xmm1,%xmm1
+	vpxor	32(%rsi),%xmm13,%xmm13
+	vpxor	48(%rsi),%xmm5,%xmm5
+	vpxor	64(%rsi),%xmm15,%xmm15
+	vpxor	80(%rsi),%xmm10,%xmm10
+	vpxor	96(%rsi),%xmm2,%xmm2
+	vpxor	112(%rsi),%xmm9,%xmm9
+	leaq	128(%rsi),%rsi
+	vpxor	0(%rsi),%xmm11,%xmm11
+	vpxor	16(%rsi),%xmm3,%xmm3
+	vpxor	32(%rsi),%xmm14,%xmm14
+	vpxor	48(%rsi),%xmm7,%xmm7
+	vpxor	64(%rsi),%xmm8,%xmm8
+	vpxor	80(%rsi),%xmm0,%xmm0
+	vpxor	96(%rsi),%xmm12,%xmm12
+	vpxor	112(%rsi),%xmm4,%xmm4
+	leaq	128(%rsi),%rsi
+
+	vmovdqu	%xmm6,0(%rdi)
+	vmovdqu	%xmm1,16(%rdi)
+	vmovdqu	%xmm13,32(%rdi)
+	vmovdqu	%xmm5,48(%rdi)
+	vmovdqu	%xmm15,64(%rdi)
+	vmovdqu	%xmm10,80(%rdi)
+	vmovdqu	%xmm2,96(%rdi)
+	vmovdqu	%xmm9,112(%rdi)
+	leaq	128(%rdi),%rdi
+	vmovdqu	%xmm11,0(%rdi)
+	vmovdqu	%xmm3,16(%rdi)
+	vmovdqu	%xmm14,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	vmovdqu	%xmm8,64(%rdi)
+	vmovdqu	%xmm0,80(%rdi)
+	vmovdqu	%xmm12,96(%rdi)
+	vmovdqu	%xmm4,112(%rdi)
+	leaq	128(%rdi),%rdi
+
+	subq	$256,%rdx
+	jnz	.Loop_outer4xop
+
+	jmp	.Ldone4xop
+
+.align	32
+.Ltail4xop:
+	cmpq	$192,%rdx
+	jae	.L192_or_more4xop
+	cmpq	$128,%rdx
+	jae	.L128_or_more4xop
+	cmpq	$64,%rdx
+	jae	.L64_or_more4xop
+
+	xorq	%r10,%r10
+	vmovdqa	%xmm6,0(%rsp)
+	vmovdqa	%xmm1,16(%rsp)
+	vmovdqa	%xmm13,32(%rsp)
+	vmovdqa	%xmm5,48(%rsp)
+	jmp	.Loop_tail4xop
+
+.align	32
+.L64_or_more4xop:
+	vpxor	0(%rsi),%xmm6,%xmm6
+	vpxor	16(%rsi),%xmm1,%xmm1
+	vpxor	32(%rsi),%xmm13,%xmm13
+	vpxor	48(%rsi),%xmm5,%xmm5
+	vmovdqu	%xmm6,0(%rdi)
+	vmovdqu	%xmm1,16(%rdi)
+	vmovdqu	%xmm13,32(%rdi)
+	vmovdqu	%xmm5,48(%rdi)
+	je	.Ldone4xop
+
+	leaq	64(%rsi),%rsi
+	vmovdqa	%xmm15,0(%rsp)
+	xorq	%r10,%r10
+	vmovdqa	%xmm10,16(%rsp)
+	leaq	64(%rdi),%rdi
+	vmovdqa	%xmm2,32(%rsp)
+	subq	$64,%rdx
+	vmovdqa	%xmm9,48(%rsp)
+	jmp	.Loop_tail4xop
+
+.align	32
+.L128_or_more4xop:
+	vpxor	0(%rsi),%xmm6,%xmm6
+	vpxor	16(%rsi),%xmm1,%xmm1
+	vpxor	32(%rsi),%xmm13,%xmm13
+	vpxor	48(%rsi),%xmm5,%xmm5
+	vpxor	64(%rsi),%xmm15,%xmm15
+	vpxor	80(%rsi),%xmm10,%xmm10
+	vpxor	96(%rsi),%xmm2,%xmm2
+	vpxor	112(%rsi),%xmm9,%xmm9
+
+	vmovdqu	%xmm6,0(%rdi)
+	vmovdqu	%xmm1,16(%rdi)
+	vmovdqu	%xmm13,32(%rdi)
+	vmovdqu	%xmm5,48(%rdi)
+	vmovdqu	%xmm15,64(%rdi)
+	vmovdqu	%xmm10,80(%rdi)
+	vmovdqu	%xmm2,96(%rdi)
+	vmovdqu	%xmm9,112(%rdi)
+	je	.Ldone4xop
+
+	leaq	128(%rsi),%rsi
+	vmovdqa	%xmm11,0(%rsp)
+	xorq	%r10,%r10
+	vmovdqa	%xmm3,16(%rsp)
+	leaq	128(%rdi),%rdi
+	vmovdqa	%xmm14,32(%rsp)
+	subq	$128,%rdx
+	vmovdqa	%xmm7,48(%rsp)
+	jmp	.Loop_tail4xop
+
+.align	32
+.L192_or_more4xop:
+	vpxor	0(%rsi),%xmm6,%xmm6
+	vpxor	16(%rsi),%xmm1,%xmm1
+	vpxor	32(%rsi),%xmm13,%xmm13
+	vpxor	48(%rsi),%xmm5,%xmm5
+	vpxor	64(%rsi),%xmm15,%xmm15
+	vpxor	80(%rsi),%xmm10,%xmm10
+	vpxor	96(%rsi),%xmm2,%xmm2
+	vpxor	112(%rsi),%xmm9,%xmm9
+	leaq	128(%rsi),%rsi
+	vpxor	0(%rsi),%xmm11,%xmm11
+	vpxor	16(%rsi),%xmm3,%xmm3
+	vpxor	32(%rsi),%xmm14,%xmm14
+	vpxor	48(%rsi),%xmm7,%xmm7
+
+	vmovdqu	%xmm6,0(%rdi)
+	vmovdqu	%xmm1,16(%rdi)
+	vmovdqu	%xmm13,32(%rdi)
+	vmovdqu	%xmm5,48(%rdi)
+	vmovdqu	%xmm15,64(%rdi)
+	vmovdqu	%xmm10,80(%rdi)
+	vmovdqu	%xmm2,96(%rdi)
+	vmovdqu	%xmm9,112(%rdi)
+	leaq	128(%rdi),%rdi
+	vmovdqu	%xmm11,0(%rdi)
+	vmovdqu	%xmm3,16(%rdi)
+	vmovdqu	%xmm14,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	je	.Ldone4xop
+
+	leaq	64(%rsi),%rsi
+	vmovdqa	%xmm8,0(%rsp)
+	xorq	%r10,%r10
+	vmovdqa	%xmm0,16(%rsp)
+	leaq	64(%rdi),%rdi
+	vmovdqa	%xmm12,32(%rsp)
+	subq	$192,%rdx
+	vmovdqa	%xmm4,48(%rsp)
+
+.Loop_tail4xop:
+	movzbl	(%rsi,%r10,1),%eax
+	movzbl	(%rsp,%r10,1),%ecx
+	leaq	1(%r10),%r10
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r10,1)
+	decq	%rdx
+	jnz	.Loop_tail4xop
+
+.Ldone4xop:
+	vzeroupper
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.L4xop_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_4xop,.-ChaCha20_4xop
+.type	ChaCha20_8x,@function
+.align	32
+ChaCha20_8x:
+.cfi_startproc	
+.LChaCha20_8x:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	subq	$0x280+8,%rsp
+	andq	$-32,%rsp
+	vzeroupper
+
+
+
+
+
+
+
+
+
+
+	vbroadcasti128	.Lsigma(%rip),%ymm11
+	vbroadcasti128	(%rcx),%ymm3
+	vbroadcasti128	16(%rcx),%ymm15
+	vbroadcasti128	(%r8),%ymm7
+	leaq	256(%rsp),%rcx
+	leaq	512(%rsp),%rax
+	leaq	.Lrot16(%rip),%r10
+	leaq	.Lrot24(%rip),%r11
+
+	vpshufd	$0x00,%ymm11,%ymm8
+	vpshufd	$0x55,%ymm11,%ymm9
+	vmovdqa	%ymm8,128-256(%rcx)
+	vpshufd	$0xaa,%ymm11,%ymm10
+	vmovdqa	%ymm9,160-256(%rcx)
+	vpshufd	$0xff,%ymm11,%ymm11
+	vmovdqa	%ymm10,192-256(%rcx)
+	vmovdqa	%ymm11,224-256(%rcx)
+
+	vpshufd	$0x00,%ymm3,%ymm0
+	vpshufd	$0x55,%ymm3,%ymm1
+	vmovdqa	%ymm0,256-256(%rcx)
+	vpshufd	$0xaa,%ymm3,%ymm2
+	vmovdqa	%ymm1,288-256(%rcx)
+	vpshufd	$0xff,%ymm3,%ymm3
+	vmovdqa	%ymm2,320-256(%rcx)
+	vmovdqa	%ymm3,352-256(%rcx)
+
+	vpshufd	$0x00,%ymm15,%ymm12
+	vpshufd	$0x55,%ymm15,%ymm13
+	vmovdqa	%ymm12,384-512(%rax)
+	vpshufd	$0xaa,%ymm15,%ymm14
+	vmovdqa	%ymm13,416-512(%rax)
+	vpshufd	$0xff,%ymm15,%ymm15
+	vmovdqa	%ymm14,448-512(%rax)
+	vmovdqa	%ymm15,480-512(%rax)
+
+	vpshufd	$0x00,%ymm7,%ymm4
+	vpshufd	$0x55,%ymm7,%ymm5
+	vpaddd	.Lincy(%rip),%ymm4,%ymm4
+	vpshufd	$0xaa,%ymm7,%ymm6
+	vmovdqa	%ymm5,544-512(%rax)
+	vpshufd	$0xff,%ymm7,%ymm7
+	vmovdqa	%ymm6,576-512(%rax)
+	vmovdqa	%ymm7,608-512(%rax)
+
+	jmp	.Loop_enter8x
+
+.align	32
+.Loop_outer8x:
+	vmovdqa	128-256(%rcx),%ymm8
+	vmovdqa	160-256(%rcx),%ymm9
+	vmovdqa	192-256(%rcx),%ymm10
+	vmovdqa	224-256(%rcx),%ymm11
+	vmovdqa	256-256(%rcx),%ymm0
+	vmovdqa	288-256(%rcx),%ymm1
+	vmovdqa	320-256(%rcx),%ymm2
+	vmovdqa	352-256(%rcx),%ymm3
+	vmovdqa	384-512(%rax),%ymm12
+	vmovdqa	416-512(%rax),%ymm13
+	vmovdqa	448-512(%rax),%ymm14
+	vmovdqa	480-512(%rax),%ymm15
+	vmovdqa	512-512(%rax),%ymm4
+	vmovdqa	544-512(%rax),%ymm5
+	vmovdqa	576-512(%rax),%ymm6
+	vmovdqa	608-512(%rax),%ymm7
+	vpaddd	.Leight(%rip),%ymm4,%ymm4
+
+.Loop_enter8x:
+	vmovdqa	%ymm14,64(%rsp)
+	vmovdqa	%ymm15,96(%rsp)
+	vbroadcasti128	(%r10),%ymm15
+	vmovdqa	%ymm4,512-512(%rax)
+	movl	$10,%eax
+	jmp	.Loop8x
+
+.align	32
+.Loop8x:
+	vpaddd	%ymm0,%ymm8,%ymm8
+	vpxor	%ymm4,%ymm8,%ymm4
+	vpshufb	%ymm15,%ymm4,%ymm4
+	vpaddd	%ymm1,%ymm9,%ymm9
+	vpxor	%ymm5,%ymm9,%ymm5
+	vpshufb	%ymm15,%ymm5,%ymm5
+	vpaddd	%ymm4,%ymm12,%ymm12
+	vpxor	%ymm0,%ymm12,%ymm0
+	vpslld	$12,%ymm0,%ymm14
+	vpsrld	$20,%ymm0,%ymm0
+	vpor	%ymm0,%ymm14,%ymm0
+	vbroadcasti128	(%r11),%ymm14
+	vpaddd	%ymm5,%ymm13,%ymm13
+	vpxor	%ymm1,%ymm13,%ymm1
+	vpslld	$12,%ymm1,%ymm15
+	vpsrld	$20,%ymm1,%ymm1
+	vpor	%ymm1,%ymm15,%ymm1
+	vpaddd	%ymm0,%ymm8,%ymm8
+	vpxor	%ymm4,%ymm8,%ymm4
+	vpshufb	%ymm14,%ymm4,%ymm4
+	vpaddd	%ymm1,%ymm9,%ymm9
+	vpxor	%ymm5,%ymm9,%ymm5
+	vpshufb	%ymm14,%ymm5,%ymm5
+	vpaddd	%ymm4,%ymm12,%ymm12
+	vpxor	%ymm0,%ymm12,%ymm0
+	vpslld	$7,%ymm0,%ymm15
+	vpsrld	$25,%ymm0,%ymm0
+	vpor	%ymm0,%ymm15,%ymm0
+	vbroadcasti128	(%r10),%ymm15
+	vpaddd	%ymm5,%ymm13,%ymm13
+	vpxor	%ymm1,%ymm13,%ymm1
+	vpslld	$7,%ymm1,%ymm14
+	vpsrld	$25,%ymm1,%ymm1
+	vpor	%ymm1,%ymm14,%ymm1
+	vmovdqa	%ymm12,0(%rsp)
+	vmovdqa	%ymm13,32(%rsp)
+	vmovdqa	64(%rsp),%ymm12
+	vmovdqa	96(%rsp),%ymm13
+	vpaddd	%ymm2,%ymm10,%ymm10
+	vpxor	%ymm6,%ymm10,%ymm6
+	vpshufb	%ymm15,%ymm6,%ymm6
+	vpaddd	%ymm3,%ymm11,%ymm11
+	vpxor	%ymm7,%ymm11,%ymm7
+	vpshufb	%ymm15,%ymm7,%ymm7
+	vpaddd	%ymm6,%ymm12,%ymm12
+	vpxor	%ymm2,%ymm12,%ymm2
+	vpslld	$12,%ymm2,%ymm14
+	vpsrld	$20,%ymm2,%ymm2
+	vpor	%ymm2,%ymm14,%ymm2
+	vbroadcasti128	(%r11),%ymm14
+	vpaddd	%ymm7,%ymm13,%ymm13
+	vpxor	%ymm3,%ymm13,%ymm3
+	vpslld	$12,%ymm3,%ymm15
+	vpsrld	$20,%ymm3,%ymm3
+	vpor	%ymm3,%ymm15,%ymm3
+	vpaddd	%ymm2,%ymm10,%ymm10
+	vpxor	%ymm6,%ymm10,%ymm6
+	vpshufb	%ymm14,%ymm6,%ymm6
+	vpaddd	%ymm3,%ymm11,%ymm11
+	vpxor	%ymm7,%ymm11,%ymm7
+	vpshufb	%ymm14,%ymm7,%ymm7
+	vpaddd	%ymm6,%ymm12,%ymm12
+	vpxor	%ymm2,%ymm12,%ymm2
+	vpslld	$7,%ymm2,%ymm15
+	vpsrld	$25,%ymm2,%ymm2
+	vpor	%ymm2,%ymm15,%ymm2
+	vbroadcasti128	(%r10),%ymm15
+	vpaddd	%ymm7,%ymm13,%ymm13
+	vpxor	%ymm3,%ymm13,%ymm3
+	vpslld	$7,%ymm3,%ymm14
+	vpsrld	$25,%ymm3,%ymm3
+	vpor	%ymm3,%ymm14,%ymm3
+	vpaddd	%ymm1,%ymm8,%ymm8
+	vpxor	%ymm7,%ymm8,%ymm7
+	vpshufb	%ymm15,%ymm7,%ymm7
+	vpaddd	%ymm2,%ymm9,%ymm9
+	vpxor	%ymm4,%ymm9,%ymm4
+	vpshufb	%ymm15,%ymm4,%ymm4
+	vpaddd	%ymm7,%ymm12,%ymm12
+	vpxor	%ymm1,%ymm12,%ymm1
+	vpslld	$12,%ymm1,%ymm14
+	vpsrld	$20,%ymm1,%ymm1
+	vpor	%ymm1,%ymm14,%ymm1
+	vbroadcasti128	(%r11),%ymm14
+	vpaddd	%ymm4,%ymm13,%ymm13
+	vpxor	%ymm2,%ymm13,%ymm2
+	vpslld	$12,%ymm2,%ymm15
+	vpsrld	$20,%ymm2,%ymm2
+	vpor	%ymm2,%ymm15,%ymm2
+	vpaddd	%ymm1,%ymm8,%ymm8
+	vpxor	%ymm7,%ymm8,%ymm7
+	vpshufb	%ymm14,%ymm7,%ymm7
+	vpaddd	%ymm2,%ymm9,%ymm9
+	vpxor	%ymm4,%ymm9,%ymm4
+	vpshufb	%ymm14,%ymm4,%ymm4
+	vpaddd	%ymm7,%ymm12,%ymm12
+	vpxor	%ymm1,%ymm12,%ymm1
+	vpslld	$7,%ymm1,%ymm15
+	vpsrld	$25,%ymm1,%ymm1
+	vpor	%ymm1,%ymm15,%ymm1
+	vbroadcasti128	(%r10),%ymm15
+	vpaddd	%ymm4,%ymm13,%ymm13
+	vpxor	%ymm2,%ymm13,%ymm2
+	vpslld	$7,%ymm2,%ymm14
+	vpsrld	$25,%ymm2,%ymm2
+	vpor	%ymm2,%ymm14,%ymm2
+	vmovdqa	%ymm12,64(%rsp)
+	vmovdqa	%ymm13,96(%rsp)
+	vmovdqa	0(%rsp),%ymm12
+	vmovdqa	32(%rsp),%ymm13
+	vpaddd	%ymm3,%ymm10,%ymm10
+	vpxor	%ymm5,%ymm10,%ymm5
+	vpshufb	%ymm15,%ymm5,%ymm5
+	vpaddd	%ymm0,%ymm11,%ymm11
+	vpxor	%ymm6,%ymm11,%ymm6
+	vpshufb	%ymm15,%ymm6,%ymm6
+	vpaddd	%ymm5,%ymm12,%ymm12
+	vpxor	%ymm3,%ymm12,%ymm3
+	vpslld	$12,%ymm3,%ymm14
+	vpsrld	$20,%ymm3,%ymm3
+	vpor	%ymm3,%ymm14,%ymm3
+	vbroadcasti128	(%r11),%ymm14
+	vpaddd	%ymm6,%ymm13,%ymm13
+	vpxor	%ymm0,%ymm13,%ymm0
+	vpslld	$12,%ymm0,%ymm15
+	vpsrld	$20,%ymm0,%ymm0
+	vpor	%ymm0,%ymm15,%ymm0
+	vpaddd	%ymm3,%ymm10,%ymm10
+	vpxor	%ymm5,%ymm10,%ymm5
+	vpshufb	%ymm14,%ymm5,%ymm5
+	vpaddd	%ymm0,%ymm11,%ymm11
+	vpxor	%ymm6,%ymm11,%ymm6
+	vpshufb	%ymm14,%ymm6,%ymm6
+	vpaddd	%ymm5,%ymm12,%ymm12
+	vpxor	%ymm3,%ymm12,%ymm3
+	vpslld	$7,%ymm3,%ymm15
+	vpsrld	$25,%ymm3,%ymm3
+	vpor	%ymm3,%ymm15,%ymm3
+	vbroadcasti128	(%r10),%ymm15
+	vpaddd	%ymm6,%ymm13,%ymm13
+	vpxor	%ymm0,%ymm13,%ymm0
+	vpslld	$7,%ymm0,%ymm14
+	vpsrld	$25,%ymm0,%ymm0
+	vpor	%ymm0,%ymm14,%ymm0
+	decl	%eax
+	jnz	.Loop8x
+
+	leaq	512(%rsp),%rax
+	vpaddd	128-256(%rcx),%ymm8,%ymm8
+	vpaddd	160-256(%rcx),%ymm9,%ymm9
+	vpaddd	192-256(%rcx),%ymm10,%ymm10
+	vpaddd	224-256(%rcx),%ymm11,%ymm11
+
+	vpunpckldq	%ymm9,%ymm8,%ymm14
+	vpunpckldq	%ymm11,%ymm10,%ymm15
+	vpunpckhdq	%ymm9,%ymm8,%ymm8
+	vpunpckhdq	%ymm11,%ymm10,%ymm10
+	vpunpcklqdq	%ymm15,%ymm14,%ymm9
+	vpunpckhqdq	%ymm15,%ymm14,%ymm14
+	vpunpcklqdq	%ymm10,%ymm8,%ymm11
+	vpunpckhqdq	%ymm10,%ymm8,%ymm8
+	vpaddd	256-256(%rcx),%ymm0,%ymm0
+	vpaddd	288-256(%rcx),%ymm1,%ymm1
+	vpaddd	320-256(%rcx),%ymm2,%ymm2
+	vpaddd	352-256(%rcx),%ymm3,%ymm3
+
+	vpunpckldq	%ymm1,%ymm0,%ymm10
+	vpunpckldq	%ymm3,%ymm2,%ymm15
+	vpunpckhdq	%ymm1,%ymm0,%ymm0
+	vpunpckhdq	%ymm3,%ymm2,%ymm2
+	vpunpcklqdq	%ymm15,%ymm10,%ymm1
+	vpunpckhqdq	%ymm15,%ymm10,%ymm10
+	vpunpcklqdq	%ymm2,%ymm0,%ymm3
+	vpunpckhqdq	%ymm2,%ymm0,%ymm0
+	vperm2i128	$0x20,%ymm1,%ymm9,%ymm15
+	vperm2i128	$0x31,%ymm1,%ymm9,%ymm1
+	vperm2i128	$0x20,%ymm10,%ymm14,%ymm9
+	vperm2i128	$0x31,%ymm10,%ymm14,%ymm10
+	vperm2i128	$0x20,%ymm3,%ymm11,%ymm14
+	vperm2i128	$0x31,%ymm3,%ymm11,%ymm3
+	vperm2i128	$0x20,%ymm0,%ymm8,%ymm11
+	vperm2i128	$0x31,%ymm0,%ymm8,%ymm0
+	vmovdqa	%ymm15,0(%rsp)
+	vmovdqa	%ymm9,32(%rsp)
+	vmovdqa	64(%rsp),%ymm15
+	vmovdqa	96(%rsp),%ymm9
+
+	vpaddd	384-512(%rax),%ymm12,%ymm12
+	vpaddd	416-512(%rax),%ymm13,%ymm13
+	vpaddd	448-512(%rax),%ymm15,%ymm15
+	vpaddd	480-512(%rax),%ymm9,%ymm9
+
+	vpunpckldq	%ymm13,%ymm12,%ymm2
+	vpunpckldq	%ymm9,%ymm15,%ymm8
+	vpunpckhdq	%ymm13,%ymm12,%ymm12
+	vpunpckhdq	%ymm9,%ymm15,%ymm15
+	vpunpcklqdq	%ymm8,%ymm2,%ymm13
+	vpunpckhqdq	%ymm8,%ymm2,%ymm2
+	vpunpcklqdq	%ymm15,%ymm12,%ymm9
+	vpunpckhqdq	%ymm15,%ymm12,%ymm12
+	vpaddd	512-512(%rax),%ymm4,%ymm4
+	vpaddd	544-512(%rax),%ymm5,%ymm5
+	vpaddd	576-512(%rax),%ymm6,%ymm6
+	vpaddd	608-512(%rax),%ymm7,%ymm7
+
+	vpunpckldq	%ymm5,%ymm4,%ymm15
+	vpunpckldq	%ymm7,%ymm6,%ymm8
+	vpunpckhdq	%ymm5,%ymm4,%ymm4
+	vpunpckhdq	%ymm7,%ymm6,%ymm6
+	vpunpcklqdq	%ymm8,%ymm15,%ymm5
+	vpunpckhqdq	%ymm8,%ymm15,%ymm15
+	vpunpcklqdq	%ymm6,%ymm4,%ymm7
+	vpunpckhqdq	%ymm6,%ymm4,%ymm4
+	vperm2i128	$0x20,%ymm5,%ymm13,%ymm8
+	vperm2i128	$0x31,%ymm5,%ymm13,%ymm5
+	vperm2i128	$0x20,%ymm15,%ymm2,%ymm13
+	vperm2i128	$0x31,%ymm15,%ymm2,%ymm15
+	vperm2i128	$0x20,%ymm7,%ymm9,%ymm2
+	vperm2i128	$0x31,%ymm7,%ymm9,%ymm7
+	vperm2i128	$0x20,%ymm4,%ymm12,%ymm9
+	vperm2i128	$0x31,%ymm4,%ymm12,%ymm4
+	vmovdqa	0(%rsp),%ymm6
+	vmovdqa	32(%rsp),%ymm12
+
+	cmpq	$512,%rdx
+	jb	.Ltail8x
+
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	leaq	128(%rsi),%rsi
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	leaq	128(%rdi),%rdi
+
+	vpxor	0(%rsi),%ymm12,%ymm12
+	vpxor	32(%rsi),%ymm13,%ymm13
+	vpxor	64(%rsi),%ymm10,%ymm10
+	vpxor	96(%rsi),%ymm15,%ymm15
+	leaq	128(%rsi),%rsi
+	vmovdqu	%ymm12,0(%rdi)
+	vmovdqu	%ymm13,32(%rdi)
+	vmovdqu	%ymm10,64(%rdi)
+	vmovdqu	%ymm15,96(%rdi)
+	leaq	128(%rdi),%rdi
+
+	vpxor	0(%rsi),%ymm14,%ymm14
+	vpxor	32(%rsi),%ymm2,%ymm2
+	vpxor	64(%rsi),%ymm3,%ymm3
+	vpxor	96(%rsi),%ymm7,%ymm7
+	leaq	128(%rsi),%rsi
+	vmovdqu	%ymm14,0(%rdi)
+	vmovdqu	%ymm2,32(%rdi)
+	vmovdqu	%ymm3,64(%rdi)
+	vmovdqu	%ymm7,96(%rdi)
+	leaq	128(%rdi),%rdi
+
+	vpxor	0(%rsi),%ymm11,%ymm11
+	vpxor	32(%rsi),%ymm9,%ymm9
+	vpxor	64(%rsi),%ymm0,%ymm0
+	vpxor	96(%rsi),%ymm4,%ymm4
+	leaq	128(%rsi),%rsi
+	vmovdqu	%ymm11,0(%rdi)
+	vmovdqu	%ymm9,32(%rdi)
+	vmovdqu	%ymm0,64(%rdi)
+	vmovdqu	%ymm4,96(%rdi)
+	leaq	128(%rdi),%rdi
+
+	subq	$512,%rdx
+	jnz	.Loop_outer8x
+
+	jmp	.Ldone8x
+
+.Ltail8x:
+	cmpq	$448,%rdx
+	jae	.L448_or_more8x
+	cmpq	$384,%rdx
+	jae	.L384_or_more8x
+	cmpq	$320,%rdx
+	jae	.L320_or_more8x
+	cmpq	$256,%rdx
+	jae	.L256_or_more8x
+	cmpq	$192,%rdx
+	jae	.L192_or_more8x
+	cmpq	$128,%rdx
+	jae	.L128_or_more8x
+	cmpq	$64,%rdx
+	jae	.L64_or_more8x
+
+	xorq	%r10,%r10
+	vmovdqa	%ymm6,0(%rsp)
+	vmovdqa	%ymm8,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L64_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	je	.Ldone8x
+
+	leaq	64(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm1,0(%rsp)
+	leaq	64(%rdi),%rdi
+	subq	$64,%rdx
+	vmovdqa	%ymm5,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L128_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	je	.Ldone8x
+
+	leaq	128(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm12,0(%rsp)
+	leaq	128(%rdi),%rdi
+	subq	$128,%rdx
+	vmovdqa	%ymm13,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L192_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vpxor	128(%rsi),%ymm12,%ymm12
+	vpxor	160(%rsi),%ymm13,%ymm13
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	vmovdqu	%ymm12,128(%rdi)
+	vmovdqu	%ymm13,160(%rdi)
+	je	.Ldone8x
+
+	leaq	192(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm10,0(%rsp)
+	leaq	192(%rdi),%rdi
+	subq	$192,%rdx
+	vmovdqa	%ymm15,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L256_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vpxor	128(%rsi),%ymm12,%ymm12
+	vpxor	160(%rsi),%ymm13,%ymm13
+	vpxor	192(%rsi),%ymm10,%ymm10
+	vpxor	224(%rsi),%ymm15,%ymm15
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	vmovdqu	%ymm12,128(%rdi)
+	vmovdqu	%ymm13,160(%rdi)
+	vmovdqu	%ymm10,192(%rdi)
+	vmovdqu	%ymm15,224(%rdi)
+	je	.Ldone8x
+
+	leaq	256(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm14,0(%rsp)
+	leaq	256(%rdi),%rdi
+	subq	$256,%rdx
+	vmovdqa	%ymm2,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L320_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vpxor	128(%rsi),%ymm12,%ymm12
+	vpxor	160(%rsi),%ymm13,%ymm13
+	vpxor	192(%rsi),%ymm10,%ymm10
+	vpxor	224(%rsi),%ymm15,%ymm15
+	vpxor	256(%rsi),%ymm14,%ymm14
+	vpxor	288(%rsi),%ymm2,%ymm2
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	vmovdqu	%ymm12,128(%rdi)
+	vmovdqu	%ymm13,160(%rdi)
+	vmovdqu	%ymm10,192(%rdi)
+	vmovdqu	%ymm15,224(%rdi)
+	vmovdqu	%ymm14,256(%rdi)
+	vmovdqu	%ymm2,288(%rdi)
+	je	.Ldone8x
+
+	leaq	320(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm3,0(%rsp)
+	leaq	320(%rdi),%rdi
+	subq	$320,%rdx
+	vmovdqa	%ymm7,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L384_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vpxor	128(%rsi),%ymm12,%ymm12
+	vpxor	160(%rsi),%ymm13,%ymm13
+	vpxor	192(%rsi),%ymm10,%ymm10
+	vpxor	224(%rsi),%ymm15,%ymm15
+	vpxor	256(%rsi),%ymm14,%ymm14
+	vpxor	288(%rsi),%ymm2,%ymm2
+	vpxor	320(%rsi),%ymm3,%ymm3
+	vpxor	352(%rsi),%ymm7,%ymm7
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	vmovdqu	%ymm12,128(%rdi)
+	vmovdqu	%ymm13,160(%rdi)
+	vmovdqu	%ymm10,192(%rdi)
+	vmovdqu	%ymm15,224(%rdi)
+	vmovdqu	%ymm14,256(%rdi)
+	vmovdqu	%ymm2,288(%rdi)
+	vmovdqu	%ymm3,320(%rdi)
+	vmovdqu	%ymm7,352(%rdi)
+	je	.Ldone8x
+
+	leaq	384(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm11,0(%rsp)
+	leaq	384(%rdi),%rdi
+	subq	$384,%rdx
+	vmovdqa	%ymm9,32(%rsp)
+	jmp	.Loop_tail8x
+
+.align	32
+.L448_or_more8x:
+	vpxor	0(%rsi),%ymm6,%ymm6
+	vpxor	32(%rsi),%ymm8,%ymm8
+	vpxor	64(%rsi),%ymm1,%ymm1
+	vpxor	96(%rsi),%ymm5,%ymm5
+	vpxor	128(%rsi),%ymm12,%ymm12
+	vpxor	160(%rsi),%ymm13,%ymm13
+	vpxor	192(%rsi),%ymm10,%ymm10
+	vpxor	224(%rsi),%ymm15,%ymm15
+	vpxor	256(%rsi),%ymm14,%ymm14
+	vpxor	288(%rsi),%ymm2,%ymm2
+	vpxor	320(%rsi),%ymm3,%ymm3
+	vpxor	352(%rsi),%ymm7,%ymm7
+	vpxor	384(%rsi),%ymm11,%ymm11
+	vpxor	416(%rsi),%ymm9,%ymm9
+	vmovdqu	%ymm6,0(%rdi)
+	vmovdqu	%ymm8,32(%rdi)
+	vmovdqu	%ymm1,64(%rdi)
+	vmovdqu	%ymm5,96(%rdi)
+	vmovdqu	%ymm12,128(%rdi)
+	vmovdqu	%ymm13,160(%rdi)
+	vmovdqu	%ymm10,192(%rdi)
+	vmovdqu	%ymm15,224(%rdi)
+	vmovdqu	%ymm14,256(%rdi)
+	vmovdqu	%ymm2,288(%rdi)
+	vmovdqu	%ymm3,320(%rdi)
+	vmovdqu	%ymm7,352(%rdi)
+	vmovdqu	%ymm11,384(%rdi)
+	vmovdqu	%ymm9,416(%rdi)
+	je	.Ldone8x
+
+	leaq	448(%rsi),%rsi
+	xorq	%r10,%r10
+	vmovdqa	%ymm0,0(%rsp)
+	leaq	448(%rdi),%rdi
+	subq	$448,%rdx
+	vmovdqa	%ymm4,32(%rsp)
+
+.Loop_tail8x:
+	movzbl	(%rsi,%r10,1),%eax
+	movzbl	(%rsp,%r10,1),%ecx
+	leaq	1(%r10),%r10
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r10,1)
+	decq	%rdx
+	jnz	.Loop_tail8x
+
+.Ldone8x:
+	vzeroall
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.L8x_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_8x,.-ChaCha20_8x
+.type	ChaCha20_avx512,@function
+.align	32
+ChaCha20_avx512:
+.cfi_startproc	
+.LChaCha20_avx512:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	cmpq	$512,%rdx
+	ja	.LChaCha20_16x
+
+	subq	$64+8,%rsp
+	vbroadcasti32x4	.Lsigma(%rip),%zmm0
+	vbroadcasti32x4	(%rcx),%zmm1
+	vbroadcasti32x4	16(%rcx),%zmm2
+	vbroadcasti32x4	(%r8),%zmm3
+
+	vmovdqa32	%zmm0,%zmm16
+	vmovdqa32	%zmm1,%zmm17
+	vmovdqa32	%zmm2,%zmm18
+	vpaddd	.Lzeroz(%rip),%zmm3,%zmm3
+	vmovdqa32	.Lfourz(%rip),%zmm20
+	movq	$10,%r8
+	vmovdqa32	%zmm3,%zmm19
+	jmp	.Loop_avx512
+
+.align	16
+.Loop_outer_avx512:
+	vmovdqa32	%zmm16,%zmm0
+	vmovdqa32	%zmm17,%zmm1
+	vmovdqa32	%zmm18,%zmm2
+	vpaddd	%zmm20,%zmm19,%zmm3
+	movq	$10,%r8
+	vmovdqa32	%zmm3,%zmm19
+	jmp	.Loop_avx512
+
+.align	32
+.Loop_avx512:
+	vpaddd	%zmm1,%zmm0,%zmm0
+	vpxord	%zmm0,%zmm3,%zmm3
+	vprold	$16,%zmm3,%zmm3
+	vpaddd	%zmm3,%zmm2,%zmm2
+	vpxord	%zmm2,%zmm1,%zmm1
+	vprold	$12,%zmm1,%zmm1
+	vpaddd	%zmm1,%zmm0,%zmm0
+	vpxord	%zmm0,%zmm3,%zmm3
+	vprold	$8,%zmm3,%zmm3
+	vpaddd	%zmm3,%zmm2,%zmm2
+	vpxord	%zmm2,%zmm1,%zmm1
+	vprold	$7,%zmm1,%zmm1
+	vpshufd	$78,%zmm2,%zmm2
+	vpshufd	$57,%zmm1,%zmm1
+	vpshufd	$147,%zmm3,%zmm3
+	vpaddd	%zmm1,%zmm0,%zmm0
+	vpxord	%zmm0,%zmm3,%zmm3
+	vprold	$16,%zmm3,%zmm3
+	vpaddd	%zmm3,%zmm2,%zmm2
+	vpxord	%zmm2,%zmm1,%zmm1
+	vprold	$12,%zmm1,%zmm1
+	vpaddd	%zmm1,%zmm0,%zmm0
+	vpxord	%zmm0,%zmm3,%zmm3
+	vprold	$8,%zmm3,%zmm3
+	vpaddd	%zmm3,%zmm2,%zmm2
+	vpxord	%zmm2,%zmm1,%zmm1
+	vprold	$7,%zmm1,%zmm1
+	vpshufd	$78,%zmm2,%zmm2
+	vpshufd	$147,%zmm1,%zmm1
+	vpshufd	$57,%zmm3,%zmm3
+	decq	%r8
+	jnz	.Loop_avx512
+	vpaddd	%zmm16,%zmm0,%zmm0
+	vpaddd	%zmm17,%zmm1,%zmm1
+	vpaddd	%zmm18,%zmm2,%zmm2
+	vpaddd	%zmm19,%zmm3,%zmm3
+
+	subq	$64,%rdx
+	jb	.Ltail64_avx512
+
+	vpxor	0(%rsi),%xmm0,%xmm4
+	vpxor	16(%rsi),%xmm1,%xmm5
+	vpxor	32(%rsi),%xmm2,%xmm6
+	vpxor	48(%rsi),%xmm3,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	jz	.Ldone_avx512
+
+	vextracti32x4	$1,%zmm0,%xmm4
+	vextracti32x4	$1,%zmm1,%xmm5
+	vextracti32x4	$1,%zmm2,%xmm6
+	vextracti32x4	$1,%zmm3,%xmm7
+
+	subq	$64,%rdx
+	jb	.Ltail_avx512
+
+	vpxor	0(%rsi),%xmm4,%xmm4
+	vpxor	16(%rsi),%xmm5,%xmm5
+	vpxor	32(%rsi),%xmm6,%xmm6
+	vpxor	48(%rsi),%xmm7,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	jz	.Ldone_avx512
+
+	vextracti32x4	$2,%zmm0,%xmm4
+	vextracti32x4	$2,%zmm1,%xmm5
+	vextracti32x4	$2,%zmm2,%xmm6
+	vextracti32x4	$2,%zmm3,%xmm7
+
+	subq	$64,%rdx
+	jb	.Ltail_avx512
+
+	vpxor	0(%rsi),%xmm4,%xmm4
+	vpxor	16(%rsi),%xmm5,%xmm5
+	vpxor	32(%rsi),%xmm6,%xmm6
+	vpxor	48(%rsi),%xmm7,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	jz	.Ldone_avx512
+
+	vextracti32x4	$3,%zmm0,%xmm4
+	vextracti32x4	$3,%zmm1,%xmm5
+	vextracti32x4	$3,%zmm2,%xmm6
+	vextracti32x4	$3,%zmm3,%xmm7
+
+	subq	$64,%rdx
+	jb	.Ltail_avx512
+
+	vpxor	0(%rsi),%xmm4,%xmm4
+	vpxor	16(%rsi),%xmm5,%xmm5
+	vpxor	32(%rsi),%xmm6,%xmm6
+	vpxor	48(%rsi),%xmm7,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	jnz	.Loop_outer_avx512
+
+	jmp	.Ldone_avx512
+
+.align	16
+.Ltail64_avx512:
+	vmovdqa	%xmm0,0(%rsp)
+	vmovdqa	%xmm1,16(%rsp)
+	vmovdqa	%xmm2,32(%rsp)
+	vmovdqa	%xmm3,48(%rsp)
+	addq	$64,%rdx
+	jmp	.Loop_tail_avx512
+
+.align	16
+.Ltail_avx512:
+	vmovdqa	%xmm4,0(%rsp)
+	vmovdqa	%xmm5,16(%rsp)
+	vmovdqa	%xmm6,32(%rsp)
+	vmovdqa	%xmm7,48(%rsp)
+	addq	$64,%rdx
+
+.Loop_tail_avx512:
+	movzbl	(%rsi,%r8,1),%eax
+	movzbl	(%rsp,%r8,1),%ecx
+	leaq	1(%r8),%r8
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r8,1)
+	decq	%rdx
+	jnz	.Loop_tail_avx512
+
+	vmovdqu32	%zmm16,0(%rsp)
+
+.Ldone_avx512:
+	vzeroall
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.Lavx512_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_avx512,.-ChaCha20_avx512
+.type	ChaCha20_avx512vl,@function
+.align	32
+ChaCha20_avx512vl:
+.cfi_startproc	
+.LChaCha20_avx512vl:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	cmpq	$128,%rdx
+	ja	.LChaCha20_8xvl
+
+	subq	$64+8,%rsp
+	vbroadcasti128	.Lsigma(%rip),%ymm0
+	vbroadcasti128	(%rcx),%ymm1
+	vbroadcasti128	16(%rcx),%ymm2
+	vbroadcasti128	(%r8),%ymm3
+
+	vmovdqa32	%ymm0,%ymm16
+	vmovdqa32	%ymm1,%ymm17
+	vmovdqa32	%ymm2,%ymm18
+	vpaddd	.Lzeroz(%rip),%ymm3,%ymm3
+	vmovdqa32	.Ltwoy(%rip),%ymm20
+	movq	$10,%r8
+	vmovdqa32	%ymm3,%ymm19
+	jmp	.Loop_avx512vl
+
+.align	16
+.Loop_outer_avx512vl:
+	vmovdqa32	%ymm18,%ymm2
+	vpaddd	%ymm20,%ymm19,%ymm3
+	movq	$10,%r8
+	vmovdqa32	%ymm3,%ymm19
+	jmp	.Loop_avx512vl
+
+.align	32
+.Loop_avx512vl:
+	vpaddd	%ymm1,%ymm0,%ymm0
+	vpxor	%ymm0,%ymm3,%ymm3
+	vprold	$16,%ymm3,%ymm3
+	vpaddd	%ymm3,%ymm2,%ymm2
+	vpxor	%ymm2,%ymm1,%ymm1
+	vprold	$12,%ymm1,%ymm1
+	vpaddd	%ymm1,%ymm0,%ymm0
+	vpxor	%ymm0,%ymm3,%ymm3
+	vprold	$8,%ymm3,%ymm3
+	vpaddd	%ymm3,%ymm2,%ymm2
+	vpxor	%ymm2,%ymm1,%ymm1
+	vprold	$7,%ymm1,%ymm1
+	vpshufd	$78,%ymm2,%ymm2
+	vpshufd	$57,%ymm1,%ymm1
+	vpshufd	$147,%ymm3,%ymm3
+	vpaddd	%ymm1,%ymm0,%ymm0
+	vpxor	%ymm0,%ymm3,%ymm3
+	vprold	$16,%ymm3,%ymm3
+	vpaddd	%ymm3,%ymm2,%ymm2
+	vpxor	%ymm2,%ymm1,%ymm1
+	vprold	$12,%ymm1,%ymm1
+	vpaddd	%ymm1,%ymm0,%ymm0
+	vpxor	%ymm0,%ymm3,%ymm3
+	vprold	$8,%ymm3,%ymm3
+	vpaddd	%ymm3,%ymm2,%ymm2
+	vpxor	%ymm2,%ymm1,%ymm1
+	vprold	$7,%ymm1,%ymm1
+	vpshufd	$78,%ymm2,%ymm2
+	vpshufd	$147,%ymm1,%ymm1
+	vpshufd	$57,%ymm3,%ymm3
+	decq	%r8
+	jnz	.Loop_avx512vl
+	vpaddd	%ymm16,%ymm0,%ymm0
+	vpaddd	%ymm17,%ymm1,%ymm1
+	vpaddd	%ymm18,%ymm2,%ymm2
+	vpaddd	%ymm19,%ymm3,%ymm3
+
+	subq	$64,%rdx
+	jb	.Ltail64_avx512vl
+
+	vpxor	0(%rsi),%xmm0,%xmm4
+	vpxor	16(%rsi),%xmm1,%xmm5
+	vpxor	32(%rsi),%xmm2,%xmm6
+	vpxor	48(%rsi),%xmm3,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	jz	.Ldone_avx512vl
+
+	vextracti128	$1,%ymm0,%xmm4
+	vextracti128	$1,%ymm1,%xmm5
+	vextracti128	$1,%ymm2,%xmm6
+	vextracti128	$1,%ymm3,%xmm7
+
+	subq	$64,%rdx
+	jb	.Ltail_avx512vl
+
+	vpxor	0(%rsi),%xmm4,%xmm4
+	vpxor	16(%rsi),%xmm5,%xmm5
+	vpxor	32(%rsi),%xmm6,%xmm6
+	vpxor	48(%rsi),%xmm7,%xmm7
+	leaq	64(%rsi),%rsi
+
+	vmovdqu	%xmm4,0(%rdi)
+	vmovdqu	%xmm5,16(%rdi)
+	vmovdqu	%xmm6,32(%rdi)
+	vmovdqu	%xmm7,48(%rdi)
+	leaq	64(%rdi),%rdi
+
+	vmovdqa32	%ymm16,%ymm0
+	vmovdqa32	%ymm17,%ymm1
+	jnz	.Loop_outer_avx512vl
+
+	jmp	.Ldone_avx512vl
+
+.align	16
+.Ltail64_avx512vl:
+	vmovdqa	%xmm0,0(%rsp)
+	vmovdqa	%xmm1,16(%rsp)
+	vmovdqa	%xmm2,32(%rsp)
+	vmovdqa	%xmm3,48(%rsp)
+	addq	$64,%rdx
+	jmp	.Loop_tail_avx512vl
+
+.align	16
+.Ltail_avx512vl:
+	vmovdqa	%xmm4,0(%rsp)
+	vmovdqa	%xmm5,16(%rsp)
+	vmovdqa	%xmm6,32(%rsp)
+	vmovdqa	%xmm7,48(%rsp)
+	addq	$64,%rdx
+
+.Loop_tail_avx512vl:
+	movzbl	(%rsi,%r8,1),%eax
+	movzbl	(%rsp,%r8,1),%ecx
+	leaq	1(%r8),%r8
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r8,1)
+	decq	%rdx
+	jnz	.Loop_tail_avx512vl
+
+	vmovdqu32	%ymm16,0(%rsp)
+	vmovdqu32	%ymm16,32(%rsp)
+
+.Ldone_avx512vl:
+	vzeroall
+	leaq	(%r9),%rsp
+.cfi_def_cfa_register	%rsp
+.Lavx512vl_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	ChaCha20_avx512vl,.-ChaCha20_avx512vl
+.type	ChaCha20_16x,@function
+.align	32
+ChaCha20_16x:
+.cfi_startproc	
+.LChaCha20_16x:
+	movq	%rsp,%r9
+.cfi_def_cfa_register	%r9
+	subq	$64+8,%rsp
+	andq	$-64,%rsp
+	vzeroupper
+
+	leaq	.Lsigma(%rip),%r10
+	vbroadcasti32x4	(%r10),%zmm3
+	vbroadcasti32x4	(%rcx),%zmm7
+	vbroadcasti32x4	16(%rcx),%zmm11
+	vbroadcasti32x4	(%r8),%zmm15
+
+	vpshufd	$0x00,%zmm3,%zmm0
+	vpshufd	$0x55,%zmm3,%zmm1
+	vpshufd	$0xaa,%zmm3,%zmm2
+	vpshufd	$0xff,%zmm3,%zmm3
+	vmovdqa64	%zmm0,%zmm16
+	vmovdqa64	%zmm1,%zmm17
+	vmovdqa64	%zmm2,%zmm18
+	vmovdqa64	%zmm3,%zmm19
+
+	vpshufd	$0x00,%zmm7,%zmm4
+	vpshufd	$0x55,%zmm7,%zmm5
+	vpshufd	$0xaa,%zmm7,%zmm6
+	vpshufd	$0xff,%zmm7,%zmm7
+	vmovdqa64	%zmm4,%zmm20
+	vmovdqa64	%zmm5,%zmm21
+	vmovdqa64	%zmm6,%zmm22
+	vmovdqa64	%zmm7,%zmm23
+
+	vpshufd	$0x00,%zmm11,%zmm8
+	vpshufd	$0x55,%zmm11,%zmm9
+	vpshufd	$0xaa,%zmm11,%zmm10
+	vpshufd	$0xff,%zmm11,%zmm11
+	vmovdqa64	%zmm8,%zmm24
+	vmovdqa64	%zmm9,%zmm25
+	vmovdqa64	%zmm10,%zmm26
+	vmovdqa64	%zmm11,%zmm27
+
+	vpshufd	$0x00,%zmm15,%zmm12
+	vpshufd	$0x55,%zmm15,%zmm13
+	vpshufd	$0xaa,%zmm15,%zmm14
+	vpshufd	$0xff,%zmm15,%zmm15
+	vpaddd	.Lincz(%rip),%zmm12,%zmm12
+	vmovdqa64	%zmm12,%zmm28
+	vmovdqa64	%zmm13,%zmm29
+	vmovdqa64	%zmm14,%zmm30
+	vmovdqa64	%zmm15,%zmm31
+
+	movl	$10,%eax
+	jmp	.Loop16x
+
+.align	32
+.Loop_outer16x:
+	vpbroadcastd	0(%r10),%zmm0
+	vpbroadcastd	4(%r10),%zmm1
+	vpbroadcastd	8(%r10),%zmm2
+	vpbroadcastd	12(%r10),%zmm3
+	vpaddd	.Lsixteen(%rip),%zmm28,%zmm28
+	vmovdqa64	%zmm20,%zmm4
+	vmovdqa64	%zmm21,%zmm5
+	vmovdqa64	%zmm22,%zmm6
+	vmovdqa64	%zmm23,%zmm7
+	vmovdqa64	%zmm24,%zmm8
+	vmovdqa64	%zmm25,%zmm9
+	vmovdqa64	%zmm26,%zmm10
+	vmovdqa64	%zmm27,%zmm11
+	vmovdqa64	%zmm28,%zmm12
+	vmovdqa64	%zmm29,%zmm13
+	vmovdqa64	%zmm30,%zmm14
+	vmovdqa64	%zmm31,%zmm15
+
+	vmovdqa64	%zmm0,%zmm16
+	vmovdqa64	%zmm1,%zmm17
+	vmovdqa64	%zmm2,%zmm18
+	vmovdqa64	%zmm3,%zmm19
+
+	movl	$10,%eax
+	jmp	.Loop16x
+
+.align	32
+.Loop16x:
+	vpaddd	%zmm4,%zmm0,%zmm0
+	vpaddd	%zmm5,%zmm1,%zmm1
+	vpaddd	%zmm6,%zmm2,%zmm2
+	vpaddd	%zmm7,%zmm3,%zmm3
+	vpxord	%zmm0,%zmm12,%zmm12
+	vpxord	%zmm1,%zmm13,%zmm13
+	vpxord	%zmm2,%zmm14,%zmm14
+	vpxord	%zmm3,%zmm15,%zmm15
+	vprold	$16,%zmm12,%zmm12
+	vprold	$16,%zmm13,%zmm13
+	vprold	$16,%zmm14,%zmm14
+	vprold	$16,%zmm15,%zmm15
+	vpaddd	%zmm12,%zmm8,%zmm8
+	vpaddd	%zmm13,%zmm9,%zmm9
+	vpaddd	%zmm14,%zmm10,%zmm10
+	vpaddd	%zmm15,%zmm11,%zmm11
+	vpxord	%zmm8,%zmm4,%zmm4
+	vpxord	%zmm9,%zmm5,%zmm5
+	vpxord	%zmm10,%zmm6,%zmm6
+	vpxord	%zmm11,%zmm7,%zmm7
+	vprold	$12,%zmm4,%zmm4
+	vprold	$12,%zmm5,%zmm5
+	vprold	$12,%zmm6,%zmm6
+	vprold	$12,%zmm7,%zmm7
+	vpaddd	%zmm4,%zmm0,%zmm0
+	vpaddd	%zmm5,%zmm1,%zmm1
+	vpaddd	%zmm6,%zmm2,%zmm2
+	vpaddd	%zmm7,%zmm3,%zmm3
+	vpxord	%zmm0,%zmm12,%zmm12
+	vpxord	%zmm1,%zmm13,%zmm13
+	vpxord	%zmm2,%zmm14,%zmm14
+	vpxord	%zmm3,%zmm15,%zmm15
+	vprold	$8,%zmm12,%zmm12
+	vprold	$8,%zmm13,%zmm13
+	vprold	$8,%zmm14,%zmm14
+	vprold	$8,%zmm15,%zmm15
+	vpaddd	%zmm12,%zmm8,%zmm8
+	vpaddd	%zmm13,%zmm9,%zmm9
+	vpaddd	%zmm14,%zmm10,%zmm10
+	vpaddd	%zmm15,%zmm11,%zmm11
+	vpxord	%zmm8,%zmm4,%zmm4
+	vpxord	%zmm9,%zmm5,%zmm5
+	vpxord	%zmm10,%zmm6,%zmm6
+	vpxord	%zmm11,%zmm7,%zmm7
+	vprold	$7,%zmm4,%zmm4
+	vprold	$7,%zmm5,%zmm5
+	vprold	$7,%zmm6,%zmm6
+	vprold	$7,%zmm7,%zmm7
+	vpaddd	%zmm5,%zmm0,%zmm0
+	vpaddd	%zmm6,%zmm1,%zmm1
+	vpaddd	%zmm7,%zmm2,%zmm2
+	vpaddd	%zmm4,%zmm3,%zmm3
+	vpxord	%zmm0,%zmm15,%zmm15
+	vpxord	%zmm1,%zmm12,%zmm12
+	vpxord	%zmm2,%zmm13,%zmm13
+	vpxord	%zmm3,%zmm14,%zmm14
+	vprold	$16,%zmm15,%zmm15
+	vprold	$16,%zmm12,%zmm12
+	vprold	$16,%zmm13,%zmm13
+	vprold	$16,%zmm14,%zmm14
+	vpaddd	%zmm15,%zmm10,%zmm10
+	vpaddd	%zmm12,%zmm11,%zmm11
+	vpaddd	%zmm13,%zmm8,%zmm8
+	vpaddd	%zmm14,%zmm9,%zmm9
+	vpxord	%zmm10,%zmm5,%zmm5
+	vpxord	%zmm11,%zmm6,%zmm6
+	vpxord	%zmm8,%zmm7,%zmm7
+	vpxord	%zmm9,%zmm4,%zmm4
+	vprold	$12,%zmm5,%zmm5
+	vprold	$12,%zmm6,%zmm6
+	vprold	$12,%zmm7,%zmm7
+	vprold	$12,%zmm4,%zmm4
+	vpaddd	%zmm5,%zmm0,%zmm0
+	vpaddd	%zmm6,%zmm1,%zmm1
+	vpaddd	%zmm7,%zmm2,%zmm2
+	vpaddd	%zmm4,%zmm3,%zmm3
+	vpxord	%zmm0,%zmm15,%zmm15
+	vpxord	%zmm1,%zmm12,%zmm12
+	vpxord	%zmm2,%zmm13,%zmm13
+	vpxord	%zmm3,%zmm14,%zmm14
+	vprold	$8,%zmm15,%zmm15
+	vprold	$8,%zmm12,%zmm12
+	vprold	$8,%zmm13,%zmm13
+	vprold	$8,%zmm14,%zmm14
+	vpaddd	%zmm15,%zmm10,%zmm10
+	vpaddd	%zmm12,%zmm11,%zmm11
+	vpaddd	%zmm13,%zmm8,%zmm8
+	vpaddd	%zmm14,%zmm9,%zmm9
+	vpxord	%zmm10,%zmm5,%zmm5
+	vpxord	%zmm11,%zmm6,%zmm6
+	vpxord	%zmm8,%zmm7,%zmm7
+	vpxord	%zmm9,%zmm4,%zmm4
+	vprold	$7,%zmm5,%zmm5
+	vprold	$7,%zmm6,%zmm6
+	vprold	$7,%zmm7,%zmm7
+	vprold	$7,%zmm4,%zmm4
+	decl	%eax
+	jnz	.Loop16x
+
+	vpaddd	%zmm16,%zmm0,%zmm0
+	vpaddd	%zmm17,%zmm1,%zmm1
+	vpaddd	%zmm18,%zmm2,%zmm2
+	vpaddd	%zmm19,%zmm3,%zmm3
+
+	vpunpckldq	%zmm1,%zmm0,%zmm18
+	vpunpckldq	%zmm3,%zmm2,%zmm19
+	vpunpckhdq	%zmm1,%zmm0,%zmm0
+	vpunpckhdq	%zmm3,%zmm2,%zmm2
+	vpunpcklqdq	%zmm19,%zmm18,%zmm1
+	vpunpckhqdq	%zmm19,%zmm18,%zmm18
+	vpunpcklqdq	%zmm2,%zmm0,%zmm3
+	vpunpckhqdq	%zmm2,%zmm0,%zmm0
+	vpaddd	%zmm20,%zmm4,%zmm4
+	vpaddd	%zmm21,%zmm5,%zmm5
+	vpaddd	%zmm22,%zmm6,%zmm6
+	vpaddd	%zmm23,%zmm7,%zmm7
+
+	vpunpckldq	%zmm5,%zmm4,%zmm2
+	vpunpckldq	%zmm7,%zmm6,%zmm19
+	vpunpckhdq	%zmm5,%zmm4,%zmm4
+	vpunpckhdq	%zmm7,%zmm6,%zmm6
+	vpunpcklqdq	%zmm19,%zmm2,%zmm5
+	vpunpckhqdq	%zmm19,%zmm2,%zmm2
+	vpunpcklqdq	%zmm6,%zmm4,%zmm7
+	vpunpckhqdq	%zmm6,%zmm4,%zmm4
+	vshufi32x4	$0x44,%zmm5,%zmm1,%zmm19
+	vshufi32x4	$0xee,%zmm5,%zmm1,%zmm5
+	vshufi32x4	$0x44,%zmm2,%zmm18,%zmm1
+	vshufi32x4	$0xee,%zmm2,%zmm18,%zmm2
+	vshufi32x4	$0x44,%zmm7,%zmm3,%zmm18
+	vshufi32x4	$0xee,%zmm7,%zmm3,%zmm7
+	vshufi32x4	$0x44,%zmm4,%zmm0,%zmm3
+	vshufi32x4	$0xee,%zmm4,%zmm0,%zmm4
+	vpaddd	%zmm24,%zmm8,%zmm8
+	vpaddd	%zmm25,%zmm9,%zmm9
+	vpaddd	%zmm26,%zmm10,%zmm10
+	vpaddd	%zmm27,%zmm11,%zmm11
+
+	vpunpckldq	%zmm9,%zmm8,%zmm6
+	vpunpckldq	%zmm11,%zmm10,%zmm0
+	vpunpckhdq	%zmm9,%zmm8,%zmm8
+	vpunpckhdq	%zmm11,%zmm10,%zmm10
+	vpunpcklqdq	%zmm0,%zmm6,%zmm9
+	vpunpckhqdq	%zmm0,%zmm6,%zmm6
+	vpunpcklqdq	%zmm10,%zmm8,%zmm11
+	vpunpckhqdq	%zmm10,%zmm8,%zmm8
+	vpaddd	%zmm28,%zmm12,%zmm12
+	vpaddd	%zmm29,%zmm13,%zmm13
+	vpaddd	%zmm30,%zmm14,%zmm14
+	vpaddd	%zmm31,%zmm15,%zmm15
+
+	vpunpckldq	%zmm13,%zmm12,%zmm10
+	vpunpckldq	%zmm15,%zmm14,%zmm0
+	vpunpckhdq	%zmm13,%zmm12,%zmm12
+	vpunpckhdq	%zmm15,%zmm14,%zmm14
+	vpunpcklqdq	%zmm0,%zmm10,%zmm13
+	vpunpckhqdq	%zmm0,%zmm10,%zmm10
+	vpunpcklqdq	%zmm14,%zmm12,%zmm15
+	vpunpckhqdq	%zmm14,%zmm12,%zmm12
+	vshufi32x4	$0x44,%zmm13,%zmm9,%zmm0
+	vshufi32x4	$0xee,%zmm13,%zmm9,%zmm13
+	vshufi32x4	$0x44,%zmm10,%zmm6,%zmm9
+	vshufi32x4	$0xee,%zmm10,%zmm6,%zmm10
+	vshufi32x4	$0x44,%zmm15,%zmm11,%zmm6
+	vshufi32x4	$0xee,%zmm15,%zmm11,%zmm15
+	vshufi32x4	$0x44,%zmm12,%zmm8,%zmm11
+	vshufi32x4	$0xee,%zmm12,%zmm8,%zmm12
+	vshufi32x4	$0x88,%zmm0,%zmm19,%zmm16
+	vshufi32x4	$0xdd,%zmm0,%zmm19,%zmm19
+	vshufi32x4	$0x88,%zmm13,%zmm5,%zmm0
+	vshufi32x4	$0xdd,%zmm13,%zmm5,%zmm13
+	vshufi32x4	$0x88,%zmm9,%zmm1,%zmm17
+	vshufi32x4	$0xdd,%zmm9,%zmm1,%zmm1
+	vshufi32x4	$0x88,%zmm10,%zmm2,%zmm9
+	vshufi32x4	$0xdd,%zmm10,%zmm2,%zmm10
+	vshufi32x4	$0x88,%zmm6,%zmm18,%zmm14
+	vshufi32x4	$0xdd,%zmm6,%zmm18,%zmm18
+	vshufi32x4	$0x88,%zmm15,%zmm7,%zmm6
+	vshufi32x4	$0xdd,%zmm15,%zmm7,%zmm15
+	vshufi32x4	$0x88,%zmm11,%zmm3,%zmm8
+	vshufi32x4	$0xdd,%zmm11,%zmm3,%zmm3
+	vshufi32x4	$0x88,%zmm12,%zmm4,%zmm11
+	vshufi32x4	$0xdd,%zmm12,%zmm4,%zmm12
+	cmpq	$1024,%rdx
+	jb	.Ltail16x
+
+	vpxord	0(%rsi),%zmm16,%zmm16
+	vpxord	64(%rsi),%zmm17,%zmm17
+	vpxord	128(%rsi),%zmm14,%zmm14
+	vpxord	192(%rsi),%zmm8,%zmm8
+	vmovdqu32	%zmm16,0(%rdi)
+	vmovdqu32	%zmm17,64(%rdi)
+	vmovdqu32	%zmm14,128(%rdi)
+	vmovdqu32	%zmm8,192(%rdi)
+
+	vpxord	256(%rsi),%zmm19,%zmm19
+	vpxord	320(%rsi),%zmm1,%zmm1
+	vpxord	384(%rsi),%zmm18,%zmm18
+	vpxord	448(%rsi),%zmm3,%zmm3
+	vmovdqu32	%zmm19,256(%rdi)
+	vmovdqu32	%zmm1,320(%rdi)
+	vmovdqu32	%zmm18,384(%rdi)
+	vmovdqu32	%zmm3,448(%rdi)
+
+	vpxord	512(%rsi),%zmm0,%zmm0
+	vpxord	576(%rsi),%zmm9,%zmm9
+	vpxord	640(%rsi),%zmm6,%zmm6
+	vpxord	704(%rsi),%zmm11,%zmm11
+	vmovdqu32	%zmm0,512(%rdi)
+	vmovdqu32	%zmm9,576(%rdi)
+	vmovdqu32	%zmm6,640(%rdi)
+	vmovdqu32	%zmm11,704(%rdi)
+
+	vpxord	768(%rsi),%zmm13,%zmm13
+	vpxord	832(%rsi),%zmm10,%zmm10
+	vpxord	896(%rsi),%zmm15,%zmm15
+	vpxord	960(%rsi),%zmm12,%zmm12
+	leaq	1024(%rsi),%rsi
+	vmovdqu32	%zmm13,768(%rdi)
+	vmovdqu32	%zmm10,832(%rdi)
+	vmovdqu32	%zmm15,896(%rdi)
+	vmovdqu32	%zmm12,960(%rdi)
+	leaq	1024(%rdi),%rdi
+
+	subq	$1024,%rdx
+	jnz	.Loop_outer16x
+
+	jmp	.Ldone16x
+
+.align	32
+.Ltail16x:
+	xorq	%r10,%r10
+	subq	%rsi,%rdi
+	cmpq	$64,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm16,%zmm16
+	vmovdqu32	%zmm16,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm17,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$128,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm17,%zmm17
+	vmovdqu32	%zmm17,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm14,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$192,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm14,%zmm14
+	vmovdqu32	%zmm14,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm8,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$256,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm8,%zmm8
+	vmovdqu32	%zmm8,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm19,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$320,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm19,%zmm19
+	vmovdqu32	%zmm19,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm1,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$384,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm1,%zmm1
+	vmovdqu32	%zmm1,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm18,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$448,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm18,%zmm18
+	vmovdqu32	%zmm18,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm3,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$512,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm3,%zmm3
+	vmovdqu32	%zmm3,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm0,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$576,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm0,%zmm0
+	vmovdqu32	%zmm0,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm9,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$640,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm9,%zmm9
+	vmovdqu32	%zmm9,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm6,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$704,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm6,%zmm6
+	vmovdqu32	%zmm6,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm11,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$768,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm11,%zmm11
+	vmovdqu32	%zmm11,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm13,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$832,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm13,%zmm13
+	vmovdqu32	%zmm13,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm10,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$896,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm10,%zmm10
+	vmovdqu32	%zmm10,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm15,%zmm16
+	leaq	64(%rsi),%rsi
+
+	cmpq	$960,%rdx
+	jb	.Less_than_64_16x
+	vpxord	(%rsi),%zmm15,%zmm15
+	vmovdqu32	%zmm15,(%rdi,%rsi,1)
+	je	.Ldone16x
+	vmovdqa32	%zmm12,%zmm16
+	leaq	64(%rsi),%rsi
+
+.Less_than_64_16x:
+	vmovdqa32	%zmm16,0(%rsp)
+	leaq	(%rdi,%rsi,1),%rdi
+	andq	$63,%rdx
+
+.Loop_tail16x:
+	movzbl	(%rsi,%r10,1),%eax
+	movzbl	(%rsp,%r10,1),%ecx
+	leaq	1(%r10),%r10
+	xorl	%ecx,%eax
+	movb	%al,-1(%rdi,%r10,1)
[… more lines in the raw message]

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#6 lore raw

PATCH v8 6/28 zinc: ChaCha20 x86_64 implementation

In reply to Jason A. Donenfeld

This ports SSSE3, AVX-2, AVX-512F, and AVX-512VL implementations for ChaCha20. The AVX-512F implementation is disabled on Skylake, due to throttling, and the VL ymm implementation is used instead. These come from Andy Polyakov's implementation, with the following modifications from Samuel Neves:

- Some cosmetic changes, like renaming labels to .Lname, constants, and other Linux conventions.

- CPU feature checking is done in C by the glue code, so that has been removed from the assembly.

- Eliminate translating certain instructions, such as pshufb, palignr, vprotd, etc, to .byte directives. This is meant for compatibility with ancient toolchains, but presumably it is unnecessary here, since the build system already does checks on what GNU as can assemble.

- When aligning the stack, the original code was saving %rsp to %r9.

    To keep objtool happy, we use instead the DRAP idiom to save %rsp
    to %r10:

      leaq    8(%rsp),%r10
      ... code here ...
      leaq    -8(%r10),%rsp

- The original code assumes the stack comes aligned to 16 bytes. This is not necessarily the case, and to avoid crashes,
andq $-alignment, %rsp was added in the prolog of a few functions.

- The original hardcodes returns as .byte 0xf3,0xc3, aka "rep ret". We replace this by "ret". "rep ret" was meant to help with AMD K8 chips, cf. http://repzret.org/p/repzret. It makes no sense to continue to use this kludge for code that won't even run on ancient AMD chips.

Cycle counts on a Core i7 6700HQ using the AVX-2 codepath, comparing this implementation ("new") to the implementation in the current crypto api ("old"):

size old new
---- ---- ----
0 62 52
16 414 376
32 410 400
48 414 422
64 362 356
80 714 666
96 714 700
112 712 718
128 692 646
144 1042 674
160 1042 694
176 1042 726
192 1018 650
208 1366 686
224 1366 696
240 1366 722
256 640 656
272 988 1246
288 988 1276
304 992 1296
320 972 1222
336 1318 1256
352 1318 1276
368 1316 1294
384 1294 1218
400 1642 1258
416 1642 1282
432 1642 1302
448 1628 1224
464 1970 1258
480 1970 1280
496 1970 1300
512 656 676
528 1010 1290
544 1010 1306
560 1010 1332
576 986 1254
592 1340 1284
608 1334 1310
624 1340 1334
640 1314 1254
656 1664 1282
672 1674 1306
688 1662 1336
704 1638 1250
720 1992 1292
736 1994 1308
752 1988 1334
768 1252 1254
784 1596 1290
800 1596 1314
816 1596 1330
832 1576 1256
848 1922 1286
864 1922 1314
880 1926 1338
896 1898 1258
912 2248 1288
928 2248 1320
944 2248 1338
960 2226 1268
976 2574 1288
992 2576 1312
1008 2574 1340

Cycle counts on a Xeon Gold 5120 using the AVX-512 codepath:

size old new
---- ---- ----
0 64 54
16 386 372
32 388 396
48 388 420
64 366 350
80 708 666
96 708 692
112 706 736
128 692 648
144 1036 682
160 1036 708
176 1036 730
192 1016 658
208 1360 684
224 1362 708
240 1360 732
256 644 500
272 990 526
288 988 556
304 988 576
320 972 500
336 1314 532
352 1316 558
368 1318 578
384 1308 506
400 1644 532
416 1644 556
432 1644 594
448 1624 508
464 1970 534
480 1970 556
496 1968 582
512 660 624
528 1016 682
544 1016 702
560 1018 728
576 998 654
592 1344 680
608 1344 708
624 1344 730
640 1326 654
656 1670 686
672 1670 708
688 1670 732
704 1652 658
720 1998 682
736 1998 710
752 1996 734
768 1256 662
784 1606 688
800 1606 714
816 1606 736
832 1584 660
848 1948 688
864 1950 714
880 1948 736
896 1912 688
912 2258 718
928 2258 744
944 2256 768
960 2238 692
976 2584 718
992 2584 744
1008 2584 770

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                             |    1 +
 lib/zinc/chacha20/chacha20-x86_64-glue.c      |  103 ++
 ...-x86_64-cryptogams.S => chacha20-x86_64.S} | 1557 ++++-------------
 lib/zinc/chacha20/chacha20.c                  |    4 +
 4 files changed, 486 insertions(+), 1179 deletions(-)
 create mode 100644 lib/zinc/chacha20/chacha20-x86_64-glue.c
 rename lib/zinc/chacha20/{chacha20-x86_64-cryptogams.S => chacha20-x86_64.S} (71%)
lib/zinc/Makefile +1 -0
index 3d80144d55a6..223a0816c918 100644
@@ -3,4 +3,5 @@ ccflags-y += -D'pr_fmt(fmt)="zinc: " fmt'
 ccflags-$(CONFIG_ZINC_DEBUG) += -DDEBUG
 
 zinc_chacha20-y := chacha20/chacha20.o
+zinc_chacha20-$(CONFIG_ZINC_ARCH_X86_64) += chacha20/chacha20-x86_64.o
 obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
lib/zinc/chacha20/chacha20-x86_64-glue.c new file +103 -0

103 lines not on this page · raw message

lib/zinc/chacha20/chacha20-x86_64.S renamed from lib/zinc/chacha20/chacha20-x86_64-cryptogams.S +378 -1179

1,941 lines not on this page · raw message

lib/zinc/chacha20/chacha20.c +4 -0
index 03209c15d1ca..22a21431c221 100644
@@ -16,6 +16,9 @@
 #include <linux/vmalloc.h>
 #include <crypto/algapi.h> // For crypto_xor_cpy.
 
+#if defined(CONFIG_ZINC_ARCH_X86_64)
+#include "chacha20-x86_64-glue.c"
+#else
 static bool *const chacha20_nobs[] __initconst = { };
 static void __init chacha20_fpu_init(void)
 {
@@ -33,6 +36,7 @@ static inline bool hchacha20_arch(u32 derived_key[CHACHA20_KEY_WORDS],
 {
 	return false;
 }
+#endif
 
 #define QUARTER_ROUND(x, a, b, c, d) ( \
 	x[a] += x[b], \
Signature · 1 line
2.19.1
author maintainer
#7 lore raw

PATCH v8 7/28 zinc: import Andy Polyakov's ChaCha20 ARM and ARM64 implementations

In reply to Jason A. Donenfeld

These NEON and non-NEON implementations come from Andy Polyakov's implementation, and are included here in raw form without modification, so that subsequent commits that fix these up for the kernel can see how it has changed.

While this is CRYPTOGAMS code, the originating code for this happens to be the same as OpenSSL's commit 87cc649f30aaf69b351701875b9dac07c29ce8a2

Based-on-code-from: Andy Polyakov <[email protected]>
Cc: Andy Polyakov <[email protected]>
Cc: Russell King <[email protected]>
Cc: [email protected]
Cc: Samuel Neves <[email protected]>
Cc: Jean-Philippe Aumasson <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 lib/zinc/chacha20/chacha20-arm-cryptogams.S   | 1440 ++++++++++++
 lib/zinc/chacha20/chacha20-arm64-cryptogams.S | 1973 +++++++++++++++++
 2 files changed, 3413 insertions(+)
 create mode 100644 lib/zinc/chacha20/chacha20-arm-cryptogams.S
 create mode 100644 lib/zinc/chacha20/chacha20-arm64-cryptogams.S
lib/zinc/chacha20/chacha20-arm-cryptogams.S new file +1440 -0
new file mode 100644
index 000000000000..05a3a9e6e93f
@@ -0,0 +1,1440 @@
+/* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
+/*
+ * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ */
+
+#include "arm_arch.h"
+
+.text
+#if defined(__thumb2__) || defined(__clang__)
+.syntax	unified
+#endif
+#if defined(__thumb2__)
+.thumb
+#else
+.code	32
+#endif
+
+#if defined(__thumb2__) || defined(__clang__)
+#define ldrhsb	ldrbhs
+#endif
+
+.align	5
+.Lsigma:
+.long	0x61707865,0x3320646e,0x79622d32,0x6b206574	@ endian-neutral
+.Lone:
+.long	1,0,0,0
+.Lrot8:
+.long	0x02010003,0x06050407
+#if __ARM_MAX_ARCH__>=7
+.LOPENSSL_armcap:
+.word   OPENSSL_armcap_P-.LChaCha20_ctr32
+#else
+.word	-1
+#endif
+
+.globl	ChaCha20_ctr32
+.type	ChaCha20_ctr32,%function
+.align	5
+ChaCha20_ctr32:
+.LChaCha20_ctr32:
+	ldr	r12,[sp,#0]		@ pull pointer to counter and nonce
+	stmdb	sp!,{r0-r2,r4-r11,lr}
+#if __ARM_ARCH__<7 && !defined(__thumb2__)
+	sub	r14,pc,#16		@ ChaCha20_ctr32
+#else
+	adr	r14,.LChaCha20_ctr32
+#endif
+	cmp	r2,#0			@ len==0?
+#ifdef	__thumb2__
+	itt	eq
+#endif
+	addeq	sp,sp,#4*3
+	beq	.Lno_data
+#if __ARM_MAX_ARCH__>=7
+	cmp	r2,#192			@ test len
+	bls	.Lshort
+	ldr	r4,[r14,#-24]
+	ldr	r4,[r14,r4]
+# ifdef	__APPLE__
+	ldr	r4,[r4]
+# endif
+	tst	r4,#ARMV7_NEON
+	bne	.LChaCha20_neon
+.Lshort:
+#endif
+	ldmia	r12,{r4-r7}		@ load counter and nonce
+	sub	sp,sp,#4*(16)		@ off-load area
+	sub	r14,r14,#64		@ .Lsigma
+	stmdb	sp!,{r4-r7}		@ copy counter and nonce
+	ldmia	r3,{r4-r11}		@ load key
+	ldmia	r14,{r0-r3}		@ load sigma
+	stmdb	sp!,{r4-r11}		@ copy key
+	stmdb	sp!,{r0-r3}		@ copy sigma
+	str	r10,[sp,#4*(16+10)]	@ off-load "rx"
+	str	r11,[sp,#4*(16+11)]	@ off-load "rx"
+	b	.Loop_outer_enter
+
+.align	4
+.Loop_outer:
+	ldmia	sp,{r0-r9}		@ load key material
+	str	r11,[sp,#4*(32+2)]	@ save len
+	str	r12,  [sp,#4*(32+1)]	@ save inp
+	str	r14,  [sp,#4*(32+0)]	@ save out
+.Loop_outer_enter:
+	ldr	r11, [sp,#4*(15)]
+	 mov	r4,r4,ror#19	@ twist b[0..3]
+	ldr	r12,[sp,#4*(12)]	@ modulo-scheduled load
+	 mov	r5,r5,ror#19
+	ldr	r10, [sp,#4*(13)]
+	 mov	r6,r6,ror#19
+	ldr	r14,[sp,#4*(14)]
+	 mov	r7,r7,ror#19
+	mov	r11,r11,ror#8	@ twist d[0..3]
+	mov	r12,r12,ror#8
+	mov	r10,r10,ror#8
+	mov	r14,r14,ror#8
+	str	r11, [sp,#4*(16+15)]
+	mov	r11,#10
+	b	.Loop
+
+.align	4
+.Loop:
+	subs	r11,r11,#1
+	add	r0,r0,r4,ror#13
+	add	r1,r1,r5,ror#13
+	eor	r12,r0,r12,ror#24
+	eor	r10,r1,r10,ror#24
+	add	r8,r8,r12,ror#16
+	add	r9,r9,r10,ror#16
+	eor	r4,r8,r4,ror#13
+	eor	r5,r9,r5,ror#13
+	add	r0,r0,r4,ror#20
+	add	r1,r1,r5,ror#20
+	eor	r12,r0,r12,ror#16
+	eor	r10,r1,r10,ror#16
+	add	r8,r8,r12,ror#24
+	str	r10,[sp,#4*(16+13)]
+	add	r9,r9,r10,ror#24
+	ldr	r10,[sp,#4*(16+15)]
+	str	r8,[sp,#4*(16+8)]
+	eor	r4,r4,r8,ror#12
+	str	r9,[sp,#4*(16+9)]
+	eor	r5,r5,r9,ror#12
+	ldr	r8,[sp,#4*(16+10)]
+	add	r2,r2,r6,ror#13
+	ldr	r9,[sp,#4*(16+11)]
+	add	r3,r3,r7,ror#13
+	eor	r14,r2,r14,ror#24
+	eor	r10,r3,r10,ror#24
+	add	r8,r8,r14,ror#16
+	add	r9,r9,r10,ror#16
+	eor	r6,r8,r6,ror#13
+	eor	r7,r9,r7,ror#13
+	add	r2,r2,r6,ror#20
+	add	r3,r3,r7,ror#20
+	eor	r14,r2,r14,ror#16
+	eor	r10,r3,r10,ror#16
+	add	r8,r8,r14,ror#24
+	add	r9,r9,r10,ror#24
+	eor	r6,r6,r8,ror#12
+	eor	r7,r7,r9,ror#12
+	add	r0,r0,r5,ror#13
+	add	r1,r1,r6,ror#13
+	eor	r10,r0,r10,ror#24
+	eor	r12,r1,r12,ror#24
+	add	r8,r8,r10,ror#16
+	add	r9,r9,r12,ror#16
+	eor	r5,r8,r5,ror#13
+	eor	r6,r9,r6,ror#13
+	add	r0,r0,r5,ror#20
+	add	r1,r1,r6,ror#20
+	eor	r10,r0,r10,ror#16
+	eor	r12,r1,r12,ror#16
+	str	r10,[sp,#4*(16+15)]
+	add	r8,r8,r10,ror#24
+	ldr	r10,[sp,#4*(16+13)]
+	add	r9,r9,r12,ror#24
+	str	r8,[sp,#4*(16+10)]
+	eor	r5,r5,r8,ror#12
+	str	r9,[sp,#4*(16+11)]
+	eor	r6,r6,r9,ror#12
+	ldr	r8,[sp,#4*(16+8)]
+	add	r2,r2,r7,ror#13
+	ldr	r9,[sp,#4*(16+9)]
+	add	r3,r3,r4,ror#13
+	eor	r10,r2,r10,ror#24
+	eor	r14,r3,r14,ror#24
+	add	r8,r8,r10,ror#16
+	add	r9,r9,r14,ror#16
+	eor	r7,r8,r7,ror#13
+	eor	r4,r9,r4,ror#13
+	add	r2,r2,r7,ror#20
+	add	r3,r3,r4,ror#20
+	eor	r10,r2,r10,ror#16
+	eor	r14,r3,r14,ror#16
+	add	r8,r8,r10,ror#24
+	add	r9,r9,r14,ror#24
+	eor	r7,r7,r8,ror#12
+	eor	r4,r4,r9,ror#12
+	bne	.Loop
+
+	ldr	r11,[sp,#4*(32+2)]	@ load len
+
+	str	r8, [sp,#4*(16+8)]	@ modulo-scheduled store
+	str	r9, [sp,#4*(16+9)]
+	str	r12,[sp,#4*(16+12)]
+	str	r10, [sp,#4*(16+13)]
+	str	r14,[sp,#4*(16+14)]
+
+	@ at this point we have first half of 512-bit result in
+	@ rx and second half at sp+4*(16+8)
+
+	cmp	r11,#64		@ done yet?
+#ifdef	__thumb2__
+	itete	lo
+#endif
+	addlo	r12,sp,#4*(0)		@ shortcut or ...
+	ldrhs	r12,[sp,#4*(32+1)]	@ ... load inp
+	addlo	r14,sp,#4*(0)		@ shortcut or ...
+	ldrhs	r14,[sp,#4*(32+0)]	@ ... load out
+
+	ldr	r8,[sp,#4*(0)]	@ load key material
+	ldr	r9,[sp,#4*(1)]
+
+#if __ARM_ARCH__>=6 || !defined(__ARMEB__)
+# if __ARM_ARCH__<7
+	orr	r10,r12,r14
+	tst	r10,#3		@ are input and output aligned?
+	ldr	r10,[sp,#4*(2)]
+	bne	.Lunaligned
+	cmp	r11,#64		@ restore flags
+# else
+	ldr	r10,[sp,#4*(2)]
+# endif
+	ldr	r11,[sp,#4*(3)]
+
+	add	r0,r0,r8	@ accumulate key material
+	add	r1,r1,r9
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r8,[r12],#16		@ load input
+	ldrhs	r9,[r12,#-12]
+
+	add	r2,r2,r10
+	add	r3,r3,r11
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r10,[r12,#-8]
+	ldrhs	r11,[r12,#-4]
+# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+# endif
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r0,r0,r8	@ xor with input
+	eorhs	r1,r1,r9
+	 add	r8,sp,#4*(4)
+	str	r0,[r14],#16		@ store output
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r2,r2,r10
+	eorhs	r3,r3,r11
+	 ldmia	r8,{r8-r11}	@ load key material
+	str	r1,[r14,#-12]
+	str	r2,[r14,#-8]
+	str	r3,[r14,#-4]
+
+	add	r4,r8,r4,ror#13 @ accumulate key material
+	add	r5,r9,r5,ror#13
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r8,[r12],#16		@ load input
+	ldrhs	r9,[r12,#-12]
+	add	r6,r10,r6,ror#13
+	add	r7,r11,r7,ror#13
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r10,[r12,#-8]
+	ldrhs	r11,[r12,#-4]
+# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+	rev	r4,r4
+	rev	r5,r5
+	rev	r6,r6
+	rev	r7,r7
+# endif
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r4,r4,r8
+	eorhs	r5,r5,r9
+	 add	r8,sp,#4*(8)
+	str	r4,[r14],#16		@ store output
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r6,r6,r10
+	eorhs	r7,r7,r11
+	str	r5,[r14,#-12]
+	 ldmia	r8,{r8-r11}	@ load key material
+	str	r6,[r14,#-8]
+	 add	r0,sp,#4*(16+8)
+	str	r7,[r14,#-4]
+
+	ldmia	r0,{r0-r7}	@ load second half
+
+	add	r0,r0,r8	@ accumulate key material
+	add	r1,r1,r9
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r8,[r12],#16		@ load input
+	ldrhs	r9,[r12,#-12]
+# ifdef	__thumb2__
+	itt	hi
+# endif
+	 strhi	r10,[sp,#4*(16+10)]	@ copy "rx" while at it
+	 strhi	r11,[sp,#4*(16+11)]	@ copy "rx" while at it
+	add	r2,r2,r10
+	add	r3,r3,r11
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r10,[r12,#-8]
+	ldrhs	r11,[r12,#-4]
+# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+# endif
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r0,r0,r8
+	eorhs	r1,r1,r9
+	 add	r8,sp,#4*(12)
+	str	r0,[r14],#16		@ store output
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r2,r2,r10
+	eorhs	r3,r3,r11
+	str	r1,[r14,#-12]
+	 ldmia	r8,{r8-r11}	@ load key material
+	str	r2,[r14,#-8]
+	str	r3,[r14,#-4]
+
+	add	r4,r8,r4,ror#24 @ accumulate key material
+	add	r5,r9,r5,ror#24
+# ifdef	__thumb2__
+	itt	hi
+# endif
+	 addhi	r8,r8,#1		@ next counter value
+	 strhi	r8,[sp,#4*(12)]	@ save next counter value
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r8,[r12],#16		@ load input
+	ldrhs	r9,[r12,#-12]
+	add	r6,r10,r6,ror#24
+	add	r7,r11,r7,ror#24
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhs	r10,[r12,#-8]
+	ldrhs	r11,[r12,#-4]
+# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+	rev	r4,r4
+	rev	r5,r5
+	rev	r6,r6
+	rev	r7,r7
+# endif
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r4,r4,r8
+	eorhs	r5,r5,r9
+# ifdef	__thumb2__
+	 it	ne
+# endif
+	 ldrne	r8,[sp,#4*(32+2)]	@ re-load len
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	eorhs	r6,r6,r10
+	eorhs	r7,r7,r11
+	str	r4,[r14],#16		@ store output
+	str	r5,[r14,#-12]
+# ifdef	__thumb2__
+	it	hs
+# endif
+	 subhs	r11,r8,#64		@ len-=64
+	str	r6,[r14,#-8]
+	str	r7,[r14,#-4]
+	bhi	.Loop_outer
+
+	beq	.Ldone
+# if __ARM_ARCH__<7
+	b	.Ltail
+
+.align	4
+.Lunaligned:				@ unaligned endian-neutral path
+	cmp	r11,#64		@ restore flags
+# endif
+#endif
+#if __ARM_ARCH__<7
+	ldr	r11,[sp,#4*(3)]
+	add	r0,r8,r0	@ accumulate key material
+	add	r1,r9,r1
+	add	r2,r10,r2
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r8,r8,r8		@ zero or ...
+	ldrhsb	r8,[r12],#16			@ ... load input
+	eorlo	r9,r9,r9
+	ldrhsb	r9,[r12,#-12]
+
+	add	r3,r11,r3
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r10,r10,r10
+	ldrhsb	r10,[r12,#-8]
+	eorlo	r11,r11,r11
+	ldrhsb	r11,[r12,#-4]
+
+	eor	r0,r8,r0		@ xor with input (or zero)
+	eor	r1,r9,r1
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-15]		@ load more input
+	ldrhsb	r9,[r12,#-11]
+	eor	r2,r10,r2
+	 strb	r0,[r14],#16		@ store output
+	eor	r3,r11,r3
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-7]
+	ldrhsb	r11,[r12,#-3]
+	 strb	r1,[r14,#-12]
+	eor	r0,r8,r0,lsr#8
+	 strb	r2,[r14,#-8]
+	eor	r1,r9,r1,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-14]		@ load more input
+	ldrhsb	r9,[r12,#-10]
+	 strb	r3,[r14,#-4]
+	eor	r2,r10,r2,lsr#8
+	 strb	r0,[r14,#-15]
+	eor	r3,r11,r3,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-6]
+	ldrhsb	r11,[r12,#-2]
+	 strb	r1,[r14,#-11]
+	eor	r0,r8,r0,lsr#8
+	 strb	r2,[r14,#-7]
+	eor	r1,r9,r1,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-13]		@ load more input
+	ldrhsb	r9,[r12,#-9]
+	 strb	r3,[r14,#-3]
+	eor	r2,r10,r2,lsr#8
+	 strb	r0,[r14,#-14]
+	eor	r3,r11,r3,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-5]
+	ldrhsb	r11,[r12,#-1]
+	 strb	r1,[r14,#-10]
+	 strb	r2,[r14,#-6]
+	eor	r0,r8,r0,lsr#8
+	 strb	r3,[r14,#-2]
+	eor	r1,r9,r1,lsr#8
+	 strb	r0,[r14,#-13]
+	eor	r2,r10,r2,lsr#8
+	 strb	r1,[r14,#-9]
+	eor	r3,r11,r3,lsr#8
+	 strb	r2,[r14,#-5]
+	 strb	r3,[r14,#-1]
+	add	r8,sp,#4*(4+0)
+	ldmia	r8,{r8-r11}		@ load key material
+	add	r0,sp,#4*(16+8)
+	add	r4,r8,r4,ror#13	@ accumulate key material
+	add	r5,r9,r5,ror#13
+	add	r6,r10,r6,ror#13
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r8,r8,r8		@ zero or ...
+	ldrhsb	r8,[r12],#16			@ ... load input
+	eorlo	r9,r9,r9
+	ldrhsb	r9,[r12,#-12]
+
+	add	r7,r11,r7,ror#13
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r10,r10,r10
+	ldrhsb	r10,[r12,#-8]
+	eorlo	r11,r11,r11
+	ldrhsb	r11,[r12,#-4]
+
+	eor	r4,r8,r4		@ xor with input (or zero)
+	eor	r5,r9,r5
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-15]		@ load more input
+	ldrhsb	r9,[r12,#-11]
+	eor	r6,r10,r6
+	 strb	r4,[r14],#16		@ store output
+	eor	r7,r11,r7
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-7]
+	ldrhsb	r11,[r12,#-3]
+	 strb	r5,[r14,#-12]
+	eor	r4,r8,r4,lsr#8
+	 strb	r6,[r14,#-8]
+	eor	r5,r9,r5,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-14]		@ load more input
+	ldrhsb	r9,[r12,#-10]
+	 strb	r7,[r14,#-4]
+	eor	r6,r10,r6,lsr#8
+	 strb	r4,[r14,#-15]
+	eor	r7,r11,r7,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-6]
+	ldrhsb	r11,[r12,#-2]
+	 strb	r5,[r14,#-11]
+	eor	r4,r8,r4,lsr#8
+	 strb	r6,[r14,#-7]
+	eor	r5,r9,r5,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-13]		@ load more input
+	ldrhsb	r9,[r12,#-9]
+	 strb	r7,[r14,#-3]
+	eor	r6,r10,r6,lsr#8
+	 strb	r4,[r14,#-14]
+	eor	r7,r11,r7,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-5]
+	ldrhsb	r11,[r12,#-1]
+	 strb	r5,[r14,#-10]
+	 strb	r6,[r14,#-6]
+	eor	r4,r8,r4,lsr#8
+	 strb	r7,[r14,#-2]
+	eor	r5,r9,r5,lsr#8
+	 strb	r4,[r14,#-13]
+	eor	r6,r10,r6,lsr#8
+	 strb	r5,[r14,#-9]
+	eor	r7,r11,r7,lsr#8
+	 strb	r6,[r14,#-5]
+	 strb	r7,[r14,#-1]
+	add	r8,sp,#4*(4+4)
+	ldmia	r8,{r8-r11}		@ load key material
+	ldmia	r0,{r0-r7}		@ load second half
+# ifdef	__thumb2__
+	itt	hi
+# endif
+	strhi	r10,[sp,#4*(16+10)]		@ copy "rx"
+	strhi	r11,[sp,#4*(16+11)]		@ copy "rx"
+	add	r0,r8,r0	@ accumulate key material
+	add	r1,r9,r1
+	add	r2,r10,r2
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r8,r8,r8		@ zero or ...
+	ldrhsb	r8,[r12],#16			@ ... load input
+	eorlo	r9,r9,r9
+	ldrhsb	r9,[r12,#-12]
+
+	add	r3,r11,r3
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r10,r10,r10
+	ldrhsb	r10,[r12,#-8]
+	eorlo	r11,r11,r11
+	ldrhsb	r11,[r12,#-4]
+
+	eor	r0,r8,r0		@ xor with input (or zero)
+	eor	r1,r9,r1
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-15]		@ load more input
+	ldrhsb	r9,[r12,#-11]
+	eor	r2,r10,r2
+	 strb	r0,[r14],#16		@ store output
+	eor	r3,r11,r3
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-7]
+	ldrhsb	r11,[r12,#-3]
+	 strb	r1,[r14,#-12]
+	eor	r0,r8,r0,lsr#8
+	 strb	r2,[r14,#-8]
+	eor	r1,r9,r1,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-14]		@ load more input
+	ldrhsb	r9,[r12,#-10]
+	 strb	r3,[r14,#-4]
+	eor	r2,r10,r2,lsr#8
+	 strb	r0,[r14,#-15]
+	eor	r3,r11,r3,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-6]
+	ldrhsb	r11,[r12,#-2]
+	 strb	r1,[r14,#-11]
+	eor	r0,r8,r0,lsr#8
+	 strb	r2,[r14,#-7]
+	eor	r1,r9,r1,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-13]		@ load more input
+	ldrhsb	r9,[r12,#-9]
+	 strb	r3,[r14,#-3]
+	eor	r2,r10,r2,lsr#8
+	 strb	r0,[r14,#-14]
+	eor	r3,r11,r3,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-5]
+	ldrhsb	r11,[r12,#-1]
+	 strb	r1,[r14,#-10]
+	 strb	r2,[r14,#-6]
+	eor	r0,r8,r0,lsr#8
+	 strb	r3,[r14,#-2]
+	eor	r1,r9,r1,lsr#8
+	 strb	r0,[r14,#-13]
+	eor	r2,r10,r2,lsr#8
+	 strb	r1,[r14,#-9]
+	eor	r3,r11,r3,lsr#8
+	 strb	r2,[r14,#-5]
+	 strb	r3,[r14,#-1]
+	add	r8,sp,#4*(4+8)
+	ldmia	r8,{r8-r11}		@ load key material
+	add	r4,r8,r4,ror#24	@ accumulate key material
+# ifdef	__thumb2__
+	itt	hi
+# endif
+	addhi	r8,r8,#1			@ next counter value
+	strhi	r8,[sp,#4*(12)]		@ save next counter value
+	add	r5,r9,r5,ror#24
+	add	r6,r10,r6,ror#24
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r8,r8,r8		@ zero or ...
+	ldrhsb	r8,[r12],#16			@ ... load input
+	eorlo	r9,r9,r9
+	ldrhsb	r9,[r12,#-12]
+
+	add	r7,r11,r7,ror#24
+# ifdef	__thumb2__
+	itete	lo
+# endif
+	eorlo	r10,r10,r10
+	ldrhsb	r10,[r12,#-8]
+	eorlo	r11,r11,r11
+	ldrhsb	r11,[r12,#-4]
+
+	eor	r4,r8,r4		@ xor with input (or zero)
+	eor	r5,r9,r5
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-15]		@ load more input
+	ldrhsb	r9,[r12,#-11]
+	eor	r6,r10,r6
+	 strb	r4,[r14],#16		@ store output
+	eor	r7,r11,r7
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-7]
+	ldrhsb	r11,[r12,#-3]
+	 strb	r5,[r14,#-12]
+	eor	r4,r8,r4,lsr#8
+	 strb	r6,[r14,#-8]
+	eor	r5,r9,r5,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-14]		@ load more input
+	ldrhsb	r9,[r12,#-10]
+	 strb	r7,[r14,#-4]
+	eor	r6,r10,r6,lsr#8
+	 strb	r4,[r14,#-15]
+	eor	r7,r11,r7,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-6]
+	ldrhsb	r11,[r12,#-2]
+	 strb	r5,[r14,#-11]
+	eor	r4,r8,r4,lsr#8
+	 strb	r6,[r14,#-7]
+	eor	r5,r9,r5,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r8,[r12,#-13]		@ load more input
+	ldrhsb	r9,[r12,#-9]
+	 strb	r7,[r14,#-3]
+	eor	r6,r10,r6,lsr#8
+	 strb	r4,[r14,#-14]
+	eor	r7,r11,r7,lsr#8
+# ifdef	__thumb2__
+	itt	hs
+# endif
+	ldrhsb	r10,[r12,#-5]
+	ldrhsb	r11,[r12,#-1]
+	 strb	r5,[r14,#-10]
+	 strb	r6,[r14,#-6]
+	eor	r4,r8,r4,lsr#8
+	 strb	r7,[r14,#-2]
+	eor	r5,r9,r5,lsr#8
+	 strb	r4,[r14,#-13]
+	eor	r6,r10,r6,lsr#8
+	 strb	r5,[r14,#-9]
+	eor	r7,r11,r7,lsr#8
+	 strb	r6,[r14,#-5]
+	 strb	r7,[r14,#-1]
+# ifdef	__thumb2__
+	it	ne
+# endif
+	ldrne	r8,[sp,#4*(32+2)]		@ re-load len
+# ifdef	__thumb2__
+	it	hs
+# endif
+	subhs	r11,r8,#64			@ len-=64
+	bhi	.Loop_outer
+
+	beq	.Ldone
+#endif
+
+.Ltail:
+	ldr	r12,[sp,#4*(32+1)]	@ load inp
+	add	r9,sp,#4*(0)
+	ldr	r14,[sp,#4*(32+0)]	@ load out
+
+.Loop_tail:
+	ldrb	r10,[r9],#1	@ read buffer on stack
+	ldrb	r11,[r12],#1		@ read input
+	subs	r8,r8,#1
+	eor	r11,r11,r10
+	strb	r11,[r14],#1		@ store output
+	bne	.Loop_tail
+
+.Ldone:
+	add	sp,sp,#4*(32+3)
+.Lno_data:
+	ldmia	sp!,{r4-r11,pc}
+.size	ChaCha20_ctr32,.-ChaCha20_ctr32
+#if __ARM_MAX_ARCH__>=7
+.arch	armv7-a
+.fpu	neon
+
+.type	ChaCha20_neon,%function
+.align	5
+ChaCha20_neon:
+	ldr		r12,[sp,#0]		@ pull pointer to counter and nonce
+	stmdb		sp!,{r0-r2,r4-r11,lr}
+.LChaCha20_neon:
+	adr		r14,.Lsigma
+	vstmdb		sp!,{d8-d15}		@ ABI spec says so
+	stmdb		sp!,{r0-r3}
+
+	vld1.32		{q1-q2},[r3]		@ load key
+	ldmia		r3,{r4-r11}		@ load key
+
+	sub		sp,sp,#4*(16+16)
+	vld1.32		{q3},[r12]		@ load counter and nonce
+	add		r12,sp,#4*8
+	ldmia		r14,{r0-r3}		@ load sigma
+	vld1.32		{q0},[r14]!		@ load sigma
+	vld1.32		{q12},[r14]!		@ one
+	@ vld1.32	{d30},[r14]		@ rot8
+	vst1.32		{q2-q3},[r12]		@ copy 1/2key|counter|nonce
+	vst1.32		{q0-q1},[sp]		@ copy sigma|1/2key
+
+	str		r10,[sp,#4*(16+10)]	@ off-load "rx"
+	str		r11,[sp,#4*(16+11)]	@ off-load "rx"
+	vshl.i32	d26,d24,#1	@ two
+	vstr		d24,[sp,#4*(16+0)]
+	vshl.i32	d28,d24,#2	@ four
+	vstr		d26,[sp,#4*(16+2)]
+	vmov		q4,q0
+	vstr		d28,[sp,#4*(16+4)]
+	vmov		q8,q0
+	@ vstr		d30,[sp,#4*(16+6)]
+	vmov		q5,q1
+	vmov		q9,q1
+	b		.Loop_neon_enter
+
+.align	4
+.Loop_neon_outer:
+	ldmia		sp,{r0-r9}		@ load key material
+	cmp		r11,#64*2		@ if len<=64*2
+	bls		.Lbreak_neon		@ switch to integer-only
+	@ vldr		d30,[sp,#4*(16+6)]	@ rot8
+	vmov		q4,q0
+	str		r11,[sp,#4*(32+2)]	@ save len
+	vmov		q8,q0
+	str		r12,  [sp,#4*(32+1)]	@ save inp
+	vmov		q5,q1
+	str		r14,  [sp,#4*(32+0)]	@ save out
+	vmov		q9,q1
+.Loop_neon_enter:
+	ldr		r11, [sp,#4*(15)]
+	 mov		r4,r4,ror#19	@ twist b[0..3]
+	vadd.i32	q7,q3,q12		@ counter+1
+	ldr		r12,[sp,#4*(12)]	@ modulo-scheduled load
+	 mov		r5,r5,ror#19
+	vmov		q6,q2
+	ldr		r10, [sp,#4*(13)]
+	 mov		r6,r6,ror#19
+	vmov		q10,q2
+	ldr		r14,[sp,#4*(14)]
+	 mov		r7,r7,ror#19
+	vadd.i32	q11,q7,q12		@ counter+2
+	add		r12,r12,#3	@ counter+3
+	mov		r11,r11,ror#8	@ twist d[0..3]
+	mov		r12,r12,ror#8
+	mov		r10,r10,ror#8
+	mov		r14,r14,ror#8
+	str		r11, [sp,#4*(16+15)]
+	mov		r11,#10
+	b		.Loop_neon
+
+.align	4
+.Loop_neon:
+	subs		r11,r11,#1
+	vadd.i32	q0,q0,q1
+	add	r0,r0,r4,ror#13
+	vadd.i32	q4,q4,q5
+	add	r1,r1,r5,ror#13
+	vadd.i32	q8,q8,q9
+	eor	r12,r0,r12,ror#24
+	veor	q3,q3,q0
+	eor	r10,r1,r10,ror#24
+	veor	q7,q7,q4
+	add	r8,r8,r12,ror#16
+	veor	q11,q11,q8
+	add	r9,r9,r10,ror#16
+	vrev32.16	q3,q3
+	eor	r4,r8,r4,ror#13
+	vrev32.16	q7,q7
+	eor	r5,r9,r5,ror#13
+	vrev32.16	q11,q11
+	add	r0,r0,r4,ror#20
+	vadd.i32	q2,q2,q3
+	add	r1,r1,r5,ror#20
+	vadd.i32	q6,q6,q7
+	eor	r12,r0,r12,ror#16
+	vadd.i32	q10,q10,q11
+	eor	r10,r1,r10,ror#16
+	veor	q12,q1,q2
+	add	r8,r8,r12,ror#24
+	veor	q13,q5,q6
+	str	r10,[sp,#4*(16+13)]
+	veor	q14,q9,q10
+	add	r9,r9,r10,ror#24
+	vshr.u32	q1,q12,#20
+	ldr	r10,[sp,#4*(16+15)]
+	vshr.u32	q5,q13,#20
+	str	r8,[sp,#4*(16+8)]
+	vshr.u32	q9,q14,#20
+	eor	r4,r4,r8,ror#12
+	vsli.32	q1,q12,#12
+	str	r9,[sp,#4*(16+9)]
+	vsli.32	q5,q13,#12
+	eor	r5,r5,r9,ror#12
+	vsli.32	q9,q14,#12
+	ldr	r8,[sp,#4*(16+10)]
+	vadd.i32	q0,q0,q1
+	add	r2,r2,r6,ror#13
+	vadd.i32	q4,q4,q5
+	ldr	r9,[sp,#4*(16+11)]
+	vadd.i32	q8,q8,q9
+	add	r3,r3,r7,ror#13
+	veor	q12,q3,q0
+	eor	r14,r2,r14,ror#24
+	veor	q13,q7,q4
+	eor	r10,r3,r10,ror#24
+	veor	q14,q11,q8
+	add	r8,r8,r14,ror#16
+	vshr.u32	q3,q12,#24
+	add	r9,r9,r10,ror#16
+	vshr.u32	q7,q13,#24
+	eor	r6,r8,r6,ror#13
+	vshr.u32	q11,q14,#24
+	eor	r7,r9,r7,ror#13
+	vsli.32	q3,q12,#8
+	add	r2,r2,r6,ror#20
+	vsli.32	q7,q13,#8
+	add	r3,r3,r7,ror#20
+	vsli.32	q11,q14,#8
+	eor	r14,r2,r14,ror#16
+	vadd.i32	q2,q2,q3
+	eor	r10,r3,r10,ror#16
+	vadd.i32	q6,q6,q7
+	add	r8,r8,r14,ror#24
+	vadd.i32	q10,q10,q11
+	add	r9,r9,r10,ror#24
+	veor	q12,q1,q2
+	eor	r6,r6,r8,ror#12
+	veor	q13,q5,q6
+	eor	r7,r7,r9,ror#12
+	veor	q14,q9,q10
+	vshr.u32	q1,q12,#25
+	vshr.u32	q5,q13,#25
+	vshr.u32	q9,q14,#25
+	vsli.32	q1,q12,#7
+	vsli.32	q5,q13,#7
+	vsli.32	q9,q14,#7
+	vext.8	q2,q2,q2,#8
+	vext.8	q6,q6,q6,#8
+	vext.8	q10,q10,q10,#8
+	vext.8	q1,q1,q1,#4
+	vext.8	q5,q5,q5,#4
+	vext.8	q9,q9,q9,#4
+	vext.8	q3,q3,q3,#12
+	vext.8	q7,q7,q7,#12
+	vext.8	q11,q11,q11,#12
+	vadd.i32	q0,q0,q1
+	add	r0,r0,r5,ror#13
+	vadd.i32	q4,q4,q5
+	add	r1,r1,r6,ror#13
+	vadd.i32	q8,q8,q9
+	eor	r10,r0,r10,ror#24
+	veor	q3,q3,q0
+	eor	r12,r1,r12,ror#24
+	veor	q7,q7,q4
+	add	r8,r8,r10,ror#16
+	veor	q11,q11,q8
+	add	r9,r9,r12,ror#16
+	vrev32.16	q3,q3
+	eor	r5,r8,r5,ror#13
+	vrev32.16	q7,q7
+	eor	r6,r9,r6,ror#13
+	vrev32.16	q11,q11
+	add	r0,r0,r5,ror#20
+	vadd.i32	q2,q2,q3
+	add	r1,r1,r6,ror#20
+	vadd.i32	q6,q6,q7
+	eor	r10,r0,r10,ror#16
+	vadd.i32	q10,q10,q11
+	eor	r12,r1,r12,ror#16
+	veor	q12,q1,q2
+	str	r10,[sp,#4*(16+15)]
+	veor	q13,q5,q6
+	add	r8,r8,r10,ror#24
+	veor	q14,q9,q10
+	ldr	r10,[sp,#4*(16+13)]
+	vshr.u32	q1,q12,#20
+	add	r9,r9,r12,ror#24
+	vshr.u32	q5,q13,#20
+	str	r8,[sp,#4*(16+10)]
+	vshr.u32	q9,q14,#20
+	eor	r5,r5,r8,ror#12
+	vsli.32	q1,q12,#12
+	str	r9,[sp,#4*(16+11)]
+	vsli.32	q5,q13,#12
+	eor	r6,r6,r9,ror#12
+	vsli.32	q9,q14,#12
+	ldr	r8,[sp,#4*(16+8)]
+	vadd.i32	q0,q0,q1
+	add	r2,r2,r7,ror#13
+	vadd.i32	q4,q4,q5
+	ldr	r9,[sp,#4*(16+9)]
+	vadd.i32	q8,q8,q9
+	add	r3,r3,r4,ror#13
+	veor	q12,q3,q0
+	eor	r10,r2,r10,ror#24
+	veor	q13,q7,q4
+	eor	r14,r3,r14,ror#24
+	veor	q14,q11,q8
+	add	r8,r8,r10,ror#16
+	vshr.u32	q3,q12,#24
+	add	r9,r9,r14,ror#16
+	vshr.u32	q7,q13,#24
+	eor	r7,r8,r7,ror#13
+	vshr.u32	q11,q14,#24
+	eor	r4,r9,r4,ror#13
+	vsli.32	q3,q12,#8
+	add	r2,r2,r7,ror#20
+	vsli.32	q7,q13,#8
+	add	r3,r3,r4,ror#20
+	vsli.32	q11,q14,#8
+	eor	r10,r2,r10,ror#16
+	vadd.i32	q2,q2,q3
+	eor	r14,r3,r14,ror#16
+	vadd.i32	q6,q6,q7
+	add	r8,r8,r10,ror#24
+	vadd.i32	q10,q10,q11
+	add	r9,r9,r14,ror#24
+	veor	q12,q1,q2
+	eor	r7,r7,r8,ror#12
+	veor	q13,q5,q6
+	eor	r4,r4,r9,ror#12
+	veor	q14,q9,q10
+	vshr.u32	q1,q12,#25
+	vshr.u32	q5,q13,#25
+	vshr.u32	q9,q14,#25
+	vsli.32	q1,q12,#7
+	vsli.32	q5,q13,#7
+	vsli.32	q9,q14,#7
+	vext.8	q2,q2,q2,#8
+	vext.8	q6,q6,q6,#8
+	vext.8	q10,q10,q10,#8
+	vext.8	q1,q1,q1,#12
+	vext.8	q5,q5,q5,#12
+	vext.8	q9,q9,q9,#12
+	vext.8	q3,q3,q3,#4
+	vext.8	q7,q7,q7,#4
+	vext.8	q11,q11,q11,#4
+	bne		.Loop_neon
+
+	add		r11,sp,#32
+	vld1.32		{q12-q13},[sp]		@ load key material
+	vld1.32		{q14-q15},[r11]
+
+	ldr		r11,[sp,#4*(32+2)]	@ load len
+
+	str		r8, [sp,#4*(16+8)]	@ modulo-scheduled store
+	str		r9, [sp,#4*(16+9)]
+	str		r12,[sp,#4*(16+12)]
+	str		r10, [sp,#4*(16+13)]
+	str		r14,[sp,#4*(16+14)]
+
+	@ at this point we have first half of 512-bit result in
+	@ rx and second half at sp+4*(16+8)
+
+	ldr		r12,[sp,#4*(32+1)]	@ load inp
+	ldr		r14,[sp,#4*(32+0)]	@ load out
+
+	vadd.i32	q0,q0,q12		@ accumulate key material
+	vadd.i32	q4,q4,q12
+	vadd.i32	q8,q8,q12
+	vldr		d24,[sp,#4*(16+0)]	@ one
+
+	vadd.i32	q1,q1,q13
+	vadd.i32	q5,q5,q13
+	vadd.i32	q9,q9,q13
+	vldr		d26,[sp,#4*(16+2)]	@ two
+
+	vadd.i32	q2,q2,q14
+	vadd.i32	q6,q6,q14
+	vadd.i32	q10,q10,q14
+	vadd.i32	d14,d14,d24	@ counter+1
+	vadd.i32	d22,d22,d26	@ counter+2
+
+	vadd.i32	q3,q3,q15
+	vadd.i32	q7,q7,q15
+	vadd.i32	q11,q11,q15
+
+	cmp		r11,#64*4
+	blo		.Ltail_neon
+
+	vld1.8		{q12-q13},[r12]!	@ load input
+	 mov		r11,sp
+	vld1.8		{q14-q15},[r12]!
+	veor		q0,q0,q12		@ xor with input
+	veor		q1,q1,q13
+	vld1.8		{q12-q13},[r12]!
+	veor		q2,q2,q14
+	veor		q3,q3,q15
+	vld1.8		{q14-q15},[r12]!
+
+	veor		q4,q4,q12
+	 vst1.8		{q0-q1},[r14]!	@ store output
+	veor		q5,q5,q13
+	vld1.8		{q12-q13},[r12]!
+	veor		q6,q6,q14
+	 vst1.8		{q2-q3},[r14]!
+	veor		q7,q7,q15
+	vld1.8		{q14-q15},[r12]!
+
+	veor		q8,q8,q12
+	 vld1.32	{q0-q1},[r11]!	@ load for next iteration
+	 veor		d25,d25,d25
+	 vldr		d24,[sp,#4*(16+4)]	@ four
+	veor		q9,q9,q13
+	 vld1.32	{q2-q3},[r11]
+	veor		q10,q10,q14
+	 vst1.8		{q4-q5},[r14]!
+	veor		q11,q11,q15
+	 vst1.8		{q6-q7},[r14]!
+
+	vadd.i32	d6,d6,d24	@ next counter value
+	vldr		d24,[sp,#4*(16+0)]	@ one
+
+	ldmia		sp,{r8-r11}	@ load key material
+	add		r0,r0,r8	@ accumulate key material
+	ldr		r8,[r12],#16		@ load input
+	 vst1.8		{q8-q9},[r14]!
+	add		r1,r1,r9
+	ldr		r9,[r12,#-12]
+	 vst1.8		{q10-q11},[r14]!
+	add		r2,r2,r10
+	ldr		r10,[r12,#-8]
+	add		r3,r3,r11
+	ldr		r11,[r12,#-4]
+# ifdef	__ARMEB__
+	rev		r0,r0
+	rev		r1,r1
+	rev		r2,r2
+	rev		r3,r3
+# endif
+	eor		r0,r0,r8	@ xor with input
+	 add		r8,sp,#4*(4)
+	eor		r1,r1,r9
+	str		r0,[r14],#16		@ store output
+	eor		r2,r2,r10
+	str		r1,[r14,#-12]
+	eor		r3,r3,r11
+	 ldmia		r8,{r8-r11}	@ load key material
+	str		r2,[r14,#-8]
+	str		r3,[r14,#-4]
+
+	add		r4,r8,r4,ror#13 @ accumulate key material
+	ldr		r8,[r12],#16		@ load input
+	add		r5,r9,r5,ror#13
+	ldr		r9,[r12,#-12]
+	add		r6,r10,r6,ror#13
+	ldr		r10,[r12,#-8]
+	add		r7,r11,r7,ror#13
+	ldr		r11,[r12,#-4]
+# ifdef	__ARMEB__
+	rev		r4,r4
+	rev		r5,r5
+	rev		r6,r6
+	rev		r7,r7
+# endif
+	eor		r4,r4,r8
+	 add		r8,sp,#4*(8)
+	eor		r5,r5,r9
+	str		r4,[r14],#16		@ store output
+	eor		r6,r6,r10
+	str		r5,[r14,#-12]
+	eor		r7,r7,r11
+	 ldmia		r8,{r8-r11}	@ load key material
+	str		r6,[r14,#-8]
+	 add		r0,sp,#4*(16+8)
+	str		r7,[r14,#-4]
+
+	ldmia		r0,{r0-r7}	@ load second half
+
+	add		r0,r0,r8	@ accumulate key material
+	ldr		r8,[r12],#16		@ load input
+	add		r1,r1,r9
+	ldr		r9,[r12,#-12]
+# ifdef	__thumb2__
+	it	hi
+# endif
+	 strhi		r10,[sp,#4*(16+10)]	@ copy "rx" while at it
+	add		r2,r2,r10
+	ldr		r10,[r12,#-8]
+# ifdef	__thumb2__
+	it	hi
+# endif
+	 strhi		r11,[sp,#4*(16+11)]	@ copy "rx" while at it
+	add		r3,r3,r11
+	ldr		r11,[r12,#-4]
+# ifdef	__ARMEB__
+	rev		r0,r0
+	rev		r1,r1
+	rev		r2,r2
+	rev		r3,r3
+# endif
+	eor		r0,r0,r8
+	 add		r8,sp,#4*(12)
+	eor		r1,r1,r9
+	str		r0,[r14],#16		@ store output
+	eor		r2,r2,r10
+	str		r1,[r14,#-12]
+	eor		r3,r3,r11
+	 ldmia		r8,{r8-r11}	@ load key material
+	str		r2,[r14,#-8]
+	str		r3,[r14,#-4]
+
+	add		r4,r8,r4,ror#24 @ accumulate key material
+	 add		r8,r8,#4		@ next counter value
+	add		r5,r9,r5,ror#24
+	 str		r8,[sp,#4*(12)]	@ save next counter value
+	ldr		r8,[r12],#16		@ load input
+	add		r6,r10,r6,ror#24
+	 add		r4,r4,#3		@ counter+3
+	ldr		r9,[r12,#-12]
+	add		r7,r11,r7,ror#24
+	ldr		r10,[r12,#-8]
+	ldr		r11,[r12,#-4]
+# ifdef	__ARMEB__
+	rev		r4,r4
+	rev		r5,r5
+	rev		r6,r6
+	rev		r7,r7
+# endif
+	eor		r4,r4,r8
+# ifdef	__thumb2__
+	it	hi
+# endif
+	 ldrhi		r8,[sp,#4*(32+2)]	@ re-load len
+	eor		r5,r5,r9
+	eor		r6,r6,r10
+	str		r4,[r14],#16		@ store output
+	eor		r7,r7,r11
+	str		r5,[r14,#-12]
+	 sub		r11,r8,#64*4	@ len-=64*4
+	str		r6,[r14,#-8]
+	str		r7,[r14,#-4]
+	bhi		.Loop_neon_outer
+
+	b		.Ldone_neon
+
+.align	4
+.Lbreak_neon:
+	@ harmonize NEON and integer-only stack frames: load data
+	@ from NEON frame, but save to integer-only one; distance
+	@ between the two is 4*(32+4+16-32)=4*(20).
+
+	str		r11, [sp,#4*(20+32+2)]	@ save len
+	 add		r11,sp,#4*(32+4)
+	str		r12,   [sp,#4*(20+32+1)]	@ save inp
+	str		r14,   [sp,#4*(20+32+0)]	@ save out
+
+	ldr		r12,[sp,#4*(16+10)]
+	ldr		r14,[sp,#4*(16+11)]
+	 vldmia		r11,{d8-d15}			@ fulfill ABI requirement
+	str		r12,[sp,#4*(20+16+10)]	@ copy "rx"
+	str		r14,[sp,#4*(20+16+11)]	@ copy "rx"
+
+	ldr		r11, [sp,#4*(15)]
+	 mov		r4,r4,ror#19		@ twist b[0..3]
+	ldr		r12,[sp,#4*(12)]		@ modulo-scheduled load
+	 mov		r5,r5,ror#19
+	ldr		r10, [sp,#4*(13)]
+	 mov		r6,r6,ror#19
+	ldr		r14,[sp,#4*(14)]
+	 mov		r7,r7,ror#19
+	mov		r11,r11,ror#8		@ twist d[0..3]
+	mov		r12,r12,ror#8
+	mov		r10,r10,ror#8
+	mov		r14,r14,ror#8
+	str		r11, [sp,#4*(20+16+15)]
+	add		r11,sp,#4*(20)
+	vst1.32		{q0-q1},[r11]!		@ copy key
+	add		sp,sp,#4*(20)			@ switch frame
+	vst1.32		{q2-q3},[r11]
+	mov		r11,#10
+	b		.Loop				@ go integer-only
+
+.align	4
+.Ltail_neon:
+	cmp		r11,#64*3
+	bhs		.L192_or_more_neon
+	cmp		r11,#64*2
+	bhs		.L128_or_more_neon
+	cmp		r11,#64*1
+	bhs		.L64_or_more_neon
+
+	add		r8,sp,#4*(8)
+	vst1.8		{q0-q1},[sp]
+	add		r10,sp,#4*(0)
+	vst1.8		{q2-q3},[r8]
+	b		.Loop_tail_neon
+
+.align	4
+.L64_or_more_neon:
+	vld1.8		{q12-q13},[r12]!
+	vld1.8		{q14-q15},[r12]!
+	veor		q0,q0,q12
+	veor		q1,q1,q13
+	veor		q2,q2,q14
+	veor		q3,q3,q15
+	vst1.8		{q0-q1},[r14]!
+	vst1.8		{q2-q3},[r14]!
+
+	beq		.Ldone_neon
+
+	add		r8,sp,#4*(8)
+	vst1.8		{q4-q5},[sp]
+	add		r10,sp,#4*(0)
+	vst1.8		{q6-q7},[r8]
+	sub		r11,r11,#64*1	@ len-=64*1
+	b		.Loop_tail_neon
+
+.align	4
+.L128_or_more_neon:
+	vld1.8		{q12-q13},[r12]!
+	vld1.8		{q14-q15},[r12]!
+	veor		q0,q0,q12
+	veor		q1,q1,q13
+	vld1.8		{q12-q13},[r12]!
+	veor		q2,q2,q14
+	veor		q3,q3,q15
+	vld1.8		{q14-q15},[r12]!
+
+	veor		q4,q4,q12
+	veor		q5,q5,q13
+	 vst1.8		{q0-q1},[r14]!
+	veor		q6,q6,q14
+	 vst1.8		{q2-q3},[r14]!
+	veor		q7,q7,q15
+	vst1.8		{q4-q5},[r14]!
+	vst1.8		{q6-q7},[r14]!
+
+	beq		.Ldone_neon
+
+	add		r8,sp,#4*(8)
+	vst1.8		{q8-q9},[sp]
+	add		r10,sp,#4*(0)
+	vst1.8		{q10-q11},[r8]
+	sub		r11,r11,#64*2	@ len-=64*2
+	b		.Loop_tail_neon
+
+.align	4
+.L192_or_more_neon:
+	vld1.8		{q12-q13},[r12]!
+	vld1.8		{q14-q15},[r12]!
+	veor		q0,q0,q12
+	veor		q1,q1,q13
+	vld1.8		{q12-q13},[r12]!
+	veor		q2,q2,q14
+	veor		q3,q3,q15
+	vld1.8		{q14-q15},[r12]!
+
+	veor		q4,q4,q12
+	veor		q5,q5,q13
+	vld1.8		{q12-q13},[r12]!
+	veor		q6,q6,q14
+	 vst1.8		{q0-q1},[r14]!
+	veor		q7,q7,q15
+	vld1.8		{q14-q15},[r12]!
+
+	veor		q8,q8,q12
+	 vst1.8		{q2-q3},[r14]!
+	veor		q9,q9,q13
+	 vst1.8		{q4-q5},[r14]!
+	veor		q10,q10,q14
+	 vst1.8		{q6-q7},[r14]!
+	veor		q11,q11,q15
+	vst1.8		{q8-q9},[r14]!
+	vst1.8		{q10-q11},[r14]!
+
+	beq		.Ldone_neon
+
+	ldmia		sp,{r8-r11}	@ load key material
+	add		r0,r0,r8	@ accumulate key material
+	 add		r8,sp,#4*(4)
+	add		r1,r1,r9
+	add		r2,r2,r10
+	add		r3,r3,r11
+	 ldmia		r8,{r8-r11}	@ load key material
+
+	add		r4,r8,r4,ror#13 @ accumulate key material
+	 add		r8,sp,#4*(8)
+	add		r5,r9,r5,ror#13
+	add		r6,r10,r6,ror#13
+	add		r7,r11,r7,ror#13
+	 ldmia		r8,{r8-r11}	@ load key material
+# ifdef	__ARMEB__
+	rev		r0,r0
+	rev		r1,r1
+	rev		r2,r2
+	rev		r3,r3
+	rev		r4,r4
+	rev		r5,r5
+	rev		r6,r6
+	rev		r7,r7
+# endif
+	stmia		sp,{r0-r7}
+	 add		r0,sp,#4*(16+8)
+
+	ldmia		r0,{r0-r7}	@ load second half
+
+	add		r0,r0,r8	@ accumulate key material
+	 add		r8,sp,#4*(12)
+	add		r1,r1,r9
+	add		r2,r2,r10
+	add		r3,r3,r11
+	 ldmia		r8,{r8-r11}	@ load key material
+
+	add		r4,r8,r4,ror#24 @ accumulate key material
+	 add		r8,sp,#4*(8)
+	add		r5,r9,r5,ror#24
+	 add		r4,r4,#3		@ counter+3
+	add		r6,r10,r6,ror#24
+	add		r7,r11,r7,ror#24
+	 ldr		r11,[sp,#4*(32+2)]	@ re-load len
+# ifdef	__ARMEB__
+	rev		r0,r0
+	rev		r1,r1
+	rev		r2,r2
+	rev		r3,r3
+	rev		r4,r4
+	rev		r5,r5
+	rev		r6,r6
+	rev		r7,r7
+# endif
+	stmia		r8,{r0-r7}
+	 add		r10,sp,#4*(0)
+	 sub		r11,r11,#64*3	@ len-=64*3
+
+.Loop_tail_neon:
+	ldrb		r8,[r10],#1	@ read buffer on stack
+	ldrb		r9,[r12],#1		@ read input
+	subs		r11,r11,#1
+	eor		r8,r8,r9
+	strb		r8,[r14],#1		@ store output
+	bne		.Loop_tail_neon
+
+.Ldone_neon:
+	add		sp,sp,#4*(32+4)
+	vldmia		sp,{d8-d15}
+	add		sp,sp,#4*(16+3)
+	ldmia		sp!,{r4-r11,pc}
+.size	ChaCha20_neon,.-ChaCha20_neon
+.comm	OPENSSL_armcap_P,4,4
+#endif
lib/zinc/chacha20/chacha20-arm64-cryptogams.S new file +1973 -0

1,560 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#8 lore raw

PATCH v8 8/28 zinc: port Andy Polyakov's ChaCha20 ARM and ARM64 implementations

In reply to Jason A. Donenfeld

These port and prepare Andy Polyakov's implementations for the kernel, but don't actually wire up any of the code yet. The wiring will be done in a subsequent commit, since we'll need to merge these implementations with another one. We make a few small changes to the assembly:

- Entries and exits use the proper kernel convention macro.
- CPU feature checking is done in C by the glue code, so that has been removed from the assembly.
- The function names have been renamed to fit kernel conventions.
- Labels have been renamed (prefixed with .L) to fit kernel conventions.
- Constants have been rearranged so that they are closer to the code that is using them. [ARM only]
- The neon code can jump to the scalar code when it makes sense to do

    so.

- The neon_512 function as a separate function has been removed, leaving the decision up to the main neon entry point. [ARM64 only]

Cc Russell King <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/chacha20/chacha20-arm-cryptogams.S   | 367 +++++++++---------
 lib/zinc/chacha20/chacha20-arm64-cryptogams.S |  75 ++--
 2 files changed, 202 insertions(+), 240 deletions(-)
lib/zinc/chacha20/chacha20-arm-cryptogams.S +180 -187
index 05a3a9e6e93f..770bab469171 100644
@@ -1,9 +1,12 @@
 /* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
 /*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
  * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ *
+ * This is based in part on Andy Polyakov's implementation from CRYPTOGAMS.
  */
 
-#include "arm_arch.h"
+#include <linux/linkage.h>
 
 .text
 #if defined(__thumb2__) || defined(__clang__)
@@ -24,48 +27,25 @@
 .long	0x61707865,0x3320646e,0x79622d32,0x6b206574	@ endian-neutral
 .Lone:
 .long	1,0,0,0
-.Lrot8:
-.long	0x02010003,0x06050407
-#if __ARM_MAX_ARCH__>=7
-.LOPENSSL_armcap:
-.word   OPENSSL_armcap_P-.LChaCha20_ctr32
-#else
 .word	-1
-#endif
 
-.globl	ChaCha20_ctr32
-.type	ChaCha20_ctr32,%function
 .align	5
-ChaCha20_ctr32:
-.LChaCha20_ctr32:
+ENTRY(chacha20_arm)
 	ldr	r12,[sp,#0]		@ pull pointer to counter and nonce
 	stmdb	sp!,{r0-r2,r4-r11,lr}
-#if __ARM_ARCH__<7 && !defined(__thumb2__)
-	sub	r14,pc,#16		@ ChaCha20_ctr32
-#else
-	adr	r14,.LChaCha20_ctr32
-#endif
 	cmp	r2,#0			@ len==0?
-#ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	eq
 #endif
 	addeq	sp,sp,#4*3
-	beq	.Lno_data
-#if __ARM_MAX_ARCH__>=7
-	cmp	r2,#192			@ test len
-	bls	.Lshort
-	ldr	r4,[r14,#-24]
-	ldr	r4,[r14,r4]
-# ifdef	__APPLE__
-	ldr	r4,[r4]
-# endif
-	tst	r4,#ARMV7_NEON
-	bne	.LChaCha20_neon
-.Lshort:
-#endif
+	beq	.Lno_data_arm
 	ldmia	r12,{r4-r7}		@ load counter and nonce
 	sub	sp,sp,#4*(16)		@ off-load area
-	sub	r14,r14,#64		@ .Lsigma
+#if __LINUX_ARM_ARCH__ < 7 && !defined(__thumb2__)
+	sub	r14,pc,#100		@ .Lsigma
+#else
+	adr	r14,.Lsigma		@ .Lsigma
+#endif
 	stmdb	sp!,{r4-r7}		@ copy counter and nonce
 	ldmia	r3,{r4-r11}		@ load key
 	ldmia	r14,{r0-r3}		@ load sigma
@@ -191,7 +171,7 @@ ChaCha20_ctr32:
 	@ rx and second half at sp+4*(16+8)
 
 	cmp	r11,#64		@ done yet?
-#ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
 #endif
 	addlo	r12,sp,#4*(0)		@ shortcut or ...
@@ -202,49 +182,49 @@ ChaCha20_ctr32:
 	ldr	r8,[sp,#4*(0)]	@ load key material
 	ldr	r9,[sp,#4*(1)]
 
-#if __ARM_ARCH__>=6 || !defined(__ARMEB__)
-# if __ARM_ARCH__<7
+#if __LINUX_ARM_ARCH__ >= 6 || !defined(__ARMEB__)
+#if __LINUX_ARM_ARCH__ < 7
 	orr	r10,r12,r14
 	tst	r10,#3		@ are input and output aligned?
 	ldr	r10,[sp,#4*(2)]
 	bne	.Lunaligned
 	cmp	r11,#64		@ restore flags
-# else
+#else
 	ldr	r10,[sp,#4*(2)]
-# endif
+#endif
 	ldr	r11,[sp,#4*(3)]
 
 	add	r0,r0,r8	@ accumulate key material
 	add	r1,r1,r9
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r8,[r12],#16		@ load input
 	ldrhs	r9,[r12,#-12]
 
 	add	r2,r2,r10
 	add	r3,r3,r11
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r10,[r12,#-8]
 	ldrhs	r11,[r12,#-4]
-# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+#if __LINUX_ARM_ARCH__ >= 6 && defined(__ARMEB__)
 	rev	r0,r0
 	rev	r1,r1
 	rev	r2,r2
 	rev	r3,r3
-# endif
-# ifdef	__thumb2__
+#endif
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r0,r0,r8	@ xor with input
 	eorhs	r1,r1,r9
 	 add	r8,sp,#4*(4)
 	str	r0,[r14],#16		@ store output
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r2,r2,r10
 	eorhs	r3,r3,r11
 	 ldmia	r8,{r8-r11}	@ load key material
@@ -254,34 +234,34 @@ ChaCha20_ctr32:
 
 	add	r4,r8,r4,ror#13 @ accumulate key material
 	add	r5,r9,r5,ror#13
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r8,[r12],#16		@ load input
 	ldrhs	r9,[r12,#-12]
 	add	r6,r10,r6,ror#13
 	add	r7,r11,r7,ror#13
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r10,[r12,#-8]
 	ldrhs	r11,[r12,#-4]
-# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+#if __LINUX_ARM_ARCH__ >= 6 && defined(__ARMEB__)
 	rev	r4,r4
 	rev	r5,r5
 	rev	r6,r6
 	rev	r7,r7
-# endif
-# ifdef	__thumb2__
+#endif
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r4,r4,r8
 	eorhs	r5,r5,r9
 	 add	r8,sp,#4*(8)
 	str	r4,[r14],#16		@ store output
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r6,r6,r10
 	eorhs	r7,r7,r11
 	str	r5,[r14,#-12]
@@ -294,39 +274,39 @@ ChaCha20_ctr32:
 
 	add	r0,r0,r8	@ accumulate key material
 	add	r1,r1,r9
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r8,[r12],#16		@ load input
 	ldrhs	r9,[r12,#-12]
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hi
-# endif
+#endif
 	 strhi	r10,[sp,#4*(16+10)]	@ copy "rx" while at it
 	 strhi	r11,[sp,#4*(16+11)]	@ copy "rx" while at it
 	add	r2,r2,r10
 	add	r3,r3,r11
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r10,[r12,#-8]
 	ldrhs	r11,[r12,#-4]
-# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+#if __LINUX_ARM_ARCH__ >= 6 && defined(__ARMEB__)
 	rev	r0,r0
 	rev	r1,r1
 	rev	r2,r2
 	rev	r3,r3
-# endif
-# ifdef	__thumb2__
+#endif
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r0,r0,r8
 	eorhs	r1,r1,r9
 	 add	r8,sp,#4*(12)
 	str	r0,[r14],#16		@ store output
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r2,r2,r10
 	eorhs	r3,r3,r11
 	str	r1,[r14,#-12]
@@ -336,79 +316,79 @@ ChaCha20_ctr32:
 
 	add	r4,r8,r4,ror#24 @ accumulate key material
 	add	r5,r9,r5,ror#24
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hi
-# endif
+#endif
 	 addhi	r8,r8,#1		@ next counter value
 	 strhi	r8,[sp,#4*(12)]	@ save next counter value
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r8,[r12],#16		@ load input
 	ldrhs	r9,[r12,#-12]
 	add	r6,r10,r6,ror#24
 	add	r7,r11,r7,ror#24
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhs	r10,[r12,#-8]
 	ldrhs	r11,[r12,#-4]
-# if __ARM_ARCH__>=6 && defined(__ARMEB__)
+#if __LINUX_ARM_ARCH__ >= 6 && defined(__ARMEB__)
 	rev	r4,r4
 	rev	r5,r5
 	rev	r6,r6
 	rev	r7,r7
-# endif
-# ifdef	__thumb2__
+#endif
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r4,r4,r8
 	eorhs	r5,r5,r9
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	 it	ne
-# endif
+#endif
 	 ldrne	r8,[sp,#4*(32+2)]	@ re-load len
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	eorhs	r6,r6,r10
 	eorhs	r7,r7,r11
 	str	r4,[r14],#16		@ store output
 	str	r5,[r14,#-12]
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	hs
-# endif
+#endif
 	 subhs	r11,r8,#64		@ len-=64
 	str	r6,[r14,#-8]
 	str	r7,[r14,#-4]
 	bhi	.Loop_outer
 
 	beq	.Ldone
-# if __ARM_ARCH__<7
+#if __LINUX_ARM_ARCH__ < 7
 	b	.Ltail
 
 .align	4
 .Lunaligned:				@ unaligned endian-neutral path
 	cmp	r11,#64		@ restore flags
-# endif
 #endif
-#if __ARM_ARCH__<7
+#endif
+#if __LINUX_ARM_ARCH__ < 7
 	ldr	r11,[sp,#4*(3)]
 	add	r0,r8,r0	@ accumulate key material
 	add	r1,r9,r1
 	add	r2,r10,r2
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r8,r8,r8		@ zero or ...
 	ldrhsb	r8,[r12],#16			@ ... load input
 	eorlo	r9,r9,r9
 	ldrhsb	r9,[r12,#-12]
 
 	add	r3,r11,r3
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r10,r10,r10
 	ldrhsb	r10,[r12,#-8]
 	eorlo	r11,r11,r11
@@ -416,53 +396,53 @@ ChaCha20_ctr32:
 
 	eor	r0,r8,r0		@ xor with input (or zero)
 	eor	r1,r9,r1
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-15]		@ load more input
 	ldrhsb	r9,[r12,#-11]
 	eor	r2,r10,r2
 	 strb	r0,[r14],#16		@ store output
 	eor	r3,r11,r3
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-7]
 	ldrhsb	r11,[r12,#-3]
 	 strb	r1,[r14,#-12]
 	eor	r0,r8,r0,lsr#8
 	 strb	r2,[r14,#-8]
 	eor	r1,r9,r1,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-14]		@ load more input
 	ldrhsb	r9,[r12,#-10]
 	 strb	r3,[r14,#-4]
 	eor	r2,r10,r2,lsr#8
 	 strb	r0,[r14,#-15]
 	eor	r3,r11,r3,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-6]
 	ldrhsb	r11,[r12,#-2]
 	 strb	r1,[r14,#-11]
 	eor	r0,r8,r0,lsr#8
 	 strb	r2,[r14,#-7]
 	eor	r1,r9,r1,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-13]		@ load more input
 	ldrhsb	r9,[r12,#-9]
 	 strb	r3,[r14,#-3]
 	eor	r2,r10,r2,lsr#8
 	 strb	r0,[r14,#-14]
 	eor	r3,r11,r3,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-5]
 	ldrhsb	r11,[r12,#-1]
 	 strb	r1,[r14,#-10]
@@ -482,18 +462,18 @@ ChaCha20_ctr32:
 	add	r4,r8,r4,ror#13	@ accumulate key material
 	add	r5,r9,r5,ror#13
 	add	r6,r10,r6,ror#13
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r8,r8,r8		@ zero or ...
 	ldrhsb	r8,[r12],#16			@ ... load input
 	eorlo	r9,r9,r9
 	ldrhsb	r9,[r12,#-12]
 
 	add	r7,r11,r7,ror#13
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r10,r10,r10
 	ldrhsb	r10,[r12,#-8]
 	eorlo	r11,r11,r11
@@ -501,53 +481,53 @@ ChaCha20_ctr32:
 
 	eor	r4,r8,r4		@ xor with input (or zero)
 	eor	r5,r9,r5
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-15]		@ load more input
 	ldrhsb	r9,[r12,#-11]
 	eor	r6,r10,r6
 	 strb	r4,[r14],#16		@ store output
 	eor	r7,r11,r7
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-7]
 	ldrhsb	r11,[r12,#-3]
 	 strb	r5,[r14,#-12]
 	eor	r4,r8,r4,lsr#8
 	 strb	r6,[r14,#-8]
 	eor	r5,r9,r5,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-14]		@ load more input
 	ldrhsb	r9,[r12,#-10]
 	 strb	r7,[r14,#-4]
 	eor	r6,r10,r6,lsr#8
 	 strb	r4,[r14,#-15]
 	eor	r7,r11,r7,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-6]
 	ldrhsb	r11,[r12,#-2]
 	 strb	r5,[r14,#-11]
 	eor	r4,r8,r4,lsr#8
 	 strb	r6,[r14,#-7]
 	eor	r5,r9,r5,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-13]		@ load more input
 	ldrhsb	r9,[r12,#-9]
 	 strb	r7,[r14,#-3]
 	eor	r6,r10,r6,lsr#8
 	 strb	r4,[r14,#-14]
 	eor	r7,r11,r7,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-5]
 	ldrhsb	r11,[r12,#-1]
 	 strb	r5,[r14,#-10]
@@ -564,26 +544,26 @@ ChaCha20_ctr32:
 	add	r8,sp,#4*(4+4)
 	ldmia	r8,{r8-r11}		@ load key material
 	ldmia	r0,{r0-r7}		@ load second half
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hi
-# endif
+#endif
 	strhi	r10,[sp,#4*(16+10)]		@ copy "rx"
 	strhi	r11,[sp,#4*(16+11)]		@ copy "rx"
 	add	r0,r8,r0	@ accumulate key material
 	add	r1,r9,r1
 	add	r2,r10,r2
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r8,r8,r8		@ zero or ...
 	ldrhsb	r8,[r12],#16			@ ... load input
 	eorlo	r9,r9,r9
 	ldrhsb	r9,[r12,#-12]
 
 	add	r3,r11,r3
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r10,r10,r10
 	ldrhsb	r10,[r12,#-8]
 	eorlo	r11,r11,r11
@@ -591,53 +571,53 @@ ChaCha20_ctr32:
 
 	eor	r0,r8,r0		@ xor with input (or zero)
 	eor	r1,r9,r1
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-15]		@ load more input
 	ldrhsb	r9,[r12,#-11]
 	eor	r2,r10,r2
 	 strb	r0,[r14],#16		@ store output
 	eor	r3,r11,r3
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-7]
 	ldrhsb	r11,[r12,#-3]
 	 strb	r1,[r14,#-12]
 	eor	r0,r8,r0,lsr#8
 	 strb	r2,[r14,#-8]
 	eor	r1,r9,r1,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-14]		@ load more input
 	ldrhsb	r9,[r12,#-10]
 	 strb	r3,[r14,#-4]
 	eor	r2,r10,r2,lsr#8
 	 strb	r0,[r14,#-15]
 	eor	r3,r11,r3,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-6]
 	ldrhsb	r11,[r12,#-2]
 	 strb	r1,[r14,#-11]
 	eor	r0,r8,r0,lsr#8
 	 strb	r2,[r14,#-7]
 	eor	r1,r9,r1,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-13]		@ load more input
 	ldrhsb	r9,[r12,#-9]
 	 strb	r3,[r14,#-3]
 	eor	r2,r10,r2,lsr#8
 	 strb	r0,[r14,#-14]
 	eor	r3,r11,r3,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-5]
 	ldrhsb	r11,[r12,#-1]
 	 strb	r1,[r14,#-10]
@@ -654,25 +634,25 @@ ChaCha20_ctr32:
 	add	r8,sp,#4*(4+8)
 	ldmia	r8,{r8-r11}		@ load key material
 	add	r4,r8,r4,ror#24	@ accumulate key material
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hi
-# endif
+#endif
 	addhi	r8,r8,#1			@ next counter value
 	strhi	r8,[sp,#4*(12)]		@ save next counter value
 	add	r5,r9,r5,ror#24
 	add	r6,r10,r6,ror#24
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r8,r8,r8		@ zero or ...
 	ldrhsb	r8,[r12],#16			@ ... load input
 	eorlo	r9,r9,r9
 	ldrhsb	r9,[r12,#-12]
 
 	add	r7,r11,r7,ror#24
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itete	lo
-# endif
+#endif
 	eorlo	r10,r10,r10
 	ldrhsb	r10,[r12,#-8]
 	eorlo	r11,r11,r11
@@ -680,53 +660,53 @@ ChaCha20_ctr32:
 
 	eor	r4,r8,r4		@ xor with input (or zero)
 	eor	r5,r9,r5
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-15]		@ load more input
 	ldrhsb	r9,[r12,#-11]
 	eor	r6,r10,r6
 	 strb	r4,[r14],#16		@ store output
 	eor	r7,r11,r7
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-7]
 	ldrhsb	r11,[r12,#-3]
 	 strb	r5,[r14,#-12]
 	eor	r4,r8,r4,lsr#8
 	 strb	r6,[r14,#-8]
 	eor	r5,r9,r5,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-14]		@ load more input
 	ldrhsb	r9,[r12,#-10]
 	 strb	r7,[r14,#-4]
 	eor	r6,r10,r6,lsr#8
 	 strb	r4,[r14,#-15]
 	eor	r7,r11,r7,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-6]
 	ldrhsb	r11,[r12,#-2]
 	 strb	r5,[r14,#-11]
 	eor	r4,r8,r4,lsr#8
 	 strb	r6,[r14,#-7]
 	eor	r5,r9,r5,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r8,[r12,#-13]		@ load more input
 	ldrhsb	r9,[r12,#-9]
 	 strb	r7,[r14,#-3]
 	eor	r6,r10,r6,lsr#8
 	 strb	r4,[r14,#-14]
 	eor	r7,r11,r7,lsr#8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	itt	hs
-# endif
+#endif
 	ldrhsb	r10,[r12,#-5]
 	ldrhsb	r11,[r12,#-1]
 	 strb	r5,[r14,#-10]
@@ -740,13 +720,13 @@ ChaCha20_ctr32:
 	eor	r7,r11,r7,lsr#8
 	 strb	r6,[r14,#-5]
 	 strb	r7,[r14,#-1]
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	ne
-# endif
+#endif
 	ldrne	r8,[sp,#4*(32+2)]		@ re-load len
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	hs
-# endif
+#endif
 	subhs	r11,r8,#64			@ len-=64
 	bhi	.Loop_outer
 
@@ -768,20 +748,33 @@ ChaCha20_ctr32:
 
 .Ldone:
 	add	sp,sp,#4*(32+3)
-.Lno_data:
+.Lno_data_arm:
 	ldmia	sp!,{r4-r11,pc}
-.size	ChaCha20_ctr32,.-ChaCha20_ctr32
-#if __ARM_MAX_ARCH__>=7
+ENDPROC(chacha20_arm)
+
+#ifdef CONFIG_KERNEL_MODE_NEON
+.align	5
+.Lsigma2:
+.long	0x61707865,0x3320646e,0x79622d32,0x6b206574	@ endian-neutral
+.Lone2:
+.long	1,0,0,0
+.word	-1
+
 .arch	armv7-a
 .fpu	neon
 
-.type	ChaCha20_neon,%function
 .align	5
-ChaCha20_neon:
+ENTRY(chacha20_neon)
 	ldr		r12,[sp,#0]		@ pull pointer to counter and nonce
 	stmdb		sp!,{r0-r2,r4-r11,lr}
-.LChaCha20_neon:
-	adr		r14,.Lsigma
+	cmp		r2,#0			@ len==0?
+#ifdef	__thumb2__
+	itt		eq
+#endif
+	addeq		sp,sp,#4*3
+	beq		.Lno_data_neon
+.Lchacha20_neon_begin:
+	adr		r14,.Lsigma2
 	vstmdb		sp!,{d8-d15}		@ ABI spec says so
 	stmdb		sp!,{r0-r3}
 
@@ -1121,12 +1114,12 @@ ChaCha20_neon:
 	ldr		r10,[r12,#-8]
 	add		r3,r3,r11
 	ldr		r11,[r12,#-4]
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r0,r0
 	rev		r1,r1
 	rev		r2,r2
 	rev		r3,r3
-# endif
+#endif
 	eor		r0,r0,r8	@ xor with input
 	 add		r8,sp,#4*(4)
 	eor		r1,r1,r9
@@ -1146,12 +1139,12 @@ ChaCha20_neon:
 	ldr		r10,[r12,#-8]
 	add		r7,r11,r7,ror#13
 	ldr		r11,[r12,#-4]
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r4,r4
 	rev		r5,r5
 	rev		r6,r6
 	rev		r7,r7
-# endif
+#endif
 	eor		r4,r4,r8
 	 add		r8,sp,#4*(8)
 	eor		r5,r5,r9
@@ -1170,24 +1163,24 @@ ChaCha20_neon:
 	ldr		r8,[r12],#16		@ load input
 	add		r1,r1,r9
 	ldr		r9,[r12,#-12]
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	hi
-# endif
+#endif
 	 strhi		r10,[sp,#4*(16+10)]	@ copy "rx" while at it
 	add		r2,r2,r10
 	ldr		r10,[r12,#-8]
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	hi
-# endif
+#endif
 	 strhi		r11,[sp,#4*(16+11)]	@ copy "rx" while at it
 	add		r3,r3,r11
 	ldr		r11,[r12,#-4]
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r0,r0
 	rev		r1,r1
 	rev		r2,r2
 	rev		r3,r3
-# endif
+#endif
 	eor		r0,r0,r8
 	 add		r8,sp,#4*(12)
 	eor		r1,r1,r9
@@ -1210,16 +1203,16 @@ ChaCha20_neon:
 	add		r7,r11,r7,ror#24
 	ldr		r10,[r12,#-8]
 	ldr		r11,[r12,#-4]
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r4,r4
 	rev		r5,r5
 	rev		r6,r6
 	rev		r7,r7
-# endif
+#endif
 	eor		r4,r4,r8
-# ifdef	__thumb2__
+#ifdef __thumb2__
 	it	hi
-# endif
+#endif
 	 ldrhi		r8,[sp,#4*(32+2)]	@ re-load len
 	eor		r5,r5,r9
 	eor		r6,r6,r10
@@ -1379,7 +1372,7 @@ ChaCha20_neon:
 	add		r6,r10,r6,ror#13
 	add		r7,r11,r7,ror#13
 	 ldmia		r8,{r8-r11}	@ load key material
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r0,r0
 	rev		r1,r1
 	rev		r2,r2
@@ -1388,7 +1381,7 @@ ChaCha20_neon:
 	rev		r5,r5
 	rev		r6,r6
 	rev		r7,r7
-# endif
+#endif
 	stmia		sp,{r0-r7}
 	 add		r0,sp,#4*(16+8)
 
@@ -1408,7 +1401,7 @@ ChaCha20_neon:
 	add		r6,r10,r6,ror#24
 	add		r7,r11,r7,ror#24
 	 ldr		r11,[sp,#4*(32+2)]	@ re-load len
-# ifdef	__ARMEB__
+#ifdef	__ARMEB__
 	rev		r0,r0
 	rev		r1,r1
 	rev		r2,r2
@@ -1417,7 +1410,7 @@ ChaCha20_neon:
 	rev		r5,r5
 	rev		r6,r6
 	rev		r7,r7
-# endif
+#endif
 	stmia		r8,{r0-r7}
 	 add		r10,sp,#4*(0)
 	 sub		r11,r11,#64*3	@ len-=64*3
@@ -1434,7 +1427,7 @@ ChaCha20_neon:
 	add		sp,sp,#4*(32+4)
 	vldmia		sp,{d8-d15}
 	add		sp,sp,#4*(16+3)
+.Lno_data_neon:
 	ldmia		sp!,{r4-r11,pc}
-.size	ChaCha20_neon,.-ChaCha20_neon
-.comm	OPENSSL_armcap_P,4,4
+ENDPROC(chacha20_neon)
 #endif
lib/zinc/chacha20/chacha20-arm64-cryptogams.S +22 -53

166 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#9 lore raw

PATCH v8 9/28 zinc: ChaCha20 ARM and ARM64 implementations

In reply to Jason A. Donenfeld

These wire Andy Polyakov's implementations up to the kernel for ARMv7,8 NEON, and introduce Eric Biggers' ultra-fast scalar implementation for CPUs without NEON or for CPUs with slow NEON (Cortex-A5,7).

This commit does the following:
- Adds the glue code for the assembly implementations.
- Renames the ARMv8 code into place, since it can at this point be used wholesale.
- Merges Andy Polyakov's ARMv7 NEON code with Eric Biggers' <=ARMv7 scalar code.

This commit delivers approximately the same or much better performance than the existing crypto API's code and has been measured to do as such on:

- ARM1176JZF-S [ARMv6]

  - Cortex-A7    [ARMv7]
  - Cortex-A8    [ARMv7]
  - Cortex-A9    [ARMv7]
  - Cortex-A17   [ARMv7]
  - Cortex-A53   [ARMv8]
  - Cortex-A55   [ARMv8]
  - Cortex-A73   [ARMv8]
  - Cortex-A75   [ARMv8]

Interestingly, Andy Polyakov's scalar code is slower than Eric Biggers', but is also significantly shorter. This has the advantage that it does not evict other code from L1 cache -- particularly on ARM11 chips -- and so in certain circumstances it can actually be faster. However, it wasn't found that this had an affect on any code existing in the kernel today.

Cc Russell King <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>

---

Notes:
Eric Biggers' scalar code is brand new, and quite possibly prematurely added to this commit, and so it may require a bit of revision. In initial evaluation and fuzzing so far, it seems fine. But we'll be looking at this a bit more as well.

 lib/zinc/Makefile                             |   2 +
 lib/zinc/chacha20/chacha20-arm-glue.c         |  98 ++++
 ...acha20-arm-cryptogams.S => chacha20-arm.S} | 503 ++++++++++++++++--
 ...20-arm64-cryptogams.S => chacha20-arm64.S} |   0
 lib/zinc/chacha20/chacha20.c                  |   2 +
 5 files changed, 567 insertions(+), 38 deletions(-)
 create mode 100644 lib/zinc/chacha20/chacha20-arm-glue.c
 rename lib/zinc/chacha20/{chacha20-arm-cryptogams.S => chacha20-arm.S} (71%)
 rename lib/zinc/chacha20/{chacha20-arm64-cryptogams.S => chacha20-arm64.S} (100%)
lib/zinc/Makefile +2 -0
index 223a0816c918..e47f64e12bbd 100644
@@ -4,4 +4,6 @@ ccflags-$(CONFIG_ZINC_DEBUG) += -DDEBUG
 
 zinc_chacha20-y := chacha20/chacha20.o
 zinc_chacha20-$(CONFIG_ZINC_ARCH_X86_64) += chacha20/chacha20-x86_64.o
+zinc_chacha20-$(CONFIG_ZINC_ARCH_ARM) += chacha20/chacha20-arm.o
+zinc_chacha20-$(CONFIG_ZINC_ARCH_ARM64) += chacha20/chacha20-arm64.o
 obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
lib/zinc/chacha20/chacha20-arm-glue.c new file +98 -0

98 lines not on this page · raw message

lib/zinc/chacha20/chacha20-arm.S renamed from lib/zinc/chacha20/chacha20-arm-cryptogams.S +465 -38

527 lines not on this page · raw message

lib/zinc/chacha20/chacha20-arm64.S renamed from lib/zinc/chacha20/chacha20-arm64-cryptogams.S +0 -0
similarity index 100%
rename from lib/zinc/chacha20/chacha20-arm64-cryptogams.S
rename to lib/zinc/chacha20/chacha20-arm64.S
lib/zinc/chacha20/chacha20.c +2 -0
index 22a21431c221..3698fcd8ae7f 100644
@@ -18,6 +18,8 @@
 
 #if defined(CONFIG_ZINC_ARCH_X86_64)
 #include "chacha20-x86_64-glue.c"
+#elif defined(CONFIG_ZINC_ARCH_ARM) || defined(CONFIG_ZINC_ARCH_ARM64)
+#include "chacha20-arm-glue.c"
 #else
 static bool *const chacha20_nobs[] __initconst = { };
 static void __init chacha20_fpu_init(void)
Signature · 1 line
2.19.1
author maintainer
#10 lore raw

PATCH v8 10/28 zinc: ChaCha20 MIPS32r2 implementation

In reply to Jason A. Donenfeld

This MIPS32r2 implementation comes from René van Dorst and me and results in a nice speedup on the usual OpenWRT targets.

Cc Ralf Baechle <[email protected]>
Cc Paul Burton <[email protected]>
Cc James Hogan <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                      |   2 +
 lib/zinc/chacha20/chacha20-mips-glue.c |  28 ++
 lib/zinc/chacha20/chacha20-mips.S      | 424 +++++++++++++++++++++++++
 lib/zinc/chacha20/chacha20.c           |   2 +
 4 files changed, 456 insertions(+)
 create mode 100644 lib/zinc/chacha20/chacha20-mips-glue.c
 create mode 100644 lib/zinc/chacha20/chacha20-mips.S
lib/zinc/Makefile +2 -0
index e47f64e12bbd..60d568cf5206 100644
@@ -6,4 +6,6 @@ zinc_chacha20-y := chacha20/chacha20.o
 zinc_chacha20-$(CONFIG_ZINC_ARCH_X86_64) += chacha20/chacha20-x86_64.o
 zinc_chacha20-$(CONFIG_ZINC_ARCH_ARM) += chacha20/chacha20-arm.o
 zinc_chacha20-$(CONFIG_ZINC_ARCH_ARM64) += chacha20/chacha20-arm64.o
+zinc_chacha20-$(CONFIG_ZINC_ARCH_MIPS) += chacha20/chacha20-mips.o
+AFLAGS_chacha20-mips.o += -O2 # This is required to fill the branch delay slots
 obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
lib/zinc/chacha20/chacha20-mips-glue.c new file +28 -0
new file mode 100644
index 000000000000..917d8fa8e3f4
@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+asmlinkage void chacha20_mips(u32 state[16], u8 *out, const u8 *in,
+			      const size_t len);
+static bool *const chacha20_nobs[] __initconst = { };
+static void __init chacha20_fpu_init(void)
+{
+}
+
+static inline bool chacha20_arch(struct chacha20_ctx *ctx, u8 *dst,
+				 const u8 *src, size_t len,
+				 simd_context_t *simd_context)
+{
+	chacha20_mips(ctx->state, dst, src, len);
+	return true;
+}
+
+
+static inline bool hchacha20_arch(u32 derived_key[CHACHA20_KEY_WORDS],
+				  const u8 nonce[HCHACHA20_NONCE_SIZE],
+				  const u8 key[HCHACHA20_KEY_SIZE],
+				  simd_context_t *simd_context)
+{
+	return false;
+}
lib/zinc/chacha20/chacha20-mips.S new file +424 -0

424 lines not on this page · raw message

lib/zinc/chacha20/chacha20.c +2 -0
index 3698fcd8ae7f..0b833310a7d8 100644
@@ -20,6 +20,8 @@
 #include "chacha20-x86_64-glue.c"
 #elif defined(CONFIG_ZINC_ARCH_ARM) || defined(CONFIG_ZINC_ARCH_ARM64)
 #include "chacha20-arm-glue.c"
+#elif defined(CONFIG_ZINC_ARCH_MIPS)
+#include "chacha20-mips-glue.c"
 #else
 static bool *const chacha20_nobs[] __initconst = { };
 static void __init chacha20_fpu_init(void)
Signature · 1 line
2.19.1
author maintainer
#11 lore raw

PATCH v8 11/28 zinc: Poly1305 generic C implementations and selftest

In reply to Jason A. Donenfeld

These two C implementations -- a 32x32 one and a 64x64 one, depending on the platform -- come from Andrew Moon's public domain poly1305-donna portable code, modified for usage in the kernel and for usage with accelerated primitives.

Information: https://cr.yp.to/mac.html

Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 include/zinc/poly1305.h              |   31 +
 lib/zinc/Kconfig                     |    3 +
 lib/zinc/Makefile                    |    3 +
 lib/zinc/poly1305/poly1305-donna32.c |  205 +++++
 lib/zinc/poly1305/poly1305-donna64.c |  182 +++++
 lib/zinc/poly1305/poly1305.c         |  154 ++++
 lib/zinc/selftest/poly1305.c         | 1107 ++++++++++++++++++++++++++
 7 files changed, 1685 insertions(+)
 create mode 100644 include/zinc/poly1305.h
 create mode 100644 lib/zinc/poly1305/poly1305-donna32.c
 create mode 100644 lib/zinc/poly1305/poly1305-donna64.c
 create mode 100644 lib/zinc/poly1305/poly1305.c
 create mode 100644 lib/zinc/selftest/poly1305.c
include/zinc/poly1305.h new file +31 -0
new file mode 100644
index 000000000000..13fe0e50fc3c
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _ZINC_POLY1305_H
+#define _ZINC_POLY1305_H
+
+#include <linux/simd.h>
+#include <linux/types.h>
+
+enum poly1305_lengths {
+	POLY1305_BLOCK_SIZE = 16,
+	POLY1305_KEY_SIZE = 32,
+	POLY1305_MAC_SIZE = 16
+};
+
+struct poly1305_ctx {
+	u8 opaque[24 * sizeof(u64)];
+	u32 nonce[4];
+	u8 data[POLY1305_BLOCK_SIZE];
+	size_t num;
+} __aligned(8);
+
+void poly1305_init(struct poly1305_ctx *ctx, const u8 key[POLY1305_KEY_SIZE]);
+void poly1305_update(struct poly1305_ctx *ctx, const u8 *input, size_t len,
+		     simd_context_t *simd_context);
+void poly1305_final(struct poly1305_ctx *ctx, u8 mac[POLY1305_MAC_SIZE],
+		    simd_context_t *simd_context);
+
+#endif /* _ZINC_POLY1305_H */
lib/zinc/Kconfig +3 -0
index d271be37cecb..f08bf1eaa2a0 100644
@@ -2,6 +2,9 @@ config ZINC_CHACHA20
 	tristate
 	select CRYPTO_ALGAPI
 
+config ZINC_POLY1305
+	tristate
+
 config ZINC_SELFTEST
 	bool "Zinc cryptography library self-tests"
 	help
lib/zinc/Makefile +3 -0
index 60d568cf5206..6fc9626c55fa 100644
@@ -9,3 +9,6 @@ zinc_chacha20-$(CONFIG_ZINC_ARCH_ARM64) += chacha20/chacha20-arm64.o
 zinc_chacha20-$(CONFIG_ZINC_ARCH_MIPS) += chacha20/chacha20-mips.o
 AFLAGS_chacha20-mips.o += -O2 # This is required to fill the branch delay slots
 obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
+
+zinc_poly1305-y := poly1305/poly1305.o
+obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
lib/zinc/poly1305/poly1305-donna32.c new file +205 -0

205 lines not on this page · raw message

lib/zinc/poly1305/poly1305-donna64.c new file +182 -0

182 lines not on this page · raw message

lib/zinc/poly1305/poly1305.c new file +154 -0

154 lines not on this page · raw message

lib/zinc/selftest/poly1305.c new file +1107 -0

1,107 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#12 lore raw

PATCH v8 12/28 zinc: import Andy Polyakov's Poly1305 x86_64 implementation

In reply to Jason A. Donenfeld

These x86_64 vectorized implementations come from Andy Polyakov's implementation, and are included here in raw form without modification, so that subsequent commits that fix these up for the kernel can see how it has changed.

While this is CRYPTOGAMS code, the originating code for this happens to be the same as OpenSSL's commit 4dfe4310c31c4483705991d9a798ce9be1ed1c68

Based-on-code-from: Andy Polyakov <[email protected]>
Cc: Andy Polyakov <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: [email protected]
Cc: Samuel Neves <[email protected]>
Cc: Jean-Philippe Aumasson <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 .../poly1305/poly1305-x86_64-cryptogams.S     | 3565 +++++++++++++++++
 1 file changed, 3565 insertions(+)
 create mode 100644 lib/zinc/poly1305/poly1305-x86_64-cryptogams.S
lib/zinc/poly1305/poly1305-x86_64-cryptogams.S new file +3565 -0
new file mode 100644
index 000000000000..ed634757354b
@@ -0,0 +1,3565 @@
+/* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
+/*
+ * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ */
+
+.text	
+
+
+
+.globl	poly1305_init
+.hidden	poly1305_init
+.globl	poly1305_blocks
+.hidden	poly1305_blocks
+.globl	poly1305_emit
+.hidden	poly1305_emit
+
+.type	poly1305_init,@function
+.align	32
+poly1305_init:
+	xorq	%rax,%rax
+	movq	%rax,0(%rdi)
+	movq	%rax,8(%rdi)
+	movq	%rax,16(%rdi)
+
+	cmpq	$0,%rsi
+	je	.Lno_key
+
+	leaq	poly1305_blocks(%rip),%r10
+	leaq	poly1305_emit(%rip),%r11
+	movq	OPENSSL_ia32cap_P+4(%rip),%r9
+	leaq	poly1305_blocks_avx(%rip),%rax
+	leaq	poly1305_emit_avx(%rip),%rcx
+	btq	$28,%r9
+	cmovcq	%rax,%r10
+	cmovcq	%rcx,%r11
+	leaq	poly1305_blocks_avx2(%rip),%rax
+	btq	$37,%r9
+	cmovcq	%rax,%r10
+	movq	$2149646336,%rax
+	shrq	$32,%r9
+	andq	%rax,%r9
+	cmpq	%rax,%r9
+	je	.Linit_base2_44
+	movq	$0x0ffffffc0fffffff,%rax
+	movq	$0x0ffffffc0ffffffc,%rcx
+	andq	0(%rsi),%rax
+	andq	8(%rsi),%rcx
+	movq	%rax,24(%rdi)
+	movq	%rcx,32(%rdi)
+	movq	%r10,0(%rdx)
+	movq	%r11,8(%rdx)
+	movl	$1,%eax
+.Lno_key:
+	.byte	0xf3,0xc3
+.size	poly1305_init,.-poly1305_init
+
+.type	poly1305_blocks,@function
+.align	32
+poly1305_blocks:
+.cfi_startproc	
+.Lblocks:
+	shrq	$4,%rdx
+	jz	.Lno_data
+
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+.Lblocks_body:
+
+	movq	%rdx,%r15
+
+	movq	24(%rdi),%r11
+	movq	32(%rdi),%r13
+
+	movq	0(%rdi),%r14
+	movq	8(%rdi),%rbx
+	movq	16(%rdi),%rbp
+
+	movq	%r13,%r12
+	shrq	$2,%r13
+	movq	%r12,%rax
+	addq	%r12,%r13
+	jmp	.Loop
+
+.align	32
+.Loop:
+	addq	0(%rsi),%r14
+	adcq	8(%rsi),%rbx
+	leaq	16(%rsi),%rsi
+	adcq	%rcx,%rbp
+	mulq	%r14
+	movq	%rax,%r9
+	movq	%r11,%rax
+	movq	%rdx,%r10
+
+	mulq	%r14
+	movq	%rax,%r14
+	movq	%r11,%rax
+	movq	%rdx,%r8
+
+	mulq	%rbx
+	addq	%rax,%r9
+	movq	%r13,%rax
+	adcq	%rdx,%r10
+
+	mulq	%rbx
+	movq	%rbp,%rbx
+	addq	%rax,%r14
+	adcq	%rdx,%r8
+
+	imulq	%r13,%rbx
+	addq	%rbx,%r9
+	movq	%r8,%rbx
+	adcq	$0,%r10
+
+	imulq	%r11,%rbp
+	addq	%r9,%rbx
+	movq	$-4,%rax
+	adcq	%rbp,%r10
+
+	andq	%r10,%rax
+	movq	%r10,%rbp
+	shrq	$2,%r10
+	andq	$3,%rbp
+	addq	%r10,%rax
+	addq	%rax,%r14
+	adcq	$0,%rbx
+	adcq	$0,%rbp
+	movq	%r12,%rax
+	decq	%r15
+	jnz	.Loop
+
+	movq	%r14,0(%rdi)
+	movq	%rbx,8(%rdi)
+	movq	%rbp,16(%rdi)
+
+	movq	0(%rsp),%r15
+.cfi_restore	%r15
+	movq	8(%rsp),%r14
+.cfi_restore	%r14
+	movq	16(%rsp),%r13
+.cfi_restore	%r13
+	movq	24(%rsp),%r12
+.cfi_restore	%r12
+	movq	32(%rsp),%rbp
+.cfi_restore	%rbp
+	movq	40(%rsp),%rbx
+.cfi_restore	%rbx
+	leaq	48(%rsp),%rsp
+.cfi_adjust_cfa_offset	-48
+.Lno_data:
+.Lblocks_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	poly1305_blocks,.-poly1305_blocks
+
+.type	poly1305_emit,@function
+.align	32
+poly1305_emit:
+.Lemit:
+	movq	0(%rdi),%r8
+	movq	8(%rdi),%r9
+	movq	16(%rdi),%r10
+
+	movq	%r8,%rax
+	addq	$5,%r8
+	movq	%r9,%rcx
+	adcq	$0,%r9
+	adcq	$0,%r10
+	shrq	$2,%r10
+	cmovnzq	%r8,%rax
+	cmovnzq	%r9,%rcx
+
+	addq	0(%rdx),%rax
+	adcq	8(%rdx),%rcx
+	movq	%rax,0(%rsi)
+	movq	%rcx,8(%rsi)
+
+	.byte	0xf3,0xc3
+.size	poly1305_emit,.-poly1305_emit
+.type	__poly1305_block,@function
+.align	32
+__poly1305_block:
+	mulq	%r14
+	movq	%rax,%r9
+	movq	%r11,%rax
+	movq	%rdx,%r10
+
+	mulq	%r14
+	movq	%rax,%r14
+	movq	%r11,%rax
+	movq	%rdx,%r8
+
+	mulq	%rbx
+	addq	%rax,%r9
+	movq	%r13,%rax
+	adcq	%rdx,%r10
+
+	mulq	%rbx
+	movq	%rbp,%rbx
+	addq	%rax,%r14
+	adcq	%rdx,%r8
+
+	imulq	%r13,%rbx
+	addq	%rbx,%r9
+	movq	%r8,%rbx
+	adcq	$0,%r10
+
+	imulq	%r11,%rbp
+	addq	%r9,%rbx
+	movq	$-4,%rax
+	adcq	%rbp,%r10
+
+	andq	%r10,%rax
+	movq	%r10,%rbp
+	shrq	$2,%r10
+	andq	$3,%rbp
+	addq	%r10,%rax
+	addq	%rax,%r14
+	adcq	$0,%rbx
+	adcq	$0,%rbp
+	.byte	0xf3,0xc3
+.size	__poly1305_block,.-__poly1305_block
+
+.type	__poly1305_init_avx,@function
+.align	32
+__poly1305_init_avx:
+	movq	%r11,%r14
+	movq	%r12,%rbx
+	xorq	%rbp,%rbp
+
+	leaq	48+64(%rdi),%rdi
+
+	movq	%r12,%rax
+	call	__poly1305_block
+
+	movl	$0x3ffffff,%eax
+	movl	$0x3ffffff,%edx
+	movq	%r14,%r8
+	andl	%r14d,%eax
+	movq	%r11,%r9
+	andl	%r11d,%edx
+	movl	%eax,-64(%rdi)
+	shrq	$26,%r8
+	movl	%edx,-60(%rdi)
+	shrq	$26,%r9
+
+	movl	$0x3ffffff,%eax
+	movl	$0x3ffffff,%edx
+	andl	%r8d,%eax
+	andl	%r9d,%edx
+	movl	%eax,-48(%rdi)
+	leal	(%rax,%rax,4),%eax
+	movl	%edx,-44(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	movl	%eax,-32(%rdi)
+	shrq	$26,%r8
+	movl	%edx,-28(%rdi)
+	shrq	$26,%r9
+
+	movq	%rbx,%rax
+	movq	%r12,%rdx
+	shlq	$12,%rax
+	shlq	$12,%rdx
+	orq	%r8,%rax
+	orq	%r9,%rdx
+	andl	$0x3ffffff,%eax
+	andl	$0x3ffffff,%edx
+	movl	%eax,-16(%rdi)
+	leal	(%rax,%rax,4),%eax
+	movl	%edx,-12(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	movl	%eax,0(%rdi)
+	movq	%rbx,%r8
+	movl	%edx,4(%rdi)
+	movq	%r12,%r9
+
+	movl	$0x3ffffff,%eax
+	movl	$0x3ffffff,%edx
+	shrq	$14,%r8
+	shrq	$14,%r9
+	andl	%r8d,%eax
+	andl	%r9d,%edx
+	movl	%eax,16(%rdi)
+	leal	(%rax,%rax,4),%eax
+	movl	%edx,20(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	movl	%eax,32(%rdi)
+	shrq	$26,%r8
+	movl	%edx,36(%rdi)
+	shrq	$26,%r9
+
+	movq	%rbp,%rax
+	shlq	$24,%rax
+	orq	%rax,%r8
+	movl	%r8d,48(%rdi)
+	leaq	(%r8,%r8,4),%r8
+	movl	%r9d,52(%rdi)
+	leaq	(%r9,%r9,4),%r9
+	movl	%r8d,64(%rdi)
+	movl	%r9d,68(%rdi)
+
+	movq	%r12,%rax
+	call	__poly1305_block
+
+	movl	$0x3ffffff,%eax
+	movq	%r14,%r8
+	andl	%r14d,%eax
+	shrq	$26,%r8
+	movl	%eax,-52(%rdi)
+
+	movl	$0x3ffffff,%edx
+	andl	%r8d,%edx
+	movl	%edx,-36(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	shrq	$26,%r8
+	movl	%edx,-20(%rdi)
+
+	movq	%rbx,%rax
+	shlq	$12,%rax
+	orq	%r8,%rax
+	andl	$0x3ffffff,%eax
+	movl	%eax,-4(%rdi)
+	leal	(%rax,%rax,4),%eax
+	movq	%rbx,%r8
+	movl	%eax,12(%rdi)
+
+	movl	$0x3ffffff,%edx
+	shrq	$14,%r8
+	andl	%r8d,%edx
+	movl	%edx,28(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	shrq	$26,%r8
+	movl	%edx,44(%rdi)
+
+	movq	%rbp,%rax
+	shlq	$24,%rax
+	orq	%rax,%r8
+	movl	%r8d,60(%rdi)
+	leaq	(%r8,%r8,4),%r8
+	movl	%r8d,76(%rdi)
+
+	movq	%r12,%rax
+	call	__poly1305_block
+
+	movl	$0x3ffffff,%eax
+	movq	%r14,%r8
+	andl	%r14d,%eax
+	shrq	$26,%r8
+	movl	%eax,-56(%rdi)
+
+	movl	$0x3ffffff,%edx
+	andl	%r8d,%edx
+	movl	%edx,-40(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	shrq	$26,%r8
+	movl	%edx,-24(%rdi)
+
+	movq	%rbx,%rax
+	shlq	$12,%rax
+	orq	%r8,%rax
+	andl	$0x3ffffff,%eax
+	movl	%eax,-8(%rdi)
+	leal	(%rax,%rax,4),%eax
+	movq	%rbx,%r8
+	movl	%eax,8(%rdi)
+
+	movl	$0x3ffffff,%edx
+	shrq	$14,%r8
+	andl	%r8d,%edx
+	movl	%edx,24(%rdi)
+	leal	(%rdx,%rdx,4),%edx
+	shrq	$26,%r8
+	movl	%edx,40(%rdi)
+
+	movq	%rbp,%rax
+	shlq	$24,%rax
+	orq	%rax,%r8
+	movl	%r8d,56(%rdi)
+	leaq	(%r8,%r8,4),%r8
+	movl	%r8d,72(%rdi)
+
+	leaq	-48-64(%rdi),%rdi
+	.byte	0xf3,0xc3
+.size	__poly1305_init_avx,.-__poly1305_init_avx
+
+.type	poly1305_blocks_avx,@function
+.align	32
+poly1305_blocks_avx:
+.cfi_startproc	
+	movl	20(%rdi),%r8d
+	cmpq	$128,%rdx
+	jae	.Lblocks_avx
+	testl	%r8d,%r8d
+	jz	.Lblocks
+
+.Lblocks_avx:
+	andq	$-16,%rdx
+	jz	.Lno_data_avx
+
+	vzeroupper
+
+	testl	%r8d,%r8d
+	jz	.Lbase2_64_avx
+
+	testq	$31,%rdx
+	jz	.Leven_avx
+
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+.Lblocks_avx_body:
+
+	movq	%rdx,%r15
+
+	movq	0(%rdi),%r8
+	movq	8(%rdi),%r9
+	movl	16(%rdi),%ebp
+
+	movq	24(%rdi),%r11
+	movq	32(%rdi),%r13
+
+
+	movl	%r8d,%r14d
+	andq	$-2147483648,%r8
+	movq	%r9,%r12
+	movl	%r9d,%ebx
+	andq	$-2147483648,%r9
+
+	shrq	$6,%r8
+	shlq	$52,%r12
+	addq	%r8,%r14
+	shrq	$12,%rbx
+	shrq	$18,%r9
+	addq	%r12,%r14
+	adcq	%r9,%rbx
+
+	movq	%rbp,%r8
+	shlq	$40,%r8
+	shrq	$24,%rbp
+	addq	%r8,%rbx
+	adcq	$0,%rbp
+
+	movq	$-4,%r9
+	movq	%rbp,%r8
+	andq	%rbp,%r9
+	shrq	$2,%r8
+	andq	$3,%rbp
+	addq	%r9,%r8
+	addq	%r8,%r14
+	adcq	$0,%rbx
+	adcq	$0,%rbp
+
+	movq	%r13,%r12
+	movq	%r13,%rax
+	shrq	$2,%r13
+	addq	%r12,%r13
+
+	addq	0(%rsi),%r14
+	adcq	8(%rsi),%rbx
+	leaq	16(%rsi),%rsi
+	adcq	%rcx,%rbp
+
+	call	__poly1305_block
+
+	testq	%rcx,%rcx
+	jz	.Lstore_base2_64_avx
+
+
+	movq	%r14,%rax
+	movq	%r14,%rdx
+	shrq	$52,%r14
+	movq	%rbx,%r11
+	movq	%rbx,%r12
+	shrq	$26,%rdx
+	andq	$0x3ffffff,%rax
+	shlq	$12,%r11
+	andq	$0x3ffffff,%rdx
+	shrq	$14,%rbx
+	orq	%r11,%r14
+	shlq	$24,%rbp
+	andq	$0x3ffffff,%r14
+	shrq	$40,%r12
+	andq	$0x3ffffff,%rbx
+	orq	%r12,%rbp
+
+	subq	$16,%r15
+	jz	.Lstore_base2_26_avx
+
+	vmovd	%eax,%xmm0
+	vmovd	%edx,%xmm1
+	vmovd	%r14d,%xmm2
+	vmovd	%ebx,%xmm3
+	vmovd	%ebp,%xmm4
+	jmp	.Lproceed_avx
+
+.align	32
+.Lstore_base2_64_avx:
+	movq	%r14,0(%rdi)
+	movq	%rbx,8(%rdi)
+	movq	%rbp,16(%rdi)
+	jmp	.Ldone_avx
+
+.align	16
+.Lstore_base2_26_avx:
+	movl	%eax,0(%rdi)
+	movl	%edx,4(%rdi)
+	movl	%r14d,8(%rdi)
+	movl	%ebx,12(%rdi)
+	movl	%ebp,16(%rdi)
+.align	16
+.Ldone_avx:
+	movq	0(%rsp),%r15
+.cfi_restore	%r15
+	movq	8(%rsp),%r14
+.cfi_restore	%r14
+	movq	16(%rsp),%r13
+.cfi_restore	%r13
+	movq	24(%rsp),%r12
+.cfi_restore	%r12
+	movq	32(%rsp),%rbp
+.cfi_restore	%rbp
+	movq	40(%rsp),%rbx
+.cfi_restore	%rbx
+	leaq	48(%rsp),%rsp
+.cfi_adjust_cfa_offset	-48
+.Lno_data_avx:
+.Lblocks_avx_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+
+.align	32
+.Lbase2_64_avx:
+.cfi_startproc	
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+.Lbase2_64_avx_body:
+
+	movq	%rdx,%r15
+
+	movq	24(%rdi),%r11
+	movq	32(%rdi),%r13
+
+	movq	0(%rdi),%r14
+	movq	8(%rdi),%rbx
+	movl	16(%rdi),%ebp
+
+	movq	%r13,%r12
+	movq	%r13,%rax
+	shrq	$2,%r13
+	addq	%r12,%r13
+
+	testq	$31,%rdx
+	jz	.Linit_avx
+
+	addq	0(%rsi),%r14
+	adcq	8(%rsi),%rbx
+	leaq	16(%rsi),%rsi
+	adcq	%rcx,%rbp
+	subq	$16,%r15
+
+	call	__poly1305_block
+
+.Linit_avx:
+
+	movq	%r14,%rax
+	movq	%r14,%rdx
+	shrq	$52,%r14
+	movq	%rbx,%r8
+	movq	%rbx,%r9
+	shrq	$26,%rdx
+	andq	$0x3ffffff,%rax
+	shlq	$12,%r8
+	andq	$0x3ffffff,%rdx
+	shrq	$14,%rbx
+	orq	%r8,%r14
+	shlq	$24,%rbp
+	andq	$0x3ffffff,%r14
+	shrq	$40,%r9
+	andq	$0x3ffffff,%rbx
+	orq	%r9,%rbp
+
+	vmovd	%eax,%xmm0
+	vmovd	%edx,%xmm1
+	vmovd	%r14d,%xmm2
+	vmovd	%ebx,%xmm3
+	vmovd	%ebp,%xmm4
+	movl	$1,20(%rdi)
+
+	call	__poly1305_init_avx
+
+.Lproceed_avx:
+	movq	%r15,%rdx
+
+	movq	0(%rsp),%r15
+.cfi_restore	%r15
+	movq	8(%rsp),%r14
+.cfi_restore	%r14
+	movq	16(%rsp),%r13
+.cfi_restore	%r13
+	movq	24(%rsp),%r12
+.cfi_restore	%r12
+	movq	32(%rsp),%rbp
+.cfi_restore	%rbp
+	movq	40(%rsp),%rbx
+.cfi_restore	%rbx
+	leaq	48(%rsp),%rax
+	leaq	48(%rsp),%rsp
+.cfi_adjust_cfa_offset	-48
+.Lbase2_64_avx_epilogue:
+	jmp	.Ldo_avx
+.cfi_endproc	
+
+.align	32
+.Leven_avx:
+.cfi_startproc	
+	vmovd	0(%rdi),%xmm0
+	vmovd	4(%rdi),%xmm1
+	vmovd	8(%rdi),%xmm2
+	vmovd	12(%rdi),%xmm3
+	vmovd	16(%rdi),%xmm4
+
+.Ldo_avx:
+	leaq	-88(%rsp),%r11
+.cfi_def_cfa	%r11,0x60
+	subq	$0x178,%rsp
+	subq	$64,%rdx
+	leaq	-32(%rsi),%rax
+	cmovcq	%rax,%rsi
+
+	vmovdqu	48(%rdi),%xmm14
+	leaq	112(%rdi),%rdi
+	leaq	.Lconst(%rip),%rcx
+
+
+
+	vmovdqu	32(%rsi),%xmm5
+	vmovdqu	48(%rsi),%xmm6
+	vmovdqa	64(%rcx),%xmm15
+
+	vpsrldq	$6,%xmm5,%xmm7
+	vpsrldq	$6,%xmm6,%xmm8
+	vpunpckhqdq	%xmm6,%xmm5,%xmm9
+	vpunpcklqdq	%xmm6,%xmm5,%xmm5
+	vpunpcklqdq	%xmm8,%xmm7,%xmm8
+
+	vpsrlq	$40,%xmm9,%xmm9
+	vpsrlq	$26,%xmm5,%xmm6
+	vpand	%xmm15,%xmm5,%xmm5
+	vpsrlq	$4,%xmm8,%xmm7
+	vpand	%xmm15,%xmm6,%xmm6
+	vpsrlq	$30,%xmm8,%xmm8
+	vpand	%xmm15,%xmm7,%xmm7
+	vpand	%xmm15,%xmm8,%xmm8
+	vpor	32(%rcx),%xmm9,%xmm9
+
+	jbe	.Lskip_loop_avx
+
+
+	vmovdqu	-48(%rdi),%xmm11
+	vmovdqu	-32(%rdi),%xmm12
+	vpshufd	$0xEE,%xmm14,%xmm13
+	vpshufd	$0x44,%xmm14,%xmm10
+	vmovdqa	%xmm13,-144(%r11)
+	vmovdqa	%xmm10,0(%rsp)
+	vpshufd	$0xEE,%xmm11,%xmm14
+	vmovdqu	-16(%rdi),%xmm10
+	vpshufd	$0x44,%xmm11,%xmm11
+	vmovdqa	%xmm14,-128(%r11)
+	vmovdqa	%xmm11,16(%rsp)
+	vpshufd	$0xEE,%xmm12,%xmm13
+	vmovdqu	0(%rdi),%xmm11
+	vpshufd	$0x44,%xmm12,%xmm12
+	vmovdqa	%xmm13,-112(%r11)
+	vmovdqa	%xmm12,32(%rsp)
+	vpshufd	$0xEE,%xmm10,%xmm14
+	vmovdqu	16(%rdi),%xmm12
+	vpshufd	$0x44,%xmm10,%xmm10
+	vmovdqa	%xmm14,-96(%r11)
+	vmovdqa	%xmm10,48(%rsp)
+	vpshufd	$0xEE,%xmm11,%xmm13
+	vmovdqu	32(%rdi),%xmm10
+	vpshufd	$0x44,%xmm11,%xmm11
+	vmovdqa	%xmm13,-80(%r11)
+	vmovdqa	%xmm11,64(%rsp)
+	vpshufd	$0xEE,%xmm12,%xmm14
+	vmovdqu	48(%rdi),%xmm11
+	vpshufd	$0x44,%xmm12,%xmm12
+	vmovdqa	%xmm14,-64(%r11)
+	vmovdqa	%xmm12,80(%rsp)
+	vpshufd	$0xEE,%xmm10,%xmm13
+	vmovdqu	64(%rdi),%xmm12
+	vpshufd	$0x44,%xmm10,%xmm10
+	vmovdqa	%xmm13,-48(%r11)
+	vmovdqa	%xmm10,96(%rsp)
+	vpshufd	$0xEE,%xmm11,%xmm14
+	vpshufd	$0x44,%xmm11,%xmm11
+	vmovdqa	%xmm14,-32(%r11)
+	vmovdqa	%xmm11,112(%rsp)
+	vpshufd	$0xEE,%xmm12,%xmm13
+	vmovdqa	0(%rsp),%xmm14
+	vpshufd	$0x44,%xmm12,%xmm12
+	vmovdqa	%xmm13,-16(%r11)
+	vmovdqa	%xmm12,128(%rsp)
+
+	jmp	.Loop_avx
+
+.align	32
+.Loop_avx:
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+	vpmuludq	%xmm5,%xmm14,%xmm10
+	vpmuludq	%xmm6,%xmm14,%xmm11
+	vmovdqa	%xmm2,32(%r11)
+	vpmuludq	%xmm7,%xmm14,%xmm12
+	vmovdqa	16(%rsp),%xmm2
+	vpmuludq	%xmm8,%xmm14,%xmm13
+	vpmuludq	%xmm9,%xmm14,%xmm14
+
+	vmovdqa	%xmm0,0(%r11)
+	vpmuludq	32(%rsp),%xmm9,%xmm0
+	vmovdqa	%xmm1,16(%r11)
+	vpmuludq	%xmm8,%xmm2,%xmm1
+	vpaddq	%xmm0,%xmm10,%xmm10
+	vpaddq	%xmm1,%xmm14,%xmm14
+	vmovdqa	%xmm3,48(%r11)
+	vpmuludq	%xmm7,%xmm2,%xmm0
+	vpmuludq	%xmm6,%xmm2,%xmm1
+	vpaddq	%xmm0,%xmm13,%xmm13
+	vmovdqa	48(%rsp),%xmm3
+	vpaddq	%xmm1,%xmm12,%xmm12
+	vmovdqa	%xmm4,64(%r11)
+	vpmuludq	%xmm5,%xmm2,%xmm2
+	vpmuludq	%xmm7,%xmm3,%xmm0
+	vpaddq	%xmm2,%xmm11,%xmm11
+
+	vmovdqa	64(%rsp),%xmm4
+	vpaddq	%xmm0,%xmm14,%xmm14
+	vpmuludq	%xmm6,%xmm3,%xmm1
+	vpmuludq	%xmm5,%xmm3,%xmm3
+	vpaddq	%xmm1,%xmm13,%xmm13
+	vmovdqa	80(%rsp),%xmm2
+	vpaddq	%xmm3,%xmm12,%xmm12
+	vpmuludq	%xmm9,%xmm4,%xmm0
+	vpmuludq	%xmm8,%xmm4,%xmm4
+	vpaddq	%xmm0,%xmm11,%xmm11
+	vmovdqa	96(%rsp),%xmm3
+	vpaddq	%xmm4,%xmm10,%xmm10
+
+	vmovdqa	128(%rsp),%xmm4
+	vpmuludq	%xmm6,%xmm2,%xmm1
+	vpmuludq	%xmm5,%xmm2,%xmm2
+	vpaddq	%xmm1,%xmm14,%xmm14
+	vpaddq	%xmm2,%xmm13,%xmm13
+	vpmuludq	%xmm9,%xmm3,%xmm0
+	vpmuludq	%xmm8,%xmm3,%xmm1
+	vpaddq	%xmm0,%xmm12,%xmm12
+	vmovdqu	0(%rsi),%xmm0
+	vpaddq	%xmm1,%xmm11,%xmm11
+	vpmuludq	%xmm7,%xmm3,%xmm3
+	vpmuludq	%xmm7,%xmm4,%xmm7
+	vpaddq	%xmm3,%xmm10,%xmm10
+
+	vmovdqu	16(%rsi),%xmm1
+	vpaddq	%xmm7,%xmm11,%xmm11
+	vpmuludq	%xmm8,%xmm4,%xmm8
+	vpmuludq	%xmm9,%xmm4,%xmm9
+	vpsrldq	$6,%xmm0,%xmm2
+	vpaddq	%xmm8,%xmm12,%xmm12
+	vpaddq	%xmm9,%xmm13,%xmm13
+	vpsrldq	$6,%xmm1,%xmm3
+	vpmuludq	112(%rsp),%xmm5,%xmm9
+	vpmuludq	%xmm6,%xmm4,%xmm5
+	vpunpckhqdq	%xmm1,%xmm0,%xmm4
+	vpaddq	%xmm9,%xmm14,%xmm14
+	vmovdqa	-144(%r11),%xmm9
+	vpaddq	%xmm5,%xmm10,%xmm10
+
+	vpunpcklqdq	%xmm1,%xmm0,%xmm0
+	vpunpcklqdq	%xmm3,%xmm2,%xmm3
+
+
+	vpsrldq	$5,%xmm4,%xmm4
+	vpsrlq	$26,%xmm0,%xmm1
+	vpand	%xmm15,%xmm0,%xmm0
+	vpsrlq	$4,%xmm3,%xmm2
+	vpand	%xmm15,%xmm1,%xmm1
+	vpand	0(%rcx),%xmm4,%xmm4
+	vpsrlq	$30,%xmm3,%xmm3
+	vpand	%xmm15,%xmm2,%xmm2
+	vpand	%xmm15,%xmm3,%xmm3
+	vpor	32(%rcx),%xmm4,%xmm4
+
+	vpaddq	0(%r11),%xmm0,%xmm0
+	vpaddq	16(%r11),%xmm1,%xmm1
+	vpaddq	32(%r11),%xmm2,%xmm2
+	vpaddq	48(%r11),%xmm3,%xmm3
+	vpaddq	64(%r11),%xmm4,%xmm4
+
+	leaq	32(%rsi),%rax
+	leaq	64(%rsi),%rsi
+	subq	$64,%rdx
+	cmovcq	%rax,%rsi
+
+
+
+
+
+
+
+
+
+
+	vpmuludq	%xmm0,%xmm9,%xmm5
+	vpmuludq	%xmm1,%xmm9,%xmm6
+	vpaddq	%xmm5,%xmm10,%xmm10
+	vpaddq	%xmm6,%xmm11,%xmm11
+	vmovdqa	-128(%r11),%xmm7
+	vpmuludq	%xmm2,%xmm9,%xmm5
+	vpmuludq	%xmm3,%xmm9,%xmm6
+	vpaddq	%xmm5,%xmm12,%xmm12
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vpmuludq	%xmm4,%xmm9,%xmm9
+	vpmuludq	-112(%r11),%xmm4,%xmm5
+	vpaddq	%xmm9,%xmm14,%xmm14
+
+	vpaddq	%xmm5,%xmm10,%xmm10
+	vpmuludq	%xmm2,%xmm7,%xmm6
+	vpmuludq	%xmm3,%xmm7,%xmm5
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vmovdqa	-96(%r11),%xmm8
+	vpaddq	%xmm5,%xmm14,%xmm14
+	vpmuludq	%xmm1,%xmm7,%xmm6
+	vpmuludq	%xmm0,%xmm7,%xmm7
+	vpaddq	%xmm6,%xmm12,%xmm12
+	vpaddq	%xmm7,%xmm11,%xmm11
+
+	vmovdqa	-80(%r11),%xmm9
+	vpmuludq	%xmm2,%xmm8,%xmm5
+	vpmuludq	%xmm1,%xmm8,%xmm6
+	vpaddq	%xmm5,%xmm14,%xmm14
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vmovdqa	-64(%r11),%xmm7
+	vpmuludq	%xmm0,%xmm8,%xmm8
+	vpmuludq	%xmm4,%xmm9,%xmm5
+	vpaddq	%xmm8,%xmm12,%xmm12
+	vpaddq	%xmm5,%xmm11,%xmm11
+	vmovdqa	-48(%r11),%xmm8
+	vpmuludq	%xmm3,%xmm9,%xmm9
+	vpmuludq	%xmm1,%xmm7,%xmm6
+	vpaddq	%xmm9,%xmm10,%xmm10
+
+	vmovdqa	-16(%r11),%xmm9
+	vpaddq	%xmm6,%xmm14,%xmm14
+	vpmuludq	%xmm0,%xmm7,%xmm7
+	vpmuludq	%xmm4,%xmm8,%xmm5
+	vpaddq	%xmm7,%xmm13,%xmm13
+	vpaddq	%xmm5,%xmm12,%xmm12
+	vmovdqu	32(%rsi),%xmm5
+	vpmuludq	%xmm3,%xmm8,%xmm7
+	vpmuludq	%xmm2,%xmm8,%xmm8
+	vpaddq	%xmm7,%xmm11,%xmm11
+	vmovdqu	48(%rsi),%xmm6
+	vpaddq	%xmm8,%xmm10,%xmm10
+
+	vpmuludq	%xmm2,%xmm9,%xmm2
+	vpmuludq	%xmm3,%xmm9,%xmm3
+	vpsrldq	$6,%xmm5,%xmm7
+	vpaddq	%xmm2,%xmm11,%xmm11
+	vpmuludq	%xmm4,%xmm9,%xmm4
+	vpsrldq	$6,%xmm6,%xmm8
+	vpaddq	%xmm3,%xmm12,%xmm2
+	vpaddq	%xmm4,%xmm13,%xmm3
+	vpmuludq	-32(%r11),%xmm0,%xmm4
+	vpmuludq	%xmm1,%xmm9,%xmm0
+	vpunpckhqdq	%xmm6,%xmm5,%xmm9
+	vpaddq	%xmm4,%xmm14,%xmm4
+	vpaddq	%xmm0,%xmm10,%xmm0
+
+	vpunpcklqdq	%xmm6,%xmm5,%xmm5
+	vpunpcklqdq	%xmm8,%xmm7,%xmm8
+
+
+	vpsrldq	$5,%xmm9,%xmm9
+	vpsrlq	$26,%xmm5,%xmm6
+	vmovdqa	0(%rsp),%xmm14
+	vpand	%xmm15,%xmm5,%xmm5
+	vpsrlq	$4,%xmm8,%xmm7
+	vpand	%xmm15,%xmm6,%xmm6
+	vpand	0(%rcx),%xmm9,%xmm9
+	vpsrlq	$30,%xmm8,%xmm8
+	vpand	%xmm15,%xmm7,%xmm7
+	vpand	%xmm15,%xmm8,%xmm8
+	vpor	32(%rcx),%xmm9,%xmm9
+
+
+
+
+
+	vpsrlq	$26,%xmm3,%xmm13
+	vpand	%xmm15,%xmm3,%xmm3
+	vpaddq	%xmm13,%xmm4,%xmm4
+
+	vpsrlq	$26,%xmm0,%xmm10
+	vpand	%xmm15,%xmm0,%xmm0
+	vpaddq	%xmm10,%xmm11,%xmm1
+
+	vpsrlq	$26,%xmm4,%xmm10
+	vpand	%xmm15,%xmm4,%xmm4
+
+	vpsrlq	$26,%xmm1,%xmm11
+	vpand	%xmm15,%xmm1,%xmm1
+	vpaddq	%xmm11,%xmm2,%xmm2
+
+	vpaddq	%xmm10,%xmm0,%xmm0
+	vpsllq	$2,%xmm10,%xmm10
+	vpaddq	%xmm10,%xmm0,%xmm0
+
+	vpsrlq	$26,%xmm2,%xmm12
+	vpand	%xmm15,%xmm2,%xmm2
+	vpaddq	%xmm12,%xmm3,%xmm3
+
+	vpsrlq	$26,%xmm0,%xmm10
+	vpand	%xmm15,%xmm0,%xmm0
+	vpaddq	%xmm10,%xmm1,%xmm1
+
+	vpsrlq	$26,%xmm3,%xmm13
+	vpand	%xmm15,%xmm3,%xmm3
+	vpaddq	%xmm13,%xmm4,%xmm4
+
+	ja	.Loop_avx
+
+.Lskip_loop_avx:
+
+
+
+	vpshufd	$0x10,%xmm14,%xmm14
+	addq	$32,%rdx
+	jnz	.Long_tail_avx
+
+	vpaddq	%xmm2,%xmm7,%xmm7
+	vpaddq	%xmm0,%xmm5,%xmm5
+	vpaddq	%xmm1,%xmm6,%xmm6
+	vpaddq	%xmm3,%xmm8,%xmm8
+	vpaddq	%xmm4,%xmm9,%xmm9
+
+.Long_tail_avx:
+	vmovdqa	%xmm2,32(%r11)
+	vmovdqa	%xmm0,0(%r11)
+	vmovdqa	%xmm1,16(%r11)
+	vmovdqa	%xmm3,48(%r11)
+	vmovdqa	%xmm4,64(%r11)
+
+
+
+
+
+
+
+	vpmuludq	%xmm7,%xmm14,%xmm12
+	vpmuludq	%xmm5,%xmm14,%xmm10
+	vpshufd	$0x10,-48(%rdi),%xmm2
+	vpmuludq	%xmm6,%xmm14,%xmm11
+	vpmuludq	%xmm8,%xmm14,%xmm13
+	vpmuludq	%xmm9,%xmm14,%xmm14
+
+	vpmuludq	%xmm8,%xmm2,%xmm0
+	vpaddq	%xmm0,%xmm14,%xmm14
+	vpshufd	$0x10,-32(%rdi),%xmm3
+	vpmuludq	%xmm7,%xmm2,%xmm1
+	vpaddq	%xmm1,%xmm13,%xmm13
+	vpshufd	$0x10,-16(%rdi),%xmm4
+	vpmuludq	%xmm6,%xmm2,%xmm0
+	vpaddq	%xmm0,%xmm12,%xmm12
+	vpmuludq	%xmm5,%xmm2,%xmm2
+	vpaddq	%xmm2,%xmm11,%xmm11
+	vpmuludq	%xmm9,%xmm3,%xmm3
+	vpaddq	%xmm3,%xmm10,%xmm10
+
+	vpshufd	$0x10,0(%rdi),%xmm2
+	vpmuludq	%xmm7,%xmm4,%xmm1
+	vpaddq	%xmm1,%xmm14,%xmm14
+	vpmuludq	%xmm6,%xmm4,%xmm0
+	vpaddq	%xmm0,%xmm13,%xmm13
+	vpshufd	$0x10,16(%rdi),%xmm3
+	vpmuludq	%xmm5,%xmm4,%xmm4
+	vpaddq	%xmm4,%xmm12,%xmm12
+	vpmuludq	%xmm9,%xmm2,%xmm1
+	vpaddq	%xmm1,%xmm11,%xmm11
+	vpshufd	$0x10,32(%rdi),%xmm4
+	vpmuludq	%xmm8,%xmm2,%xmm2
+	vpaddq	%xmm2,%xmm10,%xmm10
+
+	vpmuludq	%xmm6,%xmm3,%xmm0
+	vpaddq	%xmm0,%xmm14,%xmm14
+	vpmuludq	%xmm5,%xmm3,%xmm3
+	vpaddq	%xmm3,%xmm13,%xmm13
+	vpshufd	$0x10,48(%rdi),%xmm2
+	vpmuludq	%xmm9,%xmm4,%xmm1
+	vpaddq	%xmm1,%xmm12,%xmm12
+	vpshufd	$0x10,64(%rdi),%xmm3
+	vpmuludq	%xmm8,%xmm4,%xmm0
+	vpaddq	%xmm0,%xmm11,%xmm11
+	vpmuludq	%xmm7,%xmm4,%xmm4
+	vpaddq	%xmm4,%xmm10,%xmm10
+
+	vpmuludq	%xmm5,%xmm2,%xmm2
+	vpaddq	%xmm2,%xmm14,%xmm14
+	vpmuludq	%xmm9,%xmm3,%xmm1
+	vpaddq	%xmm1,%xmm13,%xmm13
+	vpmuludq	%xmm8,%xmm3,%xmm0
+	vpaddq	%xmm0,%xmm12,%xmm12
+	vpmuludq	%xmm7,%xmm3,%xmm1
+	vpaddq	%xmm1,%xmm11,%xmm11
+	vpmuludq	%xmm6,%xmm3,%xmm3
+	vpaddq	%xmm3,%xmm10,%xmm10
+
+	jz	.Lshort_tail_avx
+
+	vmovdqu	0(%rsi),%xmm0
+	vmovdqu	16(%rsi),%xmm1
+
+	vpsrldq	$6,%xmm0,%xmm2
+	vpsrldq	$6,%xmm1,%xmm3
+	vpunpckhqdq	%xmm1,%xmm0,%xmm4
+	vpunpcklqdq	%xmm1,%xmm0,%xmm0
+	vpunpcklqdq	%xmm3,%xmm2,%xmm3
+
+	vpsrlq	$40,%xmm4,%xmm4
+	vpsrlq	$26,%xmm0,%xmm1
+	vpand	%xmm15,%xmm0,%xmm0
+	vpsrlq	$4,%xmm3,%xmm2
+	vpand	%xmm15,%xmm1,%xmm1
+	vpsrlq	$30,%xmm3,%xmm3
+	vpand	%xmm15,%xmm2,%xmm2
+	vpand	%xmm15,%xmm3,%xmm3
+	vpor	32(%rcx),%xmm4,%xmm4
+
+	vpshufd	$0x32,-64(%rdi),%xmm9
+	vpaddq	0(%r11),%xmm0,%xmm0
+	vpaddq	16(%r11),%xmm1,%xmm1
+	vpaddq	32(%r11),%xmm2,%xmm2
+	vpaddq	48(%r11),%xmm3,%xmm3
+	vpaddq	64(%r11),%xmm4,%xmm4
+
+
+
+
+	vpmuludq	%xmm0,%xmm9,%xmm5
+	vpaddq	%xmm5,%xmm10,%xmm10
+	vpmuludq	%xmm1,%xmm9,%xmm6
+	vpaddq	%xmm6,%xmm11,%xmm11
+	vpmuludq	%xmm2,%xmm9,%xmm5
+	vpaddq	%xmm5,%xmm12,%xmm12
+	vpshufd	$0x32,-48(%rdi),%xmm7
+	vpmuludq	%xmm3,%xmm9,%xmm6
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vpmuludq	%xmm4,%xmm9,%xmm9
+	vpaddq	%xmm9,%xmm14,%xmm14
+
+	vpmuludq	%xmm3,%xmm7,%xmm5
+	vpaddq	%xmm5,%xmm14,%xmm14
+	vpshufd	$0x32,-32(%rdi),%xmm8
+	vpmuludq	%xmm2,%xmm7,%xmm6
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vpshufd	$0x32,-16(%rdi),%xmm9
+	vpmuludq	%xmm1,%xmm7,%xmm5
+	vpaddq	%xmm5,%xmm12,%xmm12
+	vpmuludq	%xmm0,%xmm7,%xmm7
+	vpaddq	%xmm7,%xmm11,%xmm11
+	vpmuludq	%xmm4,%xmm8,%xmm8
+	vpaddq	%xmm8,%xmm10,%xmm10
+
+	vpshufd	$0x32,0(%rdi),%xmm7
+	vpmuludq	%xmm2,%xmm9,%xmm6
+	vpaddq	%xmm6,%xmm14,%xmm14
+	vpmuludq	%xmm1,%xmm9,%xmm5
+	vpaddq	%xmm5,%xmm13,%xmm13
+	vpshufd	$0x32,16(%rdi),%xmm8
+	vpmuludq	%xmm0,%xmm9,%xmm9
+	vpaddq	%xmm9,%xmm12,%xmm12
+	vpmuludq	%xmm4,%xmm7,%xmm6
+	vpaddq	%xmm6,%xmm11,%xmm11
+	vpshufd	$0x32,32(%rdi),%xmm9
+	vpmuludq	%xmm3,%xmm7,%xmm7
+	vpaddq	%xmm7,%xmm10,%xmm10
+
+	vpmuludq	%xmm1,%xmm8,%xmm5
+	vpaddq	%xmm5,%xmm14,%xmm14
+	vpmuludq	%xmm0,%xmm8,%xmm8
+	vpaddq	%xmm8,%xmm13,%xmm13
+	vpshufd	$0x32,48(%rdi),%xmm7
+	vpmuludq	%xmm4,%xmm9,%xmm6
+	vpaddq	%xmm6,%xmm12,%xmm12
+	vpshufd	$0x32,64(%rdi),%xmm8
+	vpmuludq	%xmm3,%xmm9,%xmm5
+	vpaddq	%xmm5,%xmm11,%xmm11
+	vpmuludq	%xmm2,%xmm9,%xmm9
+	vpaddq	%xmm9,%xmm10,%xmm10
+
+	vpmuludq	%xmm0,%xmm7,%xmm7
+	vpaddq	%xmm7,%xmm14,%xmm14
+	vpmuludq	%xmm4,%xmm8,%xmm6
+	vpaddq	%xmm6,%xmm13,%xmm13
+	vpmuludq	%xmm3,%xmm8,%xmm5
+	vpaddq	%xmm5,%xmm12,%xmm12
+	vpmuludq	%xmm2,%xmm8,%xmm6
+	vpaddq	%xmm6,%xmm11,%xmm11
+	vpmuludq	%xmm1,%xmm8,%xmm8
+	vpaddq	%xmm8,%xmm10,%xmm10
+
+.Lshort_tail_avx:
+
+
+
+	vpsrldq	$8,%xmm14,%xmm9
+	vpsrldq	$8,%xmm13,%xmm8
+	vpsrldq	$8,%xmm11,%xmm6
+	vpsrldq	$8,%xmm10,%xmm5
+	vpsrldq	$8,%xmm12,%xmm7
+	vpaddq	%xmm8,%xmm13,%xmm13
+	vpaddq	%xmm9,%xmm14,%xmm14
+	vpaddq	%xmm5,%xmm10,%xmm10
+	vpaddq	%xmm6,%xmm11,%xmm11
+	vpaddq	%xmm7,%xmm12,%xmm12
+
+
+
+
+	vpsrlq	$26,%xmm13,%xmm3
+	vpand	%xmm15,%xmm13,%xmm13
+	vpaddq	%xmm3,%xmm14,%xmm14
+
+	vpsrlq	$26,%xmm10,%xmm0
+	vpand	%xmm15,%xmm10,%xmm10
+	vpaddq	%xmm0,%xmm11,%xmm11
+
+	vpsrlq	$26,%xmm14,%xmm4
+	vpand	%xmm15,%xmm14,%xmm14
+
+	vpsrlq	$26,%xmm11,%xmm1
+	vpand	%xmm15,%xmm11,%xmm11
+	vpaddq	%xmm1,%xmm12,%xmm12
+
+	vpaddq	%xmm4,%xmm10,%xmm10
+	vpsllq	$2,%xmm4,%xmm4
+	vpaddq	%xmm4,%xmm10,%xmm10
+
+	vpsrlq	$26,%xmm12,%xmm2
+	vpand	%xmm15,%xmm12,%xmm12
+	vpaddq	%xmm2,%xmm13,%xmm13
+
+	vpsrlq	$26,%xmm10,%xmm0
+	vpand	%xmm15,%xmm10,%xmm10
+	vpaddq	%xmm0,%xmm11,%xmm11
+
+	vpsrlq	$26,%xmm13,%xmm3
+	vpand	%xmm15,%xmm13,%xmm13
+	vpaddq	%xmm3,%xmm14,%xmm14
+
+	vmovd	%xmm10,-112(%rdi)
+	vmovd	%xmm11,-108(%rdi)
+	vmovd	%xmm12,-104(%rdi)
+	vmovd	%xmm13,-100(%rdi)
+	vmovd	%xmm14,-96(%rdi)
+	leaq	88(%r11),%rsp
+.cfi_def_cfa	%rsp,8
+	vzeroupper
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	poly1305_blocks_avx,.-poly1305_blocks_avx
+
+.type	poly1305_emit_avx,@function
+.align	32
+poly1305_emit_avx:
+	cmpl	$0,20(%rdi)
+	je	.Lemit
+
+	movl	0(%rdi),%eax
+	movl	4(%rdi),%ecx
+	movl	8(%rdi),%r8d
+	movl	12(%rdi),%r11d
+	movl	16(%rdi),%r10d
+
+	shlq	$26,%rcx
+	movq	%r8,%r9
+	shlq	$52,%r8
+	addq	%rcx,%rax
+	shrq	$12,%r9
+	addq	%rax,%r8
+	adcq	$0,%r9
+
+	shlq	$14,%r11
+	movq	%r10,%rax
+	shrq	$24,%r10
+	addq	%r11,%r9
+	shlq	$40,%rax
+	addq	%rax,%r9
+	adcq	$0,%r10
+
+	movq	%r10,%rax
+	movq	%r10,%rcx
+	andq	$3,%r10
+	shrq	$2,%rax
+	andq	$-4,%rcx
+	addq	%rcx,%rax
+	addq	%rax,%r8
+	adcq	$0,%r9
+	adcq	$0,%r10
+
+	movq	%r8,%rax
+	addq	$5,%r8
+	movq	%r9,%rcx
+	adcq	$0,%r9
+	adcq	$0,%r10
+	shrq	$2,%r10
+	cmovnzq	%r8,%rax
+	cmovnzq	%r9,%rcx
+
+	addq	0(%rdx),%rax
+	adcq	8(%rdx),%rcx
+	movq	%rax,0(%rsi)
+	movq	%rcx,8(%rsi)
+
+	.byte	0xf3,0xc3
+.size	poly1305_emit_avx,.-poly1305_emit_avx
+.type	poly1305_blocks_avx2,@function
+.align	32
+poly1305_blocks_avx2:
+.cfi_startproc	
+	movl	20(%rdi),%r8d
+	cmpq	$128,%rdx
+	jae	.Lblocks_avx2
+	testl	%r8d,%r8d
+	jz	.Lblocks
+
+.Lblocks_avx2:
+	andq	$-16,%rdx
+	jz	.Lno_data_avx2
+
+	vzeroupper
+
+	testl	%r8d,%r8d
+	jz	.Lbase2_64_avx2
+
+	testq	$63,%rdx
+	jz	.Leven_avx2
+
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+.Lblocks_avx2_body:
+
+	movq	%rdx,%r15
+
+	movq	0(%rdi),%r8
+	movq	8(%rdi),%r9
+	movl	16(%rdi),%ebp
+
+	movq	24(%rdi),%r11
+	movq	32(%rdi),%r13
+
+
+	movl	%r8d,%r14d
+	andq	$-2147483648,%r8
+	movq	%r9,%r12
+	movl	%r9d,%ebx
+	andq	$-2147483648,%r9
+
+	shrq	$6,%r8
+	shlq	$52,%r12
+	addq	%r8,%r14
+	shrq	$12,%rbx
+	shrq	$18,%r9
+	addq	%r12,%r14
+	adcq	%r9,%rbx
+
+	movq	%rbp,%r8
+	shlq	$40,%r8
+	shrq	$24,%rbp
+	addq	%r8,%rbx
+	adcq	$0,%rbp
+
+	movq	$-4,%r9
+	movq	%rbp,%r8
+	andq	%rbp,%r9
+	shrq	$2,%r8
+	andq	$3,%rbp
+	addq	%r9,%r8
+	addq	%r8,%r14
+	adcq	$0,%rbx
+	adcq	$0,%rbp
+
+	movq	%r13,%r12
+	movq	%r13,%rax
+	shrq	$2,%r13
+	addq	%r12,%r13
+
+.Lbase2_26_pre_avx2:
+	addq	0(%rsi),%r14
+	adcq	8(%rsi),%rbx
+	leaq	16(%rsi),%rsi
+	adcq	%rcx,%rbp
+	subq	$16,%r15
+
+	call	__poly1305_block
+	movq	%r12,%rax
+
+	testq	$63,%r15
+	jnz	.Lbase2_26_pre_avx2
+
+	testq	%rcx,%rcx
+	jz	.Lstore_base2_64_avx2
+
+
+	movq	%r14,%rax
+	movq	%r14,%rdx
+	shrq	$52,%r14
+	movq	%rbx,%r11
+	movq	%rbx,%r12
+	shrq	$26,%rdx
+	andq	$0x3ffffff,%rax
+	shlq	$12,%r11
+	andq	$0x3ffffff,%rdx
+	shrq	$14,%rbx
+	orq	%r11,%r14
+	shlq	$24,%rbp
+	andq	$0x3ffffff,%r14
+	shrq	$40,%r12
+	andq	$0x3ffffff,%rbx
+	orq	%r12,%rbp
+
+	testq	%r15,%r15
+	jz	.Lstore_base2_26_avx2
+
+	vmovd	%eax,%xmm0
+	vmovd	%edx,%xmm1
+	vmovd	%r14d,%xmm2
+	vmovd	%ebx,%xmm3
+	vmovd	%ebp,%xmm4
+	jmp	.Lproceed_avx2
+
+.align	32
+.Lstore_base2_64_avx2:
+	movq	%r14,0(%rdi)
+	movq	%rbx,8(%rdi)
+	movq	%rbp,16(%rdi)
+	jmp	.Ldone_avx2
+
+.align	16
+.Lstore_base2_26_avx2:
+	movl	%eax,0(%rdi)
+	movl	%edx,4(%rdi)
+	movl	%r14d,8(%rdi)
+	movl	%ebx,12(%rdi)
+	movl	%ebp,16(%rdi)
+.align	16
+.Ldone_avx2:
+	movq	0(%rsp),%r15
+.cfi_restore	%r15
+	movq	8(%rsp),%r14
+.cfi_restore	%r14
+	movq	16(%rsp),%r13
+.cfi_restore	%r13
+	movq	24(%rsp),%r12
+.cfi_restore	%r12
+	movq	32(%rsp),%rbp
+.cfi_restore	%rbp
+	movq	40(%rsp),%rbx
+.cfi_restore	%rbx
+	leaq	48(%rsp),%rsp
+.cfi_adjust_cfa_offset	-48
+.Lno_data_avx2:
+.Lblocks_avx2_epilogue:
+	.byte	0xf3,0xc3
+.cfi_endproc	
+
+.align	32
+.Lbase2_64_avx2:
+.cfi_startproc	
+	pushq	%rbx
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbx,-16
+	pushq	%rbp
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%rbp,-24
+	pushq	%r12
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r12,-32
+	pushq	%r13
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r13,-40
+	pushq	%r14
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r14,-48
+	pushq	%r15
+.cfi_adjust_cfa_offset	8
+.cfi_offset	%r15,-56
+.Lbase2_64_avx2_body:
+
+	movq	%rdx,%r15
+
+	movq	24(%rdi),%r11
+	movq	32(%rdi),%r13
+
+	movq	0(%rdi),%r14
+	movq	8(%rdi),%rbx
+	movl	16(%rdi),%ebp
+
+	movq	%r13,%r12
+	movq	%r13,%rax
+	shrq	$2,%r13
+	addq	%r12,%r13
+
+	testq	$63,%rdx
+	jz	.Linit_avx2
+
+.Lbase2_64_pre_avx2:
+	addq	0(%rsi),%r14
+	adcq	8(%rsi),%rbx
+	leaq	16(%rsi),%rsi
+	adcq	%rcx,%rbp
+	subq	$16,%r15
+
+	call	__poly1305_block
+	movq	%r12,%rax
+
+	testq	$63,%r15
+	jnz	.Lbase2_64_pre_avx2
+
+.Linit_avx2:
+
+	movq	%r14,%rax
+	movq	%r14,%rdx
+	shrq	$52,%r14
+	movq	%rbx,%r8
+	movq	%rbx,%r9
+	shrq	$26,%rdx
+	andq	$0x3ffffff,%rax
+	shlq	$12,%r8
+	andq	$0x3ffffff,%rdx
+	shrq	$14,%rbx
+	orq	%r8,%r14
+	shlq	$24,%rbp
+	andq	$0x3ffffff,%r14
+	shrq	$40,%r9
+	andq	$0x3ffffff,%rbx
+	orq	%r9,%rbp
+
+	vmovd	%eax,%xmm0
+	vmovd	%edx,%xmm1
+	vmovd	%r14d,%xmm2
+	vmovd	%ebx,%xmm3
+	vmovd	%ebp,%xmm4
+	movl	$1,20(%rdi)
+
+	call	__poly1305_init_avx
+
+.Lproceed_avx2:
+	movq	%r15,%rdx
+	movl	OPENSSL_ia32cap_P+8(%rip),%r10d
+	movl	$3221291008,%r11d
+
+	movq	0(%rsp),%r15
+.cfi_restore	%r15
+	movq	8(%rsp),%r14
+.cfi_restore	%r14
+	movq	16(%rsp),%r13
+.cfi_restore	%r13
+	movq	24(%rsp),%r12
+.cfi_restore	%r12
+	movq	32(%rsp),%rbp
+.cfi_restore	%rbp
+	movq	40(%rsp),%rbx
+.cfi_restore	%rbx
+	leaq	48(%rsp),%rax
+	leaq	48(%rsp),%rsp
+.cfi_adjust_cfa_offset	-48
+.Lbase2_64_avx2_epilogue:
+	jmp	.Ldo_avx2
+.cfi_endproc	
+
+.align	32
+.Leven_avx2:
+.cfi_startproc	
+	movl	OPENSSL_ia32cap_P+8(%rip),%r10d
+	vmovd	0(%rdi),%xmm0
+	vmovd	4(%rdi),%xmm1
+	vmovd	8(%rdi),%xmm2
+	vmovd	12(%rdi),%xmm3
+	vmovd	16(%rdi),%xmm4
+
+.Ldo_avx2:
+	cmpq	$512,%rdx
+	jb	.Lskip_avx512
+	andl	%r11d,%r10d
+	testl	$65536,%r10d
+	jnz	.Lblocks_avx512
+.Lskip_avx512:
+	leaq	-8(%rsp),%r11
+.cfi_def_cfa	%r11,16
+	subq	$0x128,%rsp
+	leaq	.Lconst(%rip),%rcx
+	leaq	48+64(%rdi),%rdi
+	vmovdqa	96(%rcx),%ymm7
+
+
+	vmovdqu	-64(%rdi),%xmm9
+	andq	$-512,%rsp
+	vmovdqu	-48(%rdi),%xmm10
+	vmovdqu	-32(%rdi),%xmm6
+	vmovdqu	-16(%rdi),%xmm11
+	vmovdqu	0(%rdi),%xmm12
+	vmovdqu	16(%rdi),%xmm13
+	leaq	144(%rsp),%rax
+	vmovdqu	32(%rdi),%xmm14
+	vpermd	%ymm9,%ymm7,%ymm9
+	vmovdqu	48(%rdi),%xmm15
+	vpermd	%ymm10,%ymm7,%ymm10
+	vmovdqu	64(%rdi),%xmm5
+	vpermd	%ymm6,%ymm7,%ymm6
+	vmovdqa	%ymm9,0(%rsp)
+	vpermd	%ymm11,%ymm7,%ymm11
+	vmovdqa	%ymm10,32-144(%rax)
+	vpermd	%ymm12,%ymm7,%ymm12
+	vmovdqa	%ymm6,64-144(%rax)
+	vpermd	%ymm13,%ymm7,%ymm13
+	vmovdqa	%ymm11,96-144(%rax)
+	vpermd	%ymm14,%ymm7,%ymm14
+	vmovdqa	%ymm12,128-144(%rax)
+	vpermd	%ymm15,%ymm7,%ymm15
+	vmovdqa	%ymm13,160-144(%rax)
+	vpermd	%ymm5,%ymm7,%ymm5
+	vmovdqa	%ymm14,192-144(%rax)
+	vmovdqa	%ymm15,224-144(%rax)
+	vmovdqa	%ymm5,256-144(%rax)
+	vmovdqa	64(%rcx),%ymm5
+
+
+
+	vmovdqu	0(%rsi),%xmm7
+	vmovdqu	16(%rsi),%xmm8
+	vinserti128	$1,32(%rsi),%ymm7,%ymm7
+	vinserti128	$1,48(%rsi),%ymm8,%ymm8
+	leaq	64(%rsi),%rsi
+
+	vpsrldq	$6,%ymm7,%ymm9
+	vpsrldq	$6,%ymm8,%ymm10
+	vpunpckhqdq	%ymm8,%ymm7,%ymm6
+	vpunpcklqdq	%ymm10,%ymm9,%ymm9
+	vpunpcklqdq	%ymm8,%ymm7,%ymm7
+
+	vpsrlq	$30,%ymm9,%ymm10
+	vpsrlq	$4,%ymm9,%ymm9
+	vpsrlq	$26,%ymm7,%ymm8
+	vpsrlq	$40,%ymm6,%ymm6
+	vpand	%ymm5,%ymm9,%ymm9
+	vpand	%ymm5,%ymm7,%ymm7
+	vpand	%ymm5,%ymm8,%ymm8
+	vpand	%ymm5,%ymm10,%ymm10
+	vpor	32(%rcx),%ymm6,%ymm6
+
+	vpaddq	%ymm2,%ymm9,%ymm2
+	subq	$64,%rdx
+	jz	.Ltail_avx2
+	jmp	.Loop_avx2
+
+.align	32
+.Loop_avx2:
+
+
+
+
+
+
+
+
+	vpaddq	%ymm0,%ymm7,%ymm0
+	vmovdqa	0(%rsp),%ymm7
+	vpaddq	%ymm1,%ymm8,%ymm1
+	vmovdqa	32(%rsp),%ymm8
+	vpaddq	%ymm3,%ymm10,%ymm3
+	vmovdqa	96(%rsp),%ymm9
+	vpaddq	%ymm4,%ymm6,%ymm4
+	vmovdqa	48(%rax),%ymm10
+	vmovdqa	112(%rax),%ymm5
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+	vpmuludq	%ymm2,%ymm7,%ymm13
+	vpmuludq	%ymm2,%ymm8,%ymm14
+	vpmuludq	%ymm2,%ymm9,%ymm15
+	vpmuludq	%ymm2,%ymm10,%ymm11
+	vpmuludq	%ymm2,%ymm5,%ymm12
+
+	vpmuludq	%ymm0,%ymm8,%ymm6
+	vpmuludq	%ymm1,%ymm8,%ymm2
+	vpaddq	%ymm6,%ymm12,%ymm12
+	vpaddq	%ymm2,%ymm13,%ymm13
+	vpmuludq	%ymm3,%ymm8,%ymm6
+	vpmuludq	64(%rsp),%ymm4,%ymm2
+	vpaddq	%ymm6,%ymm15,%ymm15
+	vpaddq	%ymm2,%ymm11,%ymm11
+	vmovdqa	-16(%rax),%ymm8
+
+	vpmuludq	%ymm0,%ymm7,%ymm6
+	vpmuludq	%ymm1,%ymm7,%ymm2
+	vpaddq	%ymm6,%ymm11,%ymm11
+	vpaddq	%ymm2,%ymm12,%ymm12
+	vpmuludq	%ymm3,%ymm7,%ymm6
+	vpmuludq	%ymm4,%ymm7,%ymm2
+	vmovdqu	0(%rsi),%xmm7
+	vpaddq	%ymm6,%ymm14,%ymm14
+	vpaddq	%ymm2,%ymm15,%ymm15
+	vinserti128	$1,32(%rsi),%ymm7,%ymm7
+
+	vpmuludq	%ymm3,%ymm8,%ymm6
+	vpmuludq	%ymm4,%ymm8,%ymm2
+	vmovdqu	16(%rsi),%xmm8
+	vpaddq	%ymm6,%ymm11,%ymm11
+	vpaddq	%ymm2,%ymm12,%ymm12
+	vmovdqa	16(%rax),%ymm2
+	vpmuludq	%ymm1,%ymm9,%ymm6
+	vpmuludq	%ymm0,%ymm9,%ymm9
+	vpaddq	%ymm6,%ymm14,%ymm14
+	vpaddq	%ymm9,%ymm13,%ymm13
+	vinserti128	$1,48(%rsi),%ymm8,%ymm8
+	leaq	64(%rsi),%rsi
+
+	vpmuludq	%ymm1,%ymm2,%ymm6
+	vpmuludq	%ymm0,%ymm2,%ymm2
+	vpsrldq	$6,%ymm7,%ymm9
+	vpaddq	%ymm6,%ymm15,%ymm15
+	vpaddq	%ymm2,%ymm14,%ymm14
+	vpmuludq	%ymm3,%ymm10,%ymm6
+	vpmuludq	%ymm4,%ymm10,%ymm2
+	vpsrldq	$6,%ymm8,%ymm10
+	vpaddq	%ymm6,%ymm12,%ymm12
+	vpaddq	%ymm2,%ymm13,%ymm13
+	vpunpckhqdq	%ymm8,%ymm7,%ymm6
+
+	vpmuludq	%ymm3,%ymm5,%ymm3
+	vpmuludq	%ymm4,%ymm5,%ymm4
+	vpunpcklqdq	%ymm8,%ymm7,%ymm7
+	vpaddq	%ymm3,%ymm13,%ymm2
+	vpaddq	%ymm4,%ymm14,%ymm3
+	vpunpcklqdq	%ymm10,%ymm9,%ymm10
+	vpmuludq	80(%rax),%ymm0,%ymm4
+	vpmuludq	%ymm1,%ymm5,%ymm0
+	vmovdqa	64(%rcx),%ymm5
+	vpaddq	%ymm4,%ymm15,%ymm4
+	vpaddq	%ymm0,%ymm11,%ymm0
+
+
+
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpaddq	%ymm11,%ymm12,%ymm1
+
+	vpsrlq	$26,%ymm4,%ymm15
+	vpand	%ymm5,%ymm4,%ymm4
+
+	vpsrlq	$4,%ymm10,%ymm9
+
+	vpsrlq	$26,%ymm1,%ymm12
+	vpand	%ymm5,%ymm1,%ymm1
+	vpaddq	%ymm12,%ymm2,%ymm2
+
+	vpaddq	%ymm15,%ymm0,%ymm0
+	vpsllq	$2,%ymm15,%ymm15
+	vpaddq	%ymm15,%ymm0,%ymm0
+
+	vpand	%ymm5,%ymm9,%ymm9
+	vpsrlq	$26,%ymm7,%ymm8
+
+	vpsrlq	$26,%ymm2,%ymm13
+	vpand	%ymm5,%ymm2,%ymm2
+	vpaddq	%ymm13,%ymm3,%ymm3
+
+	vpaddq	%ymm9,%ymm2,%ymm2
+	vpsrlq	$30,%ymm10,%ymm10
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpaddq	%ymm11,%ymm1,%ymm1
+
+	vpsrlq	$40,%ymm6,%ymm6
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	vpand	%ymm5,%ymm7,%ymm7
+	vpand	%ymm5,%ymm8,%ymm8
+	vpand	%ymm5,%ymm10,%ymm10
+	vpor	32(%rcx),%ymm6,%ymm6
+
+	subq	$64,%rdx
+	jnz	.Loop_avx2
+
+.byte	0x66,0x90
+.Ltail_avx2:
+
+
+
+
+
+
+
+	vpaddq	%ymm0,%ymm7,%ymm0
+	vmovdqu	4(%rsp),%ymm7
+	vpaddq	%ymm1,%ymm8,%ymm1
+	vmovdqu	36(%rsp),%ymm8
+	vpaddq	%ymm3,%ymm10,%ymm3
+	vmovdqu	100(%rsp),%ymm9
+	vpaddq	%ymm4,%ymm6,%ymm4
+	vmovdqu	52(%rax),%ymm10
+	vmovdqu	116(%rax),%ymm5
+
+	vpmuludq	%ymm2,%ymm7,%ymm13
+	vpmuludq	%ymm2,%ymm8,%ymm14
+	vpmuludq	%ymm2,%ymm9,%ymm15
+	vpmuludq	%ymm2,%ymm10,%ymm11
+	vpmuludq	%ymm2,%ymm5,%ymm12
+
+	vpmuludq	%ymm0,%ymm8,%ymm6
+	vpmuludq	%ymm1,%ymm8,%ymm2
+	vpaddq	%ymm6,%ymm12,%ymm12
+	vpaddq	%ymm2,%ymm13,%ymm13
+	vpmuludq	%ymm3,%ymm8,%ymm6
+	vpmuludq	68(%rsp),%ymm4,%ymm2
+	vpaddq	%ymm6,%ymm15,%ymm15
+	vpaddq	%ymm2,%ymm11,%ymm11
+
+	vpmuludq	%ymm0,%ymm7,%ymm6
+	vpmuludq	%ymm1,%ymm7,%ymm2
+	vpaddq	%ymm6,%ymm11,%ymm11
+	vmovdqu	-12(%rax),%ymm8
+	vpaddq	%ymm2,%ymm12,%ymm12
+	vpmuludq	%ymm3,%ymm7,%ymm6
+	vpmuludq	%ymm4,%ymm7,%ymm2
+	vpaddq	%ymm6,%ymm14,%ymm14
+	vpaddq	%ymm2,%ymm15,%ymm15
+
+	vpmuludq	%ymm3,%ymm8,%ymm6
+	vpmuludq	%ymm4,%ymm8,%ymm2
+	vpaddq	%ymm6,%ymm11,%ymm11
+	vpaddq	%ymm2,%ymm12,%ymm12
+	vmovdqu	20(%rax),%ymm2
+	vpmuludq	%ymm1,%ymm9,%ymm6
+	vpmuludq	%ymm0,%ymm9,%ymm9
+	vpaddq	%ymm6,%ymm14,%ymm14
+	vpaddq	%ymm9,%ymm13,%ymm13
+
+	vpmuludq	%ymm1,%ymm2,%ymm6
+	vpmuludq	%ymm0,%ymm2,%ymm2
+	vpaddq	%ymm6,%ymm15,%ymm15
+	vpaddq	%ymm2,%ymm14,%ymm14
+	vpmuludq	%ymm3,%ymm10,%ymm6
+	vpmuludq	%ymm4,%ymm10,%ymm2
+	vpaddq	%ymm6,%ymm12,%ymm12
+	vpaddq	%ymm2,%ymm13,%ymm13
+
+	vpmuludq	%ymm3,%ymm5,%ymm3
+	vpmuludq	%ymm4,%ymm5,%ymm4
+	vpaddq	%ymm3,%ymm13,%ymm2
+	vpaddq	%ymm4,%ymm14,%ymm3
+	vpmuludq	84(%rax),%ymm0,%ymm4
+	vpmuludq	%ymm1,%ymm5,%ymm0
+	vmovdqa	64(%rcx),%ymm5
+	vpaddq	%ymm4,%ymm15,%ymm4
+	vpaddq	%ymm0,%ymm11,%ymm0
+
+
+
+
+	vpsrldq	$8,%ymm12,%ymm8
+	vpsrldq	$8,%ymm2,%ymm9
+	vpsrldq	$8,%ymm3,%ymm10
+	vpsrldq	$8,%ymm4,%ymm6
+	vpsrldq	$8,%ymm0,%ymm7
+	vpaddq	%ymm8,%ymm12,%ymm12
+	vpaddq	%ymm9,%ymm2,%ymm2
+	vpaddq	%ymm10,%ymm3,%ymm3
+	vpaddq	%ymm6,%ymm4,%ymm4
+	vpaddq	%ymm7,%ymm0,%ymm0
+
+	vpermq	$0x2,%ymm3,%ymm10
+	vpermq	$0x2,%ymm4,%ymm6
+	vpermq	$0x2,%ymm0,%ymm7
+	vpermq	$0x2,%ymm12,%ymm8
+	vpermq	$0x2,%ymm2,%ymm9
+	vpaddq	%ymm10,%ymm3,%ymm3
+	vpaddq	%ymm6,%ymm4,%ymm4
+	vpaddq	%ymm7,%ymm0,%ymm0
+	vpaddq	%ymm8,%ymm12,%ymm12
+	vpaddq	%ymm9,%ymm2,%ymm2
+
+
+
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpaddq	%ymm11,%ymm12,%ymm1
+
+	vpsrlq	$26,%ymm4,%ymm15
+	vpand	%ymm5,%ymm4,%ymm4
+
+	vpsrlq	$26,%ymm1,%ymm12
+	vpand	%ymm5,%ymm1,%ymm1
+	vpaddq	%ymm12,%ymm2,%ymm2
+
+	vpaddq	%ymm15,%ymm0,%ymm0
+	vpsllq	$2,%ymm15,%ymm15
+	vpaddq	%ymm15,%ymm0,%ymm0
+
+	vpsrlq	$26,%ymm2,%ymm13
+	vpand	%ymm5,%ymm2,%ymm2
+	vpaddq	%ymm13,%ymm3,%ymm3
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpaddq	%ymm11,%ymm1,%ymm1
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	vmovd	%xmm0,-112(%rdi)
+	vmovd	%xmm1,-108(%rdi)
+	vmovd	%xmm2,-104(%rdi)
+	vmovd	%xmm3,-100(%rdi)
+	vmovd	%xmm4,-96(%rdi)
+	leaq	8(%r11),%rsp
+.cfi_def_cfa	%rsp,8
+	vzeroupper
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	poly1305_blocks_avx2,.-poly1305_blocks_avx2
+.type	poly1305_blocks_avx512,@function
+.align	32
+poly1305_blocks_avx512:
+.cfi_startproc	
+.Lblocks_avx512:
+	movl	$15,%eax
+	kmovw	%eax,%k2
+	leaq	-8(%rsp),%r11
+.cfi_def_cfa	%r11,16
+	subq	$0x128,%rsp
+	leaq	.Lconst(%rip),%rcx
+	leaq	48+64(%rdi),%rdi
+	vmovdqa	96(%rcx),%ymm9
+
+
+	vmovdqu	-64(%rdi),%xmm11
+	andq	$-512,%rsp
+	vmovdqu	-48(%rdi),%xmm12
+	movq	$0x20,%rax
+	vmovdqu	-32(%rdi),%xmm7
+	vmovdqu	-16(%rdi),%xmm13
+	vmovdqu	0(%rdi),%xmm8
+	vmovdqu	16(%rdi),%xmm14
+	vmovdqu	32(%rdi),%xmm10
+	vmovdqu	48(%rdi),%xmm15
+	vmovdqu	64(%rdi),%xmm6
+	vpermd	%zmm11,%zmm9,%zmm16
+	vpbroadcastq	64(%rcx),%zmm5
+	vpermd	%zmm12,%zmm9,%zmm17
+	vpermd	%zmm7,%zmm9,%zmm21
+	vpermd	%zmm13,%zmm9,%zmm18
+	vmovdqa64	%zmm16,0(%rsp){%k2}
+	vpsrlq	$32,%zmm16,%zmm7
+	vpermd	%zmm8,%zmm9,%zmm22
+	vmovdqu64	%zmm17,0(%rsp,%rax,1){%k2}
+	vpsrlq	$32,%zmm17,%zmm8
+	vpermd	%zmm14,%zmm9,%zmm19
+	vmovdqa64	%zmm21,64(%rsp){%k2}
+	vpermd	%zmm10,%zmm9,%zmm23
+	vpermd	%zmm15,%zmm9,%zmm20
+	vmovdqu64	%zmm18,64(%rsp,%rax,1){%k2}
+	vpermd	%zmm6,%zmm9,%zmm24
+	vmovdqa64	%zmm22,128(%rsp){%k2}
+	vmovdqu64	%zmm19,128(%rsp,%rax,1){%k2}
+	vmovdqa64	%zmm23,192(%rsp){%k2}
+	vmovdqu64	%zmm20,192(%rsp,%rax,1){%k2}
+	vmovdqa64	%zmm24,256(%rsp){%k2}
+
+
+
+
+
+
+
+
+
+
+	vpmuludq	%zmm7,%zmm16,%zmm11
+	vpmuludq	%zmm7,%zmm17,%zmm12
+	vpmuludq	%zmm7,%zmm18,%zmm13
+	vpmuludq	%zmm7,%zmm19,%zmm14
+	vpmuludq	%zmm7,%zmm20,%zmm15
+	vpsrlq	$32,%zmm18,%zmm9
+
+	vpmuludq	%zmm8,%zmm24,%zmm25
+	vpmuludq	%zmm8,%zmm16,%zmm26
+	vpmuludq	%zmm8,%zmm17,%zmm27
+	vpmuludq	%zmm8,%zmm18,%zmm28
+	vpmuludq	%zmm8,%zmm19,%zmm29
+	vpsrlq	$32,%zmm19,%zmm10
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+
+	vpmuludq	%zmm9,%zmm23,%zmm25
+	vpmuludq	%zmm9,%zmm24,%zmm26
+	vpmuludq	%zmm9,%zmm17,%zmm28
+	vpmuludq	%zmm9,%zmm18,%zmm29
+	vpmuludq	%zmm9,%zmm16,%zmm27
+	vpsrlq	$32,%zmm20,%zmm6
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpmuludq	%zmm10,%zmm22,%zmm25
+	vpmuludq	%zmm10,%zmm16,%zmm28
+	vpmuludq	%zmm10,%zmm17,%zmm29
+	vpmuludq	%zmm10,%zmm23,%zmm26
+	vpmuludq	%zmm10,%zmm24,%zmm27
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpmuludq	%zmm6,%zmm24,%zmm28
+	vpmuludq	%zmm6,%zmm16,%zmm29
+	vpmuludq	%zmm6,%zmm21,%zmm25
+	vpmuludq	%zmm6,%zmm22,%zmm26
+	vpmuludq	%zmm6,%zmm23,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+
+
+	vmovdqu64	0(%rsi),%zmm10
+	vmovdqu64	64(%rsi),%zmm6
+	leaq	128(%rsi),%rsi
+
+
+
+
+	vpsrlq	$26,%zmm14,%zmm28
+	vpandq	%zmm5,%zmm14,%zmm14
+	vpaddq	%zmm28,%zmm15,%zmm15
+
+	vpsrlq	$26,%zmm11,%zmm25
+	vpandq	%zmm5,%zmm11,%zmm11
+	vpaddq	%zmm25,%zmm12,%zmm12
+
+	vpsrlq	$26,%zmm15,%zmm29
+	vpandq	%zmm5,%zmm15,%zmm15
+
+	vpsrlq	$26,%zmm12,%zmm26
+	vpandq	%zmm5,%zmm12,%zmm12
+	vpaddq	%zmm26,%zmm13,%zmm13
+
+	vpaddq	%zmm29,%zmm11,%zmm11
+	vpsllq	$2,%zmm29,%zmm29
+	vpaddq	%zmm29,%zmm11,%zmm11
+
+	vpsrlq	$26,%zmm13,%zmm27
+	vpandq	%zmm5,%zmm13,%zmm13
+	vpaddq	%zmm27,%zmm14,%zmm14
+
+	vpsrlq	$26,%zmm11,%zmm25
+	vpandq	%zmm5,%zmm11,%zmm11
+	vpaddq	%zmm25,%zmm12,%zmm12
+
+	vpsrlq	$26,%zmm14,%zmm28
+	vpandq	%zmm5,%zmm14,%zmm14
+	vpaddq	%zmm28,%zmm15,%zmm15
+
+
+
+
+
+	vpunpcklqdq	%zmm6,%zmm10,%zmm7
+	vpunpckhqdq	%zmm6,%zmm10,%zmm6
+
+
+
+
+
+
+	vmovdqa32	128(%rcx),%zmm25
+	movl	$0x7777,%eax
+	kmovw	%eax,%k1
+
+	vpermd	%zmm16,%zmm25,%zmm16
+	vpermd	%zmm17,%zmm25,%zmm17
+	vpermd	%zmm18,%zmm25,%zmm18
+	vpermd	%zmm19,%zmm25,%zmm19
+	vpermd	%zmm20,%zmm25,%zmm20
+
+	vpermd	%zmm11,%zmm25,%zmm16{%k1}
+	vpermd	%zmm12,%zmm25,%zmm17{%k1}
+	vpermd	%zmm13,%zmm25,%zmm18{%k1}
+	vpermd	%zmm14,%zmm25,%zmm19{%k1}
+	vpermd	%zmm15,%zmm25,%zmm20{%k1}
+
+	vpslld	$2,%zmm17,%zmm21
+	vpslld	$2,%zmm18,%zmm22
+	vpslld	$2,%zmm19,%zmm23
+	vpslld	$2,%zmm20,%zmm24
+	vpaddd	%zmm17,%zmm21,%zmm21
+	vpaddd	%zmm18,%zmm22,%zmm22
+	vpaddd	%zmm19,%zmm23,%zmm23
+	vpaddd	%zmm20,%zmm24,%zmm24
+
+	vpbroadcastq	32(%rcx),%zmm30
+
+	vpsrlq	$52,%zmm7,%zmm9
+	vpsllq	$12,%zmm6,%zmm10
+	vporq	%zmm10,%zmm9,%zmm9
+	vpsrlq	$26,%zmm7,%zmm8
+	vpsrlq	$14,%zmm6,%zmm10
+	vpsrlq	$40,%zmm6,%zmm6
+	vpandq	%zmm5,%zmm9,%zmm9
+	vpandq	%zmm5,%zmm7,%zmm7
+
+
+
+
+	vpaddq	%zmm2,%zmm9,%zmm2
+	subq	$192,%rdx
+	jbe	.Ltail_avx512
+	jmp	.Loop_avx512
+
+.align	32
+.Loop_avx512:
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+	vpmuludq	%zmm2,%zmm17,%zmm14
+	vpaddq	%zmm0,%zmm7,%zmm0
+	vpmuludq	%zmm2,%zmm18,%zmm15
+	vpandq	%zmm5,%zmm8,%zmm8
+	vpmuludq	%zmm2,%zmm23,%zmm11
+	vpandq	%zmm5,%zmm10,%zmm10
+	vpmuludq	%zmm2,%zmm24,%zmm12
+	vporq	%zmm30,%zmm6,%zmm6
+	vpmuludq	%zmm2,%zmm16,%zmm13
+	vpaddq	%zmm1,%zmm8,%zmm1
+	vpaddq	%zmm3,%zmm10,%zmm3
+	vpaddq	%zmm4,%zmm6,%zmm4
+
+	vmovdqu64	0(%rsi),%zmm10
+	vmovdqu64	64(%rsi),%zmm6
+	leaq	128(%rsi),%rsi
+	vpmuludq	%zmm0,%zmm19,%zmm28
+	vpmuludq	%zmm0,%zmm20,%zmm29
+	vpmuludq	%zmm0,%zmm16,%zmm25
+	vpmuludq	%zmm0,%zmm17,%zmm26
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+
+	vpmuludq	%zmm1,%zmm18,%zmm28
+	vpmuludq	%zmm1,%zmm19,%zmm29
+	vpmuludq	%zmm1,%zmm24,%zmm25
+	vpmuludq	%zmm0,%zmm18,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpunpcklqdq	%zmm6,%zmm10,%zmm7
+	vpunpckhqdq	%zmm6,%zmm10,%zmm6
+
+	vpmuludq	%zmm3,%zmm16,%zmm28
+	vpmuludq	%zmm3,%zmm17,%zmm29
+	vpmuludq	%zmm1,%zmm16,%zmm26
+	vpmuludq	%zmm1,%zmm17,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpmuludq	%zmm4,%zmm24,%zmm28
+	vpmuludq	%zmm4,%zmm16,%zmm29
+	vpmuludq	%zmm3,%zmm22,%zmm25
+	vpmuludq	%zmm3,%zmm23,%zmm26
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpmuludq	%zmm3,%zmm24,%zmm27
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpmuludq	%zmm4,%zmm21,%zmm25
+	vpmuludq	%zmm4,%zmm22,%zmm26
+	vpmuludq	%zmm4,%zmm23,%zmm27
+	vpaddq	%zmm25,%zmm11,%zmm0
+	vpaddq	%zmm26,%zmm12,%zmm1
+	vpaddq	%zmm27,%zmm13,%zmm2
+
+
+
+
+	vpsrlq	$52,%zmm7,%zmm9
+	vpsllq	$12,%zmm6,%zmm10
+
+	vpsrlq	$26,%zmm14,%zmm3
+	vpandq	%zmm5,%zmm14,%zmm14
+	vpaddq	%zmm3,%zmm15,%zmm4
+
+	vporq	%zmm10,%zmm9,%zmm9
+
+	vpsrlq	$26,%zmm0,%zmm11
+	vpandq	%zmm5,%zmm0,%zmm0
+	vpaddq	%zmm11,%zmm1,%zmm1
+
+	vpandq	%zmm5,%zmm9,%zmm9
+
+	vpsrlq	$26,%zmm4,%zmm15
+	vpandq	%zmm5,%zmm4,%zmm4
+
+	vpsrlq	$26,%zmm1,%zmm12
+	vpandq	%zmm5,%zmm1,%zmm1
+	vpaddq	%zmm12,%zmm2,%zmm2
+
+	vpaddq	%zmm15,%zmm0,%zmm0
+	vpsllq	$2,%zmm15,%zmm15
+	vpaddq	%zmm15,%zmm0,%zmm0
+
+	vpaddq	%zmm9,%zmm2,%zmm2
+	vpsrlq	$26,%zmm7,%zmm8
+
+	vpsrlq	$26,%zmm2,%zmm13
+	vpandq	%zmm5,%zmm2,%zmm2
+	vpaddq	%zmm13,%zmm14,%zmm3
+
+	vpsrlq	$14,%zmm6,%zmm10
+
+	vpsrlq	$26,%zmm0,%zmm11
+	vpandq	%zmm5,%zmm0,%zmm0
+	vpaddq	%zmm11,%zmm1,%zmm1
+
+	vpsrlq	$40,%zmm6,%zmm6
+
+	vpsrlq	$26,%zmm3,%zmm14
+	vpandq	%zmm5,%zmm3,%zmm3
+	vpaddq	%zmm14,%zmm4,%zmm4
+
+	vpandq	%zmm5,%zmm7,%zmm7
+
+
+
+
+	subq	$128,%rdx
+	ja	.Loop_avx512
+
+.Ltail_avx512:
+
+
+
+
+
+	vpsrlq	$32,%zmm16,%zmm16
+	vpsrlq	$32,%zmm17,%zmm17
+	vpsrlq	$32,%zmm18,%zmm18
+	vpsrlq	$32,%zmm23,%zmm23
+	vpsrlq	$32,%zmm24,%zmm24
+	vpsrlq	$32,%zmm19,%zmm19
+	vpsrlq	$32,%zmm20,%zmm20
+	vpsrlq	$32,%zmm21,%zmm21
+	vpsrlq	$32,%zmm22,%zmm22
+
+
+
+	leaq	(%rsi,%rdx,1),%rsi
+
+
+	vpaddq	%zmm0,%zmm7,%zmm0
+
+	vpmuludq	%zmm2,%zmm17,%zmm14
+	vpmuludq	%zmm2,%zmm18,%zmm15
+	vpmuludq	%zmm2,%zmm23,%zmm11
+	vpandq	%zmm5,%zmm8,%zmm8
+	vpmuludq	%zmm2,%zmm24,%zmm12
+	vpandq	%zmm5,%zmm10,%zmm10
+	vpmuludq	%zmm2,%zmm16,%zmm13
+	vporq	%zmm30,%zmm6,%zmm6
+	vpaddq	%zmm1,%zmm8,%zmm1
+	vpaddq	%zmm3,%zmm10,%zmm3
+	vpaddq	%zmm4,%zmm6,%zmm4
+
+	vmovdqu	0(%rsi),%xmm7
+	vpmuludq	%zmm0,%zmm19,%zmm28
+	vpmuludq	%zmm0,%zmm20,%zmm29
+	vpmuludq	%zmm0,%zmm16,%zmm25
+	vpmuludq	%zmm0,%zmm17,%zmm26
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+
+	vmovdqu	16(%rsi),%xmm8
+	vpmuludq	%zmm1,%zmm18,%zmm28
+	vpmuludq	%zmm1,%zmm19,%zmm29
+	vpmuludq	%zmm1,%zmm24,%zmm25
+	vpmuludq	%zmm0,%zmm18,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vinserti128	$1,32(%rsi),%ymm7,%ymm7
+	vpmuludq	%zmm3,%zmm16,%zmm28
+	vpmuludq	%zmm3,%zmm17,%zmm29
+	vpmuludq	%zmm1,%zmm16,%zmm26
+	vpmuludq	%zmm1,%zmm17,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm14
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vinserti128	$1,48(%rsi),%ymm8,%ymm8
+	vpmuludq	%zmm4,%zmm24,%zmm28
+	vpmuludq	%zmm4,%zmm16,%zmm29
+	vpmuludq	%zmm3,%zmm22,%zmm25
+	vpmuludq	%zmm3,%zmm23,%zmm26
+	vpmuludq	%zmm3,%zmm24,%zmm27
+	vpaddq	%zmm28,%zmm14,%zmm3
+	vpaddq	%zmm29,%zmm15,%zmm15
+	vpaddq	%zmm25,%zmm11,%zmm11
+	vpaddq	%zmm26,%zmm12,%zmm12
+	vpaddq	%zmm27,%zmm13,%zmm13
+
+	vpmuludq	%zmm4,%zmm21,%zmm25
+	vpmuludq	%zmm4,%zmm22,%zmm26
+	vpmuludq	%zmm4,%zmm23,%zmm27
+	vpaddq	%zmm25,%zmm11,%zmm0
+	vpaddq	%zmm26,%zmm12,%zmm1
+	vpaddq	%zmm27,%zmm13,%zmm2
+
+
+
+
+	movl	$1,%eax
+	vpermq	$0xb1,%zmm3,%zmm14
+	vpermq	$0xb1,%zmm15,%zmm4
+	vpermq	$0xb1,%zmm0,%zmm11
+	vpermq	$0xb1,%zmm1,%zmm12
+	vpermq	$0xb1,%zmm2,%zmm13
+	vpaddq	%zmm14,%zmm3,%zmm3
+	vpaddq	%zmm15,%zmm4,%zmm4
+	vpaddq	%zmm11,%zmm0,%zmm0
+	vpaddq	%zmm12,%zmm1,%zmm1
+	vpaddq	%zmm13,%zmm2,%zmm2
+
+	kmovw	%eax,%k3
+	vpermq	$0x2,%zmm3,%zmm14
+	vpermq	$0x2,%zmm4,%zmm15
+	vpermq	$0x2,%zmm0,%zmm11
+	vpermq	$0x2,%zmm1,%zmm12
+	vpermq	$0x2,%zmm2,%zmm13
+	vpaddq	%zmm14,%zmm3,%zmm3
+	vpaddq	%zmm15,%zmm4,%zmm4
+	vpaddq	%zmm11,%zmm0,%zmm0
+	vpaddq	%zmm12,%zmm1,%zmm1
+	vpaddq	%zmm13,%zmm2,%zmm2
+
+	vextracti64x4	$0x1,%zmm3,%ymm14
+	vextracti64x4	$0x1,%zmm4,%ymm15
+	vextracti64x4	$0x1,%zmm0,%ymm11
+	vextracti64x4	$0x1,%zmm1,%ymm12
+	vextracti64x4	$0x1,%zmm2,%ymm13
+	vpaddq	%zmm14,%zmm3,%zmm3{%k3}{z}
+	vpaddq	%zmm15,%zmm4,%zmm4{%k3}{z}
+	vpaddq	%zmm11,%zmm0,%zmm0{%k3}{z}
+	vpaddq	%zmm12,%zmm1,%zmm1{%k3}{z}
+	vpaddq	%zmm13,%zmm2,%zmm2{%k3}{z}
+
+
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpsrldq	$6,%ymm7,%ymm9
+	vpsrldq	$6,%ymm8,%ymm10
+	vpunpckhqdq	%ymm8,%ymm7,%ymm6
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpunpcklqdq	%ymm10,%ymm9,%ymm9
+	vpunpcklqdq	%ymm8,%ymm7,%ymm7
+	vpaddq	%ymm11,%ymm1,%ymm1
+
+	vpsrlq	$26,%ymm4,%ymm15
+	vpand	%ymm5,%ymm4,%ymm4
+
+	vpsrlq	$26,%ymm1,%ymm12
+	vpand	%ymm5,%ymm1,%ymm1
+	vpsrlq	$30,%ymm9,%ymm10
+	vpsrlq	$4,%ymm9,%ymm9
+	vpaddq	%ymm12,%ymm2,%ymm2
+
+	vpaddq	%ymm15,%ymm0,%ymm0
+	vpsllq	$2,%ymm15,%ymm15
+	vpsrlq	$26,%ymm7,%ymm8
+	vpsrlq	$40,%ymm6,%ymm6
+	vpaddq	%ymm15,%ymm0,%ymm0
+
+	vpsrlq	$26,%ymm2,%ymm13
+	vpand	%ymm5,%ymm2,%ymm2
+	vpand	%ymm5,%ymm9,%ymm9
+	vpand	%ymm5,%ymm7,%ymm7
+	vpaddq	%ymm13,%ymm3,%ymm3
+
+	vpsrlq	$26,%ymm0,%ymm11
+	vpand	%ymm5,%ymm0,%ymm0
+	vpaddq	%ymm2,%ymm9,%ymm2
+	vpand	%ymm5,%ymm8,%ymm8
+	vpaddq	%ymm11,%ymm1,%ymm1
+
+	vpsrlq	$26,%ymm3,%ymm14
+	vpand	%ymm5,%ymm3,%ymm3
+	vpand	%ymm5,%ymm10,%ymm10
+	vpor	32(%rcx),%ymm6,%ymm6
+	vpaddq	%ymm14,%ymm4,%ymm4
+
+	leaq	144(%rsp),%rax
+	addq	$64,%rdx
+	jnz	.Ltail_avx2
+
+	vpsubq	%ymm9,%ymm2,%ymm2
+	vmovd	%xmm0,-112(%rdi)
+	vmovd	%xmm1,-108(%rdi)
+	vmovd	%xmm2,-104(%rdi)
+	vmovd	%xmm3,-100(%rdi)
+	vmovd	%xmm4,-96(%rdi)
+	vzeroall
+	leaq	8(%r11),%rsp
+.cfi_def_cfa	%rsp,8
+	.byte	0xf3,0xc3
+.cfi_endproc	
+.size	poly1305_blocks_avx512,.-poly1305_blocks_avx512
+.type	poly1305_init_base2_44,@function
+.align	32
+poly1305_init_base2_44:
+	xorq	%rax,%rax
+	movq	%rax,0(%rdi)
+	movq	%rax,8(%rdi)
+	movq	%rax,16(%rdi)
+
+.Linit_base2_44:
+	leaq	poly1305_blocks_vpmadd52(%rip),%r10
+	leaq	poly1305_emit_base2_44(%rip),%r11
+
+	movq	$0x0ffffffc0fffffff,%rax
+	movq	$0x0ffffffc0ffffffc,%rcx
+	andq	0(%rsi),%rax
+	movq	$0x00000fffffffffff,%r8
+	andq	8(%rsi),%rcx
+	movq	$0x00000fffffffffff,%r9
+	andq	%rax,%r8
+	shrdq	$44,%rcx,%rax
+	movq	%r8,40(%rdi)
+	andq	%r9,%rax
+	shrq	$24,%rcx
+	movq	%rax,48(%rdi)
+	leaq	(%rax,%rax,4),%rax
+	movq	%rcx,56(%rdi)
+	shlq	$2,%rax
+	leaq	(%rcx,%rcx,4),%rcx
+	shlq	$2,%rcx
+	movq	%rax,24(%rdi)
+	movq	%rcx,32(%rdi)
+	movq	$-1,64(%rdi)
+	movq	%r10,0(%rdx)
+	movq	%r11,8(%rdx)
+	movl	$1,%eax
+	.byte	0xf3,0xc3
+.size	poly1305_init_base2_44,.-poly1305_init_base2_44
+.type	poly1305_blocks_vpmadd52,@function
+.align	32
+poly1305_blocks_vpmadd52:
+	shrq	$4,%rdx
+	jz	.Lno_data_vpmadd52
+
+	shlq	$40,%rcx
+	movq	64(%rdi),%r8
+
+
+
+
+
+
+	movq	$3,%rax
+	movq	$1,%r10
+	cmpq	$4,%rdx
+	cmovaeq	%r10,%rax
+	testq	%r8,%r8
+	cmovnsq	%r10,%rax
+
+	andq	%rdx,%rax
+	jz	.Lblocks_vpmadd52_4x
+
+	subq	%rax,%rdx
+	movl	$7,%r10d
+	movl	$1,%r11d
+	kmovw	%r10d,%k7
+	leaq	.L2_44_inp_permd(%rip),%r10
+	kmovw	%r11d,%k1
+
+	vmovq	%rcx,%xmm21
+	vmovdqa64	0(%r10),%ymm19
+	vmovdqa64	32(%r10),%ymm20
+	vpermq	$0xcf,%ymm21,%ymm21
+	vmovdqa64	64(%r10),%ymm22
+
+	vmovdqu64	0(%rdi),%ymm16{%k7}{z}
+	vmovdqu64	40(%rdi),%ymm3{%k7}{z}
+	vmovdqu64	32(%rdi),%ymm4{%k7}{z}
+	vmovdqu64	24(%rdi),%ymm5{%k7}{z}
+
+	vmovdqa64	96(%r10),%ymm23
+	vmovdqa64	128(%r10),%ymm24
+
+	jmp	.Loop_vpmadd52
+
+.align	32
+.Loop_vpmadd52:
+	vmovdqu32	0(%rsi),%xmm18
+	leaq	16(%rsi),%rsi
+
+	vpermd	%ymm18,%ymm19,%ymm18
+	vpsrlvq	%ymm20,%ymm18,%ymm18
+	vpandq	%ymm22,%ymm18,%ymm18
+	vporq	%ymm21,%ymm18,%ymm18
+
+	vpaddq	%ymm18,%ymm16,%ymm16
+
+	vpermq	$0,%ymm16,%ymm0{%k7}{z}
+	vpermq	$85,%ymm16,%ymm1{%k7}{z}
+	vpermq	$170,%ymm16,%ymm2{%k7}{z}
+
+	vpxord	%ymm16,%ymm16,%ymm16
+	vpxord	%ymm17,%ymm17,%ymm17
+
+	vpmadd52luq	%ymm3,%ymm0,%ymm16
+	vpmadd52huq	%ymm3,%ymm0,%ymm17
+
+	vpmadd52luq	%ymm4,%ymm1,%ymm16
+	vpmadd52huq	%ymm4,%ymm1,%ymm17
+
+	vpmadd52luq	%ymm5,%ymm2,%ymm16
+	vpmadd52huq	%ymm5,%ymm2,%ymm17
+
+	vpsrlvq	%ymm23,%ymm16,%ymm18
+	vpsllvq	%ymm24,%ymm17,%ymm17
+	vpandq	%ymm22,%ymm16,%ymm16
+
+	vpaddq	%ymm18,%ymm17,%ymm17
+
+	vpermq	$147,%ymm17,%ymm17
+
+	vpaddq	%ymm17,%ymm16,%ymm16
+
+	vpsrlvq	%ymm23,%ymm16,%ymm18
+	vpandq	%ymm22,%ymm16,%ymm16
+
+	vpermq	$147,%ymm18,%ymm18
+
+	vpaddq	%ymm18,%ymm16,%ymm16
+
+	vpermq	$147,%ymm16,%ymm18{%k1}{z}
+
+	vpaddq	%ymm18,%ymm16,%ymm16
+	vpsllq	$2,%ymm18,%ymm18
+
+	vpaddq	%ymm18,%ymm16,%ymm16
+
+	decq	%rax
+	jnz	.Loop_vpmadd52
+
+	vmovdqu64	%ymm16,0(%rdi){%k7}
+
+	testq	%rdx,%rdx
+	jnz	.Lblocks_vpmadd52_4x
+
+.Lno_data_vpmadd52:
+	.byte	0xf3,0xc3
+.size	poly1305_blocks_vpmadd52,.-poly1305_blocks_vpmadd52
+.type	poly1305_blocks_vpmadd52_4x,@function
+.align	32
+poly1305_blocks_vpmadd52_4x:
+	shrq	$4,%rdx
+	jz	.Lno_data_vpmadd52_4x
+
+	shlq	$40,%rcx
+	movq	64(%rdi),%r8
+
+.Lblocks_vpmadd52_4x:
+	vpbroadcastq	%rcx,%ymm31
+
+	vmovdqa64	.Lx_mask44(%rip),%ymm28
+	movl	$5,%eax
+	vmovdqa64	.Lx_mask42(%rip),%ymm29
+	kmovw	%eax,%k1
+
+	testq	%r8,%r8
+	js	.Linit_vpmadd52
+
+	vmovq	0(%rdi),%xmm0
+	vmovq	8(%rdi),%xmm1
+	vmovq	16(%rdi),%xmm2
+
+	testq	$3,%rdx
+	jnz	.Lblocks_vpmadd52_2x_do
+
+.Lblocks_vpmadd52_4x_do:
+	vpbroadcastq	64(%rdi),%ymm3
+	vpbroadcastq	96(%rdi),%ymm4
+	vpbroadcastq	128(%rdi),%ymm5
+	vpbroadcastq	160(%rdi),%ymm16
+
+.Lblocks_vpmadd52_4x_key_loaded:
+	vpsllq	$2,%ymm5,%ymm17
+	vpaddq	%ymm5,%ymm17,%ymm17
+	vpsllq	$2,%ymm17,%ymm17
+
+	testq	$7,%rdx
+	jz	.Lblocks_vpmadd52_8x
+
+	vmovdqu64	0(%rsi),%ymm26
+	vmovdqu64	32(%rsi),%ymm27
+	leaq	64(%rsi),%rsi
+
+	vpunpcklqdq	%ymm27,%ymm26,%ymm25
+	vpunpckhqdq	%ymm27,%ymm26,%ymm27
+
+
+
+	vpsrlq	$24,%ymm27,%ymm26
+	vporq	%ymm31,%ymm26,%ymm26
+	vpaddq	%ymm26,%ymm2,%ymm2
+	vpandq	%ymm28,%ymm25,%ymm24
+	vpsrlq	$44,%ymm25,%ymm25
+	vpsllq	$20,%ymm27,%ymm27
+	vporq	%ymm27,%ymm25,%ymm25
+	vpandq	%ymm28,%ymm25,%ymm25
+
+	subq	$4,%rdx
+	jz	.Ltail_vpmadd52_4x
+	jmp	.Loop_vpmadd52_4x
+	ud2
+
+.align	32
+.Linit_vpmadd52:
+	vmovq	24(%rdi),%xmm16
+	vmovq	56(%rdi),%xmm2
+	vmovq	32(%rdi),%xmm17
+	vmovq	40(%rdi),%xmm3
+	vmovq	48(%rdi),%xmm4
+
+	vmovdqa	%ymm3,%ymm0
+	vmovdqa	%ymm4,%ymm1
+	vmovdqa	%ymm2,%ymm5
+
+	movl	$2,%eax
+
+.Lmul_init_vpmadd52:
+	vpxorq	%ymm18,%ymm18,%ymm18
+	vpmadd52luq	%ymm2,%ymm16,%ymm18
+	vpxorq	%ymm19,%ymm19,%ymm19
+	vpmadd52huq	%ymm2,%ymm16,%ymm19
+	vpxorq	%ymm20,%ymm20,%ymm20
+	vpmadd52luq	%ymm2,%ymm17,%ymm20
+	vpxorq	%ymm21,%ymm21,%ymm21
+	vpmadd52huq	%ymm2,%ymm17,%ymm21
+	vpxorq	%ymm22,%ymm22,%ymm22
+	vpmadd52luq	%ymm2,%ymm3,%ymm22
+	vpxorq	%ymm23,%ymm23,%ymm23
+	vpmadd52huq	%ymm2,%ymm3,%ymm23
+
+	vpmadd52luq	%ymm0,%ymm3,%ymm18
+	vpmadd52huq	%ymm0,%ymm3,%ymm19
+	vpmadd52luq	%ymm0,%ymm4,%ymm20
+	vpmadd52huq	%ymm0,%ymm4,%ymm21
+	vpmadd52luq	%ymm0,%ymm5,%ymm22
+	vpmadd52huq	%ymm0,%ymm5,%ymm23
+
+	vpmadd52luq	%ymm1,%ymm17,%ymm18
+	vpmadd52huq	%ymm1,%ymm17,%ymm19
+	vpmadd52luq	%ymm1,%ymm3,%ymm20
+	vpmadd52huq	%ymm1,%ymm3,%ymm21
+	vpmadd52luq	%ymm1,%ymm4,%ymm22
+	vpmadd52huq	%ymm1,%ymm4,%ymm23
+
+
+
+	vpsrlq	$44,%ymm18,%ymm30
+	vpsllq	$8,%ymm19,%ymm19
+	vpandq	%ymm28,%ymm18,%ymm0
+	vpaddq	%ymm30,%ymm19,%ymm19
+
+	vpaddq	%ymm19,%ymm20,%ymm20
+
+	vpsrlq	$44,%ymm20,%ymm30
+	vpsllq	$8,%ymm21,%ymm21
+	vpandq	%ymm28,%ymm20,%ymm1
+	vpaddq	%ymm30,%ymm21,%ymm21
+
+	vpaddq	%ymm21,%ymm22,%ymm22
+
+	vpsrlq	$42,%ymm22,%ymm30
+	vpsllq	$10,%ymm23,%ymm23
+	vpandq	%ymm29,%ymm22,%ymm2
+	vpaddq	%ymm30,%ymm23,%ymm23
+
+	vpaddq	%ymm23,%ymm0,%ymm0
+	vpsllq	$2,%ymm23,%ymm23
+
+	vpaddq	%ymm23,%ymm0,%ymm0
+
+	vpsrlq	$44,%ymm0,%ymm30
+	vpandq	%ymm28,%ymm0,%ymm0
+
+	vpaddq	%ymm30,%ymm1,%ymm1
+
+	decl	%eax
+	jz	.Ldone_init_vpmadd52
+
+	vpunpcklqdq	%ymm4,%ymm1,%ymm4
+	vpbroadcastq	%xmm1,%xmm1
+	vpunpcklqdq	%ymm5,%ymm2,%ymm5
+	vpbroadcastq	%xmm2,%xmm2
+	vpunpcklqdq	%ymm3,%ymm0,%ymm3
+	vpbroadcastq	%xmm0,%xmm0
+
+	vpsllq	$2,%ymm4,%ymm16
+	vpsllq	$2,%ymm5,%ymm17
+	vpaddq	%ymm4,%ymm16,%ymm16
+	vpaddq	%ymm5,%ymm17,%ymm17
+	vpsllq	$2,%ymm16,%ymm16
+	vpsllq	$2,%ymm17,%ymm17
+
+	jmp	.Lmul_init_vpmadd52
+	ud2
+
+.align	32
+.Ldone_init_vpmadd52:
+	vinserti128	$1,%xmm4,%ymm1,%ymm4
+	vinserti128	$1,%xmm5,%ymm2,%ymm5
+	vinserti128	$1,%xmm3,%ymm0,%ymm3
+
+	vpermq	$216,%ymm4,%ymm4
+	vpermq	$216,%ymm5,%ymm5
+	vpermq	$216,%ymm3,%ymm3
+
+	vpsllq	$2,%ymm4,%ymm16
+	vpaddq	%ymm4,%ymm16,%ymm16
+	vpsllq	$2,%ymm16,%ymm16
+
+	vmovq	0(%rdi),%xmm0
+	vmovq	8(%rdi),%xmm1
+	vmovq	16(%rdi),%xmm2
+
+	testq	$3,%rdx
+	jnz	.Ldone_init_vpmadd52_2x
+
+	vmovdqu64	%ymm3,64(%rdi)
+	vpbroadcastq	%xmm3,%ymm3
+	vmovdqu64	%ymm4,96(%rdi)
+	vpbroadcastq	%xmm4,%ymm4
+	vmovdqu64	%ymm5,128(%rdi)
+	vpbroadcastq	%xmm5,%ymm5
+	vmovdqu64	%ymm16,160(%rdi)
+	vpbroadcastq	%xmm16,%ymm16
+
+	jmp	.Lblocks_vpmadd52_4x_key_loaded
+	ud2
+
+.align	32
+.Ldone_init_vpmadd52_2x:
+	vmovdqu64	%ymm3,64(%rdi)
+	vpsrldq	$8,%ymm3,%ymm3
+	vmovdqu64	%ymm4,96(%rdi)
+	vpsrldq	$8,%ymm4,%ymm4
+	vmovdqu64	%ymm5,128(%rdi)
+	vpsrldq	$8,%ymm5,%ymm5
+	vmovdqu64	%ymm16,160(%rdi)
+	vpsrldq	$8,%ymm16,%ymm16
+	jmp	.Lblocks_vpmadd52_2x_key_loaded
+	ud2
+
+.align	32
+.Lblocks_vpmadd52_2x_do:
+	vmovdqu64	128+8(%rdi),%ymm5{%k1}{z}
+	vmovdqu64	160+8(%rdi),%ymm16{%k1}{z}
+	vmovdqu64	64+8(%rdi),%ymm3{%k1}{z}
+	vmovdqu64	96+8(%rdi),%ymm4{%k1}{z}
+
+.Lblocks_vpmadd52_2x_key_loaded:
+	vmovdqu64	0(%rsi),%ymm26
+	vpxorq	%ymm27,%ymm27,%ymm27
+	leaq	32(%rsi),%rsi
+
+	vpunpcklqdq	%ymm27,%ymm26,%ymm25
+	vpunpckhqdq	%ymm27,%ymm26,%ymm27
+
+
+
+	vpsrlq	$24,%ymm27,%ymm26
+	vporq	%ymm31,%ymm26,%ymm26
+	vpaddq	%ymm26,%ymm2,%ymm2
+	vpandq	%ymm28,%ymm25,%ymm24
+	vpsrlq	$44,%ymm25,%ymm25
+	vpsllq	$20,%ymm27,%ymm27
+	vporq	%ymm27,%ymm25,%ymm25
+	vpandq	%ymm28,%ymm25,%ymm25
+
+	jmp	.Ltail_vpmadd52_2x
+	ud2
+
+.align	32
+.Loop_vpmadd52_4x:
+
+	vpaddq	%ymm24,%ymm0,%ymm0
+	vpaddq	%ymm25,%ymm1,%ymm1
+
+	vpxorq	%ymm18,%ymm18,%ymm18
+	vpmadd52luq	%ymm2,%ymm16,%ymm18
+	vpxorq	%ymm19,%ymm19,%ymm19
+	vpmadd52huq	%ymm2,%ymm16,%ymm19
+	vpxorq	%ymm20,%ymm20,%ymm20
+	vpmadd52luq	%ymm2,%ymm17,%ymm20
+	vpxorq	%ymm21,%ymm21,%ymm21
+	vpmadd52huq	%ymm2,%ymm17,%ymm21
+	vpxorq	%ymm22,%ymm22,%ymm22
+	vpmadd52luq	%ymm2,%ymm3,%ymm22
+	vpxorq	%ymm23,%ymm23,%ymm23
+	vpmadd52huq	%ymm2,%ymm3,%ymm23
+
+	vmovdqu64	0(%rsi),%ymm26
+	vmovdqu64	32(%rsi),%ymm27
+	leaq	64(%rsi),%rsi
+	vpmadd52luq	%ymm0,%ymm3,%ymm18
+	vpmadd52huq	%ymm0,%ymm3,%ymm19
+	vpmadd52luq	%ymm0,%ymm4,%ymm20
+	vpmadd52huq	%ymm0,%ymm4,%ymm21
+	vpmadd52luq	%ymm0,%ymm5,%ymm22
+	vpmadd52huq	%ymm0,%ymm5,%ymm23
+
+	vpunpcklqdq	%ymm27,%ymm26,%ymm25
+	vpunpckhqdq	%ymm27,%ymm26,%ymm27
+	vpmadd52luq	%ymm1,%ymm17,%ymm18
+	vpmadd52huq	%ymm1,%ymm17,%ymm19
+	vpmadd52luq	%ymm1,%ymm3,%ymm20
+	vpmadd52huq	%ymm1,%ymm3,%ymm21
+	vpmadd52luq	%ymm1,%ymm4,%ymm22
+	vpmadd52huq	%ymm1,%ymm4,%ymm23
+
+
+
+	vpsrlq	$44,%ymm18,%ymm30
+	vpsllq	$8,%ymm19,%ymm19
+	vpandq	%ymm28,%ymm18,%ymm0
+	vpaddq	%ymm30,%ymm19,%ymm19
+
+	vpsrlq	$24,%ymm27,%ymm26
+	vporq	%ymm31,%ymm26,%ymm26
+	vpaddq	%ymm19,%ymm20,%ymm20
+
+	vpsrlq	$44,%ymm20,%ymm30
+	vpsllq	$8,%ymm21,%ymm21
+	vpandq	%ymm28,%ymm20,%ymm1
+	vpaddq	%ymm30,%ymm21,%ymm21
+
+	vpandq	%ymm28,%ymm25,%ymm24
+	vpsrlq	$44,%ymm25,%ymm25
+	vpsllq	$20,%ymm27,%ymm27
+	vpaddq	%ymm21,%ymm22,%ymm22
+
+	vpsrlq	$42,%ymm22,%ymm30
+	vpsllq	$10,%ymm23,%ymm23
+	vpandq	%ymm29,%ymm22,%ymm2
+	vpaddq	%ymm30,%ymm23,%ymm23
+
+	vpaddq	%ymm26,%ymm2,%ymm2
+	vpaddq	%ymm23,%ymm0,%ymm0
+	vpsllq	$2,%ymm23,%ymm23
+
+	vpaddq	%ymm23,%ymm0,%ymm0
+	vporq	%ymm27,%ymm25,%ymm25
+	vpandq	%ymm28,%ymm25,%ymm25
+
+	vpsrlq	$44,%ymm0,%ymm30
+	vpandq	%ymm28,%ymm0,%ymm0
+
+	vpaddq	%ymm30,%ymm1,%ymm1
+
+	subq	$4,%rdx
+	jnz	.Loop_vpmadd52_4x
+
+.Ltail_vpmadd52_4x:
+	vmovdqu64	128(%rdi),%ymm5
+	vmovdqu64	160(%rdi),%ymm16
+	vmovdqu64	64(%rdi),%ymm3
+	vmovdqu64	96(%rdi),%ymm4
+
+.Ltail_vpmadd52_2x:
+	vpsllq	$2,%ymm5,%ymm17
+	vpaddq	%ymm5,%ymm17,%ymm17
+	vpsllq	$2,%ymm17,%ymm17
+
+
+	vpaddq	%ymm24,%ymm0,%ymm0
+	vpaddq	%ymm25,%ymm1,%ymm1
+
+	vpxorq	%ymm18,%ymm18,%ymm18
+	vpmadd52luq	%ymm2,%ymm16,%ymm18
+	vpxorq	%ymm19,%ymm19,%ymm19
+	vpmadd52huq	%ymm2,%ymm16,%ymm19
+	vpxorq	%ymm20,%ymm20,%ymm20
+	vpmadd52luq	%ymm2,%ymm17,%ymm20
+	vpxorq	%ymm21,%ymm21,%ymm21
+	vpmadd52huq	%ymm2,%ymm17,%ymm21
+	vpxorq	%ymm22,%ymm22,%ymm22
+	vpmadd52luq	%ymm2,%ymm3,%ymm22
+	vpxorq	%ymm23,%ymm23,%ymm23
+	vpmadd52huq	%ymm2,%ymm3,%ymm23
+
+	vpmadd52luq	%ymm0,%ymm3,%ymm18
+	vpmadd52huq	%ymm0,%ymm3,%ymm19
+	vpmadd52luq	%ymm0,%ymm4,%ymm20
+	vpmadd52huq	%ymm0,%ymm4,%ymm21
+	vpmadd52luq	%ymm0,%ymm5,%ymm22
+	vpmadd52huq	%ymm0,%ymm5,%ymm23
+
+	vpmadd52luq	%ymm1,%ymm17,%ymm18
+	vpmadd52huq	%ymm1,%ymm17,%ymm19
+	vpmadd52luq	%ymm1,%ymm3,%ymm20
+	vpmadd52huq	%ymm1,%ymm3,%ymm21
+	vpmadd52luq	%ymm1,%ymm4,%ymm22
+	vpmadd52huq	%ymm1,%ymm4,%ymm23
+
+
+
+
+	movl	$1,%eax
+	kmovw	%eax,%k1
+	vpsrldq	$8,%ymm18,%ymm24
+	vpsrldq	$8,%ymm19,%ymm0
+	vpsrldq	$8,%ymm20,%ymm25
+	vpsrldq	$8,%ymm21,%ymm1
+	vpaddq	%ymm24,%ymm18,%ymm18
+	vpaddq	%ymm0,%ymm19,%ymm19
+	vpsrldq	$8,%ymm22,%ymm26
+	vpsrldq	$8,%ymm23,%ymm2
+	vpaddq	%ymm25,%ymm20,%ymm20
+	vpaddq	%ymm1,%ymm21,%ymm21
+	vpermq	$0x2,%ymm18,%ymm24
+	vpermq	$0x2,%ymm19,%ymm0
[… more lines in the raw message]

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#13 lore raw

PATCH v8 13/28 zinc: Poly1305 x86_64 implementation

In reply to Jason A. Donenfeld

This ports AVX, AVX-2, and AVX-512F implementations for Poly1305. The AVX-512F implementation is disabled on Skylake, due to throttling. These come from Andy Polyakov's implementation, with the following modifications from Samuel Neves:

- Some cosmetic changes, like renaming labels to .Lname, constants, and other Linux conventions.

- CPU feature checking is done in C by the glue code, so that has been removed from the assembly.

- poly1305_blocks_avx512 jumped to the middle of the poly1305_blocks_avx2 for the final blocks. To appease objtool, the relevant tail avx2 code was duplicated for the avx512 function.

- The original uses %rbp as a scratch register. However, the kernel expects %rbp to be a valid frame pointer at any given time in order to do proper unwinding. Thus we need to alter the code in order to preserve it. The most straightforward manner in which this was accomplished was by replacing $d3, formerly %r10, by %rdi, and replacing %rbp by %r10. Because %rdi, a pointer to the context structure, does not change and is not used by poly1305_iteration, it is safe to use it here, and the overhead of saving and restoring it should be minimal.

- The original hardcodes returns as .byte 0xf3,0xc3, aka "rep ret". We replace this by "ret". "rep ret" was meant to help with AMD K8 chips, cf. http://repzret.org/p/repzret. It makes no sense to continue to use this kludge for code that won't even run on ancient AMD chips.

The AVX code uses base 2^26, while the scalar code uses base 2^64. If we hit the unfortunate situation of using AVX and then having to go back to scalar -- because the user is silly and has called the update function from two separate contexts -- then we need to convert back to the original base before proceeding. It is possible to reason that the initial reduction below is sufficient given the implementation invariants. However, for an avoidance of doubt and because this is not performance critical, we do the full reduction anyway. This conversion is found in the glue code, and a proof of correctness may be easily obtained from Z3: <https://xn--4db.cc/ltPtHCKN/py>.

Cycle counts on a Core i7 6700HQ using the AVX-2 codepath, comparing this implementation ("new") to the implementation in the current crypto api ("old"):

size old new
---- ---- ----
0 70 68
16 92 90
32 134 104
48 172 120
64 218 136
80 254 158
96 298 174
112 342 192
128 388 212
144 428 228
160 466 246
176 510 264
192 550 282
208 594 302
224 628 316
240 676 334
256 716 354
272 764 374
288 802 352
304 420 366
320 428 360
336 484 378
352 426 384
368 478 400
384 488 394
400 542 408
416 486 416
432 534 430
448 544 422
464 600 438
480 540 448
496 594 464
512 602 456
528 656 476
544 600 480
560 650 494
576 664 490
592 714 508
608 656 514
624 708 532
640 716 524
656 770 536
672 716 548
688 770 562
704 774 552
720 826 568
736 768 574
752 822 592
768 830 584
784 884 602
800 828 610
816 884 628
832 888 618
848 942 632
864 884 644
880 936 660
896 948 652
912 1000 664
928 942 676
944 994 690
960 1002 680
976 1054 694
992 1002 706
1008 1052 720

Cycle counts on a Xeon Gold 5120 using the AVX-512 codepath:

size old new
---- ---- ----
0 74 70
16 96 92
32 136 106
48 184 124
64 218 138
80 260 160
96 300 176
112 342 194
128 384 212
144 420 226
160 464 248
176 504 264
192 544 282
208 582 300
224 624 318
240 662 338
256 708 358
272 748 372
288 788 358
304 422 370
320 432 364
336 486 380
352 434 390
368 480 408
384 490 398
400 542 412
416 492 426
432 538 436
448 546 432
464 600 448
480 548 456
496 594 476
512 606 470
528 656 480
544 606 498
560 652 512
576 662 508
592 716 522
608 664 538
624 710 552
640 720 516
656 772 526
672 722 544
688 768 556
704 778 556
720 832 568
736 780 584
752 826 600
768 836 560
784 888 572
800 838 588
816 884 604
832 894 598
848 946 612
864 896 628
880 942 644
896 952 608
912 1004 616
928 954 634
944 1000 646
960 1008 646
976 1062 658
992 1012 674
1008 1058 690

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                             |    1 +
 lib/zinc/poly1305/poly1305-x86_64-glue.c      |  154 ++
 ...-x86_64-cryptogams.S => poly1305-x86_64.S} | 2459 ++++++-----------
 lib/zinc/poly1305/poly1305.c                  |    4 +
 4 files changed, 1002 insertions(+), 1616 deletions(-)
 create mode 100644 lib/zinc/poly1305/poly1305-x86_64-glue.c
 rename lib/zinc/poly1305/{poly1305-x86_64-cryptogams.S => poly1305-x86_64.S} (58%)
lib/zinc/Makefile +1 -0
index 6fc9626c55fa..a8943d960b6a 100644
@@ -11,4 +11,5 @@ AFLAGS_chacha20-mips.o += -O2 # This is required to fill the branch delay slots
 obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
 
 zinc_poly1305-y := poly1305/poly1305.o
+zinc_poly1305-$(CONFIG_ZINC_ARCH_X86_64) += poly1305/poly1305-x86_64.o
 obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
lib/zinc/poly1305/poly1305-x86_64-glue.c new file +154 -0

154 lines not on this page · raw message

lib/zinc/poly1305/poly1305-x86_64.S renamed from lib/zinc/poly1305/poly1305-x86_64-cryptogams.S +843 -1616

2,841 lines not on this page · raw message

lib/zinc/poly1305/poly1305.c +4 -0
index d4922d230c4f..0a580f1328fc 100644

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#14 lore raw

PATCH v8 14/28 zinc: import Andy Polyakov's Poly1305 ARM and ARM64 implementations

In reply to Jason A. Donenfeld

These NEON and non-NEON implementations come from Andy Polyakov's implementation, and are included here in raw form without modification, so that subsequent commits that fix these up for the kernel can see how it has changed.

While this is CRYPTOGAMS code, the originating code for this happens to be the same as OpenSSL's commit 5bb1cd2292b388263a0cc05392bb99141212aa53

Based-on-code-from: Andy Polyakov <[email protected]>
Cc: Andy Polyakov <[email protected]>
Cc: Russell King <[email protected]>
Cc: [email protected]
Cc: Samuel Neves <[email protected]>
Cc: Jean-Philippe Aumasson <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 lib/zinc/poly1305/poly1305-arm-cryptogams.S   | 1172 +++++++++++++++++
 lib/zinc/poly1305/poly1305-arm64-cryptogams.S |  869 ++++++++++++
 2 files changed, 2041 insertions(+)
 create mode 100644 lib/zinc/poly1305/poly1305-arm-cryptogams.S
 create mode 100644 lib/zinc/poly1305/poly1305-arm64-cryptogams.S
lib/zinc/poly1305/poly1305-arm-cryptogams.S new file +1172 -0
new file mode 100644
index 000000000000..884b465030e4
@@ -0,0 +1,1172 @@
+/* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
+/*
+ * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ */
+
+#include "arm_arch.h"
+
+.text
+#if defined(__thumb2__)
+.syntax	unified
+.thumb
+#else
+.code	32
+#endif
+
+.globl	poly1305_emit
+.globl	poly1305_blocks
+.globl	poly1305_init
+.type	poly1305_init,%function
+.align	5
+poly1305_init:
+.Lpoly1305_init:
+	stmdb	sp!,{r4-r11}
+
+	eor	r3,r3,r3
+	cmp	r1,#0
+	str	r3,[r0,#0]		@ zero hash value
+	str	r3,[r0,#4]
+	str	r3,[r0,#8]
+	str	r3,[r0,#12]
+	str	r3,[r0,#16]
+	str	r3,[r0,#36]		@ is_base2_26
+	add	r0,r0,#20
+
+#ifdef	__thumb2__
+	it	eq
+#endif
+	moveq	r0,#0
+	beq	.Lno_key
+
+#if	__ARM_MAX_ARCH__>=7
+	adr	r11,.Lpoly1305_init
+	ldr	r12,.LOPENSSL_armcap
+#endif
+	ldrb	r4,[r1,#0]
+	mov	r10,#0x0fffffff
+	ldrb	r5,[r1,#1]
+	and	r3,r10,#-4		@ 0x0ffffffc
+	ldrb	r6,[r1,#2]
+	ldrb	r7,[r1,#3]
+	orr	r4,r4,r5,lsl#8
+	ldrb	r5,[r1,#4]
+	orr	r4,r4,r6,lsl#16
+	ldrb	r6,[r1,#5]
+	orr	r4,r4,r7,lsl#24
+	ldrb	r7,[r1,#6]
+	and	r4,r4,r10
+
+#if	__ARM_MAX_ARCH__>=7
+	ldr	r12,[r11,r12]		@ OPENSSL_armcap_P
+# ifdef	__APPLE__
+	ldr	r12,[r12]
+# endif
+#endif
+	ldrb	r8,[r1,#7]
+	orr	r5,r5,r6,lsl#8
+	ldrb	r6,[r1,#8]
+	orr	r5,r5,r7,lsl#16
+	ldrb	r7,[r1,#9]
+	orr	r5,r5,r8,lsl#24
+	ldrb	r8,[r1,#10]
+	and	r5,r5,r3
+
+#if	__ARM_MAX_ARCH__>=7
+	tst	r12,#ARMV7_NEON		@ check for NEON
+# ifdef	__APPLE__
+	adr	r9,poly1305_blocks_neon
+	adr	r11,poly1305_blocks
+#  ifdef __thumb2__
+	it	ne
+#  endif
+	movne	r11,r9
+	adr	r12,poly1305_emit
+	adr	r10,poly1305_emit_neon
+#  ifdef __thumb2__
+	it	ne
+#  endif
+	movne	r12,r10
+# else
+#  ifdef __thumb2__
+	itete	eq
+#  endif
+	addeq	r12,r11,#(poly1305_emit-.Lpoly1305_init)
+	addne	r12,r11,#(poly1305_emit_neon-.Lpoly1305_init)
+	addeq	r11,r11,#(poly1305_blocks-.Lpoly1305_init)
+	addne	r11,r11,#(poly1305_blocks_neon-.Lpoly1305_init)
+# endif
+# ifdef	__thumb2__
+	orr	r12,r12,#1	@ thumb-ify address
+	orr	r11,r11,#1
+# endif
+#endif
+	ldrb	r9,[r1,#11]
+	orr	r6,r6,r7,lsl#8
+	ldrb	r7,[r1,#12]
+	orr	r6,r6,r8,lsl#16
+	ldrb	r8,[r1,#13]
+	orr	r6,r6,r9,lsl#24
+	ldrb	r9,[r1,#14]
+	and	r6,r6,r3
+
+	ldrb	r10,[r1,#15]
+	orr	r7,r7,r8,lsl#8
+	str	r4,[r0,#0]
+	orr	r7,r7,r9,lsl#16
+	str	r5,[r0,#4]
+	orr	r7,r7,r10,lsl#24
+	str	r6,[r0,#8]
+	and	r7,r7,r3
+	str	r7,[r0,#12]
+#if	__ARM_MAX_ARCH__>=7
+	stmia	r2,{r11,r12}		@ fill functions table
+	mov	r0,#1
+#else
+	mov	r0,#0
+#endif
+.Lno_key:
+	ldmia	sp!,{r4-r11}
+#if	__ARM_ARCH__>=5
+	bx	lr				@ bx	lr
+#else
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	.word	0xe12fff1e			@ interoperable with Thumb ISA:-)
+#endif
+.size	poly1305_init,.-poly1305_init
+.type	poly1305_blocks,%function
+.align	5
+poly1305_blocks:
+.Lpoly1305_blocks:
+	stmdb	sp!,{r3-r11,lr}
+
+	ands	r2,r2,#-16
+	beq	.Lno_data
+
+	cmp	r3,#0
+	add	r2,r2,r1		@ end pointer
+	sub	sp,sp,#32
+
+	ldmia	r0,{r4-r12}		@ load context
+
+	str	r0,[sp,#12]		@ offload stuff
+	mov	lr,r1
+	str	r2,[sp,#16]
+	str	r10,[sp,#20]
+	str	r11,[sp,#24]
+	str	r12,[sp,#28]
+	b	.Loop
+
+.Loop:
+#if __ARM_ARCH__<7
+	ldrb	r0,[lr],#16		@ load input
+# ifdef	__thumb2__
+	it	hi
+# endif
+	addhi	r8,r8,#1		@ 1<<128
+	ldrb	r1,[lr,#-15]
+	ldrb	r2,[lr,#-14]
+	ldrb	r3,[lr,#-13]
+	orr	r1,r0,r1,lsl#8
+	ldrb	r0,[lr,#-12]
+	orr	r2,r1,r2,lsl#16
+	ldrb	r1,[lr,#-11]
+	orr	r3,r2,r3,lsl#24
+	ldrb	r2,[lr,#-10]
+	adds	r4,r4,r3		@ accumulate input
+
+	ldrb	r3,[lr,#-9]
+	orr	r1,r0,r1,lsl#8
+	ldrb	r0,[lr,#-8]
+	orr	r2,r1,r2,lsl#16
+	ldrb	r1,[lr,#-7]
+	orr	r3,r2,r3,lsl#24
+	ldrb	r2,[lr,#-6]
+	adcs	r5,r5,r3
+
+	ldrb	r3,[lr,#-5]
+	orr	r1,r0,r1,lsl#8
+	ldrb	r0,[lr,#-4]
+	orr	r2,r1,r2,lsl#16
+	ldrb	r1,[lr,#-3]
+	orr	r3,r2,r3,lsl#24
+	ldrb	r2,[lr,#-2]
+	adcs	r6,r6,r3
+
+	ldrb	r3,[lr,#-1]
+	orr	r1,r0,r1,lsl#8
+	str	lr,[sp,#8]		@ offload input pointer
+	orr	r2,r1,r2,lsl#16
+	add	r10,r10,r10,lsr#2
+	orr	r3,r2,r3,lsl#24
+#else
+	ldr	r0,[lr],#16		@ load input
+# ifdef	__thumb2__
+	it	hi
+# endif
+	addhi	r8,r8,#1		@ padbit
+	ldr	r1,[lr,#-12]
+	ldr	r2,[lr,#-8]
+	ldr	r3,[lr,#-4]
+# ifdef	__ARMEB__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+# endif
+	adds	r4,r4,r0		@ accumulate input
+	str	lr,[sp,#8]		@ offload input pointer
+	adcs	r5,r5,r1
+	add	r10,r10,r10,lsr#2
+	adcs	r6,r6,r2
+#endif
+	add	r11,r11,r11,lsr#2
+	adcs	r7,r7,r3
+	add	r12,r12,r12,lsr#2
+
+	umull	r2,r3,r5,r9
+	 adc	r8,r8,#0
+	umull	r0,r1,r4,r9
+	umlal	r2,r3,r8,r10
+	umlal	r0,r1,r7,r10
+	ldr	r10,[sp,#20]		@ reload r10
+	umlal	r2,r3,r6,r12
+	umlal	r0,r1,r5,r12
+	umlal	r2,r3,r7,r11
+	umlal	r0,r1,r6,r11
+	umlal	r2,r3,r4,r10
+	str	r0,[sp,#0]		@ future r4
+	 mul	r0,r11,r8
+	ldr	r11,[sp,#24]		@ reload r11
+	adds	r2,r2,r1		@ d1+=d0>>32
+	 eor	r1,r1,r1
+	adc	lr,r3,#0		@ future r6
+	str	r2,[sp,#4]		@ future r5
+
+	mul	r2,r12,r8
+	eor	r3,r3,r3
+	umlal	r0,r1,r7,r12
+	ldr	r12,[sp,#28]		@ reload r12
+	umlal	r2,r3,r7,r9
+	umlal	r0,r1,r6,r9
+	umlal	r2,r3,r6,r10
+	umlal	r0,r1,r5,r10
+	umlal	r2,r3,r5,r11
+	umlal	r0,r1,r4,r11
+	umlal	r2,r3,r4,r12
+	ldr	r4,[sp,#0]
+	mul	r8,r9,r8
+	ldr	r5,[sp,#4]
+
+	adds	r6,lr,r0		@ d2+=d1>>32
+	ldr	lr,[sp,#8]		@ reload input pointer
+	adc	r1,r1,#0
+	adds	r7,r2,r1		@ d3+=d2>>32
+	ldr	r0,[sp,#16]		@ reload end pointer
+	adc	r3,r3,#0
+	add	r8,r8,r3		@ h4+=d3>>32
+
+	and	r1,r8,#-4
+	and	r8,r8,#3
+	add	r1,r1,r1,lsr#2		@ *=5
+	adds	r4,r4,r1
+	adcs	r5,r5,#0
+	adcs	r6,r6,#0
+	adcs	r7,r7,#0
+	adc	r8,r8,#0
+
+	cmp	r0,lr			@ done yet?
+	bhi	.Loop
+
+	ldr	r0,[sp,#12]
+	add	sp,sp,#32
+	stmia	r0,{r4-r8}		@ store the result
+
+.Lno_data:
+#if	__ARM_ARCH__>=5
+	ldmia	sp!,{r3-r11,pc}
+#else
+	ldmia	sp!,{r3-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	.word	0xe12fff1e			@ interoperable with Thumb ISA:-)
+#endif
+.size	poly1305_blocks,.-poly1305_blocks
+.type	poly1305_emit,%function
+.align	5
+poly1305_emit:
+	stmdb	sp!,{r4-r11}
+.Lpoly1305_emit_enter:
+
+	ldmia	r0,{r3-r7}
+	adds	r8,r3,#5		@ compare to modulus
+	adcs	r9,r4,#0
+	adcs	r10,r5,#0
+	adcs	r11,r6,#0
+	adc	r7,r7,#0
+	tst	r7,#4			@ did it carry/borrow?
+
+#ifdef	__thumb2__
+	it	ne
+#endif
+	movne	r3,r8
+	ldr	r8,[r2,#0]
+#ifdef	__thumb2__
+	it	ne
+#endif
+	movne	r4,r9
+	ldr	r9,[r2,#4]
+#ifdef	__thumb2__
+	it	ne
+#endif
+	movne	r5,r10
+	ldr	r10,[r2,#8]
+#ifdef	__thumb2__
+	it	ne
+#endif
+	movne	r6,r11
+	ldr	r11,[r2,#12]
+
+	adds	r3,r3,r8
+	adcs	r4,r4,r9
+	adcs	r5,r5,r10
+	adc	r6,r6,r11
+
+#if __ARM_ARCH__>=7
+# ifdef __ARMEB__
+	rev	r3,r3
+	rev	r4,r4
+	rev	r5,r5
+	rev	r6,r6
+# endif
+	str	r3,[r1,#0]
+	str	r4,[r1,#4]
+	str	r5,[r1,#8]
+	str	r6,[r1,#12]
+#else
+	strb	r3,[r1,#0]
+	mov	r3,r3,lsr#8
+	strb	r4,[r1,#4]
+	mov	r4,r4,lsr#8
+	strb	r5,[r1,#8]
+	mov	r5,r5,lsr#8
+	strb	r6,[r1,#12]
+	mov	r6,r6,lsr#8
+
+	strb	r3,[r1,#1]
+	mov	r3,r3,lsr#8
+	strb	r4,[r1,#5]
+	mov	r4,r4,lsr#8
+	strb	r5,[r1,#9]
+	mov	r5,r5,lsr#8
+	strb	r6,[r1,#13]
+	mov	r6,r6,lsr#8
+
+	strb	r3,[r1,#2]
+	mov	r3,r3,lsr#8
+	strb	r4,[r1,#6]
+	mov	r4,r4,lsr#8
+	strb	r5,[r1,#10]
+	mov	r5,r5,lsr#8
+	strb	r6,[r1,#14]
+	mov	r6,r6,lsr#8
+
+	strb	r3,[r1,#3]
+	strb	r4,[r1,#7]
+	strb	r5,[r1,#11]
+	strb	r6,[r1,#15]
+#endif
+	ldmia	sp!,{r4-r11}
+#if	__ARM_ARCH__>=5
+	bx	lr				@ bx	lr
+#else
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	.word	0xe12fff1e			@ interoperable with Thumb ISA:-)
+#endif
+.size	poly1305_emit,.-poly1305_emit
+#if	__ARM_MAX_ARCH__>=7
+.fpu	neon
+
+.type	poly1305_init_neon,%function
+.align	5
+poly1305_init_neon:
+	ldr	r4,[r0,#20]		@ load key base 2^32
+	ldr	r5,[r0,#24]
+	ldr	r6,[r0,#28]
+	ldr	r7,[r0,#32]
+
+	and	r2,r4,#0x03ffffff	@ base 2^32 -> base 2^26
+	mov	r3,r4,lsr#26
+	mov	r4,r5,lsr#20
+	orr	r3,r3,r5,lsl#6
+	mov	r5,r6,lsr#14
+	orr	r4,r4,r6,lsl#12
+	mov	r6,r7,lsr#8
+	orr	r5,r5,r7,lsl#18
+	and	r3,r3,#0x03ffffff
+	and	r4,r4,#0x03ffffff
+	and	r5,r5,#0x03ffffff
+
+	vdup.32	d0,r2			@ r^1 in both lanes
+	add	r2,r3,r3,lsl#2		@ *5
+	vdup.32	d1,r3
+	add	r3,r4,r4,lsl#2
+	vdup.32	d2,r2
+	vdup.32	d3,r4
+	add	r4,r5,r5,lsl#2
+	vdup.32	d4,r3
+	vdup.32	d5,r5
+	add	r5,r6,r6,lsl#2
+	vdup.32	d6,r4
+	vdup.32	d7,r6
+	vdup.32	d8,r5
+
+	mov	r5,#2		@ counter
+
+.Lsquare_neon:
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ d0 = h0*r0 + h4*5*r1 + h3*5*r2 + h2*5*r3 + h1*5*r4
+	@ d1 = h1*r0 + h0*r1   + h4*5*r2 + h3*5*r3 + h2*5*r4
+	@ d2 = h2*r0 + h1*r1   + h0*r2   + h4*5*r3 + h3*5*r4
+	@ d3 = h3*r0 + h2*r1   + h1*r2   + h0*r3   + h4*5*r4
+	@ d4 = h4*r0 + h3*r1   + h2*r2   + h1*r3   + h0*r4
+
+	vmull.u32	q5,d0,d0[1]
+	vmull.u32	q6,d1,d0[1]
+	vmull.u32	q7,d3,d0[1]
+	vmull.u32	q8,d5,d0[1]
+	vmull.u32	q9,d7,d0[1]
+
+	vmlal.u32	q5,d7,d2[1]
+	vmlal.u32	q6,d0,d1[1]
+	vmlal.u32	q7,d1,d1[1]
+	vmlal.u32	q8,d3,d1[1]
+	vmlal.u32	q9,d5,d1[1]
+
+	vmlal.u32	q5,d5,d4[1]
+	vmlal.u32	q6,d7,d4[1]
+	vmlal.u32	q8,d1,d3[1]
+	vmlal.u32	q7,d0,d3[1]
+	vmlal.u32	q9,d3,d3[1]
+
+	vmlal.u32	q5,d3,d6[1]
+	vmlal.u32	q8,d0,d5[1]
+	vmlal.u32	q6,d5,d6[1]
+	vmlal.u32	q7,d7,d6[1]
+	vmlal.u32	q9,d1,d5[1]
+
+	vmlal.u32	q8,d7,d8[1]
+	vmlal.u32	q5,d1,d8[1]
+	vmlal.u32	q6,d3,d8[1]
+	vmlal.u32	q7,d5,d8[1]
+	vmlal.u32	q9,d0,d7[1]
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ lazy reduction as discussed in "NEON crypto" by D.J. Bernstein
+	@ and P. Schwabe
+	@
+	@ H0>>+H1>>+H2>>+H3>>+H4
+	@ H3>>+H4>>*5+H0>>+H1
+	@
+	@ Trivia.
+	@
+	@ Result of multiplication of n-bit number by m-bit number is
+	@ n+m bits wide. However! Even though 2^n is a n+1-bit number,
+	@ m-bit number multiplied by 2^n is still n+m bits wide.
+	@
+	@ Sum of two n-bit numbers is n+1 bits wide, sum of three - n+2,
+	@ and so is sum of four. Sum of 2^m n-m-bit numbers and n-bit
+	@ one is n+1 bits wide.
+	@
+	@ >>+ denotes Hnext += Hn>>26, Hn &= 0x3ffffff. This means that
+	@ H0, H2, H3 are guaranteed to be 26 bits wide, while H1 and H4
+	@ can be 27. However! In cases when their width exceeds 26 bits
+	@ they are limited by 2^26+2^6. This in turn means that *sum*
+	@ of the products with these values can still be viewed as sum
+	@ of 52-bit numbers as long as the amount of addends is not a
+	@ power of 2. For example,
+	@
+	@ H4 = H4*R0 + H3*R1 + H2*R2 + H1*R3 + H0 * R4,
+	@
+	@ which can't be larger than 5 * (2^26 + 2^6) * (2^26 + 2^6), or
+	@ 5 * (2^52 + 2*2^32 + 2^12), which in turn is smaller than
+	@ 8 * (2^52) or 2^55. However, the value is then multiplied by
+	@ by 5, so we should be looking at 5 * 5 * (2^52 + 2^33 + 2^12),
+	@ which is less than 32 * (2^52) or 2^57. And when processing
+	@ data we are looking at triple as many addends...
+	@
+	@ In key setup procedure pre-reduced H0 is limited by 5*4+1 and
+	@ 5*H4 - by 5*5 52-bit addends, or 57 bits. But when hashing the
+	@ input H0 is limited by (5*4+1)*3 addends, or 58 bits, while
+	@ 5*H4 by 5*5*3, or 59[!] bits. How is this relevant? vmlal.u32
+	@ instruction accepts 2x32-bit input and writes 2x64-bit result.
+	@ This means that result of reduction have to be compressed upon
+	@ loop wrap-around. This can be done in the process of reduction
+	@ to minimize amount of instructions [as well as amount of
+	@ 128-bit instructions, which benefits low-end processors], but
+	@ one has to watch for H2 (which is narrower than H0) and 5*H4
+	@ not being wider than 58 bits, so that result of right shift
+	@ by 26 bits fits in 32 bits. This is also useful on x86,
+	@ because it allows to use paddd in place for paddq, which
+	@ benefits Atom, where paddq is ridiculously slow.
+
+	vshr.u64	q15,q8,#26
+	vmovn.i64	d16,q8
+	 vshr.u64	q4,q5,#26
+	 vmovn.i64	d10,q5
+	vadd.i64	q9,q9,q15		@ h3 -> h4
+	vbic.i32	d16,#0xfc000000	@ &=0x03ffffff
+	 vadd.i64	q6,q6,q4		@ h0 -> h1
+	 vbic.i32	d10,#0xfc000000
+
+	vshrn.u64	d30,q9,#26
+	vmovn.i64	d18,q9
+	 vshr.u64	q4,q6,#26
+	 vmovn.i64	d12,q6
+	 vadd.i64	q7,q7,q4		@ h1 -> h2
+	vbic.i32	d18,#0xfc000000
+	 vbic.i32	d12,#0xfc000000
+
+	vadd.i32	d10,d10,d30
+	vshl.u32	d30,d30,#2
+	 vshrn.u64	d8,q7,#26
+	 vmovn.i64	d14,q7
+	vadd.i32	d10,d10,d30	@ h4 -> h0
+	 vadd.i32	d16,d16,d8	@ h2 -> h3
+	 vbic.i32	d14,#0xfc000000
+
+	vshr.u32	d30,d10,#26
+	vbic.i32	d10,#0xfc000000
+	 vshr.u32	d8,d16,#26
+	 vbic.i32	d16,#0xfc000000
+	vadd.i32	d12,d12,d30	@ h0 -> h1
+	 vadd.i32	d18,d18,d8	@ h3 -> h4
+
+	subs		r5,r5,#1
+	beq		.Lsquare_break_neon
+
+	add		r6,r0,#(48+0*9*4)
+	add		r7,r0,#(48+1*9*4)
+
+	vtrn.32		d0,d10		@ r^2:r^1
+	vtrn.32		d3,d14
+	vtrn.32		d5,d16
+	vtrn.32		d1,d12
+	vtrn.32		d7,d18
+
+	vshl.u32	d4,d3,#2		@ *5
+	vshl.u32	d6,d5,#2
+	vshl.u32	d2,d1,#2
+	vshl.u32	d8,d7,#2
+	vadd.i32	d4,d4,d3
+	vadd.i32	d2,d2,d1
+	vadd.i32	d6,d6,d5
+	vadd.i32	d8,d8,d7
+
+	vst4.32		{d0[0],d1[0],d2[0],d3[0]},[r6]!
+	vst4.32		{d0[1],d1[1],d2[1],d3[1]},[r7]!
+	vst4.32		{d4[0],d5[0],d6[0],d7[0]},[r6]!
+	vst4.32		{d4[1],d5[1],d6[1],d7[1]},[r7]!
+	vst1.32		{d8[0]},[r6,:32]
+	vst1.32		{d8[1]},[r7,:32]
+
+	b		.Lsquare_neon
+
+.align	4
+.Lsquare_break_neon:
+	add		r6,r0,#(48+2*4*9)
+	add		r7,r0,#(48+3*4*9)
+
+	vmov		d0,d10		@ r^4:r^3
+	vshl.u32	d2,d12,#2		@ *5
+	vmov		d1,d12
+	vshl.u32	d4,d14,#2
+	vmov		d3,d14
+	vshl.u32	d6,d16,#2
+	vmov		d5,d16
+	vshl.u32	d8,d18,#2
+	vmov		d7,d18
+	vadd.i32	d2,d2,d12
+	vadd.i32	d4,d4,d14
+	vadd.i32	d6,d6,d16
+	vadd.i32	d8,d8,d18
+
+	vst4.32		{d0[0],d1[0],d2[0],d3[0]},[r6]!
+	vst4.32		{d0[1],d1[1],d2[1],d3[1]},[r7]!
+	vst4.32		{d4[0],d5[0],d6[0],d7[0]},[r6]!
+	vst4.32		{d4[1],d5[1],d6[1],d7[1]},[r7]!
+	vst1.32		{d8[0]},[r6]
+	vst1.32		{d8[1]},[r7]
+
+	bx	lr				@ bx	lr
+.size	poly1305_init_neon,.-poly1305_init_neon
+
+.type	poly1305_blocks_neon,%function
+.align	5
+poly1305_blocks_neon:
+	ldr	ip,[r0,#36]		@ is_base2_26
+	ands	r2,r2,#-16
+	beq	.Lno_data_neon
+
+	cmp	r2,#64
+	bhs	.Lenter_neon
+	tst	ip,ip			@ is_base2_26?
+	beq	.Lpoly1305_blocks
+
+.Lenter_neon:
+	stmdb	sp!,{r4-r7}
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+
+	tst	ip,ip			@ is_base2_26?
+	bne	.Lbase2_26_neon
+
+	stmdb	sp!,{r1-r3,lr}
+	bl	poly1305_init_neon
+
+	ldr	r4,[r0,#0]		@ load hash value base 2^32
+	ldr	r5,[r0,#4]
+	ldr	r6,[r0,#8]
+	ldr	r7,[r0,#12]
+	ldr	ip,[r0,#16]
+
+	and	r2,r4,#0x03ffffff	@ base 2^32 -> base 2^26
+	mov	r3,r4,lsr#26
+	 veor	d10,d10,d10
+	mov	r4,r5,lsr#20
+	orr	r3,r3,r5,lsl#6
+	 veor	d12,d12,d12
+	mov	r5,r6,lsr#14
+	orr	r4,r4,r6,lsl#12
+	 veor	d14,d14,d14
+	mov	r6,r7,lsr#8
+	orr	r5,r5,r7,lsl#18
+	 veor	d16,d16,d16
+	and	r3,r3,#0x03ffffff
+	orr	r6,r6,ip,lsl#24
+	 veor	d18,d18,d18
+	and	r4,r4,#0x03ffffff
+	mov	r1,#1
+	and	r5,r5,#0x03ffffff
+	str	r1,[r0,#36]		@ is_base2_26
+
+	vmov.32	d10[0],r2
+	vmov.32	d12[0],r3
+	vmov.32	d14[0],r4
+	vmov.32	d16[0],r5
+	vmov.32	d18[0],r6
+	adr	r5,.Lzeros
+
+	ldmia	sp!,{r1-r3,lr}
+	b	.Lbase2_32_neon
+
+.align	4
+.Lbase2_26_neon:
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ load hash value
+
+	veor		d10,d10,d10
+	veor		d12,d12,d12
+	veor		d14,d14,d14
+	veor		d16,d16,d16
+	veor		d18,d18,d18
+	vld4.32		{d10[0],d12[0],d14[0],d16[0]},[r0]!
+	adr		r5,.Lzeros
+	vld1.32		{d18[0]},[r0]
+	sub		r0,r0,#16		@ rewind
+
+.Lbase2_32_neon:
+	add		r4,r1,#32
+	mov		r3,r3,lsl#24
+	tst		r2,#31
+	beq		.Leven
+
+	vld4.32		{d20[0],d22[0],d24[0],d26[0]},[r1]!
+	vmov.32		d28[0],r3
+	sub		r2,r2,#16
+	add		r4,r1,#32
+
+# ifdef	__ARMEB__
+	vrev32.8	q10,q10
+	vrev32.8	q13,q13
+	vrev32.8	q11,q11
+	vrev32.8	q12,q12
+# endif
+	vsri.u32	d28,d26,#8	@ base 2^32 -> base 2^26
+	vshl.u32	d26,d26,#18
+
+	vsri.u32	d26,d24,#14
+	vshl.u32	d24,d24,#12
+	vadd.i32	d29,d28,d18	@ add hash value and move to #hi
+
+	vbic.i32	d26,#0xfc000000
+	vsri.u32	d24,d22,#20
+	vshl.u32	d22,d22,#6
+
+	vbic.i32	d24,#0xfc000000
+	vsri.u32	d22,d20,#26
+	vadd.i32	d27,d26,d16
+
+	vbic.i32	d20,#0xfc000000
+	vbic.i32	d22,#0xfc000000
+	vadd.i32	d25,d24,d14
+
+	vadd.i32	d21,d20,d10
+	vadd.i32	d23,d22,d12
+
+	mov		r7,r5
+	add		r6,r0,#48
+
+	cmp		r2,r2
+	b		.Long_tail
+
+.align	4
+.Leven:
+	subs		r2,r2,#64
+	it		lo
+	movlo		r4,r5
+
+	vmov.i32	q14,#1<<24		@ padbit, yes, always
+	vld4.32		{d20,d22,d24,d26},[r1]	@ inp[0:1]
+	add		r1,r1,#64
+	vld4.32		{d21,d23,d25,d27},[r4]	@ inp[2:3] (or 0)
+	add		r4,r4,#64
+	itt		hi
+	addhi		r7,r0,#(48+1*9*4)
+	addhi		r6,r0,#(48+3*9*4)
+
+# ifdef	__ARMEB__
+	vrev32.8	q10,q10
+	vrev32.8	q13,q13
+	vrev32.8	q11,q11
+	vrev32.8	q12,q12
+# endif
+	vsri.u32	q14,q13,#8		@ base 2^32 -> base 2^26
+	vshl.u32	q13,q13,#18
+
+	vsri.u32	q13,q12,#14
+	vshl.u32	q12,q12,#12
+
+	vbic.i32	q13,#0xfc000000
+	vsri.u32	q12,q11,#20
+	vshl.u32	q11,q11,#6
+
+	vbic.i32	q12,#0xfc000000
+	vsri.u32	q11,q10,#26
+
+	vbic.i32	q10,#0xfc000000
+	vbic.i32	q11,#0xfc000000
+
+	bls		.Lskip_loop
+
+	vld4.32		{d0[1],d1[1],d2[1],d3[1]},[r7]!	@ load r^2
+	vld4.32		{d0[0],d1[0],d2[0],d3[0]},[r6]!	@ load r^4
+	vld4.32		{d4[1],d5[1],d6[1],d7[1]},[r7]!
+	vld4.32		{d4[0],d5[0],d6[0],d7[0]},[r6]!
+	b		.Loop_neon
+
+.align	5
+.Loop_neon:
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ ((inp[0]*r^4+inp[2]*r^2+inp[4])*r^4+inp[6]*r^2
+	@ ((inp[1]*r^4+inp[3]*r^2+inp[5])*r^3+inp[7]*r
+	@   ___________________/
+	@ ((inp[0]*r^4+inp[2]*r^2+inp[4])*r^4+inp[6]*r^2+inp[8])*r^2
+	@ ((inp[1]*r^4+inp[3]*r^2+inp[5])*r^4+inp[7]*r^2+inp[9])*r
+	@   ___________________/ ____________________/
+	@
+	@ Note that we start with inp[2:3]*r^2. This is because it
+	@ doesn't depend on reduction in previous iteration.
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ d4 = h4*r0 + h3*r1   + h2*r2   + h1*r3   + h0*r4
+	@ d3 = h3*r0 + h2*r1   + h1*r2   + h0*r3   + h4*5*r4
+	@ d2 = h2*r0 + h1*r1   + h0*r2   + h4*5*r3 + h3*5*r4
+	@ d1 = h1*r0 + h0*r1   + h4*5*r2 + h3*5*r3 + h2*5*r4
+	@ d0 = h0*r0 + h4*5*r1 + h3*5*r2 + h2*5*r3 + h1*5*r4
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ inp[2:3]*r^2
+
+	vadd.i32	d24,d24,d14	@ accumulate inp[0:1]
+	vmull.u32	q7,d25,d0[1]
+	vadd.i32	d20,d20,d10
+	vmull.u32	q5,d21,d0[1]
+	vadd.i32	d26,d26,d16
+	vmull.u32	q8,d27,d0[1]
+	vmlal.u32	q7,d23,d1[1]
+	vadd.i32	d22,d22,d12
+	vmull.u32	q6,d23,d0[1]
+
+	vadd.i32	d28,d28,d18
+	vmull.u32	q9,d29,d0[1]
+	subs		r2,r2,#64
+	vmlal.u32	q5,d29,d2[1]
+	it		lo
+	movlo		r4,r5
+	vmlal.u32	q8,d25,d1[1]
+	vld1.32		d8[1],[r7,:32]
+	vmlal.u32	q6,d21,d1[1]
+	vmlal.u32	q9,d27,d1[1]
+
+	vmlal.u32	q5,d27,d4[1]
+	vmlal.u32	q8,d23,d3[1]
+	vmlal.u32	q9,d25,d3[1]
+	vmlal.u32	q6,d29,d4[1]
+	vmlal.u32	q7,d21,d3[1]
+
+	vmlal.u32	q8,d21,d5[1]
+	vmlal.u32	q5,d25,d6[1]
+	vmlal.u32	q9,d23,d5[1]
+	vmlal.u32	q6,d27,d6[1]
+	vmlal.u32	q7,d29,d6[1]
+
+	vmlal.u32	q8,d29,d8[1]
+	vmlal.u32	q5,d23,d8[1]
+	vmlal.u32	q9,d21,d7[1]
+	vmlal.u32	q6,d25,d8[1]
+	vmlal.u32	q7,d27,d8[1]
+
+	vld4.32		{d21,d23,d25,d27},[r4]	@ inp[2:3] (or 0)
+	add		r4,r4,#64
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ (hash+inp[0:1])*r^4 and accumulate
+
+	vmlal.u32	q8,d26,d0[0]
+	vmlal.u32	q5,d20,d0[0]
+	vmlal.u32	q9,d28,d0[0]
+	vmlal.u32	q6,d22,d0[0]
+	vmlal.u32	q7,d24,d0[0]
+	vld1.32		d8[0],[r6,:32]
+
+	vmlal.u32	q8,d24,d1[0]
+	vmlal.u32	q5,d28,d2[0]
+	vmlal.u32	q9,d26,d1[0]
+	vmlal.u32	q6,d20,d1[0]
+	vmlal.u32	q7,d22,d1[0]
+
+	vmlal.u32	q8,d22,d3[0]
+	vmlal.u32	q5,d26,d4[0]
+	vmlal.u32	q9,d24,d3[0]
+	vmlal.u32	q6,d28,d4[0]
+	vmlal.u32	q7,d20,d3[0]
+
+	vmlal.u32	q8,d20,d5[0]
+	vmlal.u32	q5,d24,d6[0]
+	vmlal.u32	q9,d22,d5[0]
+	vmlal.u32	q6,d26,d6[0]
+	vmlal.u32	q8,d28,d8[0]
+
+	vmlal.u32	q7,d28,d6[0]
+	vmlal.u32	q5,d22,d8[0]
+	vmlal.u32	q9,d20,d7[0]
+	vmov.i32	q14,#1<<24		@ padbit, yes, always
+	vmlal.u32	q6,d24,d8[0]
+	vmlal.u32	q7,d26,d8[0]
+
+	vld4.32		{d20,d22,d24,d26},[r1]	@ inp[0:1]
+	add		r1,r1,#64
+# ifdef	__ARMEB__
+	vrev32.8	q10,q10
+	vrev32.8	q11,q11
+	vrev32.8	q12,q12
+	vrev32.8	q13,q13
+# endif
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ lazy reduction interleaved with base 2^32 -> base 2^26 of
+	@ inp[0:3] previously loaded to q10-q13 and smashed to q10-q14.
+
+	vshr.u64	q15,q8,#26
+	vmovn.i64	d16,q8
+	 vshr.u64	q4,q5,#26
+	 vmovn.i64	d10,q5
+	vadd.i64	q9,q9,q15		@ h3 -> h4
+	vbic.i32	d16,#0xfc000000
+	  vsri.u32	q14,q13,#8		@ base 2^32 -> base 2^26
+	 vadd.i64	q6,q6,q4		@ h0 -> h1
+	  vshl.u32	q13,q13,#18
+	 vbic.i32	d10,#0xfc000000
+
+	vshrn.u64	d30,q9,#26
+	vmovn.i64	d18,q9
+	 vshr.u64	q4,q6,#26
+	 vmovn.i64	d12,q6
+	 vadd.i64	q7,q7,q4		@ h1 -> h2
+	  vsri.u32	q13,q12,#14
+	vbic.i32	d18,#0xfc000000
+	  vshl.u32	q12,q12,#12
+	 vbic.i32	d12,#0xfc000000
+
+	vadd.i32	d10,d10,d30
+	vshl.u32	d30,d30,#2
+	  vbic.i32	q13,#0xfc000000
+	 vshrn.u64	d8,q7,#26
+	 vmovn.i64	d14,q7
+	vaddl.u32	q5,d10,d30	@ h4 -> h0 [widen for a sec]
+	  vsri.u32	q12,q11,#20
+	 vadd.i32	d16,d16,d8	@ h2 -> h3
+	  vshl.u32	q11,q11,#6
+	 vbic.i32	d14,#0xfc000000
+	  vbic.i32	q12,#0xfc000000
+
+	vshrn.u64	d30,q5,#26		@ re-narrow
+	vmovn.i64	d10,q5
+	  vsri.u32	q11,q10,#26
+	  vbic.i32	q10,#0xfc000000
+	 vshr.u32	d8,d16,#26
+	 vbic.i32	d16,#0xfc000000
+	vbic.i32	d10,#0xfc000000
+	vadd.i32	d12,d12,d30	@ h0 -> h1
+	 vadd.i32	d18,d18,d8	@ h3 -> h4
+	  vbic.i32	q11,#0xfc000000
+
+	bhi		.Loop_neon
+
+.Lskip_loop:
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ multiply (inp[0:1]+hash) or inp[2:3] by r^2:r^1
+
+	add		r7,r0,#(48+0*9*4)
+	add		r6,r0,#(48+1*9*4)
+	adds		r2,r2,#32
+	it		ne
+	movne		r2,#0
+	bne		.Long_tail
+
+	vadd.i32	d25,d24,d14	@ add hash value and move to #hi
+	vadd.i32	d21,d20,d10
+	vadd.i32	d27,d26,d16
+	vadd.i32	d23,d22,d12
+	vadd.i32	d29,d28,d18
+
+.Long_tail:
+	vld4.32		{d0[1],d1[1],d2[1],d3[1]},[r7]!	@ load r^1
+	vld4.32		{d0[0],d1[0],d2[0],d3[0]},[r6]!	@ load r^2
+
+	vadd.i32	d24,d24,d14	@ can be redundant
+	vmull.u32	q7,d25,d0
+	vadd.i32	d20,d20,d10
+	vmull.u32	q5,d21,d0
+	vadd.i32	d26,d26,d16
+	vmull.u32	q8,d27,d0
+	vadd.i32	d22,d22,d12
+	vmull.u32	q6,d23,d0
+	vadd.i32	d28,d28,d18
+	vmull.u32	q9,d29,d0
+
+	vmlal.u32	q5,d29,d2
+	vld4.32		{d4[1],d5[1],d6[1],d7[1]},[r7]!
+	vmlal.u32	q8,d25,d1
+	vld4.32		{d4[0],d5[0],d6[0],d7[0]},[r6]!
+	vmlal.u32	q6,d21,d1
+	vmlal.u32	q9,d27,d1
+	vmlal.u32	q7,d23,d1
+
+	vmlal.u32	q8,d23,d3
+	vld1.32		d8[1],[r7,:32]
+	vmlal.u32	q5,d27,d4
+	vld1.32		d8[0],[r6,:32]
+	vmlal.u32	q9,d25,d3
+	vmlal.u32	q6,d29,d4
+	vmlal.u32	q7,d21,d3
+
+	vmlal.u32	q8,d21,d5
+	 it		ne
+	 addne		r7,r0,#(48+2*9*4)
+	vmlal.u32	q5,d25,d6
+	 it		ne
+	 addne		r6,r0,#(48+3*9*4)
+	vmlal.u32	q9,d23,d5
+	vmlal.u32	q6,d27,d6
+	vmlal.u32	q7,d29,d6
+
+	vmlal.u32	q8,d29,d8
+	 vorn		q0,q0,q0	@ all-ones, can be redundant
+	vmlal.u32	q5,d23,d8
+	 vshr.u64	q0,q0,#38
+	vmlal.u32	q9,d21,d7
+	vmlal.u32	q6,d25,d8
+	vmlal.u32	q7,d27,d8
+
+	beq		.Lshort_tail
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ (hash+inp[0:1])*r^4:r^3 and accumulate
+
+	vld4.32		{d0[1],d1[1],d2[1],d3[1]},[r7]!	@ load r^3
+	vld4.32		{d0[0],d1[0],d2[0],d3[0]},[r6]!	@ load r^4
+
+	vmlal.u32	q7,d24,d0
+	vmlal.u32	q5,d20,d0
+	vmlal.u32	q8,d26,d0
+	vmlal.u32	q6,d22,d0
+	vmlal.u32	q9,d28,d0
+
+	vmlal.u32	q5,d28,d2
+	vld4.32		{d4[1],d5[1],d6[1],d7[1]},[r7]!
+	vmlal.u32	q8,d24,d1
+	vld4.32		{d4[0],d5[0],d6[0],d7[0]},[r6]!
+	vmlal.u32	q6,d20,d1
+	vmlal.u32	q9,d26,d1
+	vmlal.u32	q7,d22,d1
+
+	vmlal.u32	q8,d22,d3
+	vld1.32		d8[1],[r7,:32]
+	vmlal.u32	q5,d26,d4
+	vld1.32		d8[0],[r6,:32]
+	vmlal.u32	q9,d24,d3
+	vmlal.u32	q6,d28,d4
+	vmlal.u32	q7,d20,d3
+
+	vmlal.u32	q8,d20,d5
+	vmlal.u32	q5,d24,d6
+	vmlal.u32	q9,d22,d5
+	vmlal.u32	q6,d26,d6
+	vmlal.u32	q7,d28,d6
+
+	vmlal.u32	q8,d28,d8
+	 vorn		q0,q0,q0	@ all-ones
+	vmlal.u32	q5,d22,d8
+	 vshr.u64	q0,q0,#38
+	vmlal.u32	q9,d20,d7
+	vmlal.u32	q6,d24,d8
+	vmlal.u32	q7,d26,d8
+
+.Lshort_tail:
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ horizontal addition
+
+	vadd.i64	d16,d16,d17
+	vadd.i64	d10,d10,d11
+	vadd.i64	d18,d18,d19
+	vadd.i64	d12,d12,d13
+	vadd.i64	d14,d14,d15
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ lazy reduction, but without narrowing
+
+	vshr.u64	q15,q8,#26
+	vand.i64	q8,q8,q0
+	 vshr.u64	q4,q5,#26
+	 vand.i64	q5,q5,q0
+	vadd.i64	q9,q9,q15		@ h3 -> h4
+	 vadd.i64	q6,q6,q4		@ h0 -> h1
+
+	vshr.u64	q15,q9,#26
+	vand.i64	q9,q9,q0
+	 vshr.u64	q4,q6,#26
+	 vand.i64	q6,q6,q0
+	 vadd.i64	q7,q7,q4		@ h1 -> h2
+
+	vadd.i64	q5,q5,q15
+	vshl.u64	q15,q15,#2
+	 vshr.u64	q4,q7,#26
+	 vand.i64	q7,q7,q0
+	vadd.i64	q5,q5,q15		@ h4 -> h0
+	 vadd.i64	q8,q8,q4		@ h2 -> h3
+
+	vshr.u64	q15,q5,#26
+	vand.i64	q5,q5,q0
+	 vshr.u64	q4,q8,#26
+	 vand.i64	q8,q8,q0
+	vadd.i64	q6,q6,q15		@ h0 -> h1
+	 vadd.i64	q9,q9,q4		@ h3 -> h4
+
+	cmp		r2,#0
+	bne		.Leven
+
+	@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
+	@ store hash value
+
+	vst4.32		{d10[0],d12[0],d14[0],d16[0]},[r0]!
+	vst1.32		{d18[0]},[r0]
+
+	vldmia	sp!,{d8-d15}			@ epilogue
+	ldmia	sp!,{r4-r7}
+.Lno_data_neon:
+	bx	lr					@ bx	lr
+.size	poly1305_blocks_neon,.-poly1305_blocks_neon
+
+.type	poly1305_emit_neon,%function
+.align	5
+poly1305_emit_neon:
+	ldr	ip,[r0,#36]		@ is_base2_26
+
+	stmdb	sp!,{r4-r11}
+
+	tst	ip,ip
+	beq	.Lpoly1305_emit_enter
+
+	ldmia	r0,{r3-r7}
+	eor	r8,r8,r8
+
+	adds	r3,r3,r4,lsl#26	@ base 2^26 -> base 2^32
+	mov	r4,r4,lsr#6
+	adcs	r4,r4,r5,lsl#20
+	mov	r5,r5,lsr#12
+	adcs	r5,r5,r6,lsl#14
+	mov	r6,r6,lsr#18
+	adcs	r6,r6,r7,lsl#8
+	adc	r7,r8,r7,lsr#24	@ can be partially reduced ...
+
+	and	r8,r7,#-4		@ ... so reduce
+	and	r7,r6,#3
+	add	r8,r8,r8,lsr#2	@ *= 5
+	adds	r3,r3,r8
+	adcs	r4,r4,#0
+	adcs	r5,r5,#0
+	adcs	r6,r6,#0
+	adc	r7,r7,#0
+
+	adds	r8,r3,#5		@ compare to modulus
+	adcs	r9,r4,#0
+	adcs	r10,r5,#0
+	adcs	r11,r6,#0
+	adc	r7,r7,#0
+	tst	r7,#4			@ did it carry/borrow?
+
+	it	ne
+	movne	r3,r8
+	ldr	r8,[r2,#0]
+	it	ne
+	movne	r4,r9
+	ldr	r9,[r2,#4]
+	it	ne
+	movne	r5,r10
+	ldr	r10,[r2,#8]
+	it	ne
+	movne	r6,r11
+	ldr	r11,[r2,#12]
+
+	adds	r3,r3,r8		@ accumulate nonce
+	adcs	r4,r4,r9
+	adcs	r5,r5,r10
+	adc	r6,r6,r11
+
+# ifdef __ARMEB__
+	rev	r3,r3
+	rev	r4,r4
+	rev	r5,r5
+	rev	r6,r6
+# endif
+	str	r3,[r1,#0]		@ store the result
+	str	r4,[r1,#4]
+	str	r5,[r1,#8]
+	str	r6,[r1,#12]
+
+	ldmia	sp!,{r4-r11}
+	bx	lr				@ bx	lr
+.size	poly1305_emit_neon,.-poly1305_emit_neon
+
+.align	5
+.Lzeros:
+.long	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-.Lpoly1305_init
+#endif
+.asciz	"Poly1305 for ARMv4/NEON, CRYPTOGAMS by <[email protected]>"
+.align	2
+#if	__ARM_MAX_ARCH__>=7
+.comm   OPENSSL_armcap_P,4,4
+#endif
lib/zinc/poly1305/poly1305-arm64-cryptogams.S new file +869 -0

869 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#15 lore raw

PATCH v8 15/28 zinc: Poly1305 ARM and ARM64 implementations

In reply to Jason A. Donenfeld

These wire Andy Polyakov's implementations up to the kernel. We make a few small changes to the assembly:

- Entries and exits use the proper kernel convention macro.
- CPU feature checking is done in C by the glue code, so that has been removed from the assembly.
- The function names have been renamed to fit kernel conventions.
- Labels have been renamed to fit kernel conventions.
- The neon code can jump to the scalar code when it makes sense to do so.

The NEON code uses base 2^26, while the scalar code uses base 2^64 on 64-bit and base 2^32 on 32-bit. If we hit the unfortunate situation of using NEON and then having to go back to scalar -- because the user is silly and has called the update function from two separate contexts -- then we need to convert back to the original base before proceeding. It is possible to reason that the initial reduction below is sufficient given the implementation invariants. However, for an avoidance of doubt and because this is not performance critical, we do the full reduction anyway. This conversion is found in the glue code, and a proof of correctness may be easily obtained from Z3: <https://xn--4db.cc/ltPtHCKN/py>.

Cc Russell King <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                             |   2 +
 lib/zinc/poly1305/poly1305-arm-glue.c         | 140 +++++++++++++++++
 ...ly1305-arm-cryptogams.S => poly1305-arm.S} | 147 ++++++------------
 ...05-arm64-cryptogams.S => poly1305-arm64.S} | 127 +++++----------
 lib/zinc/poly1305/poly1305.c                  |   2 +
 5 files changed, 231 insertions(+), 187 deletions(-)
 create mode 100644 lib/zinc/poly1305/poly1305-arm-glue.c
 rename lib/zinc/poly1305/{poly1305-arm-cryptogams.S => poly1305-arm.S} (91%)
 rename lib/zinc/poly1305/{poly1305-arm64-cryptogams.S => poly1305-arm64.S} (89%)
lib/zinc/Makefile +2 -0
index a8943d960b6a..c09fd3de60f9 100644
@@ -12,4 +12,6 @@ obj-$(CONFIG_ZINC_CHACHA20) += zinc_chacha20.o
 
 zinc_poly1305-y := poly1305/poly1305.o
 zinc_poly1305-$(CONFIG_ZINC_ARCH_X86_64) += poly1305/poly1305-x86_64.o
+zinc_poly1305-$(CONFIG_ZINC_ARCH_ARM) += poly1305/poly1305-arm.o
+zinc_poly1305-$(CONFIG_ZINC_ARCH_ARM64) += poly1305/poly1305-arm64.o
 obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
lib/zinc/poly1305/poly1305-arm-glue.c new file +140 -0

140 lines not on this page · raw message

lib/zinc/poly1305/poly1305-arm.S renamed from lib/zinc/poly1305/poly1305-arm-cryptogams.S +46 -101

334 lines not on this page · raw message

lib/zinc/poly1305/poly1305-arm64.S renamed from lib/zinc/poly1305/poly1305-arm64-cryptogams.S +41 -86

295 lines not on this page · raw message

lib/zinc/poly1305/poly1305.c +2 -0
index 0a580f1328fc..21d9a0b9c11c 100644
@@ -18,6 +18,8 @@
 
 #if defined(CONFIG_ZINC_ARCH_X86_64)
 #include "poly1305-x86_64-glue.c"
+#elif defined(CONFIG_ZINC_ARCH_ARM) || defined(CONFIG_ZINC_ARCH_ARM64)
+#include "poly1305-arm-glue.c"
 #else
 static inline bool poly1305_init_arch(void *ctx,
 				      const u8 key[POLY1305_KEY_SIZE])
Signature · 1 line
2.19.1
author maintainer
#16 lore raw

PATCH v8 16/28 zinc: import Andy Polyakov's Poly1305 MIPS64 implementation

In reply to Jason A. Donenfeld

This MIPS64 accelerated implementation comes from Andy Polyakov's implementation, and is included here in raw form without modification, so that subsequent commits that fix these up for the kernel can see how it has changed.

While this is CRYPTOGAMS code, the originating code for this happens to be the same as OpenSSL's commit 947716c1872d210828122212d076d503ae68b928

Based-on-code-from: Andy Polyakov <[email protected]>
Cc: Andy Polyakov <[email protected]>
Cc: Ralf Baechle <[email protected]>
Cc: Paul Burton <[email protected]>
Cc: James Hogan <[email protected]>
Cc: [email protected]
Cc: Samuel Neves <[email protected]>
Cc: Jean-Philippe Aumasson <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 .../poly1305/poly1305-mips64-cryptogams.S     | 338 ++++++++++++++++++
 1 file changed, 338 insertions(+)
 create mode 100644 lib/zinc/poly1305/poly1305-mips64-cryptogams.S
lib/zinc/poly1305/poly1305-mips64-cryptogams.S new file +338 -0
new file mode 100644
index 000000000000..24a6005884c3
@@ -0,0 +1,338 @@
+/* SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause */
+/*
+ * Copyright (C) 2006-2017 CRYPTOGAMS by <[email protected]>. All Rights Reserved.
+ */
+
+#include "mips_arch.h"
+
+#ifdef MIPSEB
+# define MSB 0
+# define LSB 7
+#else
+# define MSB 7
+# define LSB 0
+#endif
+
+.text
+.set	noat
+.set	noreorder
+
+.align	5
+.globl	poly1305_init
+.ent	poly1305_init
+poly1305_init:
+	.frame	$29,0,$31
+	.set	reorder
+
+	sd	$0,0($4)
+	sd	$0,8($4)
+	sd	$0,16($4)
+
+	beqz	$5,.Lno_key
+
+#if defined(_MIPS_ARCH_MIPS64R6)
+	ld	$8,0($5)
+	ld	$9,8($5)
+#else
+	ldl	$8,0+MSB($5)
+	ldl	$9,8+MSB($5)
+	ldr	$8,0+LSB($5)
+	ldr	$9,8+LSB($5)
+#endif
+#ifdef	MIPSEB
+# if defined(_MIPS_ARCH_MIPS64R2)
+	dsbh	$8,$8		# byte swap
+	 dsbh	$9,$9
+	dshd	$8,$8
+	 dshd	$9,$9
+# else
+	ori	$10,$0,0xFF
+	dsll	$1,$10,32
+	or	$10,$1		# 0x000000FF000000FF
+
+	and	$11,$8,$10	# byte swap
+	 and	$2,$9,$10
+	dsrl	$1,$8,24
+	 dsrl	$24,$9,24
+	dsll	$11,24
+	 dsll	$2,24
+	and	$1,$10
+	 and	$24,$10
+	dsll	$10,8			# 0x0000FF000000FF00
+	or	$11,$1
+	 or	$2,$24
+	and	$1,$8,$10
+	 and	$24,$9,$10
+	dsrl	$8,8
+	 dsrl	$9,8
+	dsll	$1,8
+	 dsll	$24,8
+	and	$8,$10
+	 and	$9,$10
+	or	$11,$1
+	 or	$2,$24
+	or	$8,$11
+	 or	$9,$2
+	dsrl	$11,$8,32
+	 dsrl	$2,$9,32
+	dsll	$8,32
+	 dsll	$9,32
+	or	$8,$11
+	 or	$9,$2
+# endif
+#endif
+	li	$10,1
+	dsll	$10,32
+	daddiu	$10,-63
+	dsll	$10,28
+	daddiu	$10,-1		# 0ffffffc0fffffff
+
+	and	$8,$10
+	daddiu	$10,-3		# 0ffffffc0ffffffc
+	and	$9,$10
+
+	sd	$8,24($4)
+	dsrl	$10,$9,2
+	sd	$9,32($4)
+	daddu	$10,$9		# s1 = r1 + (r1 >> 2)
+	sd	$10,40($4)
+
+.Lno_key:
+	li	$2,0			# return 0
+	jr	$31
+.end	poly1305_init
+.align	5
+.globl	poly1305_blocks
+.ent	poly1305_blocks
+poly1305_blocks:
+	.set	noreorder
+	dsrl	$6,4			# number of complete blocks
+	bnez	$6,poly1305_blocks_internal
+	nop
+	jr	$31
+	nop
+.end	poly1305_blocks
+
+.align	5
+.ent	poly1305_blocks_internal
+poly1305_blocks_internal:
+	.frame	$29,6*8,$31
+	.mask	0x00030000,-8
+	.set	noreorder
+	dsubu	$29,6*8
+	sd	$17,40($29)
+	sd	$16,32($29)
+	.set	reorder
+
+	ld	$12,0($4)		# load hash value
+	ld	$13,8($4)
+	ld	$14,16($4)
+
+	ld	$15,24($4)		# load key
+	ld	$16,32($4)
+	ld	$17,40($4)
+
+.Loop:
+#if defined(_MIPS_ARCH_MIPS64R6)
+	ld	$8,0($5)		# load input
+	ld	$9,8($5)
+#else
+	ldl	$8,0+MSB($5)	# load input
+	ldl	$9,8+MSB($5)
+	ldr	$8,0+LSB($5)
+	ldr	$9,8+LSB($5)
+#endif
+	daddiu	$6,-1
+	daddiu	$5,16
+#ifdef	MIPSEB
+# if defined(_MIPS_ARCH_MIPS64R2)
+	dsbh	$8,$8		# byte swap
+	 dsbh	$9,$9
+	dshd	$8,$8
+	 dshd	$9,$9
+# else
+	ori	$10,$0,0xFF
+	dsll	$1,$10,32
+	or	$10,$1		# 0x000000FF000000FF
+
+	and	$11,$8,$10	# byte swap
+	 and	$2,$9,$10
+	dsrl	$1,$8,24
+	 dsrl	$24,$9,24
+	dsll	$11,24
+	 dsll	$2,24
+	and	$1,$10
+	 and	$24,$10
+	dsll	$10,8			# 0x0000FF000000FF00
+	or	$11,$1
+	 or	$2,$24
+	and	$1,$8,$10
+	 and	$24,$9,$10
+	dsrl	$8,8
+	 dsrl	$9,8
+	dsll	$1,8
+	 dsll	$24,8
+	and	$8,$10
+	 and	$9,$10
+	or	$11,$1
+	 or	$2,$24
+	or	$8,$11
+	 or	$9,$2
+	dsrl	$11,$8,32
+	 dsrl	$2,$9,32
+	dsll	$8,32
+	 dsll	$9,32
+	or	$8,$11
+	 or	$9,$2
+# endif
+#endif
+	daddu	$12,$8		# accumulate input
+	daddu	$13,$9
+	sltu	$10,$12,$8
+	sltu	$11,$13,$9
+	daddu	$13,$10
+
+	dmultu	($15,$12)		# h0*r0
+	 daddu	$14,$7
+	 sltu	$10,$13,$10
+	mflo	($8,$15,$12)
+	mfhi	($9,$15,$12)
+
+	dmultu	($17,$13)		# h1*5*r1
+	 daddu	$10,$11
+	 daddu	$14,$10
+	mflo	($10,$17,$13)
+	mfhi	($11,$17,$13)
+
+	dmultu	($16,$12)		# h0*r1
+	 daddu	$8,$10
+	 daddu	$9,$11
+	mflo	($1,$16,$12)
+	mfhi	($25,$16,$12)
+	 sltu	$10,$8,$10
+	 daddu	$9,$10
+
+	dmultu	($15,$13)		# h1*r0
+	 daddu	$9,$1
+	 sltu	$1,$9,$1
+	mflo	($10,$15,$13)
+	mfhi	($11,$15,$13)
+	 daddu	$25,$1
+
+	dmultu	($17,$14)		# h2*5*r1
+	 daddu	$9,$10
+	 daddu	$25,$11
+	mflo	($1,$17,$14)
+
+	dmultu	($15,$14)		# h2*r0
+	 sltu	$10,$9,$10
+	 daddu	$25,$10
+	mflo	($2,$15,$14)
+
+	daddu	$9,$1
+	daddu	$25,$2
+	sltu	$1,$9,$1
+	daddu	$25,$1
+
+	li	$10,-4		# final reduction
+	and	$10,$25
+	dsrl	$11,$25,2
+	andi	$14,$25,3
+	daddu	$10,$11
+	daddu	$12,$8,$10
+	sltu	$10,$12,$10
+	daddu	$13,$9,$10
+	sltu	$10,$13,$10
+	daddu	$14,$14,$10
+
+	bnez	$6,.Loop
+
+	sd	$12,0($4)		# store hash value
+	sd	$13,8($4)
+	sd	$14,16($4)
+
+	.set	noreorder
+	ld	$17,40($29)		# epilogue
+	ld	$16,32($29)
+	jr	$31
+	daddu	$29,6*8
+.end	poly1305_blocks_internal
+.align	5
+.globl	poly1305_emit
+.ent	poly1305_emit
+poly1305_emit:
+	.frame	$29,0,$31
+	.set	reorder
+
+	ld	$10,0($4)
+	ld	$11,8($4)
+	ld	$1,16($4)
+
+	daddiu	$8,$10,5		# compare to modulus
+	sltiu	$2,$8,5
+	daddu	$9,$11,$2
+	sltu	$2,$9,$2
+	daddu	$1,$1,$2
+
+	dsrl	$1,2			# see if it carried/borrowed
+	dsubu	$1,$0,$1
+	nor	$2,$0,$1
+
+	and	$8,$1
+	and	$10,$2
+	and	$9,$1
+	and	$11,$2
+	or	$8,$10
+	or	$9,$11
+
+	lwu	$10,0($6)		# load nonce
+	lwu	$11,4($6)
+	lwu	$1,8($6)
+	lwu	$2,12($6)
+	dsll	$11,32
+	dsll	$2,32
+	or	$10,$11
+	or	$1,$2
+
+	daddu	$8,$10		# accumulate nonce
+	daddu	$9,$1
+	sltu	$10,$8,$10
+	daddu	$9,$10
+
+	dsrl	$10,$8,8		# write mac value
+	dsrl	$11,$8,16
+	dsrl	$1,$8,24
+	sb	$8,0($5)
+	dsrl	$2,$8,32
+	sb	$10,1($5)
+	dsrl	$10,$8,40
+	sb	$11,2($5)
+	dsrl	$11,$8,48
+	sb	$1,3($5)
+	dsrl	$1,$8,56
+	sb	$2,4($5)
+	dsrl	$2,$9,8
+	sb	$10,5($5)
+	dsrl	$10,$9,16
+	sb	$11,6($5)
+	dsrl	$11,$9,24
+	sb	$1,7($5)
+
+	sb	$9,8($5)
+	dsrl	$1,$9,32
+	sb	$2,9($5)
+	dsrl	$2,$9,40
+	sb	$10,10($5)
+	dsrl	$10,$9,48
+	sb	$11,11($5)
+	dsrl	$11,$9,56
+	sb	$1,12($5)
+	sb	$2,13($5)
+	sb	$10,14($5)
+	sb	$11,15($5)
+
+	jr	$31
+.end	poly1305_emit
+.rdata
+.asciiz	"Poly1305 for MIPS64, CRYPTOGAMS by <[email protected]>"
+.align	2
Signature · 1 line
2.19.1
author maintainer
#17 lore raw

PATCH v8 17/28 zinc: Poly1305 MIPS32r2 and MIPS64 implementations

In reply to Jason A. Donenfeld

This MIPS32r2 implementation comes from René van Dorst and me and results in a nice speedup on the usual OpenWRT targets. The MIPS64 implementation from Andy Polyakov ported here results in a nice speedup on commodity Octeon hardware, and has been modified slightly from the original:

- The function names have been renamed to fit kernel conventions.
- A comment has been added.

Cc Ralf Baechle <[email protected]>
Cc Paul Burton <[email protected]>
Cc James Hogan <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                             |   3 +
 lib/zinc/poly1305/poly1305-mips-glue.c        |  37 ++
 lib/zinc/poly1305/poly1305-mips.S             | 407 ++++++++++++++++++
 ...-mips64-cryptogams.S => poly1305-mips64.S} |  80 ++--
 lib/zinc/poly1305/poly1305.c                  |   2 +
 5 files changed, 500 insertions(+), 29 deletions(-)
 create mode 100644 lib/zinc/poly1305/poly1305-mips-glue.c
 create mode 100644 lib/zinc/poly1305/poly1305-mips.S
 rename lib/zinc/poly1305/{poly1305-mips64-cryptogams.S => poly1305-mips64.S} (75%)
lib/zinc/Makefile +3 -0
index c09fd3de60f9..5c4b1d51cb03 100644
@@ -14,4 +14,7 @@ zinc_poly1305-y := poly1305/poly1305.o
 zinc_poly1305-$(CONFIG_ZINC_ARCH_X86_64) += poly1305/poly1305-x86_64.o
 zinc_poly1305-$(CONFIG_ZINC_ARCH_ARM) += poly1305/poly1305-arm.o
 zinc_poly1305-$(CONFIG_ZINC_ARCH_ARM64) += poly1305/poly1305-arm64.o
+zinc_poly1305-$(CONFIG_ZINC_ARCH_MIPS) += poly1305/poly1305-mips.o
+AFLAGS_poly1305-mips.o += -O2 # This is required to fill the branch delay slots
+zinc_poly1305-$(CONFIG_ZINC_ARCH_MIPS64) += poly1305/poly1305-mips64.o
 obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
lib/zinc/poly1305/poly1305-mips-glue.c new file +37 -0
new file mode 100644
index 000000000000..1eba9512a05c
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+asmlinkage void poly1305_init_mips(void *ctx, const u8 key[16]);
+asmlinkage void poly1305_blocks_mips(void *ctx, const u8 *inp, const size_t len,
+				     const u32 padbit);
+asmlinkage void poly1305_emit_mips(void *ctx, u8 mac[16], const u32 nonce[4]);
+
+static bool *const poly1305_nobs[] __initconst = { };
+static void __init poly1305_fpu_init(void)
+{
+}
+
+static inline bool poly1305_init_arch(void *ctx,
+				      const u8 key[POLY1305_KEY_SIZE])
+{
+	poly1305_init_mips(ctx, key);
+	return true;
+}
+
+static inline bool poly1305_blocks_arch(void *ctx, const u8 *inp,
+					size_t len, const u32 padbit,
+					simd_context_t *simd_context)
+{
+	poly1305_blocks_mips(ctx, inp, len, padbit);
+	return true;
+}
+
+static inline bool poly1305_emit_arch(void *ctx, u8 mac[POLY1305_MAC_SIZE],
+				      const u32 nonce[4],
+				      simd_context_t *simd_context)
+{
+	poly1305_emit_mips(ctx, mac, nonce);
+	return true;
+}
lib/zinc/poly1305/poly1305-mips.S new file +407 -0

407 lines not on this page · raw message

lib/zinc/poly1305/poly1305-mips64.S renamed from lib/zinc/poly1305/poly1305-mips64-cryptogams.S +51 -29

152 lines not on this page · raw message

lib/zinc/poly1305/poly1305.c +2 -0
index 21d9a0b9c11c..69e634c8a8d9 100644
@@ -20,6 +20,8 @@
 #include "poly1305-x86_64-glue.c"
 #elif defined(CONFIG_ZINC_ARCH_ARM) || defined(CONFIG_ZINC_ARCH_ARM64)
 #include "poly1305-arm-glue.c"
+#elif defined(CONFIG_ZINC_ARCH_MIPS) || defined(CONFIG_ZINC_ARCH_MIPS64)
+#include "poly1305-mips-glue.c"
 #else
 static inline bool poly1305_init_arch(void *ctx,
 				      const u8 key[POLY1305_KEY_SIZE])
Signature · 1 line
2.19.1
author maintainer
#18 lore raw

PATCH v8 18/28 zinc: ChaCha20Poly1305 construction and selftest

In reply to Jason A. Donenfeld

This is an implementation of the ChaCha20Poly1305 AEAD, with an easy API for encrypting either contiguous buffers or scatter gather lists (such as those created from skb_to_sgvec).

Information: https://tools.ietf.org/html/rfc8439

Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 include/zinc/chacha20poly1305.h      |   50 +
 lib/zinc/Kconfig                     |    6 +
 lib/zinc/Makefile                    |    3 +
 lib/zinc/chacha20poly1305.c          |  362 ++
 lib/zinc/selftest/chacha20poly1305.c | 9034 ++++++++++++++++++++++++++
 5 files changed, 9455 insertions(+)
 create mode 100644 include/zinc/chacha20poly1305.h
 create mode 100644 lib/zinc/chacha20poly1305.c
 create mode 100644 lib/zinc/selftest/chacha20poly1305.c
include/zinc/chacha20poly1305.h new file +50 -0
new file mode 100644
index 000000000000..d2753c5bada5
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _ZINC_CHACHA20POLY1305_H
+#define _ZINC_CHACHA20POLY1305_H
+
+#include <linux/simd.h>
+#include <linux/types.h>
+
+struct scatterlist;
+
+enum chacha20poly1305_lengths {
+	XCHACHA20POLY1305_NONCE_SIZE = 24,
+	CHACHA20POLY1305_KEY_SIZE = 32,
+	CHACHA20POLY1305_AUTHTAG_SIZE = 16
+};
+
+void chacha20poly1305_encrypt(u8 *dst, const u8 *src, const size_t src_len,
+			      const u8 *ad, const size_t ad_len,
+			      const u64 nonce,
+			      const u8 key[CHACHA20POLY1305_KEY_SIZE]);
+
+bool __must_check chacha20poly1305_encrypt_sg(
+	struct scatterlist *dst, struct scatterlist *src, const size_t src_len,
+	const u8 *ad, const size_t ad_len, const u64 nonce,
+	const u8 key[CHACHA20POLY1305_KEY_SIZE], simd_context_t *simd_context);
+
+bool __must_check
+chacha20poly1305_decrypt(u8 *dst, const u8 *src, const size_t src_len,
+			 const u8 *ad, const size_t ad_len, const u64 nonce,
+			 const u8 key[CHACHA20POLY1305_KEY_SIZE]);
+
+bool __must_check chacha20poly1305_decrypt_sg(
+	struct scatterlist *dst, struct scatterlist *src, const size_t src_len,
+	const u8 *ad, const size_t ad_len, const u64 nonce,
+	const u8 key[CHACHA20POLY1305_KEY_SIZE], simd_context_t *simd_context);
+
+void xchacha20poly1305_encrypt(u8 *dst, const u8 *src, const size_t src_len,
+			       const u8 *ad, const size_t ad_len,
+			       const u8 nonce[XCHACHA20POLY1305_NONCE_SIZE],
+			       const u8 key[CHACHA20POLY1305_KEY_SIZE]);
+
+bool __must_check xchacha20poly1305_decrypt(
+	u8 *dst, const u8 *src, const size_t src_len, const u8 *ad,
+	const size_t ad_len, const u8 nonce[XCHACHA20POLY1305_NONCE_SIZE],
+	const u8 key[CHACHA20POLY1305_KEY_SIZE]);
+
+#endif /* _ZINC_CHACHA20POLY1305_H */
lib/zinc/Kconfig +6 -0
index f08bf1eaa2a0..765eba3267c9 100644
@@ -5,6 +5,12 @@ config ZINC_CHACHA20
 config ZINC_POLY1305
 	tristate
 
+config ZINC_CHACHA20POLY1305
+	tristate
+	select ZINC_CHACHA20
+	select ZINC_POLY1305
+	select CRYPTO_BLKCIPHER
+
 config ZINC_SELFTEST
 	bool "Zinc cryptography library self-tests"
 	help
lib/zinc/Makefile +3 -0
index 5c4b1d51cb03..c31186b491e8 100644
@@ -18,3 +18,6 @@ zinc_poly1305-$(CONFIG_ZINC_ARCH_MIPS) += poly1305/poly1305-mips.o
 AFLAGS_poly1305-mips.o += -O2 # This is required to fill the branch delay slots
 zinc_poly1305-$(CONFIG_ZINC_ARCH_MIPS64) += poly1305/poly1305-mips64.o
 obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
+
+zinc_chacha20poly1305-y := chacha20poly1305.o
+obj-$(CONFIG_ZINC_CHACHA20POLY1305) += zinc_chacha20poly1305.o
lib/zinc/chacha20poly1305.c new file +362 -0

362 lines not on this page · raw message

lib/zinc/selftest/chacha20poly1305.c new file +9034 -0

2,570 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#19 lore raw

PATCH v8 19/28 zinc: BLAKE2s generic C implementation and selftest

In reply to Jason A. Donenfeld

The C implementation was originally based on Samuel Neves' public domain reference implementation but has since been heavily modified for the kernel. We're able to do compile-time optimizations by moving some scaffolding around the final function into the header file.

Information: https://blake2.net/

---
 include/zinc/blake2s.h      |   56 +
 lib/zinc/Kconfig            |    3 +
 lib/zinc/Makefile           |    3 +
 lib/zinc/blake2s/blake2s.c  |  295 +++++
 lib/zinc/selftest/blake2s.c | 2090 +++++++++++++++++++++++++++++++++++
 5 files changed, 2447 insertions(+)
 create mode 100644 include/zinc/blake2s.h
 create mode 100644 lib/zinc/blake2s/blake2s.c
 create mode 100644 lib/zinc/selftest/blake2s.c
include/zinc/blake2s.h new file +56 -0
new file mode 100644
index 000000000000..701a08ba47c1
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _ZINC_BLAKE2S_H
+#define _ZINC_BLAKE2S_H
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <asm/bug.h>
+
+enum blake2s_lengths {
+	BLAKE2S_BLOCK_SIZE = 64,
+	BLAKE2S_HASH_SIZE = 32,
+	BLAKE2S_KEY_SIZE = 32
+};
+
+struct blake2s_state {
+	u32 h[8];
+	u32 t[2];
+	u32 f[2];
+	u8 buf[BLAKE2S_BLOCK_SIZE];
+	size_t buflen;
+	u8 last_node;
+};
+
+void blake2s_init(struct blake2s_state *state, const size_t outlen);
+void blake2s_init_key(struct blake2s_state *state, const size_t outlen,
+		      const void *key, const size_t keylen);
+void blake2s_update(struct blake2s_state *state, const u8 *in, size_t inlen);
+void blake2s_final(struct blake2s_state *state, u8 *out, const size_t outlen);
+
+static inline void blake2s(u8 *out, const u8 *in, const u8 *key,
+			   const size_t outlen, const size_t inlen,
+			   const size_t keylen)
+{
+	struct blake2s_state state;
+
+	WARN_ON(IS_ENABLED(DEBUG) && ((!in && inlen > 0) || !out || !outlen ||
+		outlen > BLAKE2S_HASH_SIZE || keylen > BLAKE2S_KEY_SIZE ||
+		(!key && keylen)));
+
+	if (keylen)
+		blake2s_init_key(&state, outlen, key, keylen);
+	else
+		blake2s_init(&state, outlen);
+
+	blake2s_update(&state, in, inlen);
+	blake2s_final(&state, out, outlen);
+}
+
+void blake2s_hmac(u8 *out, const u8 *in, const u8 *key, const size_t outlen,
+		  const size_t inlen, const size_t keylen);
+
+#endif /* _ZINC_BLAKE2S_H */
lib/zinc/Kconfig +3 -0
index 765eba3267c9..9fc21f93ee9f 100644
@@ -11,6 +11,9 @@ config ZINC_CHACHA20POLY1305
 	select ZINC_POLY1305
 	select CRYPTO_BLKCIPHER
 
+config ZINC_BLAKE2S
+	tristate
+
 config ZINC_SELFTEST
 	bool "Zinc cryptography library self-tests"
 	help
lib/zinc/Makefile +3 -0
index c31186b491e8..d2ec55c33ef0 100644
@@ -21,3 +21,6 @@ obj-$(CONFIG_ZINC_POLY1305) += zinc_poly1305.o
 
 zinc_chacha20poly1305-y := chacha20poly1305.o
 obj-$(CONFIG_ZINC_CHACHA20POLY1305) += zinc_chacha20poly1305.o
+
+zinc_blake2s-y := blake2s/blake2s.o
+obj-$(CONFIG_ZINC_BLAKE2S) += zinc_blake2s.o
lib/zinc/blake2s/blake2s.c new file +295 -0

295 lines not on this page · raw message

lib/zinc/selftest/blake2s.c new file +2090 -0

2,090 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#20 lore raw

PATCH v8 20/28 zinc: BLAKE2s x86_64 implementation

In reply to Jason A. Donenfeld

These implementations from Samuel Neves support AVX and AVX-512VL. Originally this used AVX-512F, but Skylake thermal throttling made AVX-512VL more attractive and possible to do with negligable difference.

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                      |   1 +
 lib/zinc/blake2s/blake2s-x86_64-glue.c |  71 +++
 lib/zinc/blake2s/blake2s-x86_64.S      | 685 +++++++++++++++++++++++++
 lib/zinc/blake2s/blake2s.c             |   4 +
 4 files changed, 761 insertions(+)
 create mode 100644 lib/zinc/blake2s/blake2s-x86_64-glue.c
 create mode 100644 lib/zinc/blake2s/blake2s-x86_64.S
lib/zinc/Makefile +1 -0
index d2ec55c33ef0..67ad837c822c 100644
@@ -23,4 +23,5 @@ zinc_chacha20poly1305-y := chacha20poly1305.o
 obj-$(CONFIG_ZINC_CHACHA20POLY1305) += zinc_chacha20poly1305.o
 
 zinc_blake2s-y := blake2s/blake2s.o
+zinc_blake2s-$(CONFIG_ZINC_ARCH_X86_64) += blake2s/blake2s-x86_64.o
 obj-$(CONFIG_ZINC_BLAKE2S) += zinc_blake2s.o
lib/zinc/blake2s/blake2s-x86_64-glue.c new file +71 -0

71 lines not on this page · raw message

lib/zinc/blake2s/blake2s-x86_64.S new file +685 -0

685 lines not on this page · raw message

lib/zinc/blake2s/blake2s.c +4 -0
index 58d7e9378bd4..59db1ce2f7ef 100644
@@ -110,6 +110,9 @@ void blake2s_init_key(struct blake2s_state *state, const size_t outlen,
 }
 EXPORT_SYMBOL(blake2s_init_key);
 
+#if defined(CONFIG_ZINC_ARCH_X86_64)
+#include "blake2s-x86_64-glue.c"
+#else
 static bool *const blake2s_nobs[] __initconst = { };
 static void __init blake2s_fpu_init(void)
 {
@@ -120,6 +123,7 @@ static inline bool blake2s_compress_arch(struct blake2s_state *state,
 {
 	return false;
 }
+#endif
 
 static inline void blake2s_compress(struct blake2s_state *state,
 				    const u8 *block, size_t nblocks,
Signature · 1 line
2.19.1
author maintainer
#21 lore raw

PATCH v8 21/28 zinc: Curve25519 generic C implementations and selftest

In reply to Jason A. Donenfeld

This contains two formally verified C implementations of the Curve25519 scalar multiplication function, one for 32-bit systems, and one for 64-bit systems whose compiler supports efficient 128-bit integer types. Not only are these implementations formally verified, but they are also the fastest available C implementations. They have been modified to be friendly to kernel space and to be generally less horrendous looking, but still an effort has been made to retain their formally verified characteristic, and so the C might look slightly unidiomatic.

The 64-bit version comes from HACL*: https://github.com/project-everest/hacl-star
The 32-bit version comes from Fiat: https://github.com/mit-plv/fiat-crypto

Information: https://cr.yp.to/ecdh.html

Cc Karthikeyan Bhargavan <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 include/zinc/curve25519.h               |   22 +
 lib/zinc/Kconfig                        |    4 +
 lib/zinc/Makefile                       |    3 +
 lib/zinc/curve25519/curve25519-fiat32.c |  860 +++++++++++++++
 lib/zinc/curve25519/curve25519-hacl64.c |  784 ++++++++++++++
 lib/zinc/curve25519/curve25519.c        |  108 ++
 lib/zinc/selftest/curve25519.c          | 1315 +++++++++++++++++++++++
 7 files changed, 3096 insertions(+)
 create mode 100644 include/zinc/curve25519.h
 create mode 100644 lib/zinc/curve25519/curve25519-fiat32.c
 create mode 100644 lib/zinc/curve25519/curve25519-hacl64.c
 create mode 100644 lib/zinc/curve25519/curve25519.c
 create mode 100644 lib/zinc/selftest/curve25519.c
include/zinc/curve25519.h new file +22 -0
new file mode 100644
index 000000000000..def173e736fc
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _ZINC_CURVE25519_H
+#define _ZINC_CURVE25519_H
+
+#include <linux/types.h>
+
+enum curve25519_lengths {
+	CURVE25519_KEY_SIZE = 32
+};
+
+bool __must_check curve25519(u8 mypublic[CURVE25519_KEY_SIZE],
+			     const u8 secret[CURVE25519_KEY_SIZE],
+			     const u8 basepoint[CURVE25519_KEY_SIZE]);
+void curve25519_generate_secret(u8 secret[CURVE25519_KEY_SIZE]);
+bool __must_check curve25519_generate_public(
+	u8 pub[CURVE25519_KEY_SIZE], const u8 secret[CURVE25519_KEY_SIZE]);
+
+#endif /* _ZINC_CURVE25519_H */
lib/zinc/Kconfig +4 -0
index 9fc21f93ee9f..f1840c4e9fde 100644
@@ -14,6 +14,10 @@ config ZINC_CHACHA20POLY1305
 config ZINC_BLAKE2S
 	tristate
 
+config ZINC_CURVE25519
+	tristate
+	select CONFIG_CRYPTO
+
 config ZINC_SELFTEST
 	bool "Zinc cryptography library self-tests"
 	help
lib/zinc/Makefile +3 -0
index 67ad837c822c..65440438c6e5 100644
@@ -25,3 +25,6 @@ obj-$(CONFIG_ZINC_CHACHA20POLY1305) += zinc_chacha20poly1305.o
 zinc_blake2s-y := blake2s/blake2s.o
 zinc_blake2s-$(CONFIG_ZINC_ARCH_X86_64) += blake2s/blake2s-x86_64.o
 obj-$(CONFIG_ZINC_BLAKE2S) += zinc_blake2s.o
+
+zinc_curve25519-y := curve25519/curve25519.o
+obj-$(CONFIG_ZINC_CURVE25519) += zinc_curve25519.o
lib/zinc/curve25519/curve25519-fiat32.c new file +860 -0

860 lines not on this page · raw message

lib/zinc/curve25519/curve25519-hacl64.c new file +784 -0

784 lines not on this page · raw message

lib/zinc/curve25519/curve25519.c new file +108 -0

108 lines not on this page · raw message

lib/zinc/selftest/curve25519.c new file +1315 -0

1,210 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

author maintainer
#22 lore raw

PATCH v8 22/28 zinc: Curve25519 x86_64 implementation

In reply to Jason A. Donenfeld

This implementation is the fastest available x86_64 implementation, and unlike Sandy2x, it doesn't requie use of the floating point registers at all. Instead it makes use of BMI2 and ADX, available on recent microarchitectures. The implementation was written by Armando Faz-Hernández with contributions (upstream) from Samuel Neves and me, in addition to further changes in the kernel implementation from us.

Cc Armando Faz-Hernández <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/curve25519/curve25519-x86_64-glue.c |   48 +
 lib/zinc/curve25519/curve25519-x86_64.c      | 2333 ++++++++++++++++++
 lib/zinc/curve25519/curve25519.c             |    4 +
 3 files changed, 2385 insertions(+)
 create mode 100644 lib/zinc/curve25519/curve25519-x86_64-glue.c
 create mode 100644 lib/zinc/curve25519/curve25519-x86_64.c
lib/zinc/curve25519/curve25519-x86_64-glue.c new file +48 -0
new file mode 100644
index 000000000000..f79c1422a106
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#include <asm/cpufeature.h>
+#include <asm/processor.h>
+
+#include "curve25519-x86_64.c"
+
+static bool curve25519_use_bmi2 __ro_after_init;
+static bool curve25519_use_adx __ro_after_init;
+static bool *const curve25519_nobs[] __initconst = {
+	&curve25519_use_bmi2, &curve25519_use_adx };
+
+static void __init curve25519_fpu_init(void)
+{
+	curve25519_use_bmi2 = boot_cpu_has(X86_FEATURE_BMI2);
+	curve25519_use_adx = boot_cpu_has(X86_FEATURE_BMI2) &&
+			     boot_cpu_has(X86_FEATURE_ADX);
+}
+
+static inline bool curve25519_arch(u8 mypublic[CURVE25519_KEY_SIZE],
+				   const u8 secret[CURVE25519_KEY_SIZE],
+				   const u8 basepoint[CURVE25519_KEY_SIZE])
+{
+	if (curve25519_use_adx) {
+		curve25519_adx(mypublic, secret, basepoint);
+		return true;
+	} else if (curve25519_use_bmi2) {
+		curve25519_bmi2(mypublic, secret, basepoint);
+		return true;
+	}
+	return false;
+}
+
+static inline bool curve25519_base_arch(u8 pub[CURVE25519_KEY_SIZE],
+					const u8 secret[CURVE25519_KEY_SIZE])
+{
+	if (curve25519_use_adx) {
+		curve25519_adx_base(pub, secret);
+		return true;
+	} else if (curve25519_use_bmi2) {
+		curve25519_bmi2_base(pub, secret);
+		return true;
+	}
+	return false;
+}
lib/zinc/curve25519/curve25519-x86_64.c new file +2333 -0

2,333 lines not on this page · raw message

lib/zinc/curve25519/curve25519.c +4 -0
index 8d9d8fc0ca38..074fbb1c2149 100644
@@ -20,6 +20,9 @@
 #include <linux/init.h>
 #include <crypto/algapi.h> // For crypto_memneq.
 
+#if defined(CONFIG_ZINC_ARCH_X86_64)
+#include "curve25519-x86_64-glue.c"
+#else
 static bool *const curve25519_nobs[] __initconst = { };
 static void __init curve25519_fpu_init(void)
 {
@@ -35,6 +38,7 @@ static inline bool curve25519_base_arch(u8 pub[CURVE25519_KEY_SIZE],
 {
 	return false;
 }
+#endif
 
 static __always_inline void normalize_secret(u8 secret[CURVE25519_KEY_SIZE])
 {
Signature · 1 line
2.19.1
author maintainer
#23 lore raw

PATCH v8 23/28 zinc: import Bernstein and Schwabe's Curve25519 ARM implementation

In reply to Jason A. Donenfeld

This comes from Dan Bernstein and Peter Schwabe's public domain NEON code, and is included here in raw form so that subsequent commits that fix these up for the kernel can see how it has changed. This code does have some entirely cosmetic formatting differences, adding indentation and so forth, so that when we actually port it for use in the kernel in the subsequent commit, it's obvious what's changed in the process.

This code originates from SUPERCOP 20180818, available at
<https://bench.cr.yp.to/supercop.html>.

Cc Russell King <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/curve25519/curve25519-arm-supercop.S | 2105 +++++++++++++++++
 1 file changed, 2105 insertions(+)
 create mode 100644 lib/zinc/curve25519/curve25519-arm-supercop.S
lib/zinc/curve25519/curve25519-arm-supercop.S new file +2105 -0
new file mode 100644
index 000000000000..f33b85fef382
@@ -0,0 +1,2105 @@
+/*
+ * Public domain code from Daniel J. Bernstein and Peter Schwabe, from
+ * SUPERCOP's curve25519/neon2/scalarmult.s.
+ */
+
+.fpu neon
+.text
+.align 4
+.global _crypto_scalarmult_curve25519_neon2
+.global crypto_scalarmult_curve25519_neon2
+.type _crypto_scalarmult_curve25519_neon2 STT_FUNC
+.type crypto_scalarmult_curve25519_neon2 STT_FUNC
+	_crypto_scalarmult_curve25519_neon2:
+	crypto_scalarmult_curve25519_neon2:
+	vpush		{q4, q5, q6, q7}
+	mov		r12, sp
+	sub		sp, sp, #736
+	and		sp, sp, #0xffffffe0
+	strd		r4, [sp, #0]
+	strd		r6, [sp, #8]
+	strd		r8, [sp, #16]
+	strd		r10, [sp, #24]
+	str		r12, [sp, #480]
+	str		r14, [sp, #484]
+	mov		r0, r0
+	mov		r1, r1
+	mov		r2, r2
+	add		r3, sp, #32
+	ldr		r4, =0
+	ldr		r5, =254
+	vmov.i32	q0, #1
+	vshr.u64	q1, q0, #7
+	vshr.u64	q0, q0, #8
+	vmov.i32	d4, #19
+	vmov.i32	d5, #38
+	add		r6, sp, #512
+	vst1.8		{d2-d3}, [r6, : 128]
+	add		r6, sp, #528
+	vst1.8		{d0-d1}, [r6, : 128]
+	add		r6, sp, #544
+	vst1.8		{d4-d5}, [r6, : 128]
+	add		r6, r3, #0
+	vmov.i32	q2, #0
+	vst1.8		{d4-d5}, [r6, : 128]!
+	vst1.8		{d4-d5}, [r6, : 128]!
+	vst1.8		d4, [r6, : 64]
+	add		r6, r3, #0
+	ldr		r7, =960
+	sub		r7, r7, #2
+	neg		r7, r7
+	sub		r7, r7, r7, LSL #7
+	str		r7, [r6]
+	add		r6, sp, #704
+	vld1.8		{d4-d5}, [r1]!
+	vld1.8		{d6-d7}, [r1]
+	vst1.8		{d4-d5}, [r6, : 128]!
+	vst1.8		{d6-d7}, [r6, : 128]
+	sub		r1, r6, #16
+	ldrb		r6, [r1]
+	and		r6, r6, #248
+	strb		r6, [r1]
+	ldrb		r6, [r1, #31]
+	and		r6, r6, #127
+	orr		r6, r6, #64
+	strb		r6, [r1, #31]
+	vmov.i64	q2, #0xffffffff
+	vshr.u64	q3, q2, #7
+	vshr.u64	q2, q2, #6
+	vld1.8		{d8}, [r2]
+	vld1.8		{d10}, [r2]
+	add		r2, r2, #6
+	vld1.8		{d12}, [r2]
+	vld1.8		{d14}, [r2]
+	add		r2, r2, #6
+	vld1.8		{d16}, [r2]
+	add		r2, r2, #4
+	vld1.8		{d18}, [r2]
+	vld1.8		{d20}, [r2]
+	add		r2, r2, #6
+	vld1.8		{d22}, [r2]
+	add		r2, r2, #2
+	vld1.8		{d24}, [r2]
+	vld1.8		{d26}, [r2]
+	vshr.u64	q5, q5, #26
+	vshr.u64	q6, q6, #3
+	vshr.u64	q7, q7, #29
+	vshr.u64	q8, q8, #6
+	vshr.u64	q10, q10, #25
+	vshr.u64	q11, q11, #3
+	vshr.u64	q12, q12, #12
+	vshr.u64	q13, q13, #38
+	vand		q4, q4, q2
+	vand		q6, q6, q2
+	vand		q8, q8, q2
+	vand		q10, q10, q2
+	vand		q2, q12, q2
+	vand		q5, q5, q3
+	vand		q7, q7, q3
+	vand		q9, q9, q3
+	vand		q11, q11, q3
+	vand		q3, q13, q3
+	add		r2, r3, #48
+	vadd.i64	q12, q4, q1
+	vadd.i64	q13, q10, q1
+	vshr.s64	q12, q12, #26
+	vshr.s64	q13, q13, #26
+	vadd.i64	q5, q5, q12
+	vshl.i64	q12, q12, #26
+	vadd.i64	q14, q5, q0
+	vadd.i64	q11, q11, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q15, q11, q0
+	vsub.i64	q4, q4, q12
+	vshr.s64	q12, q14, #25
+	vsub.i64	q10, q10, q13
+	vshr.s64	q13, q15, #25
+	vadd.i64	q6, q6, q12
+	vshl.i64	q12, q12, #25
+	vadd.i64	q14, q6, q1
+	vadd.i64	q2, q2, q13
+	vsub.i64	q5, q5, q12
+	vshr.s64	q12, q14, #26
+	vshl.i64	q13, q13, #25
+	vadd.i64	q14, q2, q1
+	vadd.i64	q7, q7, q12
+	vshl.i64	q12, q12, #26
+	vadd.i64	q15, q7, q0
+	vsub.i64	q11, q11, q13
+	vshr.s64	q13, q14, #26
+	vsub.i64	q6, q6, q12
+	vshr.s64	q12, q15, #25
+	vadd.i64	q3, q3, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q14, q3, q0
+	vadd.i64	q8, q8, q12
+	vshl.i64	q12, q12, #25
+	vadd.i64	q15, q8, q1
+	add		r2, r2, #8
+	vsub.i64	q2, q2, q13
+	vshr.s64	q13, q14, #25
+	vsub.i64	q7, q7, q12
+	vshr.s64	q12, q15, #26
+	vadd.i64	q14, q13, q13
+	vadd.i64	q9, q9, q12
+	vtrn.32		d12, d14
+	vshl.i64	q12, q12, #26
+	vtrn.32		d13, d15
+	vadd.i64	q0, q9, q0
+	vadd.i64	q4, q4, q14
+	vst1.8		d12, [r2, : 64]!
+	vshl.i64	q6, q13, #4
+	vsub.i64	q7, q8, q12
+	vshr.s64	q0, q0, #25
+	vadd.i64	q4, q4, q6
+	vadd.i64	q6, q10, q0
+	vshl.i64	q0, q0, #25
+	vadd.i64	q8, q6, q1
+	vadd.i64	q4, q4, q13
+	vshl.i64	q10, q13, #25
+	vadd.i64	q1, q4, q1
+	vsub.i64	q0, q9, q0
+	vshr.s64	q8, q8, #26
+	vsub.i64	q3, q3, q10
+	vtrn.32		d14, d0
+	vshr.s64	q1, q1, #26
+	vtrn.32		d15, d1
+	vadd.i64	q0, q11, q8
+	vst1.8		d14, [r2, : 64]
+	vshl.i64	q7, q8, #26
+	vadd.i64	q5, q5, q1
+	vtrn.32		d4, d6
+	vshl.i64	q1, q1, #26
+	vtrn.32		d5, d7
+	vsub.i64	q3, q6, q7
+	add		r2, r2, #16
+	vsub.i64	q1, q4, q1
+	vst1.8		d4, [r2, : 64]
+	vtrn.32		d6, d0
+	vtrn.32		d7, d1
+	sub		r2, r2, #8
+	vtrn.32		d2, d10
+	vtrn.32		d3, d11
+	vst1.8		d6, [r2, : 64]
+	sub		r2, r2, #24
+	vst1.8		d2, [r2, : 64]
+	add		r2, r3, #96
+	vmov.i32	q0, #0
+	vmov.i64	d2, #0xff
+	vmov.i64	d3, #0
+	vshr.u32	q1, q1, #7
+	vst1.8		{d2-d3}, [r2, : 128]!
+	vst1.8		{d0-d1}, [r2, : 128]!
+	vst1.8		d0, [r2, : 64]
+	add		r2, r3, #144
+	vmov.i32	q0, #0
+	vst1.8		{d0-d1}, [r2, : 128]!
+	vst1.8		{d0-d1}, [r2, : 128]!
+	vst1.8		d0, [r2, : 64]
+	add		r2, r3, #240
+	vmov.i32	q0, #0
+	vmov.i64	d2, #0xff
+	vmov.i64	d3, #0
+	vshr.u32	q1, q1, #7
+	vst1.8		{d2-d3}, [r2, : 128]!
+	vst1.8		{d0-d1}, [r2, : 128]!
+	vst1.8		d0, [r2, : 64]
+	add		r2, r3, #48
+	add		r6, r3, #192
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d4}, [r2, : 64]
+	vst1.8		{d0-d1}, [r6, : 128]!
+	vst1.8		{d2-d3}, [r6, : 128]!
+	vst1.8		d4, [r6, : 64]
+._mainloop:
+	mov		r2, r5, LSR #3
+	and		r6, r5, #7
+	ldrb		r2, [r1, r2]
+	mov		r2, r2, LSR r6
+	and		r2, r2, #1
+	str		r5, [sp, #488]
+	eor		r4, r4, r2
+	str		r2, [sp, #492]
+	neg		r2, r4
+	add		r4, r3, #96
+	add		r5, r3, #192
+	add		r6, r3, #144
+	vld1.8		{d8-d9}, [r4, : 128]!
+	add		r7, r3, #240
+	vld1.8		{d10-d11}, [r5, : 128]!
+	veor		q6, q4, q5
+	vld1.8		{d14-d15}, [r6, : 128]!
+	vdup.i32	q8, r2
+	vld1.8		{d18-d19}, [r7, : 128]!
+	veor		q10, q7, q9
+	vld1.8		{d22-d23}, [r4, : 128]!
+	vand		q6, q6, q8
+	vld1.8		{d24-d25}, [r5, : 128]!
+	vand		q10, q10, q8
+	vld1.8		{d26-d27}, [r6, : 128]!
+	veor		q4, q4, q6
+	vld1.8		{d28-d29}, [r7, : 128]!
+	veor		q5, q5, q6
+	vld1.8		{d0}, [r4, : 64]
+	veor		q6, q7, q10
+	vld1.8		{d2}, [r5, : 64]
+	veor		q7, q9, q10
+	vld1.8		{d4}, [r6, : 64]
+	veor		q9, q11, q12
+	vld1.8		{d6}, [r7, : 64]
+	veor		q10, q0, q1
+	sub		r2, r4, #32
+	vand		q9, q9, q8
+	sub		r4, r5, #32
+	vand		q10, q10, q8
+	sub		r5, r6, #32
+	veor		q11, q11, q9
+	sub		r6, r7, #32
+	veor		q0, q0, q10
+	veor		q9, q12, q9
+	veor		q1, q1, q10
+	veor		q10, q13, q14
+	veor		q12, q2, q3
+	vand		q10, q10, q8
+	vand		q8, q12, q8
+	veor		q12, q13, q10
+	veor		q2, q2, q8
+	veor		q10, q14, q10
+	veor		q3, q3, q8
+	vadd.i32	q8, q4, q6
+	vsub.i32	q4, q4, q6
+	vst1.8		{d16-d17}, [r2, : 128]!
+	vadd.i32	q6, q11, q12
+	vst1.8		{d8-d9}, [r5, : 128]!
+	vsub.i32	q4, q11, q12
+	vst1.8		{d12-d13}, [r2, : 128]!
+	vadd.i32	q6, q0, q2
+	vst1.8		{d8-d9}, [r5, : 128]!
+	vsub.i32	q0, q0, q2
+	vst1.8		d12, [r2, : 64]
+	vadd.i32	q2, q5, q7
+	vst1.8		d0, [r5, : 64]
+	vsub.i32	q0, q5, q7
+	vst1.8		{d4-d5}, [r4, : 128]!
+	vadd.i32	q2, q9, q10
+	vst1.8		{d0-d1}, [r6, : 128]!
+	vsub.i32	q0, q9, q10
+	vst1.8		{d4-d5}, [r4, : 128]!
+	vadd.i32	q2, q1, q3
+	vst1.8		{d0-d1}, [r6, : 128]!
+	vsub.i32	q0, q1, q3
+	vst1.8		d4, [r4, : 64]
+	vst1.8		d0, [r6, : 64]
+	add		r2, sp, #544
+	add		r4, r3, #96
+	add		r5, r3, #144
+	vld1.8		{d0-d1}, [r2, : 128]
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vld1.8		{d4-d5}, [r5, : 128]!
+	vzip.i32	q1, q2
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vld1.8		{d8-d9}, [r5, : 128]!
+	vshl.i32	q5, q1, #1
+	vzip.i32	q3, q4
+	vshl.i32	q6, q2, #1
+	vld1.8		{d14}, [r4, : 64]
+	vshl.i32	q8, q3, #1
+	vld1.8		{d15}, [r5, : 64]
+	vshl.i32	q9, q4, #1
+	vmul.i32	d21, d7, d1
+	vtrn.32		d14, d15
+	vmul.i32	q11, q4, q0
+	vmul.i32	q0, q7, q0
+	vmull.s32	q12, d2, d2
+	vmlal.s32	q12, d11, d1
+	vmlal.s32	q12, d12, d0
+	vmlal.s32	q12, d13, d23
+	vmlal.s32	q12, d16, d22
+	vmlal.s32	q12, d7, d21
+	vmull.s32	q10, d2, d11
+	vmlal.s32	q10, d4, d1
+	vmlal.s32	q10, d13, d0
+	vmlal.s32	q10, d6, d23
+	vmlal.s32	q10, d17, d22
+	vmull.s32	q13, d10, d4
+	vmlal.s32	q13, d11, d3
+	vmlal.s32	q13, d13, d1
+	vmlal.s32	q13, d16, d0
+	vmlal.s32	q13, d17, d23
+	vmlal.s32	q13, d8, d22
+	vmull.s32	q1, d10, d5
+	vmlal.s32	q1, d11, d4
+	vmlal.s32	q1, d6, d1
+	vmlal.s32	q1, d17, d0
+	vmlal.s32	q1, d8, d23
+	vmull.s32	q14, d10, d6
+	vmlal.s32	q14, d11, d13
+	vmlal.s32	q14, d4, d4
+	vmlal.s32	q14, d17, d1
+	vmlal.s32	q14, d18, d0
+	vmlal.s32	q14, d9, d23
+	vmull.s32	q11, d10, d7
+	vmlal.s32	q11, d11, d6
+	vmlal.s32	q11, d12, d5
+	vmlal.s32	q11, d8, d1
+	vmlal.s32	q11, d19, d0
+	vmull.s32	q15, d10, d8
+	vmlal.s32	q15, d11, d17
+	vmlal.s32	q15, d12, d6
+	vmlal.s32	q15, d13, d5
+	vmlal.s32	q15, d19, d1
+	vmlal.s32	q15, d14, d0
+	vmull.s32	q2, d10, d9
+	vmlal.s32	q2, d11, d8
+	vmlal.s32	q2, d12, d7
+	vmlal.s32	q2, d13, d6
+	vmlal.s32	q2, d14, d1
+	vmull.s32	q0, d15, d1
+	vmlal.s32	q0, d10, d14
+	vmlal.s32	q0, d11, d19
+	vmlal.s32	q0, d12, d8
+	vmlal.s32	q0, d13, d17
+	vmlal.s32	q0, d6, d6
+	add		r2, sp, #512
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmull.s32	q3, d16, d7
+	vmlal.s32	q3, d10, d15
+	vmlal.s32	q3, d11, d14
+	vmlal.s32	q3, d12, d9
+	vmlal.s32	q3, d13, d8
+	add		r2, sp, #528
+	vld1.8		{d8-d9}, [r2, : 128]
+	vadd.i64	q5, q12, q9
+	vadd.i64	q6, q15, q9
+	vshr.s64	q5, q5, #26
+	vshr.s64	q6, q6, #26
+	vadd.i64	q7, q10, q5
+	vshl.i64	q5, q5, #26
+	vadd.i64	q8, q7, q4
+	vadd.i64	q2, q2, q6
+	vshl.i64	q6, q6, #26
+	vadd.i64	q10, q2, q4
+	vsub.i64	q5, q12, q5
+	vshr.s64	q8, q8, #25
+	vsub.i64	q6, q15, q6
+	vshr.s64	q10, q10, #25
+	vadd.i64	q12, q13, q8
+	vshl.i64	q8, q8, #25
+	vadd.i64	q13, q12, q9
+	vadd.i64	q0, q0, q10
+	vsub.i64	q7, q7, q8
+	vshr.s64	q8, q13, #26
+	vshl.i64	q10, q10, #25
+	vadd.i64	q13, q0, q9
+	vadd.i64	q1, q1, q8
+	vshl.i64	q8, q8, #26
+	vadd.i64	q15, q1, q4
+	vsub.i64	q2, q2, q10
+	vshr.s64	q10, q13, #26
+	vsub.i64	q8, q12, q8
+	vshr.s64	q12, q15, #25
+	vadd.i64	q3, q3, q10
+	vshl.i64	q10, q10, #26
+	vadd.i64	q13, q3, q4
+	vadd.i64	q14, q14, q12
+	add		r2, r3, #288
+	vshl.i64	q12, q12, #25
+	add		r4, r3, #336
+	vadd.i64	q15, q14, q9
+	add		r2, r2, #8
+	vsub.i64	q0, q0, q10
+	add		r4, r4, #8
+	vshr.s64	q10, q13, #25
+	vsub.i64	q1, q1, q12
+	vshr.s64	q12, q15, #26
+	vadd.i64	q13, q10, q10
+	vadd.i64	q11, q11, q12
+	vtrn.32		d16, d2
+	vshl.i64	q12, q12, #26
+	vtrn.32		d17, d3
+	vadd.i64	q1, q11, q4
+	vadd.i64	q4, q5, q13
+	vst1.8		d16, [r2, : 64]!
+	vshl.i64	q5, q10, #4
+	vst1.8		d17, [r4, : 64]!
+	vsub.i64	q8, q14, q12
+	vshr.s64	q1, q1, #25
+	vadd.i64	q4, q4, q5
+	vadd.i64	q5, q6, q1
+	vshl.i64	q1, q1, #25
+	vadd.i64	q6, q5, q9
+	vadd.i64	q4, q4, q10
+	vshl.i64	q10, q10, #25
+	vadd.i64	q9, q4, q9
+	vsub.i64	q1, q11, q1
+	vshr.s64	q6, q6, #26
+	vsub.i64	q3, q3, q10
+	vtrn.32		d16, d2
+	vshr.s64	q9, q9, #26
+	vtrn.32		d17, d3
+	vadd.i64	q1, q2, q6
+	vst1.8		d16, [r2, : 64]
+	vshl.i64	q2, q6, #26
+	vst1.8		d17, [r4, : 64]
+	vadd.i64	q6, q7, q9
+	vtrn.32		d0, d6
+	vshl.i64	q7, q9, #26
+	vtrn.32		d1, d7
+	vsub.i64	q2, q5, q2
+	add		r2, r2, #16
+	vsub.i64	q3, q4, q7
+	vst1.8		d0, [r2, : 64]
+	add		r4, r4, #16
+	vst1.8		d1, [r4, : 64]
+	vtrn.32		d4, d2
+	vtrn.32		d5, d3
+	sub		r2, r2, #8
+	sub		r4, r4, #8
+	vtrn.32		d6, d12
+	vtrn.32		d7, d13
+	vst1.8		d4, [r2, : 64]
+	vst1.8		d5, [r4, : 64]
+	sub		r2, r2, #24
+	sub		r4, r4, #24
+	vst1.8		d6, [r2, : 64]
+	vst1.8		d7, [r4, : 64]
+	add		r2, r3, #240
+	add		r4, r3, #96
+	vld1.8		{d0-d1}, [r4, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vld1.8		{d4}, [r4, : 64]
+	add		r4, r3, #144
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vtrn.32		q0, q3
+	vld1.8		{d8-d9}, [r4, : 128]!
+	vshl.i32	q5, q0, #4
+	vtrn.32		q1, q4
+	vshl.i32	q6, q3, #4
+	vadd.i32	q5, q5, q0
+	vadd.i32	q6, q6, q3
+	vshl.i32	q7, q1, #4
+	vld1.8		{d5}, [r4, : 64]
+	vshl.i32	q8, q4, #4
+	vtrn.32		d4, d5
+	vadd.i32	q7, q7, q1
+	vadd.i32	q8, q8, q4
+	vld1.8		{d18-d19}, [r2, : 128]!
+	vshl.i32	q10, q2, #4
+	vld1.8		{d22-d23}, [r2, : 128]!
+	vadd.i32	q10, q10, q2
+	vld1.8		{d24}, [r2, : 64]
+	vadd.i32	q5, q5, q0
+	add		r2, r3, #192
+	vld1.8		{d26-d27}, [r2, : 128]!
+	vadd.i32	q6, q6, q3
+	vld1.8		{d28-d29}, [r2, : 128]!
+	vadd.i32	q8, q8, q4
+	vld1.8		{d25}, [r2, : 64]
+	vadd.i32	q10, q10, q2
+	vtrn.32		q9, q13
+	vadd.i32	q7, q7, q1
+	vadd.i32	q5, q5, q0
+	vtrn.32		q11, q14
+	vadd.i32	q6, q6, q3
+	add		r2, sp, #560
+	vadd.i32	q10, q10, q2
+	vtrn.32		d24, d25
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q6, q13, #1
+	add		r2, sp, #576
+	vst1.8		{d20-d21}, [r2, : 128]
+	vshl.i32	q10, q14, #1
+	add		r2, sp, #592
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q15, q12, #1
+	vadd.i32	q8, q8, q4
+	vext.32		d10, d31, d30, #0
+	vadd.i32	q7, q7, q1
+	add		r2, sp, #608
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q8, d18, d5
+	vmlal.s32	q8, d26, d4
+	vmlal.s32	q8, d19, d9
+	vmlal.s32	q8, d27, d3
+	vmlal.s32	q8, d22, d8
+	vmlal.s32	q8, d28, d2
+	vmlal.s32	q8, d23, d7
+	vmlal.s32	q8, d29, d1
+	vmlal.s32	q8, d24, d6
+	vmlal.s32	q8, d25, d0
+	add		r2, sp, #624
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q2, d18, d4
+	vmlal.s32	q2, d12, d9
+	vmlal.s32	q2, d13, d8
+	vmlal.s32	q2, d19, d3
+	vmlal.s32	q2, d22, d2
+	vmlal.s32	q2, d23, d1
+	vmlal.s32	q2, d24, d0
+	add		r2, sp, #640
+	vst1.8		{d20-d21}, [r2, : 128]
+	vmull.s32	q7, d18, d9
+	vmlal.s32	q7, d26, d3
+	vmlal.s32	q7, d19, d8
+	vmlal.s32	q7, d27, d2
+	vmlal.s32	q7, d22, d7
+	vmlal.s32	q7, d28, d1
+	vmlal.s32	q7, d23, d6
+	vmlal.s32	q7, d29, d0
+	add		r2, sp, #656
+	vst1.8		{d10-d11}, [r2, : 128]
+	vmull.s32	q5, d18, d3
+	vmlal.s32	q5, d19, d2
+	vmlal.s32	q5, d22, d1
+	vmlal.s32	q5, d23, d0
+	vmlal.s32	q5, d12, d8
+	add		r2, sp, #672
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q4, d18, d8
+	vmlal.s32	q4, d26, d2
+	vmlal.s32	q4, d19, d7
+	vmlal.s32	q4, d27, d1
+	vmlal.s32	q4, d22, d6
+	vmlal.s32	q4, d28, d0
+	vmull.s32	q8, d18, d7
+	vmlal.s32	q8, d26, d1
+	vmlal.s32	q8, d19, d6
+	vmlal.s32	q8, d27, d0
+	add		r2, sp, #576
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q7, d24, d21
+	vmlal.s32	q7, d25, d20
+	vmlal.s32	q4, d23, d21
+	vmlal.s32	q4, d29, d20
+	vmlal.s32	q8, d22, d21
+	vmlal.s32	q8, d28, d20
+	vmlal.s32	q5, d24, d20
+	add		r2, sp, #576
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q7, d18, d6
+	vmlal.s32	q7, d26, d0
+	add		r2, sp, #656
+	vld1.8		{d30-d31}, [r2, : 128]
+	vmlal.s32	q2, d30, d21
+	vmlal.s32	q7, d19, d21
+	vmlal.s32	q7, d27, d20
+	add		r2, sp, #624
+	vld1.8		{d26-d27}, [r2, : 128]
+	vmlal.s32	q4, d25, d27
+	vmlal.s32	q8, d29, d27
+	vmlal.s32	q8, d25, d26
+	vmlal.s32	q7, d28, d27
+	vmlal.s32	q7, d29, d26
+	add		r2, sp, #608
+	vld1.8		{d28-d29}, [r2, : 128]
+	vmlal.s32	q4, d24, d29
+	vmlal.s32	q8, d23, d29
+	vmlal.s32	q8, d24, d28
+	vmlal.s32	q7, d22, d29
+	vmlal.s32	q7, d23, d28
+	add		r2, sp, #608
+	vst1.8		{d8-d9}, [r2, : 128]
+	add		r2, sp, #560
+	vld1.8		{d8-d9}, [r2, : 128]
+	vmlal.s32	q7, d24, d9
+	vmlal.s32	q7, d25, d31
+	vmull.s32	q1, d18, d2
+	vmlal.s32	q1, d19, d1
+	vmlal.s32	q1, d22, d0
+	vmlal.s32	q1, d24, d27
+	vmlal.s32	q1, d23, d20
+	vmlal.s32	q1, d12, d7
+	vmlal.s32	q1, d13, d6
+	vmull.s32	q6, d18, d1
+	vmlal.s32	q6, d19, d0
+	vmlal.s32	q6, d23, d27
+	vmlal.s32	q6, d22, d20
+	vmlal.s32	q6, d24, d26
+	vmull.s32	q0, d18, d0
+	vmlal.s32	q0, d22, d27
+	vmlal.s32	q0, d23, d26
+	vmlal.s32	q0, d24, d31
+	vmlal.s32	q0, d19, d20
+	add		r2, sp, #640
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q2, d18, d7
+	vmlal.s32	q2, d19, d6
+	vmlal.s32	q5, d18, d6
+	vmlal.s32	q5, d19, d21
+	vmlal.s32	q1, d18, d21
+	vmlal.s32	q1, d19, d29
+	vmlal.s32	q0, d18, d28
+	vmlal.s32	q0, d19, d9
+	vmlal.s32	q6, d18, d29
+	vmlal.s32	q6, d19, d28
+	add		r2, sp, #592
+	vld1.8		{d18-d19}, [r2, : 128]
+	add		r2, sp, #512
+	vld1.8		{d22-d23}, [r2, : 128]
+	vmlal.s32	q5, d19, d7
+	vmlal.s32	q0, d18, d21
+	vmlal.s32	q0, d19, d29
+	vmlal.s32	q6, d18, d6
+	add		r2, sp, #528
+	vld1.8		{d6-d7}, [r2, : 128]
+	vmlal.s32	q6, d19, d21
+	add		r2, sp, #576
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q0, d30, d8
+	add		r2, sp, #672
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q5, d30, d29
+	add		r2, sp, #608
+	vld1.8		{d24-d25}, [r2, : 128]
+	vmlal.s32	q1, d30, d28
+	vadd.i64	q13, q0, q11
+	vadd.i64	q14, q5, q11
+	vmlal.s32	q6, d30, d9
+	vshr.s64	q4, q13, #26
+	vshr.s64	q13, q14, #26
+	vadd.i64	q7, q7, q4
+	vshl.i64	q4, q4, #26
+	vadd.i64	q14, q7, q3
+	vadd.i64	q9, q9, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q15, q9, q3
+	vsub.i64	q0, q0, q4
+	vshr.s64	q4, q14, #25
+	vsub.i64	q5, q5, q13
+	vshr.s64	q13, q15, #25
+	vadd.i64	q6, q6, q4
+	vshl.i64	q4, q4, #25
+	vadd.i64	q14, q6, q11
+	vadd.i64	q2, q2, q13
+	vsub.i64	q4, q7, q4
+	vshr.s64	q7, q14, #26
+	vshl.i64	q13, q13, #25
+	vadd.i64	q14, q2, q11
+	vadd.i64	q8, q8, q7
+	vshl.i64	q7, q7, #26
+	vadd.i64	q15, q8, q3
+	vsub.i64	q9, q9, q13
+	vshr.s64	q13, q14, #26
+	vsub.i64	q6, q6, q7
+	vshr.s64	q7, q15, #25
+	vadd.i64	q10, q10, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q14, q10, q3
+	vadd.i64	q1, q1, q7
+	add		r2, r3, #144
+	vshl.i64	q7, q7, #25
+	add		r4, r3, #96
+	vadd.i64	q15, q1, q11
+	add		r2, r2, #8
+	vsub.i64	q2, q2, q13
+	add		r4, r4, #8
+	vshr.s64	q13, q14, #25
+	vsub.i64	q7, q8, q7
+	vshr.s64	q8, q15, #26
+	vadd.i64	q14, q13, q13
+	vadd.i64	q12, q12, q8
+	vtrn.32		d12, d14
+	vshl.i64	q8, q8, #26
+	vtrn.32		d13, d15
+	vadd.i64	q3, q12, q3
+	vadd.i64	q0, q0, q14
+	vst1.8		d12, [r2, : 64]!
+	vshl.i64	q7, q13, #4
+	vst1.8		d13, [r4, : 64]!
+	vsub.i64	q1, q1, q8
+	vshr.s64	q3, q3, #25
+	vadd.i64	q0, q0, q7
+	vadd.i64	q5, q5, q3
+	vshl.i64	q3, q3, #25
+	vadd.i64	q6, q5, q11
+	vadd.i64	q0, q0, q13
+	vshl.i64	q7, q13, #25
+	vadd.i64	q8, q0, q11
+	vsub.i64	q3, q12, q3
+	vshr.s64	q6, q6, #26
+	vsub.i64	q7, q10, q7
+	vtrn.32		d2, d6
+	vshr.s64	q8, q8, #26
+	vtrn.32		d3, d7
+	vadd.i64	q3, q9, q6
+	vst1.8		d2, [r2, : 64]
+	vshl.i64	q6, q6, #26
+	vst1.8		d3, [r4, : 64]
+	vadd.i64	q1, q4, q8
+	vtrn.32		d4, d14
+	vshl.i64	q4, q8, #26
+	vtrn.32		d5, d15
+	vsub.i64	q5, q5, q6
+	add		r2, r2, #16
+	vsub.i64	q0, q0, q4
+	vst1.8		d4, [r2, : 64]
+	add		r4, r4, #16
+	vst1.8		d5, [r4, : 64]
+	vtrn.32		d10, d6
+	vtrn.32		d11, d7
+	sub		r2, r2, #8
+	sub		r4, r4, #8
+	vtrn.32		d0, d2
+	vtrn.32		d1, d3
+	vst1.8		d10, [r2, : 64]
+	vst1.8		d11, [r4, : 64]
+	sub		r2, r2, #24
+	sub		r4, r4, #24
+	vst1.8		d0, [r2, : 64]
+	vst1.8		d1, [r4, : 64]
+	add		r2, r3, #288
+	add		r4, r3, #336
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vsub.i32	q0, q0, q1
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d4-d5}, [r4, : 128]!
+	vsub.i32	q1, q1, q2
+	add		r5, r3, #240
+	vld1.8		{d4}, [r2, : 64]
+	vld1.8		{d6}, [r4, : 64]
+	vsub.i32	q2, q2, q3
+	vst1.8		{d0-d1}, [r5, : 128]!
+	vst1.8		{d2-d3}, [r5, : 128]!
+	vst1.8		d4, [r5, : 64]
+	add		r2, r3, #144
+	add		r4, r3, #96
+	add		r5, r3, #144
+	add		r6, r3, #192
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vsub.i32	q2, q0, q1
+	vadd.i32	q0, q0, q1
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vsub.i32	q4, q1, q3
+	vadd.i32	q1, q1, q3
+	vld1.8		{d6}, [r2, : 64]
+	vld1.8		{d10}, [r4, : 64]
+	vsub.i32	q6, q3, q5
+	vadd.i32	q3, q3, q5
+	vst1.8		{d4-d5}, [r5, : 128]!
+	vst1.8		{d0-d1}, [r6, : 128]!
+	vst1.8		{d8-d9}, [r5, : 128]!
+	vst1.8		{d2-d3}, [r6, : 128]!
+	vst1.8		d12, [r5, : 64]
+	vst1.8		d6, [r6, : 64]
+	add		r2, r3, #0
+	add		r4, r3, #240
+	vld1.8		{d0-d1}, [r4, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vld1.8		{d4}, [r4, : 64]
+	add		r4, r3, #336
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vtrn.32		q0, q3
+	vld1.8		{d8-d9}, [r4, : 128]!
+	vshl.i32	q5, q0, #4
+	vtrn.32		q1, q4
+	vshl.i32	q6, q3, #4
+	vadd.i32	q5, q5, q0
+	vadd.i32	q6, q6, q3
+	vshl.i32	q7, q1, #4
+	vld1.8		{d5}, [r4, : 64]
+	vshl.i32	q8, q4, #4
+	vtrn.32		d4, d5
+	vadd.i32	q7, q7, q1
+	vadd.i32	q8, q8, q4
+	vld1.8		{d18-d19}, [r2, : 128]!
+	vshl.i32	q10, q2, #4
+	vld1.8		{d22-d23}, [r2, : 128]!
+	vadd.i32	q10, q10, q2
+	vld1.8		{d24}, [r2, : 64]
+	vadd.i32	q5, q5, q0
+	add		r2, r3, #288
+	vld1.8		{d26-d27}, [r2, : 128]!
+	vadd.i32	q6, q6, q3
+	vld1.8		{d28-d29}, [r2, : 128]!
+	vadd.i32	q8, q8, q4
+	vld1.8		{d25}, [r2, : 64]
+	vadd.i32	q10, q10, q2
+	vtrn.32		q9, q13
+	vadd.i32	q7, q7, q1
+	vadd.i32	q5, q5, q0
+	vtrn.32		q11, q14
+	vadd.i32	q6, q6, q3
+	add		r2, sp, #560
+	vadd.i32	q10, q10, q2
+	vtrn.32		d24, d25
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q6, q13, #1
+	add		r2, sp, #576
+	vst1.8		{d20-d21}, [r2, : 128]
+	vshl.i32	q10, q14, #1
+	add		r2, sp, #592
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q15, q12, #1
+	vadd.i32	q8, q8, q4
+	vext.32		d10, d31, d30, #0
+	vadd.i32	q7, q7, q1
+	add		r2, sp, #608
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q8, d18, d5
+	vmlal.s32	q8, d26, d4
+	vmlal.s32	q8, d19, d9
+	vmlal.s32	q8, d27, d3
+	vmlal.s32	q8, d22, d8
+	vmlal.s32	q8, d28, d2
+	vmlal.s32	q8, d23, d7
+	vmlal.s32	q8, d29, d1
+	vmlal.s32	q8, d24, d6
+	vmlal.s32	q8, d25, d0
+	add		r2, sp, #624
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q2, d18, d4
+	vmlal.s32	q2, d12, d9
+	vmlal.s32	q2, d13, d8
+	vmlal.s32	q2, d19, d3
+	vmlal.s32	q2, d22, d2
+	vmlal.s32	q2, d23, d1
+	vmlal.s32	q2, d24, d0
+	add		r2, sp, #640
+	vst1.8		{d20-d21}, [r2, : 128]
+	vmull.s32	q7, d18, d9
+	vmlal.s32	q7, d26, d3
+	vmlal.s32	q7, d19, d8
+	vmlal.s32	q7, d27, d2
+	vmlal.s32	q7, d22, d7
+	vmlal.s32	q7, d28, d1
+	vmlal.s32	q7, d23, d6
+	vmlal.s32	q7, d29, d0
+	add		r2, sp, #656
+	vst1.8		{d10-d11}, [r2, : 128]
+	vmull.s32	q5, d18, d3
+	vmlal.s32	q5, d19, d2
+	vmlal.s32	q5, d22, d1
+	vmlal.s32	q5, d23, d0
+	vmlal.s32	q5, d12, d8
+	add		r2, sp, #672
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q4, d18, d8
+	vmlal.s32	q4, d26, d2
+	vmlal.s32	q4, d19, d7
+	vmlal.s32	q4, d27, d1
+	vmlal.s32	q4, d22, d6
+	vmlal.s32	q4, d28, d0
+	vmull.s32	q8, d18, d7
+	vmlal.s32	q8, d26, d1
+	vmlal.s32	q8, d19, d6
+	vmlal.s32	q8, d27, d0
+	add		r2, sp, #576
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q7, d24, d21
+	vmlal.s32	q7, d25, d20
+	vmlal.s32	q4, d23, d21
+	vmlal.s32	q4, d29, d20
+	vmlal.s32	q8, d22, d21
+	vmlal.s32	q8, d28, d20
+	vmlal.s32	q5, d24, d20
+	add		r2, sp, #576
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q7, d18, d6
+	vmlal.s32	q7, d26, d0
+	add		r2, sp, #656
+	vld1.8		{d30-d31}, [r2, : 128]
+	vmlal.s32	q2, d30, d21
+	vmlal.s32	q7, d19, d21
+	vmlal.s32	q7, d27, d20
+	add		r2, sp, #624
+	vld1.8		{d26-d27}, [r2, : 128]
+	vmlal.s32	q4, d25, d27
+	vmlal.s32	q8, d29, d27
+	vmlal.s32	q8, d25, d26
+	vmlal.s32	q7, d28, d27
+	vmlal.s32	q7, d29, d26
+	add		r2, sp, #608
+	vld1.8		{d28-d29}, [r2, : 128]
+	vmlal.s32	q4, d24, d29
+	vmlal.s32	q8, d23, d29
+	vmlal.s32	q8, d24, d28
+	vmlal.s32	q7, d22, d29
+	vmlal.s32	q7, d23, d28
+	add		r2, sp, #608
+	vst1.8		{d8-d9}, [r2, : 128]
+	add		r2, sp, #560
+	vld1.8		{d8-d9}, [r2, : 128]
+	vmlal.s32	q7, d24, d9
+	vmlal.s32	q7, d25, d31
+	vmull.s32	q1, d18, d2
+	vmlal.s32	q1, d19, d1
+	vmlal.s32	q1, d22, d0
+	vmlal.s32	q1, d24, d27
+	vmlal.s32	q1, d23, d20
+	vmlal.s32	q1, d12, d7
+	vmlal.s32	q1, d13, d6
+	vmull.s32	q6, d18, d1
+	vmlal.s32	q6, d19, d0
+	vmlal.s32	q6, d23, d27
+	vmlal.s32	q6, d22, d20
+	vmlal.s32	q6, d24, d26
+	vmull.s32	q0, d18, d0
+	vmlal.s32	q0, d22, d27
+	vmlal.s32	q0, d23, d26
+	vmlal.s32	q0, d24, d31
+	vmlal.s32	q0, d19, d20
+	add		r2, sp, #640
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q2, d18, d7
+	vmlal.s32	q2, d19, d6
+	vmlal.s32	q5, d18, d6
+	vmlal.s32	q5, d19, d21
+	vmlal.s32	q1, d18, d21
+	vmlal.s32	q1, d19, d29
+	vmlal.s32	q0, d18, d28
+	vmlal.s32	q0, d19, d9
+	vmlal.s32	q6, d18, d29
+	vmlal.s32	q6, d19, d28
+	add		r2, sp, #592
+	vld1.8		{d18-d19}, [r2, : 128]
+	add		r2, sp, #512
+	vld1.8		{d22-d23}, [r2, : 128]
+	vmlal.s32	q5, d19, d7
+	vmlal.s32	q0, d18, d21
+	vmlal.s32	q0, d19, d29
+	vmlal.s32	q6, d18, d6
+	add		r2, sp, #528
+	vld1.8		{d6-d7}, [r2, : 128]
+	vmlal.s32	q6, d19, d21
+	add		r2, sp, #576
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q0, d30, d8
+	add		r2, sp, #672
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q5, d30, d29
+	add		r2, sp, #608
+	vld1.8		{d24-d25}, [r2, : 128]
+	vmlal.s32	q1, d30, d28
+	vadd.i64	q13, q0, q11
+	vadd.i64	q14, q5, q11
+	vmlal.s32	q6, d30, d9
+	vshr.s64	q4, q13, #26
+	vshr.s64	q13, q14, #26
+	vadd.i64	q7, q7, q4
+	vshl.i64	q4, q4, #26
+	vadd.i64	q14, q7, q3
+	vadd.i64	q9, q9, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q15, q9, q3
+	vsub.i64	q0, q0, q4
+	vshr.s64	q4, q14, #25
+	vsub.i64	q5, q5, q13
+	vshr.s64	q13, q15, #25
+	vadd.i64	q6, q6, q4
+	vshl.i64	q4, q4, #25
+	vadd.i64	q14, q6, q11
+	vadd.i64	q2, q2, q13
+	vsub.i64	q4, q7, q4
+	vshr.s64	q7, q14, #26
+	vshl.i64	q13, q13, #25
+	vadd.i64	q14, q2, q11
+	vadd.i64	q8, q8, q7
+	vshl.i64	q7, q7, #26
+	vadd.i64	q15, q8, q3
+	vsub.i64	q9, q9, q13
+	vshr.s64	q13, q14, #26
+	vsub.i64	q6, q6, q7
+	vshr.s64	q7, q15, #25
+	vadd.i64	q10, q10, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q14, q10, q3
+	vadd.i64	q1, q1, q7
+	add		r2, r3, #288
+	vshl.i64	q7, q7, #25
+	add		r4, r3, #96
+	vadd.i64	q15, q1, q11
+	add		r2, r2, #8
+	vsub.i64	q2, q2, q13
+	add		r4, r4, #8
+	vshr.s64	q13, q14, #25
+	vsub.i64	q7, q8, q7
+	vshr.s64	q8, q15, #26
+	vadd.i64	q14, q13, q13
+	vadd.i64	q12, q12, q8
+	vtrn.32		d12, d14
+	vshl.i64	q8, q8, #26
+	vtrn.32		d13, d15
+	vadd.i64	q3, q12, q3
+	vadd.i64	q0, q0, q14
+	vst1.8		d12, [r2, : 64]!
+	vshl.i64	q7, q13, #4
+	vst1.8		d13, [r4, : 64]!
+	vsub.i64	q1, q1, q8
+	vshr.s64	q3, q3, #25
+	vadd.i64	q0, q0, q7
+	vadd.i64	q5, q5, q3
+	vshl.i64	q3, q3, #25
+	vadd.i64	q6, q5, q11
+	vadd.i64	q0, q0, q13
+	vshl.i64	q7, q13, #25
+	vadd.i64	q8, q0, q11
+	vsub.i64	q3, q12, q3
+	vshr.s64	q6, q6, #26
+	vsub.i64	q7, q10, q7
+	vtrn.32		d2, d6
+	vshr.s64	q8, q8, #26
+	vtrn.32		d3, d7
+	vadd.i64	q3, q9, q6
+	vst1.8		d2, [r2, : 64]
+	vshl.i64	q6, q6, #26
+	vst1.8		d3, [r4, : 64]
+	vadd.i64	q1, q4, q8
+	vtrn.32		d4, d14
+	vshl.i64	q4, q8, #26
+	vtrn.32		d5, d15
+	vsub.i64	q5, q5, q6
+	add		r2, r2, #16
+	vsub.i64	q0, q0, q4
+	vst1.8		d4, [r2, : 64]
+	add		r4, r4, #16
+	vst1.8		d5, [r4, : 64]
+	vtrn.32		d10, d6
+	vtrn.32		d11, d7
+	sub		r2, r2, #8
+	sub		r4, r4, #8
+	vtrn.32		d0, d2
+	vtrn.32		d1, d3
+	vst1.8		d10, [r2, : 64]
+	vst1.8		d11, [r4, : 64]
+	sub		r2, r2, #24
+	sub		r4, r4, #24
+	vst1.8		d0, [r2, : 64]
+	vst1.8		d1, [r4, : 64]
+	add		r2, sp, #544
+	add		r4, r3, #144
+	add		r5, r3, #192
+	vld1.8		{d0-d1}, [r2, : 128]
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vld1.8		{d4-d5}, [r5, : 128]!
+	vzip.i32	q1, q2
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vld1.8		{d8-d9}, [r5, : 128]!
+	vshl.i32	q5, q1, #1
+	vzip.i32	q3, q4
+	vshl.i32	q6, q2, #1
+	vld1.8		{d14}, [r4, : 64]
+	vshl.i32	q8, q3, #1
+	vld1.8		{d15}, [r5, : 64]
+	vshl.i32	q9, q4, #1
+	vmul.i32	d21, d7, d1
+	vtrn.32		d14, d15
+	vmul.i32	q11, q4, q0
+	vmul.i32	q0, q7, q0
+	vmull.s32	q12, d2, d2
+	vmlal.s32	q12, d11, d1
+	vmlal.s32	q12, d12, d0
+	vmlal.s32	q12, d13, d23
+	vmlal.s32	q12, d16, d22
+	vmlal.s32	q12, d7, d21
+	vmull.s32	q10, d2, d11
+	vmlal.s32	q10, d4, d1
+	vmlal.s32	q10, d13, d0
+	vmlal.s32	q10, d6, d23
+	vmlal.s32	q10, d17, d22
+	vmull.s32	q13, d10, d4
+	vmlal.s32	q13, d11, d3
+	vmlal.s32	q13, d13, d1
+	vmlal.s32	q13, d16, d0
+	vmlal.s32	q13, d17, d23
+	vmlal.s32	q13, d8, d22
+	vmull.s32	q1, d10, d5
+	vmlal.s32	q1, d11, d4
+	vmlal.s32	q1, d6, d1
+	vmlal.s32	q1, d17, d0
+	vmlal.s32	q1, d8, d23
+	vmull.s32	q14, d10, d6
+	vmlal.s32	q14, d11, d13
+	vmlal.s32	q14, d4, d4
+	vmlal.s32	q14, d17, d1
+	vmlal.s32	q14, d18, d0
+	vmlal.s32	q14, d9, d23
+	vmull.s32	q11, d10, d7
+	vmlal.s32	q11, d11, d6
+	vmlal.s32	q11, d12, d5
+	vmlal.s32	q11, d8, d1
+	vmlal.s32	q11, d19, d0
+	vmull.s32	q15, d10, d8
+	vmlal.s32	q15, d11, d17
+	vmlal.s32	q15, d12, d6
+	vmlal.s32	q15, d13, d5
+	vmlal.s32	q15, d19, d1
+	vmlal.s32	q15, d14, d0
+	vmull.s32	q2, d10, d9
+	vmlal.s32	q2, d11, d8
+	vmlal.s32	q2, d12, d7
+	vmlal.s32	q2, d13, d6
+	vmlal.s32	q2, d14, d1
+	vmull.s32	q0, d15, d1
+	vmlal.s32	q0, d10, d14
+	vmlal.s32	q0, d11, d19
+	vmlal.s32	q0, d12, d8
+	vmlal.s32	q0, d13, d17
+	vmlal.s32	q0, d6, d6
+	add		r2, sp, #512
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmull.s32	q3, d16, d7
+	vmlal.s32	q3, d10, d15
+	vmlal.s32	q3, d11, d14
+	vmlal.s32	q3, d12, d9
+	vmlal.s32	q3, d13, d8
+	add		r2, sp, #528
+	vld1.8		{d8-d9}, [r2, : 128]
+	vadd.i64	q5, q12, q9
+	vadd.i64	q6, q15, q9
+	vshr.s64	q5, q5, #26
+	vshr.s64	q6, q6, #26
+	vadd.i64	q7, q10, q5
+	vshl.i64	q5, q5, #26
+	vadd.i64	q8, q7, q4
+	vadd.i64	q2, q2, q6
+	vshl.i64	q6, q6, #26
+	vadd.i64	q10, q2, q4
+	vsub.i64	q5, q12, q5
+	vshr.s64	q8, q8, #25
+	vsub.i64	q6, q15, q6
+	vshr.s64	q10, q10, #25
+	vadd.i64	q12, q13, q8
+	vshl.i64	q8, q8, #25
+	vadd.i64	q13, q12, q9
+	vadd.i64	q0, q0, q10
+	vsub.i64	q7, q7, q8
+	vshr.s64	q8, q13, #26
+	vshl.i64	q10, q10, #25
+	vadd.i64	q13, q0, q9
+	vadd.i64	q1, q1, q8
+	vshl.i64	q8, q8, #26
+	vadd.i64	q15, q1, q4
+	vsub.i64	q2, q2, q10
+	vshr.s64	q10, q13, #26
+	vsub.i64	q8, q12, q8
+	vshr.s64	q12, q15, #25
+	vadd.i64	q3, q3, q10
+	vshl.i64	q10, q10, #26
+	vadd.i64	q13, q3, q4
+	vadd.i64	q14, q14, q12
+	add		r2, r3, #144
+	vshl.i64	q12, q12, #25
+	add		r4, r3, #192
+	vadd.i64	q15, q14, q9
+	add		r2, r2, #8
+	vsub.i64	q0, q0, q10
+	add		r4, r4, #8
+	vshr.s64	q10, q13, #25
+	vsub.i64	q1, q1, q12
+	vshr.s64	q12, q15, #26
+	vadd.i64	q13, q10, q10
+	vadd.i64	q11, q11, q12
+	vtrn.32		d16, d2
+	vshl.i64	q12, q12, #26
+	vtrn.32		d17, d3
+	vadd.i64	q1, q11, q4
+	vadd.i64	q4, q5, q13
+	vst1.8		d16, [r2, : 64]!
+	vshl.i64	q5, q10, #4
+	vst1.8		d17, [r4, : 64]!
+	vsub.i64	q8, q14, q12
+	vshr.s64	q1, q1, #25
+	vadd.i64	q4, q4, q5
+	vadd.i64	q5, q6, q1
+	vshl.i64	q1, q1, #25
+	vadd.i64	q6, q5, q9
+	vadd.i64	q4, q4, q10
+	vshl.i64	q10, q10, #25
+	vadd.i64	q9, q4, q9
+	vsub.i64	q1, q11, q1
+	vshr.s64	q6, q6, #26
+	vsub.i64	q3, q3, q10
+	vtrn.32		d16, d2
+	vshr.s64	q9, q9, #26
+	vtrn.32		d17, d3
+	vadd.i64	q1, q2, q6
+	vst1.8		d16, [r2, : 64]
+	vshl.i64	q2, q6, #26
+	vst1.8		d17, [r4, : 64]
+	vadd.i64	q6, q7, q9
+	vtrn.32		d0, d6
+	vshl.i64	q7, q9, #26
+	vtrn.32		d1, d7
+	vsub.i64	q2, q5, q2
+	add		r2, r2, #16
+	vsub.i64	q3, q4, q7
+	vst1.8		d0, [r2, : 64]
+	add		r4, r4, #16
+	vst1.8		d1, [r4, : 64]
+	vtrn.32		d4, d2
+	vtrn.32		d5, d3
+	sub		r2, r2, #8
+	sub		r4, r4, #8
+	vtrn.32		d6, d12
+	vtrn.32		d7, d13
+	vst1.8		d4, [r2, : 64]
+	vst1.8		d5, [r4, : 64]
+	sub		r2, r2, #24
+	sub		r4, r4, #24
+	vst1.8		d6, [r2, : 64]
+	vst1.8		d7, [r4, : 64]
+	add		r2, r3, #336
+	add		r4, r3, #288
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vadd.i32	q0, q0, q1
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d4-d5}, [r4, : 128]!
+	vadd.i32	q1, q1, q2
+	add		r5, r3, #288
+	vld1.8		{d4}, [r2, : 64]
+	vld1.8		{d6}, [r4, : 64]
+	vadd.i32	q2, q2, q3
+	vst1.8		{d0-d1}, [r5, : 128]!
+	vst1.8		{d2-d3}, [r5, : 128]!
+	vst1.8		d4, [r5, : 64]
+	add		r2, r3, #48
+	add		r4, r3, #144
+	vld1.8		{d0-d1}, [r4, : 128]!
+	vld1.8		{d2-d3}, [r4, : 128]!
+	vld1.8		{d4}, [r4, : 64]
+	add		r4, r3, #288
+	vld1.8		{d6-d7}, [r4, : 128]!
+	vtrn.32		q0, q3
+	vld1.8		{d8-d9}, [r4, : 128]!
+	vshl.i32	q5, q0, #4
+	vtrn.32		q1, q4
+	vshl.i32	q6, q3, #4
+	vadd.i32	q5, q5, q0
+	vadd.i32	q6, q6, q3
+	vshl.i32	q7, q1, #4
+	vld1.8		{d5}, [r4, : 64]
+	vshl.i32	q8, q4, #4
+	vtrn.32		d4, d5
+	vadd.i32	q7, q7, q1
+	vadd.i32	q8, q8, q4
+	vld1.8		{d18-d19}, [r2, : 128]!
+	vshl.i32	q10, q2, #4
+	vld1.8		{d22-d23}, [r2, : 128]!
+	vadd.i32	q10, q10, q2
+	vld1.8		{d24}, [r2, : 64]
+	vadd.i32	q5, q5, q0
+	add		r2, r3, #240
+	vld1.8		{d26-d27}, [r2, : 128]!
+	vadd.i32	q6, q6, q3
+	vld1.8		{d28-d29}, [r2, : 128]!
+	vadd.i32	q8, q8, q4
+	vld1.8		{d25}, [r2, : 64]
+	vadd.i32	q10, q10, q2
+	vtrn.32		q9, q13
+	vadd.i32	q7, q7, q1
+	vadd.i32	q5, q5, q0
+	vtrn.32		q11, q14
+	vadd.i32	q6, q6, q3
+	add		r2, sp, #560
+	vadd.i32	q10, q10, q2
+	vtrn.32		d24, d25
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q6, q13, #1
+	add		r2, sp, #576
+	vst1.8		{d20-d21}, [r2, : 128]
+	vshl.i32	q10, q14, #1
+	add		r2, sp, #592
+	vst1.8		{d12-d13}, [r2, : 128]
+	vshl.i32	q15, q12, #1
+	vadd.i32	q8, q8, q4
+	vext.32		d10, d31, d30, #0
+	vadd.i32	q7, q7, q1
+	add		r2, sp, #608
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q8, d18, d5
+	vmlal.s32	q8, d26, d4
+	vmlal.s32	q8, d19, d9
+	vmlal.s32	q8, d27, d3
+	vmlal.s32	q8, d22, d8
+	vmlal.s32	q8, d28, d2
+	vmlal.s32	q8, d23, d7
+	vmlal.s32	q8, d29, d1
+	vmlal.s32	q8, d24, d6
+	vmlal.s32	q8, d25, d0
+	add		r2, sp, #624
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q2, d18, d4
+	vmlal.s32	q2, d12, d9
+	vmlal.s32	q2, d13, d8
+	vmlal.s32	q2, d19, d3
+	vmlal.s32	q2, d22, d2
+	vmlal.s32	q2, d23, d1
+	vmlal.s32	q2, d24, d0
+	add		r2, sp, #640
+	vst1.8		{d20-d21}, [r2, : 128]
+	vmull.s32	q7, d18, d9
+	vmlal.s32	q7, d26, d3
+	vmlal.s32	q7, d19, d8
+	vmlal.s32	q7, d27, d2
+	vmlal.s32	q7, d22, d7
+	vmlal.s32	q7, d28, d1
+	vmlal.s32	q7, d23, d6
+	vmlal.s32	q7, d29, d0
+	add		r2, sp, #656
+	vst1.8		{d10-d11}, [r2, : 128]
+	vmull.s32	q5, d18, d3
+	vmlal.s32	q5, d19, d2
+	vmlal.s32	q5, d22, d1
+	vmlal.s32	q5, d23, d0
+	vmlal.s32	q5, d12, d8
+	add		r2, sp, #672
+	vst1.8		{d16-d17}, [r2, : 128]
+	vmull.s32	q4, d18, d8
+	vmlal.s32	q4, d26, d2
+	vmlal.s32	q4, d19, d7
+	vmlal.s32	q4, d27, d1
+	vmlal.s32	q4, d22, d6
+	vmlal.s32	q4, d28, d0
+	vmull.s32	q8, d18, d7
+	vmlal.s32	q8, d26, d1
+	vmlal.s32	q8, d19, d6
+	vmlal.s32	q8, d27, d0
+	add		r2, sp, #576
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q7, d24, d21
+	vmlal.s32	q7, d25, d20
+	vmlal.s32	q4, d23, d21
+	vmlal.s32	q4, d29, d20
+	vmlal.s32	q8, d22, d21
+	vmlal.s32	q8, d28, d20
+	vmlal.s32	q5, d24, d20
+	add		r2, sp, #576
+	vst1.8		{d14-d15}, [r2, : 128]
+	vmull.s32	q7, d18, d6
+	vmlal.s32	q7, d26, d0
+	add		r2, sp, #656
+	vld1.8		{d30-d31}, [r2, : 128]
+	vmlal.s32	q2, d30, d21
+	vmlal.s32	q7, d19, d21
+	vmlal.s32	q7, d27, d20
+	add		r2, sp, #624
+	vld1.8		{d26-d27}, [r2, : 128]
+	vmlal.s32	q4, d25, d27
+	vmlal.s32	q8, d29, d27
+	vmlal.s32	q8, d25, d26
+	vmlal.s32	q7, d28, d27
+	vmlal.s32	q7, d29, d26
+	add		r2, sp, #608
+	vld1.8		{d28-d29}, [r2, : 128]
+	vmlal.s32	q4, d24, d29
+	vmlal.s32	q8, d23, d29
+	vmlal.s32	q8, d24, d28
+	vmlal.s32	q7, d22, d29
+	vmlal.s32	q7, d23, d28
+	add		r2, sp, #608
+	vst1.8		{d8-d9}, [r2, : 128]
+	add		r2, sp, #560
+	vld1.8		{d8-d9}, [r2, : 128]
+	vmlal.s32	q7, d24, d9
+	vmlal.s32	q7, d25, d31
+	vmull.s32	q1, d18, d2
+	vmlal.s32	q1, d19, d1
+	vmlal.s32	q1, d22, d0
+	vmlal.s32	q1, d24, d27
+	vmlal.s32	q1, d23, d20
+	vmlal.s32	q1, d12, d7
+	vmlal.s32	q1, d13, d6
+	vmull.s32	q6, d18, d1
+	vmlal.s32	q6, d19, d0
+	vmlal.s32	q6, d23, d27
+	vmlal.s32	q6, d22, d20
+	vmlal.s32	q6, d24, d26
+	vmull.s32	q0, d18, d0
+	vmlal.s32	q0, d22, d27
+	vmlal.s32	q0, d23, d26
+	vmlal.s32	q0, d24, d31
+	vmlal.s32	q0, d19, d20
+	add		r2, sp, #640
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q2, d18, d7
+	vmlal.s32	q2, d19, d6
+	vmlal.s32	q5, d18, d6
+	vmlal.s32	q5, d19, d21
+	vmlal.s32	q1, d18, d21
+	vmlal.s32	q1, d19, d29
+	vmlal.s32	q0, d18, d28
+	vmlal.s32	q0, d19, d9
+	vmlal.s32	q6, d18, d29
+	vmlal.s32	q6, d19, d28
+	add		r2, sp, #592
+	vld1.8		{d18-d19}, [r2, : 128]
+	add		r2, sp, #512
+	vld1.8		{d22-d23}, [r2, : 128]
+	vmlal.s32	q5, d19, d7
+	vmlal.s32	q0, d18, d21
+	vmlal.s32	q0, d19, d29
+	vmlal.s32	q6, d18, d6
+	add		r2, sp, #528
+	vld1.8		{d6-d7}, [r2, : 128]
+	vmlal.s32	q6, d19, d21
+	add		r2, sp, #576
+	vld1.8		{d18-d19}, [r2, : 128]
+	vmlal.s32	q0, d30, d8
+	add		r2, sp, #672
+	vld1.8		{d20-d21}, [r2, : 128]
+	vmlal.s32	q5, d30, d29
+	add		r2, sp, #608
+	vld1.8		{d24-d25}, [r2, : 128]
+	vmlal.s32	q1, d30, d28
+	vadd.i64	q13, q0, q11
+	vadd.i64	q14, q5, q11
+	vmlal.s32	q6, d30, d9
+	vshr.s64	q4, q13, #26
+	vshr.s64	q13, q14, #26
+	vadd.i64	q7, q7, q4
+	vshl.i64	q4, q4, #26
+	vadd.i64	q14, q7, q3
+	vadd.i64	q9, q9, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q15, q9, q3
+	vsub.i64	q0, q0, q4
+	vshr.s64	q4, q14, #25
+	vsub.i64	q5, q5, q13
+	vshr.s64	q13, q15, #25
+	vadd.i64	q6, q6, q4
+	vshl.i64	q4, q4, #25
+	vadd.i64	q14, q6, q11
+	vadd.i64	q2, q2, q13
+	vsub.i64	q4, q7, q4
+	vshr.s64	q7, q14, #26
+	vshl.i64	q13, q13, #25
+	vadd.i64	q14, q2, q11
+	vadd.i64	q8, q8, q7
+	vshl.i64	q7, q7, #26
+	vadd.i64	q15, q8, q3
+	vsub.i64	q9, q9, q13
+	vshr.s64	q13, q14, #26
+	vsub.i64	q6, q6, q7
+	vshr.s64	q7, q15, #25
+	vadd.i64	q10, q10, q13
+	vshl.i64	q13, q13, #26
+	vadd.i64	q14, q10, q3
+	vadd.i64	q1, q1, q7
+	add		r2, r3, #240
+	vshl.i64	q7, q7, #25
+	add		r4, r3, #144
+	vadd.i64	q15, q1, q11
+	add		r2, r2, #8
+	vsub.i64	q2, q2, q13
+	add		r4, r4, #8
+	vshr.s64	q13, q14, #25
+	vsub.i64	q7, q8, q7
+	vshr.s64	q8, q15, #26
+	vadd.i64	q14, q13, q13
+	vadd.i64	q12, q12, q8
+	vtrn.32		d12, d14
+	vshl.i64	q8, q8, #26
+	vtrn.32		d13, d15
+	vadd.i64	q3, q12, q3
+	vadd.i64	q0, q0, q14
+	vst1.8		d12, [r2, : 64]!
+	vshl.i64	q7, q13, #4
+	vst1.8		d13, [r4, : 64]!
+	vsub.i64	q1, q1, q8
+	vshr.s64	q3, q3, #25
+	vadd.i64	q0, q0, q7
+	vadd.i64	q5, q5, q3
+	vshl.i64	q3, q3, #25
+	vadd.i64	q6, q5, q11
+	vadd.i64	q0, q0, q13
+	vshl.i64	q7, q13, #25
+	vadd.i64	q8, q0, q11
+	vsub.i64	q3, q12, q3
+	vshr.s64	q6, q6, #26
+	vsub.i64	q7, q10, q7
+	vtrn.32		d2, d6
+	vshr.s64	q8, q8, #26
+	vtrn.32		d3, d7
+	vadd.i64	q3, q9, q6
+	vst1.8		d2, [r2, : 64]
+	vshl.i64	q6, q6, #26
+	vst1.8		d3, [r4, : 64]
+	vadd.i64	q1, q4, q8
+	vtrn.32		d4, d14
+	vshl.i64	q4, q8, #26
+	vtrn.32		d5, d15
+	vsub.i64	q5, q5, q6
+	add		r2, r2, #16
+	vsub.i64	q0, q0, q4
+	vst1.8		d4, [r2, : 64]
+	add		r4, r4, #16
+	vst1.8		d5, [r4, : 64]
+	vtrn.32		d10, d6
+	vtrn.32		d11, d7
+	sub		r2, r2, #8
+	sub		r4, r4, #8
+	vtrn.32		d0, d2
+	vtrn.32		d1, d3
+	vst1.8		d10, [r2, : 64]
+	vst1.8		d11, [r4, : 64]
+	sub		r2, r2, #24
+	sub		r4, r4, #24
+	vst1.8		d0, [r2, : 64]
+	vst1.8		d1, [r4, : 64]
+	ldr		r2, [sp, #488]
+	ldr		r4, [sp, #492]
+	subs		r5, r2, #1
+	bge		._mainloop
+	add		r1, r3, #144
+	add		r2, r3, #336
+	vld1.8		{d0-d1}, [r1, : 128]!
+	vld1.8		{d2-d3}, [r1, : 128]!
+	vld1.8		{d4}, [r1, : 64]
+	vst1.8		{d0-d1}, [r2, : 128]!
+	vst1.8		{d2-d3}, [r2, : 128]!
+	vst1.8		d4, [r2, : 64]
+	ldr		r1, =0
+._invertloop:
+	add		r2, r3, #144
+	ldr		r4, =0
+	ldr		r5, =2
+	cmp		r1, #1
+	ldreq		r5, =1
+	addeq		r2, r3, #336
+	addeq		r4, r3, #48
+	cmp		r1, #2
+	ldreq		r5, =1
+	addeq		r2, r3, #48
+	cmp		r1, #3
+	ldreq		r5, =5
+	addeq		r4, r3, #336
+	cmp		r1, #4
+	ldreq		r5, =10
+	cmp		r1, #5
+	ldreq		r5, =20
+	cmp		r1, #6
+	ldreq		r5, =10
+	addeq		r2, r3, #336
+	addeq		r4, r3, #336
+	cmp		r1, #7
+	ldreq		r5, =50
+	cmp		r1, #8
+	ldreq		r5, =100
+	cmp		r1, #9
+	ldreq		r5, =50
+	addeq		r2, r3, #336
+	cmp		r1, #10
+	ldreq		r5, =5
+	addeq		r2, r3, #48
+	cmp		r1, #11
+	ldreq		r5, =0
+	addeq		r2, r3, #96
+	add		r6, r3, #144
+	add		r7, r3, #288
+	vld1.8		{d0-d1}, [r6, : 128]!
+	vld1.8		{d2-d3}, [r6, : 128]!
+	vld1.8		{d4}, [r6, : 64]
+	vst1.8		{d0-d1}, [r7, : 128]!
+	vst1.8		{d2-d3}, [r7, : 128]!
+	vst1.8		d4, [r7, : 64]
+	cmp		r5, #0
+	beq		._skipsquaringloop
+._squaringloop:
+	add		r6, r3, #288
+	add		r7, r3, #288
+	add		r8, r3, #288
+	vmov.i32	q0, #19
+	vmov.i32	q1, #0
+	vmov.i32	q2, #1
+	vzip.i32	q1, q2
+	vld1.8		{d4-d5}, [r7, : 128]!
+	vld1.8		{d6-d7}, [r7, : 128]!
+	vld1.8		{d9}, [r7, : 64]
+	vld1.8		{d10-d11}, [r6, : 128]!
+	add		r7, sp, #416
+	vld1.8		{d12-d13}, [r6, : 128]!
+	vmul.i32	q7, q2, q0
+	vld1.8		{d8}, [r6, : 64]
+	vext.32		d17, d11, d10, #1
+	vmul.i32	q9, q3, q0
+	vext.32		d16, d10, d8, #1
+	vshl.u32	q10, q5, q1
+	vext.32		d22, d14, d4, #1
+	vext.32		d24, d18, d6, #1
+	vshl.u32	q13, q6, q1
+	vshl.u32	d28, d8, d2
+	vrev64.i32	d22, d22
+	vmul.i32	d1, d9, d1
+	vrev64.i32	d24, d24
+	vext.32		d29, d8, d13, #1
+	vext.32		d0, d1, d9, #1
+	vrev64.i32	d0, d0
+	vext.32		d2, d9, d1, #1
+	vext.32		d23, d15, d5, #1
+	vmull.s32	q4, d20, d4
+	vrev64.i32	d23, d23
+	vmlal.s32	q4, d21, d1
+	vrev64.i32	d2, d2
+	vmlal.s32	q4, d26, d19
+	vext.32		d3, d5, d15, #1
+	vmlal.s32	q4, d27, d18
+	vrev64.i32	d3, d3
+	vmlal.s32	q4, d28, d15
+	vext.32		d14, d12, d11, #1
+	vmull.s32	q5, d16, d23
+	vext.32		d15, d13, d12, #1
+	vmlal.s32	q5, d17, d4
+	vst1.8		d8, [r7, : 64]!
+	vmlal.s32	q5, d14, d1
+	vext.32		d12, d9, d8, #0
+	vmlal.s32	q5, d15, d19
+	vmov.i64	d13, #0
+	vmlal.s32	q5, d29, d18
+	vext.32		d25, d19, d7, #1
+	vmlal.s32	q6, d20, d5
+	vrev64.i32	d25, d25
+	vmlal.s32	q6, d21, d4
+	vst1.8		d11, [r7, : 64]!
+	vmlal.s32	q6, d26, d1
+	vext.32		d9, d10, d10, #0
+	vmlal.s32	q6, d27, d19
+	vmov.i64	d8, #0
+	vmlal.s32	q6, d28, d18
+	vmlal.s32	q4, d16, d24
+	vmlal.s32	q4, d17, d5
+	vmlal.s32	q4, d14, d4
+	vst1.8		d12, [r7, : 64]!
+	vmlal.s32	q4, d15, d1
+	vext.32		d10, d13, d12, #0
+	vmlal.s32	q4, d29, d19
+	vmov.i64	d11, #0
+	vmlal.s32	q5, d20, d6
+	vmlal.s32	q5, d21, d5
+	vmlal.s32	q5, d26, d4
+	vext.32		d13, d8, d8, #0
+	vmlal.s32	q5, d27, d1
+	vmov.i64	d12, #0
+	vmlal.s32	q5, d28, d19
+	vst1.8		d9, [r7, : 64]!
+	vmlal.s32	q6, d16, d25
+	vmlal.s32	q6, d17, d6
+	vst1.8		d10, [r7, : 64]
+	vmlal.s32	q6, d14, d5
+	vext.32		d8, d11, d10, #0
+	vmlal.s32	q6, d15, d4
+	vmov.i64	d9, #0
+	vmlal.s32	q6, d29, d1
+	vmlal.s32	q4, d20, d7
+	vmlal.s32	q4, d21, d6
+	vmlal.s32	q4, d26, d5
+	vext.32		d11, d12, d12, #0
+	vmlal.s32	q4, d27, d4
+	vmov.i64	d10, #0
+	vmlal.s32	q4, d28, d1
+	vmlal.s32	q5, d16, d0
+	sub		r6, r7, #32
+	vmlal.s32	q5, d17, d7
+	vmlal.s32	q5, d14, d6
+	vext.32		d30, d9, d8, #0
+	vmlal.s32	q5, d15, d5
+	vld1.8		{d31}, [r6, : 64]!
+	vmlal.s32	q5, d29, d4
+	vmlal.s32	q15, d20, d0
+	vext.32		d0, d6, d18, #1
+	vmlal.s32	q15, d21, d25
+	vrev64.i32	d0, d0
+	vmlal.s32	q15, d26, d24
+	vext.32		d1, d7, d19, #1
+	vext.32		d7, d10, d10, #0
+	vmlal.s32	q15, d27, d23
+	vrev64.i32	d1, d1
+	vld1.8		{d6}, [r6, : 64]
+	vmlal.s32	q15, d28, d22
+	vmlal.s32	q3, d16, d4
+	add		r6, r6, #24
+	vmlal.s32	q3, d17, d2
+	vext.32		d4, d31, d30, #0
+	vmov		d17, d11
+	vmlal.s32	q3, d14, d1
+	vext.32		d11, d13, d13, #0
+	vext.32		d13, d30, d30, #0
+	vmlal.s32	q3, d15, d0
+	vext.32		d1, d8, d8, #0
+	vmlal.s32	q3, d29, d3
+	vld1.8		{d5}, [r6, : 64]
+	sub		r6, r6, #16
+	vext.32		d10, d6, d6, #0
+	vmov.i32	q1, #0xffffffff
+	vshl.i64	q4, q1, #25
+	add		r7, sp, #512
+	vld1.8		{d14-d15}, [r7, : 128]
+	vadd.i64	q9, q2, q7
+	vshl.i64	q1, q1, #26
+	vshr.s64	q10, q9, #26
+	vld1.8		{d0}, [r6, : 64]!
+	vadd.i64	q5, q5, q10
+	vand		q9, q9, q1
+	vld1.8		{d16}, [r6, : 64]!
+	add		r6, sp, #528
+	vld1.8		{d20-d21}, [r6, : 128]
+	vadd.i64	q11, q5, q10
+	vsub.i64	q2, q2, q9
+	vshr.s64	q9, q11, #25
+	vext.32		d12, d5, d4, #0
+	vand		q11, q11, q4
+	vadd.i64	q0, q0, q9
+	vmov		d19, d7
+	vadd.i64	q3, q0, q7
+	vsub.i64	q5, q5, q11
+	vshr.s64	q11, q3, #26
+	vext.32		d18, d11, d10, #0
+	vand		q3, q3, q1
+	vadd.i64	q8, q8, q11
+	vadd.i64	q11, q8, q10
+	vsub.i64	q0, q0, q3
+	vshr.s64	q3, q11, #25
+	vand		q11, q11, q4
+	vadd.i64	q3, q6, q3
+	vadd.i64	q6, q3, q7
+	vsub.i64	q8, q8, q11
+	vshr.s64	q11, q6, #26
+	vand		q6, q6, q1
+	vadd.i64	q9, q9, q11
+	vadd.i64	d25, d19, d21
+	vsub.i64	q3, q3, q6
+	vshr.s64	d23, d25, #25
+	vand		q4, q12, q4
+	vadd.i64	d21, d23, d23
+	vshl.i64	d25, d23, #4
+	vadd.i64	d21, d21, d23
+	vadd.i64	d25, d25, d21
+	vadd.i64	d4, d4, d25
+	vzip.i32	q0, q8
+	vadd.i64	d12, d4, d14
+	add		r6, r8, #8
+	vst1.8		d0, [r6, : 64]
+	vsub.i64	d19, d19, d9
+	add		r6, r6, #16
+	vst1.8		d16, [r6, : 64]
+	vshr.s64	d22, d12, #26
+	vand		q0, q6, q1
+	vadd.i64	d10, d10, d22
+	vzip.i32	q3, q9
+	vsub.i64	d4, d4, d0
+	sub		r6, r6, #8
+	vst1.8		d6, [r6, : 64]
+	add		r6, r6, #16
+	vst1.8		d18, [r6, : 64]
+	vzip.i32	q2, q5
+	sub		r6, r6, #32
+	vst1.8		d4, [r6, : 64]
+	subs		r5, r5, #1
+	bhi		._squaringloop
+._skipsquaringloop:
+	mov		r2, r2
+	add		r5, r3, #288
+	add		r6, r3, #144
+	vmov.i32	q0, #19
+	vmov.i32	q1, #0
+	vmov.i32	q2, #1
+	vzip.i32	q1, q2
+	vld1.8		{d4-d5}, [r5, : 128]!
+	vld1.8		{d6-d7}, [r5, : 128]!
+	vld1.8		{d9}, [r5, : 64]
+	vld1.8		{d10-d11}, [r2, : 128]!
+	add		r5, sp, #416
+	vld1.8		{d12-d13}, [r2, : 128]!
+	vmul.i32	q7, q2, q0
+	vld1.8		{d8}, [r2, : 64]
+	vext.32		d17, d11, d10, #1
+	vmul.i32	q9, q3, q0
+	vext.32		d16, d10, d8, #1
+	vshl.u32	q10, q5, q1
+	vext.32		d22, d14, d4, #1
+	vext.32		d24, d18, d6, #1
+	vshl.u32	q13, q6, q1
+	vshl.u32	d28, d8, d2
+	vrev64.i32	d22, d22
+	vmul.i32	d1, d9, d1
+	vrev64.i32	d24, d24
+	vext.32		d29, d8, d13, #1
+	vext.32		d0, d1, d9, #1
+	vrev64.i32	d0, d0
+	vext.32		d2, d9, d1, #1
+	vext.32		d23, d15, d5, #1
+	vmull.s32	q4, d20, d4
+	vrev64.i32	d23, d23
+	vmlal.s32	q4, d21, d1
+	vrev64.i32	d2, d2
+	vmlal.s32	q4, d26, d19
+	vext.32		d3, d5, d15, #1
+	vmlal.s32	q4, d27, d18
+	vrev64.i32	d3, d3
+	vmlal.s32	q4, d28, d15
+	vext.32		d14, d12, d11, #1
+	vmull.s32	q5, d16, d23
+	vext.32		d15, d13, d12, #1
+	vmlal.s32	q5, d17, d4
+	vst1.8		d8, [r5, : 64]!
+	vmlal.s32	q5, d14, d1
+	vext.32		d12, d9, d8, #0
+	vmlal.s32	q5, d15, d19
+	vmov.i64	d13, #0
+	vmlal.s32	q5, d29, d18
+	vext.32		d25, d19, d7, #1
+	vmlal.s32	q6, d20, d5
+	vrev64.i32	d25, d25
+	vmlal.s32	q6, d21, d4
+	vst1.8		d11, [r5, : 64]!
+	vmlal.s32	q6, d26, d1
+	vext.32		d9, d10, d10, #0
+	vmlal.s32	q6, d27, d19
+	vmov.i64	d8, #0
+	vmlal.s32	q6, d28, d18
+	vmlal.s32	q4, d16, d24
+	vmlal.s32	q4, d17, d5
+	vmlal.s32	q4, d14, d4
+	vst1.8		d12, [r5, : 64]!
+	vmlal.s32	q4, d15, d1
+	vext.32		d10, d13, d12, #0
+	vmlal.s32	q4, d29, d19
+	vmov.i64	d11, #0
+	vmlal.s32	q5, d20, d6
+	vmlal.s32	q5, d21, d5
+	vmlal.s32	q5, d26, d4
+	vext.32		d13, d8, d8, #0
+	vmlal.s32	q5, d27, d1
+	vmov.i64	d12, #0
+	vmlal.s32	q5, d28, d19
+	vst1.8		d9, [r5, : 64]!
+	vmlal.s32	q6, d16, d25
+	vmlal.s32	q6, d17, d6
+	vst1.8		d10, [r5, : 64]
+	vmlal.s32	q6, d14, d5
+	vext.32		d8, d11, d10, #0
+	vmlal.s32	q6, d15, d4
+	vmov.i64	d9, #0
+	vmlal.s32	q6, d29, d1
+	vmlal.s32	q4, d20, d7
+	vmlal.s32	q4, d21, d6
+	vmlal.s32	q4, d26, d5
+	vext.32		d11, d12, d12, #0
+	vmlal.s32	q4, d27, d4
+	vmov.i64	d10, #0
+	vmlal.s32	q4, d28, d1
+	vmlal.s32	q5, d16, d0
+	sub		r2, r5, #32
+	vmlal.s32	q5, d17, d7
+	vmlal.s32	q5, d14, d6
+	vext.32		d30, d9, d8, #0
+	vmlal.s32	q5, d15, d5
+	vld1.8		{d31}, [r2, : 64]!
+	vmlal.s32	q5, d29, d4
+	vmlal.s32	q15, d20, d0
+	vext.32		d0, d6, d18, #1
+	vmlal.s32	q15, d21, d25
+	vrev64.i32	d0, d0
+	vmlal.s32	q15, d26, d24
+	vext.32		d1, d7, d19, #1
+	vext.32		d7, d10, d10, #0
+	vmlal.s32	q15, d27, d23
+	vrev64.i32	d1, d1
+	vld1.8		{d6}, [r2, : 64]
+	vmlal.s32	q15, d28, d22
+	vmlal.s32	q3, d16, d4
+	add		r2, r2, #24
+	vmlal.s32	q3, d17, d2
+	vext.32		d4, d31, d30, #0
+	vmov		d17, d11
+	vmlal.s32	q3, d14, d1
+	vext.32		d11, d13, d13, #0
+	vext.32		d13, d30, d30, #0
+	vmlal.s32	q3, d15, d0
+	vext.32		d1, d8, d8, #0
+	vmlal.s32	q3, d29, d3
+	vld1.8		{d5}, [r2, : 64]
+	sub		r2, r2, #16
+	vext.32		d10, d6, d6, #0
+	vmov.i32	q1, #0xffffffff
+	vshl.i64	q4, q1, #25
+	add		r5, sp, #512
+	vld1.8		{d14-d15}, [r5, : 128]
+	vadd.i64	q9, q2, q7
+	vshl.i64	q1, q1, #26
+	vshr.s64	q10, q9, #26
+	vld1.8		{d0}, [r2, : 64]!
+	vadd.i64	q5, q5, q10
+	vand		q9, q9, q1
+	vld1.8		{d16}, [r2, : 64]!
+	add		r2, sp, #528
+	vld1.8		{d20-d21}, [r2, : 128]
+	vadd.i64	q11, q5, q10
+	vsub.i64	q2, q2, q9
+	vshr.s64	q9, q11, #25
+	vext.32		d12, d5, d4, #0
+	vand		q11, q11, q4
+	vadd.i64	q0, q0, q9
+	vmov		d19, d7
+	vadd.i64	q3, q0, q7
+	vsub.i64	q5, q5, q11
+	vshr.s64	q11, q3, #26
+	vext.32		d18, d11, d10, #0
+	vand		q3, q3, q1
+	vadd.i64	q8, q8, q11
+	vadd.i64	q11, q8, q10
+	vsub.i64	q0, q0, q3
+	vshr.s64	q3, q11, #25
+	vand		q11, q11, q4
+	vadd.i64	q3, q6, q3
+	vadd.i64	q6, q3, q7
+	vsub.i64	q8, q8, q11
+	vshr.s64	q11, q6, #26
+	vand		q6, q6, q1
+	vadd.i64	q9, q9, q11
+	vadd.i64	d25, d19, d21
+	vsub.i64	q3, q3, q6
+	vshr.s64	d23, d25, #25
+	vand		q4, q12, q4
+	vadd.i64	d21, d23, d23
+	vshl.i64	d25, d23, #4
+	vadd.i64	d21, d21, d23
+	vadd.i64	d25, d25, d21
+	vadd.i64	d4, d4, d25
+	vzip.i32	q0, q8
+	vadd.i64	d12, d4, d14
+	add		r2, r6, #8
+	vst1.8		d0, [r2, : 64]
+	vsub.i64	d19, d19, d9
+	add		r2, r2, #16
+	vst1.8		d16, [r2, : 64]
+	vshr.s64	d22, d12, #26
+	vand		q0, q6, q1
+	vadd.i64	d10, d10, d22
+	vzip.i32	q3, q9
+	vsub.i64	d4, d4, d0
+	sub		r2, r2, #8
+	vst1.8		d6, [r2, : 64]
+	add		r2, r2, #16
+	vst1.8		d18, [r2, : 64]
+	vzip.i32	q2, q5
+	sub		r2, r2, #32
+	vst1.8		d4, [r2, : 64]
+	cmp		r4, #0
+	beq		._skippostcopy
+	add		r2, r3, #144
+	mov		r4, r4
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d4}, [r2, : 64]
+	vst1.8		{d0-d1}, [r4, : 128]!
+	vst1.8		{d2-d3}, [r4, : 128]!
+	vst1.8		d4, [r4, : 64]
+._skippostcopy:
+	cmp		r1, #1
+	bne		._skipfinalcopy
+	add		r2, r3, #288
+	add		r4, r3, #144
+	vld1.8		{d0-d1}, [r2, : 128]!
+	vld1.8		{d2-d3}, [r2, : 128]!
+	vld1.8		{d4}, [r2, : 64]
+	vst1.8		{d0-d1}, [r4, : 128]!
+	vst1.8		{d2-d3}, [r4, : 128]!
+	vst1.8		d4, [r4, : 64]
+._skipfinalcopy:
+	add		r1, r1, #1
+	cmp		r1, #12
+	blo		._invertloop
+	add		r1, r3, #144
+	ldr		r2, [r1], #4
+	ldr		r3, [r1], #4
+	ldr		r4, [r1], #4
+	ldr		r5, [r1], #4
+	ldr		r6, [r1], #4
+	ldr		r7, [r1], #4
+	ldr		r8, [r1], #4
+	ldr		r9, [r1], #4
+	ldr		r10, [r1], #4
+	ldr		r1, [r1]
+	add		r11, r1, r1, LSL #4
+	add		r11, r11, r1, LSL #1
+	add		r11, r11, #16777216
+	mov		r11, r11, ASR #25
+	add		r11, r11, r2
+	mov		r11, r11, ASR #26
+	add		r11, r11, r3
+	mov		r11, r11, ASR #25
+	add		r11, r11, r4
+	mov		r11, r11, ASR #26
+	add		r11, r11, r5
+	mov		r11, r11, ASR #25
+	add		r11, r11, r6
+	mov		r11, r11, ASR #26
+	add		r11, r11, r7
+	mov		r11, r11, ASR #25
+	add		r11, r11, r8
+	mov		r11, r11, ASR #26
+	add		r11, r11, r9
+	mov		r11, r11, ASR #25
+	add		r11, r11, r10
+	mov		r11, r11, ASR #26
+	add		r11, r11, r1
+	mov		r11, r11, ASR #25
+	add		r2, r2, r11
+	add		r2, r2, r11, LSL #1
+	add		r2, r2, r11, LSL #4
+	mov		r11, r2, ASR #26
+	add		r3, r3, r11
+	sub		r2, r2, r11, LSL #26
+	mov		r11, r3, ASR #25
+	add		r4, r4, r11
+	sub		r3, r3, r11, LSL #25
+	mov		r11, r4, ASR #26
+	add		r5, r5, r11
+	sub		r4, r4, r11, LSL #26
+	mov		r11, r5, ASR #25
+	add		r6, r6, r11
+	sub		r5, r5, r11, LSL #25
+	mov		r11, r6, ASR #26
+	add		r7, r7, r11
+	sub		r6, r6, r11, LSL #26
+	mov		r11, r7, ASR #25
+	add		r8, r8, r11
+	sub		r7, r7, r11, LSL #25
+	mov		r11, r8, ASR #26
+	add		r9, r9, r11
+	sub		r8, r8, r11, LSL #26
+	mov		r11, r9, ASR #25
+	add		r10, r10, r11
+	sub		r9, r9, r11, LSL #25
+	mov		r11, r10, ASR #26
+	add		r1, r1, r11
+	sub		r10, r10, r11, LSL #26
+	mov		r11, r1, ASR #25
+	sub		r1, r1, r11, LSL #25
+	add		r2, r2, r3, LSL #26
+	mov		r3, r3, LSR #6
+	add		r3, r3, r4, LSL #19
+	mov		r4, r4, LSR #13
+	add		r4, r4, r5, LSL #13
+	mov		r5, r5, LSR #19
+	add		r5, r5, r6, LSL #6
+	add		r6, r7, r8, LSL #25
+	mov		r7, r8, LSR #7
+	add		r7, r7, r9, LSL #19
+	mov		r8, r9, LSR #13
+	add		r8, r8, r10, LSL #12
+	mov		r9, r10, LSR #20
+	add		r1, r9, r1, LSL #6
+	str		r2, [r0], #4
+	str		r3, [r0], #4
+	str		r4, [r0], #4
+	str		r5, [r0], #4
+	str		r6, [r0], #4
+	str		r7, [r0], #4
+	str		r8, [r0], #4
+	str		r1, [r0]
+	ldrd		r4, [sp, #0]
+	ldrd		r6, [sp, #8]
+	ldrd		r8, [sp, #16]
+	ldrd		r10, [sp, #24]
+	ldr		r12, [sp, #480]
+	ldr		r14, [sp, #484]
+	ldr		r0, =0
+	mov		sp, r12
+	vpop		{q4, q5, q6, q7}
+	bx		lr
Signature · 1 line
2.19.1
author maintainer
#24 lore raw

PATCH v8 24/28 zinc: Curve25519 ARM implementation

In reply to Jason A. Donenfeld

This ports the SUPERCOP implementation for usage in kernel space. In addition to the usual header, macro, and style changes required for kernel space, it makes a few small changes to the code:

- The stack alignment is relaxed to 16 bytes.
- Superfluous mov statements have been removed.
- ldr for constants has been replaced with movw.
- ldreq has been replaced with moveq.
- The str epilogue has been made more idiomatic.
- SIMD registers are not pushed and popped at the beginning and end.
- The prologue and epilogue have been made idiomatic.
- A hole has been removed from the stack, saving 32 bytes.
- We write-back the base register whenever possible for vld1.8.
- Some multiplications have been reordered for better A7 performance.

There are more opportunities for cleanup, since this code is from qhasm, which doesn't always do the most opportune thing. But even prior to extensive hand optimizations, this code delivers significant performance improvements (given in get_cycles() per call):

		      ----------- -------------
	             | generic C | this commit |
	 ------------ ----------- -------------
	| Cortex-A7  |     49136 |       22395 |
	 ------------ ----------- -------------
	| Cortex-A17 |     17326 |        4983 |
	 ------------ ----------- -------------
Cc Russell King <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 lib/zinc/Makefile                             |   1 +
 lib/zinc/curve25519/curve25519-arm-glue.c     |  43 +++
 ...e25519-arm-supercop.S => curve25519-arm.S} | 349 ++++++++----------
 lib/zinc/curve25519/curve25519.c              |   2 +
 4 files changed, 200 insertions(+), 195 deletions(-)
 create mode 100644 lib/zinc/curve25519/curve25519-arm-glue.c
 rename lib/zinc/curve25519/{curve25519-arm-supercop.S => curve25519-arm.S} (92%)
lib/zinc/Makefile +1 -0
index 65440438c6e5..be73c342f9ba 100644
@@ -27,4 +27,5 @@ zinc_blake2s-$(CONFIG_ZINC_ARCH_X86_64) += blake2s/blake2s-x86_64.o
 obj-$(CONFIG_ZINC_BLAKE2S) += zinc_blake2s.o
 
 zinc_curve25519-y := curve25519/curve25519.o
+zinc_curve25519-$(CONFIG_ZINC_ARCH_ARM) += curve25519/curve25519-arm.o
 obj-$(CONFIG_ZINC_CURVE25519) += zinc_curve25519.o
lib/zinc/curve25519/curve25519-arm-glue.c new file +43 -0

43 lines not on this page · raw message

lib/zinc/curve25519/curve25519-arm.S renamed from lib/zinc/curve25519/curve25519-arm-supercop.S +154 -195

805 lines not on this page · raw message

lib/zinc/curve25519/curve25519.c +2 -0
index 074fbb1c2149..0ebe16f8c638 100644
@@ -22,6 +22,8 @@
 
 #if defined(CONFIG_ZINC_ARCH_X86_64)
 #include "curve25519-x86_64-glue.c"
+#elif defined(CONFIG_ZINC_ARCH_ARM)
+#include "curve25519-arm-glue.c"
 #else
 static bool *const curve25519_nobs[] __initconst = { };
 static void __init curve25519_fpu_init(void)
Signature · 1 line
2.19.1
author maintainer
#25 lore raw

PATCH v8 25/28 crypto: port Poly1305 to Zinc

In reply to Jason A. Donenfeld

Now that Poly1305 is in Zinc, we can have the crypto API code simply call into it. We have to do a little bit of book keeping here, because the crypto API receives the key in the first few calls to update.

---
 arch/x86/crypto/Makefile               |   3 -
 arch/x86/crypto/poly1305-avx2-x86_64.S | 388 ----------------
 arch/x86/crypto/poly1305-sse2-x86_64.S | 584 -------------------------
 arch/x86/crypto/poly1305_glue.c        | 205 ---------
 crypto/Kconfig                         |  15 +-
 crypto/Makefile                        |   2 +-
 crypto/chacha20poly1305.c              |  12 +-
 crypto/poly1305_generic.c              | 304 -------------
 crypto/poly1305_zinc.c                 | 101 +++++
 include/crypto/poly1305.h              |  40 --
 10 files changed, 110 insertions(+), 1544 deletions(-)
 delete mode 100644 arch/x86/crypto/poly1305-avx2-x86_64.S
 delete mode 100644 arch/x86/crypto/poly1305-sse2-x86_64.S
 delete mode 100644 arch/x86/crypto/poly1305_glue.c
 delete mode 100644 crypto/poly1305_generic.c
 create mode 100644 crypto/poly1305_zinc.c
 delete mode 100644 include/crypto/poly1305.h
arch/x86/crypto/Makefile +0 -3
index a450ad573dcb..cf830219846b 100644
@@ -34,7 +34,6 @@ obj-$(CONFIG_CRYPTO_CRC32_PCLMUL) += crc32-pclmul.o
 obj-$(CONFIG_CRYPTO_SHA256_SSSE3) += sha256-ssse3.o
 obj-$(CONFIG_CRYPTO_SHA512_SSSE3) += sha512-ssse3.o
 obj-$(CONFIG_CRYPTO_CRCT10DIF_PCLMUL) += crct10dif-pclmul.o
-obj-$(CONFIG_CRYPTO_POLY1305_X86_64) += poly1305-x86_64.o
 
 obj-$(CONFIG_CRYPTO_AEGIS128_AESNI_SSE2) += aegis128-aesni.o
 obj-$(CONFIG_CRYPTO_AEGIS128L_AESNI_SSE2) += aegis128l-aesni.o
@@ -110,10 +109,8 @@ aesni-intel-y := aesni-intel_asm.o aesni-intel_glue.o fpu.o
 aesni-intel-$(CONFIG_64BIT) += aesni-intel_avx-x86_64.o aes_ctrby8_avx-x86_64.o
 ghash-clmulni-intel-y := ghash-clmulni-intel_asm.o ghash-clmulni-intel_glue.o
 sha1-ssse3-y := sha1_ssse3_asm.o sha1_ssse3_glue.o
-poly1305-x86_64-y := poly1305-sse2-x86_64.o poly1305_glue.o
 ifeq ($(avx2_supported),yes)
 sha1-ssse3-y += sha1_avx2_x86_64_asm.o
-poly1305-x86_64-y += poly1305-avx2-x86_64.o
 endif
 ifeq ($(sha1_ni_supported),yes)
 sha1-ssse3-y += sha1_ni_asm.o
arch/x86/crypto/poly1305-avx2-x86_64.S deleted +0 -388

388 lines not on this page · raw message

arch/x86/crypto/poly1305-sse2-x86_64.S deleted +0 -584

584 lines not on this page · raw message

arch/x86/crypto/poly1305_glue.c deleted +0 -205

205 lines not on this page · raw message

crypto/Kconfig +2 -13
index f3e40ac56d93..47859a0f8052 100644
@@ -656,24 +656,13 @@ config CRYPTO_GHASH
 config CRYPTO_POLY1305
 	tristate "Poly1305 authenticator algorithm"
 	select CRYPTO_HASH
+	select ZINC_POLY1305
 	help
 	  Poly1305 authenticator algorithm, RFC7539.
 
 	  Poly1305 is an authenticator algorithm designed by Daniel J. Bernstein.
 	  It is used for the ChaCha20-Poly1305 AEAD, specified in RFC7539 for use
-	  in IETF protocols. This is the portable C implementation of Poly1305.
-
-config CRYPTO_POLY1305_X86_64
-	tristate "Poly1305 authenticator algorithm (x86_64/SSE2/AVX2)"
-	depends on X86 && 64BIT
-	select CRYPTO_POLY1305
-	help
-	  Poly1305 authenticator algorithm, RFC7539.
-
-	  Poly1305 is an authenticator algorithm designed by Daniel J. Bernstein.
-	  It is used for the ChaCha20-Poly1305 AEAD, specified in RFC7539 for use
-	  in IETF protocols. This is the x86_64 assembler implementation using SIMD
-	  instructions.
+	  in IETF protocols.
 
 config CRYPTO_MD4
 	tristate "MD4 digest algorithm"
crypto/Makefile +1 -1
index 6d1d40eeb964..5e60348d02e2 100644
@@ -118,7 +118,7 @@ obj-$(CONFIG_CRYPTO_SEED) += seed.o
 obj-$(CONFIG_CRYPTO_SPECK) += speck.o
 obj-$(CONFIG_CRYPTO_SALSA20) += salsa20_generic.o
 obj-$(CONFIG_CRYPTO_CHACHA20) += chacha20_generic.o
-obj-$(CONFIG_CRYPTO_POLY1305) += poly1305_generic.o
+obj-$(CONFIG_CRYPTO_POLY1305) += poly1305_zinc.o
 obj-$(CONFIG_CRYPTO_DEFLATE) += deflate.o
 obj-$(CONFIG_CRYPTO_MICHAEL_MIC) += michael_mic.o
 obj-$(CONFIG_CRYPTO_CRC32C) += crc32c_generic.o
crypto/chacha20poly1305.c +6 -6
index 600afa99941f..bf523797bef3 100644
@@ -14,7 +14,7 @@
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/chacha20.h>
-#include <crypto/poly1305.h>
+#include <zinc/poly1305.h>
 #include <linux/err.h>
 #include <linux/init.h>
 #include <linux/kernel.h>
@@ -62,7 +62,7 @@ struct chachapoly_req_ctx {
 	/* the key we generate for Poly1305 using Chacha20 */
 	u8 key[POLY1305_KEY_SIZE];
 	/* calculated Poly1305 tag */
-	u8 tag[POLY1305_DIGEST_SIZE];
+	u8 tag[POLY1305_MAC_SIZE];
 	/* length of data to en/decrypt, without ICV */
 	unsigned int cryptlen;
 	/* Actual AD, excluding IV */
@@ -471,7 +471,7 @@ static int chachapoly_decrypt(struct aead_request *req)
 {
 	struct chachapoly_req_ctx *rctx = aead_request_ctx(req);
 
-	rctx->cryptlen = req->cryptlen - POLY1305_DIGEST_SIZE;
+	rctx->cryptlen = req->cryptlen - POLY1305_MAC_SIZE;
 
 	/* decrypt call chain:
 	 * - poly_genkey/done()
@@ -513,7 +513,7 @@ static int chachapoly_setkey(struct crypto_aead *aead, const u8 *key,
 static int chachapoly_setauthsize(struct crypto_aead *tfm,
 				  unsigned int authsize)
 {
-	if (authsize != POLY1305_DIGEST_SIZE)
+	if (authsize != POLY1305_MAC_SIZE)
 		return -EINVAL;
 
 	return 0;
@@ -613,7 +613,7 @@ static int chachapoly_create(struct crypto_template *tmpl, struct rtattr **tb,
 	poly_hash = __crypto_hash_alg_common(poly);
 
 	err = -EINVAL;
-	if (poly_hash->digestsize != POLY1305_DIGEST_SIZE)
+	if (poly_hash->digestsize != POLY1305_MAC_SIZE)
 		goto out_put_poly;
 
 	err = -ENOMEM;
@@ -666,7 +666,7 @@ static int chachapoly_create(struct crypto_template *tmpl, struct rtattr **tb,
 				     ctx->saltlen;
 	inst->alg.ivsize = ivsize;
 	inst->alg.chunksize = crypto_skcipher_alg_chunksize(chacha);
-	inst->alg.maxauthsize = POLY1305_DIGEST_SIZE;
+	inst->alg.maxauthsize = POLY1305_MAC_SIZE;
 	inst->alg.init = chachapoly_init;
 	inst->alg.exit = chachapoly_exit;
 	inst->alg.encrypt = chachapoly_encrypt;
crypto/poly1305_generic.c deleted +0 -304

304 lines not on this page · raw message

crypto/poly1305_zinc.c new file +101 -0

101 lines not on this page · raw message

include/crypto/poly1305.h deleted +0 -40
deleted file mode 100644
index f718a19da82f..000000000000
@@ -1,40 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 */
-/*
- * Common values for the Poly1305 algorithm
- */
-
-#ifndef _CRYPTO_POLY1305_H
-#define _CRYPTO_POLY1305_H
-
-#include <linux/types.h>
-#include <linux/crypto.h>
-
-#define POLY1305_BLOCK_SIZE	16
-#define POLY1305_KEY_SIZE	32
-#define POLY1305_DIGEST_SIZE	16
-
-struct poly1305_desc_ctx {
-	/* key */
-	u32 r[5];
-	/* finalize key */
-	u32 s[4];
-	/* accumulator */
-	u32 h[5];
-	/* partial buffer */
-	u8 buf[POLY1305_BLOCK_SIZE];
-	/* bytes used in partial buffer */
-	unsigned int buflen;
-	/* r key has been set */
-	bool rset;
-	/* s key has been set */
-	bool sset;
-};
-
-int crypto_poly1305_init(struct shash_desc *desc);
-unsigned int crypto_poly1305_setdesckey(struct poly1305_desc_ctx *dctx,
-					const u8 *src, unsigned int srclen);
-int crypto_poly1305_update(struct shash_desc *desc,
-			   const u8 *src, unsigned int srclen);
-int crypto_poly1305_final(struct shash_desc *desc, u8 *dst);
-
-#endif
Signature · 1 line
2.19.1
author maintainer
#26 lore raw

PATCH v8 26/28 crypto: port ChaCha20 to Zinc

In reply to Jason A. Donenfeld

Now that ChaCha20 is in Zinc, we can have the crypto API code simply call into it. The crypto API expects to have a stored key per instance and independent nonces, so we follow suite and store the key and initialize the nonce independently.

---
 arch/arm/configs/exynos_defconfig       |   1 -
 arch/arm/configs/multi_v7_defconfig     |   1 -
 arch/arm/configs/omap2plus_defconfig    |   1 -
 arch/arm/crypto/Kconfig                 |   6 -
 arch/arm/crypto/Makefile                |   2 -
 arch/arm/crypto/chacha20-neon-core.S    | 521 --------------------
 arch/arm/crypto/chacha20-neon-glue.c    | 127 -----
 arch/arm64/configs/defconfig            |   1 -
 arch/arm64/crypto/Kconfig               |   6 -
 arch/arm64/crypto/Makefile              |   3 -
 arch/arm64/crypto/chacha20-neon-core.S  | 450 -----------------
 arch/arm64/crypto/chacha20-neon-glue.c  | 133 -----
 arch/x86/crypto/Makefile                |   3 -
 arch/x86/crypto/chacha20-avx2-x86_64.S  | 448 -----------------
 arch/x86/crypto/chacha20-ssse3-x86_64.S | 630 ------------------------
 arch/x86/crypto/chacha20_glue.c         | 146 ------
 crypto/Kconfig                          |  17 +-
 crypto/Makefile                         |   2 +-
 crypto/chacha20_generic.c               | 136 -----
 crypto/chacha20_zinc.c                  |  90 ++++
 crypto/chacha20poly1305.c               |   8 +-
 include/crypto/chacha20.h               |  12 -
 22 files changed, 96 insertions(+), 2648 deletions(-)
 delete mode 100644 arch/arm/crypto/chacha20-neon-core.S
 delete mode 100644 arch/arm/crypto/chacha20-neon-glue.c
 delete mode 100644 arch/arm64/crypto/chacha20-neon-core.S
 delete mode 100644 arch/arm64/crypto/chacha20-neon-glue.c
 delete mode 100644 arch/x86/crypto/chacha20-avx2-x86_64.S
 delete mode 100644 arch/x86/crypto/chacha20-ssse3-x86_64.S
 delete mode 100644 arch/x86/crypto/chacha20_glue.c
 delete mode 100644 crypto/chacha20_generic.c
 create mode 100644 crypto/chacha20_zinc.c
arch/arm/configs/exynos_defconfig +0 -1
index 27ea6dfcf2f2..95929b5e7b10 100644
@@ -350,7 +350,6 @@ CONFIG_CRYPTO_SHA1_ARM_NEON=m
 CONFIG_CRYPTO_SHA256_ARM=m
 CONFIG_CRYPTO_SHA512_ARM=m
 CONFIG_CRYPTO_AES_ARM_BS=m
-CONFIG_CRYPTO_CHACHA20_NEON=m
 CONFIG_CRC_CCITT=y
 CONFIG_FONTS=y
 CONFIG_FONT_7x14=y
arch/arm/configs/multi_v7_defconfig +0 -1
index fc33444e94f0..63be07724db3 100644
@@ -1000,4 +1000,3 @@ CONFIG_CRYPTO_AES_ARM_BS=m
 CONFIG_CRYPTO_AES_ARM_CE=m
 CONFIG_CRYPTO_GHASH_ARM_CE=m
 CONFIG_CRYPTO_CRC32_ARM_CE=m
-CONFIG_CRYPTO_CHACHA20_NEON=m
arch/arm/configs/omap2plus_defconfig +0 -1
index 6491419b1dad..f585a8ecc336 100644
@@ -547,7 +547,6 @@ CONFIG_CRYPTO_SHA512_ARM=m
 CONFIG_CRYPTO_AES_ARM=m
 CONFIG_CRYPTO_AES_ARM_BS=m
 CONFIG_CRYPTO_GHASH_ARM_CE=m
-CONFIG_CRYPTO_CHACHA20_NEON=m
 CONFIG_CRC_CCITT=y
 CONFIG_CRC_T10DIF=y
 CONFIG_CRC_ITU_T=y
arch/arm/crypto/Kconfig +0 -6
index 925d1364727a..fb80fd89f0e7 100644
@@ -115,12 +115,6 @@ config CRYPTO_CRC32_ARM_CE
 	depends on KERNEL_MODE_NEON && CRC32
 	select CRYPTO_HASH
 
-config CRYPTO_CHACHA20_NEON
-	tristate "NEON accelerated ChaCha20 symmetric cipher"
-	depends on KERNEL_MODE_NEON
-	select CRYPTO_BLKCIPHER
-	select CRYPTO_CHACHA20
-
 config CRYPTO_SPECK_NEON
 	tristate "NEON accelerated Speck cipher algorithms"
 	depends on KERNEL_MODE_NEON
arch/arm/crypto/Makefile +0 -2
index 8de542c48ade..bbfa98447063 100644
@@ -9,7 +9,6 @@ obj-$(CONFIG_CRYPTO_SHA1_ARM) += sha1-arm.o
 obj-$(CONFIG_CRYPTO_SHA1_ARM_NEON) += sha1-arm-neon.o
 obj-$(CONFIG_CRYPTO_SHA256_ARM) += sha256-arm.o
 obj-$(CONFIG_CRYPTO_SHA512_ARM) += sha512-arm.o
-obj-$(CONFIG_CRYPTO_CHACHA20_NEON) += chacha20-neon.o
 obj-$(CONFIG_CRYPTO_SPECK_NEON) += speck-neon.o
 
 ce-obj-$(CONFIG_CRYPTO_AES_ARM_CE) += aes-arm-ce.o
@@ -53,7 +52,6 @@ aes-arm-ce-y	:= aes-ce-core.o aes-ce-glue.o
 ghash-arm-ce-y	:= ghash-ce-core.o ghash-ce-glue.o
 crct10dif-arm-ce-y	:= crct10dif-ce-core.o crct10dif-ce-glue.o
 crc32-arm-ce-y:= crc32-ce-core.o crc32-ce-glue.o
-chacha20-neon-y := chacha20-neon-core.o chacha20-neon-glue.o
 speck-neon-y := speck-neon-core.o speck-neon-glue.o
 
 ifdef REGENERATE_ARM_CRYPTO
arch/arm/crypto/chacha20-neon-core.S deleted +0 -521

521 lines not on this page · raw message

arch/arm/crypto/chacha20-neon-glue.c deleted +0 -127

127 lines not on this page · raw message

arch/arm64/configs/defconfig +0 -1
index db8d364f8476..6cc3c8a0ad88 100644
@@ -709,5 +709,4 @@ CONFIG_CRYPTO_CRCT10DIF_ARM64_CE=m
 CONFIG_CRYPTO_CRC32_ARM64_CE=m
 CONFIG_CRYPTO_AES_ARM64_CE_CCM=y
 CONFIG_CRYPTO_AES_ARM64_CE_BLK=y
-CONFIG_CRYPTO_CHACHA20_NEON=m
 CONFIG_CRYPTO_AES_ARM64_BS=m
arch/arm64/crypto/Kconfig +0 -6
index e3fdb0fd6f70..9db6d775a880 100644
@@ -105,12 +105,6 @@ config CRYPTO_AES_ARM64_NEON_BLK
 	select CRYPTO_AES
 	select CRYPTO_SIMD
 
-config CRYPTO_CHACHA20_NEON
-	tristate "NEON accelerated ChaCha20 symmetric cipher"
-	depends on KERNEL_MODE_NEON
-	select CRYPTO_BLKCIPHER
-	select CRYPTO_CHACHA20
-
 config CRYPTO_AES_ARM64_BS
 	tristate "AES in ECB/CBC/CTR/XTS modes using bit-sliced NEON algorithm"
 	depends on KERNEL_MODE_NEON
arch/arm64/crypto/Makefile +0 -3
index bcafd016618e..507c4bfb86e3 100644
@@ -53,9 +53,6 @@ sha256-arm64-y := sha256-glue.o sha256-core.o
 obj-$(CONFIG_CRYPTO_SHA512_ARM64) += sha512-arm64.o
 sha512-arm64-y := sha512-glue.o sha512-core.o
 
-obj-$(CONFIG_CRYPTO_CHACHA20_NEON) += chacha20-neon.o
-chacha20-neon-y := chacha20-neon-core.o chacha20-neon-glue.o
-
 obj-$(CONFIG_CRYPTO_SPECK_NEON) += speck-neon.o
 speck-neon-y := speck-neon-core.o speck-neon-glue.o
 
arch/arm64/crypto/chacha20-neon-core.S deleted +0 -450

450 lines not on this page · raw message

arch/arm64/crypto/chacha20-neon-glue.c deleted +0 -133

133 lines not on this page · raw message

arch/x86/crypto/Makefile +0 -3
index cf830219846b..419212c31246 100644
@@ -23,7 +23,6 @@ obj-$(CONFIG_CRYPTO_CAMELLIA_X86_64) += camellia-x86_64.o
 obj-$(CONFIG_CRYPTO_BLOWFISH_X86_64) += blowfish-x86_64.o
 obj-$(CONFIG_CRYPTO_TWOFISH_X86_64) += twofish-x86_64.o
 obj-$(CONFIG_CRYPTO_TWOFISH_X86_64_3WAY) += twofish-x86_64-3way.o
-obj-$(CONFIG_CRYPTO_CHACHA20_X86_64) += chacha20-x86_64.o
 obj-$(CONFIG_CRYPTO_SERPENT_SSE2_X86_64) += serpent-sse2-x86_64.o
 obj-$(CONFIG_CRYPTO_AES_NI_INTEL) += aesni-intel.o
 obj-$(CONFIG_CRYPTO_GHASH_CLMUL_NI_INTEL) += ghash-clmulni-intel.o
@@ -76,7 +75,6 @@ camellia-x86_64-y := camellia-x86_64-asm_64.o camellia_glue.o
 blowfish-x86_64-y := blowfish-x86_64-asm_64.o blowfish_glue.o
 twofish-x86_64-y := twofish-x86_64-asm_64.o twofish_glue.o
 twofish-x86_64-3way-y := twofish-x86_64-asm_64-3way.o twofish_glue_3way.o
-chacha20-x86_64-y := chacha20-ssse3-x86_64.o chacha20_glue.o
 serpent-sse2-x86_64-y := serpent-sse2-x86_64-asm_64.o serpent_sse2_glue.o
 
 aegis128-aesni-y := aegis128-aesni-asm.o aegis128-aesni-glue.o
@@ -99,7 +97,6 @@ endif
 
 ifeq ($(avx2_supported),yes)
 	camellia-aesni-avx2-y := camellia-aesni-avx2-asm_64.o camellia_aesni_avx2_glue.o
-	chacha20-x86_64-y += chacha20-avx2-x86_64.o
 	serpent-avx2-y := serpent-avx2-asm_64.o serpent_avx2_glue.o
 
 	morus1280-avx2-y := morus1280-avx2-asm.o morus1280-avx2-glue.o
arch/x86/crypto/chacha20-avx2-x86_64.S deleted +0 -448

448 lines not on this page · raw message

arch/x86/crypto/chacha20-ssse3-x86_64.S deleted +0 -630

630 lines not on this page · raw message

arch/x86/crypto/chacha20_glue.c deleted +0 -146

146 lines not on this page · raw message

crypto/Kconfig +1 -16
index 47859a0f8052..42dc48aa9b81 100644
@@ -1428,27 +1428,12 @@ config CRYPTO_SALSA20
 config CRYPTO_CHACHA20
 	tristate "ChaCha20 cipher algorithm"
 	select CRYPTO_BLKCIPHER
+	select ZINC_CHACHA20
 	help
 	  ChaCha20 cipher algorithm, RFC7539.
 
 	  ChaCha20 is a 256-bit high-speed stream cipher designed by Daniel J.
 	  Bernstein and further specified in RFC7539 for use in IETF protocols.
-	  This is the portable C implementation of ChaCha20.
-
-	  See also:
-	  <http://cr.yp.to/chacha/chacha-20080128.pdf>
-
-config CRYPTO_CHACHA20_X86_64
-	tristate "ChaCha20 cipher algorithm (x86_64/SSSE3/AVX2)"
-	depends on X86 && 64BIT
-	select CRYPTO_BLKCIPHER
-	select CRYPTO_CHACHA20
-	help
-	  ChaCha20 cipher algorithm, RFC7539.
-
-	  ChaCha20 is a 256-bit high-speed stream cipher designed by Daniel J.
-	  Bernstein and further specified in RFC7539 for use in IETF protocols.
-	  This is the x86_64 assembler implementation using SIMD instructions.
 
 	  See also:
 	  <http://cr.yp.to/chacha/chacha-20080128.pdf>
crypto/Makefile +1 -1
index 5e60348d02e2..587103b87890 100644
@@ -117,7 +117,7 @@ obj-$(CONFIG_CRYPTO_ANUBIS) += anubis.o
 obj-$(CONFIG_CRYPTO_SEED) += seed.o
 obj-$(CONFIG_CRYPTO_SPECK) += speck.o
 obj-$(CONFIG_CRYPTO_SALSA20) += salsa20_generic.o
-obj-$(CONFIG_CRYPTO_CHACHA20) += chacha20_generic.o
+obj-$(CONFIG_CRYPTO_CHACHA20) += chacha20_zinc.o
 obj-$(CONFIG_CRYPTO_POLY1305) += poly1305_zinc.o
 obj-$(CONFIG_CRYPTO_DEFLATE) += deflate.o
 obj-$(CONFIG_CRYPTO_MICHAEL_MIC) += michael_mic.o
crypto/chacha20_generic.c deleted +0 -136

136 lines not on this page · raw message

crypto/chacha20_zinc.c new file +90 -0

90 lines not on this page · raw message

crypto/chacha20poly1305.c +4 -4
index bf523797bef3..585c7ef4f543 100644
@@ -13,7 +13,7 @@
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/chacha20.h>
+#include <zinc/chacha20.h>
 #include <zinc/poly1305.h>
 #include <linux/err.h>
 #include <linux/init.h>
@@ -51,7 +51,7 @@ struct poly_req {
 };
 
 struct chacha_req {
-	u8 iv[CHACHA20_IV_SIZE];
+	u8 iv[CHACHA20_NONCE_SIZE];
 	struct scatterlist src[1];
 	struct skcipher_request req; /* must be last member */
 };
@@ -91,7 +91,7 @@ static void chacha_iv(u8 *iv, struct aead_request *req, u32 icb)
 	memcpy(iv, &leicb, sizeof(leicb));
 	memcpy(iv + sizeof(leicb), ctx->salt, ctx->saltlen);
 	memcpy(iv + sizeof(leicb) + ctx->saltlen, req->iv,
-	       CHACHA20_IV_SIZE - sizeof(leicb) - ctx->saltlen);
+	       CHACHA20_NONCE_SIZE - sizeof(leicb) - ctx->saltlen);
 }
 
 static int poly_verify_tag(struct aead_request *req)
@@ -639,7 +639,7 @@ static int chachapoly_create(struct crypto_template *tmpl, struct rtattr **tb,
 
 	err = -EINVAL;
 	/* Need 16-byte IV size, including Initial Block Counter value */
-	if (crypto_skcipher_alg_ivsize(chacha) != CHACHA20_IV_SIZE)
+	if (crypto_skcipher_alg_ivsize(chacha) != CHACHA20_NONCE_SIZE)
 		goto out_drop_chacha;
 	/* Not a stream cipher? */
 	if (chacha->base.cra_blocksize != 1)
include/crypto/chacha20.h +0 -12
index b83d66073db0..3b92f58f3891 100644
@@ -6,23 +6,11 @@
 #ifndef _CRYPTO_CHACHA20_H
 #define _CRYPTO_CHACHA20_H
 
-#include <crypto/skcipher.h>
-#include <linux/types.h>
-#include <linux/crypto.h>
-
 #define CHACHA20_IV_SIZE	16
 #define CHACHA20_KEY_SIZE	32
 #define CHACHA20_BLOCK_SIZE	64
 #define CHACHA20_BLOCK_WORDS	(CHACHA20_BLOCK_SIZE / sizeof(u32))
 
-struct chacha20_ctx {
-	u32 key[8];
-};
-
 void chacha20_block(u32 *state, u32 *stream);
-void crypto_chacha20_init(u32 *state, struct chacha20_ctx *ctx, u8 *iv);
-int crypto_chacha20_setkey(struct crypto_skcipher *tfm, const u8 *key,
-			   unsigned int keysize);
-int crypto_chacha20_crypt(struct skcipher_request *req);
 
 #endif
Signature · 1 line
2.19.1
author maintainer
#27 lore raw

PATCH v8 27/28 security/keys: rewrite big_key crypto to use Zinc

In reply to Jason A. Donenfeld

A while back, I noticed that the crypto and crypto API usage in big_keys were entirely broken in multiple ways, so I rewrote it. Now, I'm rewriting it again, but this time using Zinc's ChaCha20Poly1305 function. This makes the file considerably more simple; the diffstat alone should justify this commit. It also should be faster, since it no longer requires a mutex around the "aead api object" (nor allocations), allowing us to encrypt multiple items in parallel. We also benefit from being able to pass any type of pointer to Zinc, so we can get rid of the ridiculously complex custom page allocator that big_key really doesn't need.

Cc David Howells <[email protected]>
Cc Samuel Neves <[email protected]>
Cc Jean-Philippe Aumasson <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Greg KH <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 security/keys/Kconfig   |   4 +-
 security/keys/big_key.c | 230 +++++-----------------------------------
 2 files changed, 28 insertions(+), 206 deletions(-)
security/keys/Kconfig +1 -3
index 6462e6654ccf..66ff26298fb3 100644
@@ -45,9 +45,7 @@ config BIG_KEYS
 	bool "Large payload keys"
 	depends on KEYS
 	depends on TMPFS
-	select CRYPTO
-	select CRYPTO_AES
-	select CRYPTO_GCM
+	select ZINC_CHACHA20POLY1305
 	help
 	  This option provides support for holding large keys within the kernel
 	  (for example Kerberos ticket caches).  The data may be stored out to
security/keys/big_key.c +27 -203

347 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#28 lore raw

PATCH v8 28/28 net: WireGuard secure network tunnel

In reply to Jason A. Donenfeld

WireGuard is a layer 3 secure networking tunnel made specifically for the kernel, that aims to be much simpler and easier to audit than IPsec. Extensive documentation and description of the protocol and considerations, along with formal proofs of the cryptography, are available at:

* https://www.wireguard.com/
* https://www.wireguard.com/papers/wireguard.pdf

This commit implements WireGuard as a simple network device driver, accessible in the usual RTNL way used by virtual network drivers. It makes use of the udp_tunnel APIs, GRO, GSO, NAPI, and the usual set of networking subsystem APIs. It has a somewhat novel multicore queueing system designed for maximum throughput and minimal latency of encryption operations, but it is implemented modestly using workqueues and NAPI. Configuration is done via generic Netlink, and following a review from the Netlink maintainer a year ago, several high profile userspace have already implemented the API.

This commit also comes with several different tests, both in-kernel tests and out-of-kernel tests based on network namespaces, taking profit of the fact that sockets used by WireGuard intentionally stay in the namespace the WireGuard interface was originally created, exactly like the semantics of userspace tun devices. See wireguard.com/netns/ for pictures and examples.

The source code is fairly short, but rather than combining everything into a single file, WireGuard is developed as cleanly separable files, making auditing and comprehension easier. Things are laid out as follows:

* noise.[ch], cookie.[ch], messages.h: These implement the bulk of the cryptographic aspects of the protocol, and are mostly data-only in nature, taking in buffers of bytes and spitting out buffers of bytes. They also handle reference counting for their various shared pieces of data, like keys and key lists.

* ratelimiter.[ch]: Used as an integral part of cookie.[ch] for ratelimiting certain types of cryptographic operations in accordance with particular WireGuard semantics.

* allowedips.[ch], hashtables.[ch]: The main lookup structures of WireGuard, the former being trie-like with particular semantics, an integral part of the design of the protocol, and the latter just being nice helper functions around the specific hashtables we use.

* device.[ch]: Implementation of functions for the netdevice and for rtnl, responsible for maintaining the life of a given interface and wiring it up to the rest of WireGuard.

* peer.[ch]: Each interface has a list of peers, with helper functions available here for creation, destruction, and reference counting.

* socket.[ch]: Implementation of functions related to udp_socket and the general set of kernel socket APIs, for sending and receiving ciphertext UDP packets, and taking care of WireGuard-specific sticky socket routing semantics for the automatic roaming.

* netlink.[ch]: Userspace API entry point for configuring WireGuard peers and devices. The API has been implemented by several userspace tools and network management utility, and the WireGuard project distributes the basic wg(8) tool.

* queueing.[ch]: Shared function on the rx and tx path for handling the various queues used in the multicore algorithms.

* send.c: Handles encrypting outgoing packets in parallel on multiple cores, before sending them in order on a single core, via workqueues and ring buffers. Also handles sending handshake and cookie messages as part of the protocol, in parallel.

* receive.c: Handles decrypting incoming packets in parallel on multiple cores, before passing them off in order to be ingested via the rest of the networking subsystem with GRO via the typical NAPI poll function. Also handles receiving handshake and cookie messages as part of the protocol, in parallel.

* timers.[ch]: Uses the timer wheel to implement protocol particular event timeouts, and gives a set of very simple event-driven entry point functions for callers.

* main.c, version.h: Initialization and deinitialization of the module.

* selftest/*.h: Runtime unit tests for some of the most security sensitive functions.

* tools/testing/selftests/wireguard/netns.sh: Aforementioned testing script using network namespaces.

This commit aims to be as self-contained as possible, implementing WireGuard as a standalone module not needing much special handling or coordination from the network subsystem. I expect for future optimizations to the network stack to positively improve WireGuard, and vice-versa, but for the time being, this exists as intentionally standalone.

We introduce a menu option for CONFIG_WIREGUARD, as well as providing a verbose debug log and self-tests via CONFIG_WIREGUARD_DEBUG.

---
 MAINTAINERS                                  |   8 +
 drivers/net/Kconfig                          |  30 +
 drivers/net/Makefile                         |   1 +
 drivers/net/wireguard/Makefile               |  18 +
 drivers/net/wireguard/allowedips.c           | 413 ++++++++++
 drivers/net/wireguard/allowedips.h           |  54 ++
 drivers/net/wireguard/cookie.c               | 236 ++++++
 drivers/net/wireguard/cookie.h               |  59 ++
 drivers/net/wireguard/device.c               | 441 +++++++++++
 drivers/net/wireguard/device.h               |  65 ++
 drivers/net/wireguard/hashtables.c           | 209 +++++
 drivers/net/wireguard/hashtables.h           |  64 ++
 drivers/net/wireguard/main.c                 |  64 ++
 drivers/net/wireguard/messages.h             | 128 +++
 drivers/net/wireguard/netlink.c              | 599 ++++++++++++++
 drivers/net/wireguard/netlink.h              |  12 +
 drivers/net/wireguard/noise.c                | 788 +++++++++++++++++++
 drivers/net/wireguard/noise.h                | 131 +++
 drivers/net/wireguard/peer.c                 | 193 +++++
 drivers/net/wireguard/peer.h                 |  82 ++
 drivers/net/wireguard/queueing.c             |  53 ++
 drivers/net/wireguard/queueing.h             | 196 +++++
 drivers/net/wireguard/ratelimiter.c          | 220 ++++++
 drivers/net/wireguard/ratelimiter.h          |  19 +
 drivers/net/wireguard/receive.c              | 597 ++++++++++++++
 drivers/net/wireguard/selftest/allowedips.c  | 687 ++++++++++++++++
 drivers/net/wireguard/selftest/counter.c     | 104 +++
 drivers/net/wireguard/selftest/ratelimiter.c | 229 ++++++
 drivers/net/wireguard/send.c                 | 423 ++++++++++
 drivers/net/wireguard/socket.c               | 433 ++++++++++
 drivers/net/wireguard/socket.h               |  44 ++
 drivers/net/wireguard/timers.c               | 271 +++++++
 drivers/net/wireguard/timers.h               |  31 +
 drivers/net/wireguard/version.h              |   1 +
 include/uapi/linux/wireguard.h               | 190 +++++
 tools/testing/selftests/wireguard/netns.sh   | 499 ++++++++++++
 36 files changed, 7592 insertions(+)
 create mode 100644 drivers/net/wireguard/Makefile
 create mode 100644 drivers/net/wireguard/allowedips.c
 create mode 100644 drivers/net/wireguard/allowedips.h
 create mode 100644 drivers/net/wireguard/cookie.c
 create mode 100644 drivers/net/wireguard/cookie.h
 create mode 100644 drivers/net/wireguard/device.c
 create mode 100644 drivers/net/wireguard/device.h
 create mode 100644 drivers/net/wireguard/hashtables.c
 create mode 100644 drivers/net/wireguard/hashtables.h
 create mode 100644 drivers/net/wireguard/main.c
 create mode 100644 drivers/net/wireguard/messages.h
 create mode 100644 drivers/net/wireguard/netlink.c
 create mode 100644 drivers/net/wireguard/netlink.h
 create mode 100644 drivers/net/wireguard/noise.c
 create mode 100644 drivers/net/wireguard/noise.h
 create mode 100644 drivers/net/wireguard/peer.c
 create mode 100644 drivers/net/wireguard/peer.h
 create mode 100644 drivers/net/wireguard/queueing.c
 create mode 100644 drivers/net/wireguard/queueing.h
 create mode 100644 drivers/net/wireguard/ratelimiter.c
 create mode 100644 drivers/net/wireguard/ratelimiter.h
 create mode 100644 drivers/net/wireguard/receive.c
 create mode 100644 drivers/net/wireguard/selftest/allowedips.c
 create mode 100644 drivers/net/wireguard/selftest/counter.c
 create mode 100644 drivers/net/wireguard/selftest/ratelimiter.c
 create mode 100644 drivers/net/wireguard/send.c
 create mode 100644 drivers/net/wireguard/socket.c
 create mode 100644 drivers/net/wireguard/socket.h
 create mode 100644 drivers/net/wireguard/timers.c
 create mode 100644 drivers/net/wireguard/timers.h
 create mode 100644 drivers/net/wireguard/version.h
 create mode 100644 include/uapi/linux/wireguard.h
 create mode 100755 tools/testing/selftests/wireguard/netns.sh
MAINTAINERS +8 -0
index c0d078d4dbd8..59606218b750 100644
@@ -15873,6 +15873,14 @@ L:	[email protected]
 S:	Maintained
 F:	drivers/gpio/gpio-ws16c48.c
 
+WIREGUARD SECURE NETWORK TUNNEL
+M:	Jason A. Donenfeld <[email protected]>
+S:	Maintained
+F:	drivers/net/wireguard/
+F:	tools/testing/selftests/wireguard/
+L:	[email protected]
+L:	[email protected]
+
 WISTRON LAPTOP BUTTON DRIVER
 M:	Miloslav Trmac <[email protected]>
 S:	Maintained
drivers/net/Kconfig +30 -0
index d03775100f7d..d4b2568f86f9 100644
@@ -70,6 +70,36 @@ config DUMMY
 	  To compile this driver as a module, choose M here: the module
 	  will be called dummy.
 
+config WIREGUARD
+	tristate "WireGuard secure network tunnel"
+	depends on NET && INET
+	depends on IPV6 || !IPV6
+	select NET_UDP_TUNNEL
+	select DST_CACHE
+	select ZINC_CHACHA20POLY1305
+	select ZINC_BLAKE2S
+	select ZINC_CURVE25519
+	help
+	  WireGuard is a secure, fast, and easy to use replacement for IPSec
+	  that uses modern cryptography and clever networking tricks. It's
+	  designed to be fairly general purpose and abstract enough to fit most
+	  use cases, while at the same time remaining extremely simple to
+	  configure. See www.wireguard.com for more info.
+
+	  It's safe to say Y or M here, as the driver is very lightweight and
+	  is only in use when an administrator chooses to add an interface.
+
+config WIREGUARD_DEBUG
+	bool "Debugging checks and verbose messages"
+	depends on WIREGUARD
+	help
+	  This will write log messages for handshake and other events
+	  that occur for a WireGuard interface. It will also perform some
+	  extra validation checks and unit tests at various points. This is
+	  only useful for debugging.
+
+	  Say N here unless you know what you're doing.
+
 config EQUALIZER
 	tristate "EQL (serial line load balancing) support"
 	---help---
drivers/net/Makefile +1 -0
index 21cde7e78621..f0acd11a143d 100644
@@ -24,6 +24,7 @@ obj-$(CONFIG_RIONET) += rionet.o
 obj-$(CONFIG_NET_TEAM) += team/
 obj-$(CONFIG_TUN) += tun.o
 obj-$(CONFIG_TAP) += tap.o
+obj-$(CONFIG_WIREGUARD) += wireguard/
 obj-$(CONFIG_VETH) += veth.o
 obj-$(CONFIG_VIRTIO_NET) += virtio_net.o
 obj-$(CONFIG_VXLAN) += vxlan.o
drivers/net/wireguard/Makefile new file +18 -0
new file mode 100644
index 000000000000..d8856255bc9d
@@ -0,0 +1,18 @@
+ccflags-y := -O3
+ccflags-y += -D'pr_fmt(fmt)=KBUILD_MODNAME ": " fmt'
+ccflags-$(CONFIG_WIREGUARD_DEBUG) += -DDEBUG
+wireguard-y := main.o
+wireguard-y += noise.o
+wireguard-y += device.o
+wireguard-y += peer.o
+wireguard-y += timers.o
+wireguard-y += queueing.o
+wireguard-y += send.o
+wireguard-y += receive.o
+wireguard-y += socket.o
+wireguard-y += hashtables.o
+wireguard-y += allowedips.o
+wireguard-y += ratelimiter.o
+wireguard-y += cookie.o
+wireguard-y += netlink.o
+obj-$(CONFIG_WIREGUARD) := wireguard.o
drivers/net/wireguard/allowedips.c new file +413 -0

413 lines not on this page · raw message

drivers/net/wireguard/allowedips.h new file +54 -0

54 lines not on this page · raw message

drivers/net/wireguard/cookie.c new file +236 -0

236 lines not on this page · raw message

drivers/net/wireguard/cookie.h new file +59 -0

59 lines not on this page · raw message

drivers/net/wireguard/device.c new file +441 -0

441 lines not on this page · raw message

drivers/net/wireguard/device.h new file +65 -0

65 lines not on this page · raw message

drivers/net/wireguard/hashtables.c new file +209 -0

209 lines not on this page · raw message

drivers/net/wireguard/hashtables.h new file +64 -0

64 lines not on this page · raw message

drivers/net/wireguard/main.c new file +64 -0

64 lines not on this page · raw message

drivers/net/wireguard/messages.h new file +128 -0

128 lines not on this page · raw message

drivers/net/wireguard/netlink.c new file +599 -0

599 lines not on this page · raw message

drivers/net/wireguard/netlink.h new file +12 -0
new file mode 100644
index 000000000000..1dc6a67bc719
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2015-2018 Jason A. Donenfeld <[email protected]>. All Rights Reserved.
+ */
+
+#ifndef _WG_NETLINK_H
+#define _WG_NETLINK_H
+
+int wg_genetlink_init(void);
+void wg_genetlink_uninit(void);
+
+#endif /* _WG_NETLINK_H */
drivers/net/wireguard/noise.c new file +788 -0

581 lines not on this page · raw message

drivers/net/wireguard/noise.h new file +131 -0

131 lines not on this page · raw message

drivers/net/wireguard/peer.c new file +193 -0

193 lines not on this page · raw message

drivers/net/wireguard/peer.h new file +82 -0

82 lines not on this page · raw message

drivers/net/wireguard/queueing.c new file +53 -0

53 lines not on this page · raw message

drivers/net/wireguard/queueing.h new file +196 -0

196 lines not on this page · raw message

drivers/net/wireguard/ratelimiter.c new file +220 -0

220 lines not on this page · raw message

drivers/net/wireguard/ratelimiter.h new file +19 -0
new file mode 100644
index 000000000000..0325d10b1f76

Lines of this file are not shown. View the raw message.

drivers/net/wireguard/receive.c new file +597 -0

597 lines not on this page · raw message

drivers/net/wireguard/selftest/allowedips.c new file +687 -0

687 lines not on this page · raw message

drivers/net/wireguard/selftest/counter.c new file +104 -0

104 lines not on this page · raw message

drivers/net/wireguard/selftest/ratelimiter.c new file +229 -0

229 lines not on this page · raw message

drivers/net/wireguard/send.c new file +423 -0

423 lines not on this page · raw message

drivers/net/wireguard/socket.c new file +433 -0

433 lines not on this page · raw message

drivers/net/wireguard/socket.h new file +44 -0

44 lines not on this page · raw message

drivers/net/wireguard/timers.c new file +271 -0

271 lines not on this page · raw message

drivers/net/wireguard/timers.h new file +31 -0
new file mode 100644
index 000000000000..5f294612df23

Lines of this file are not shown. View the raw message.

drivers/net/wireguard/version.h new file +1 -0
new file mode 100644
index 000000000000..bbf71294d163

Lines of this file are not shown. View the raw message.

include/uapi/linux/wireguard.h new file +190 -0

190 lines not on this page · raw message

tools/testing/selftests/wireguard/netns.sh new file +499 -0

499 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.19.1

Part of the diff is not shown here. View the raw message.

#29 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

drivers/net/wireguard/allowedips.c line 44

Quoted 2 lines
drivers/net/wireguard/allowedips.c new file +2 -0
+#define choose_node(parent, key)                                               \
+	parent->bit[(key[parent->bit_at_a] >> parent->bit_at_b) & 1]

Hi Jason

This should be a function, not a macro.

drivers/net/wireguard/allowedips.c line 57

Quoted 13 lines
drivers/net/wireguard/allowedips.c new file +13 -0
+
+static void node_free_rcu(struct rcu_head *rcu)
+{
+	kfree(container_of(rcu, struct allowedips_node, rcu));
+}
+
+#define push_rcu(stack, p, len) ({                                             \
+		if (rcu_access_pointer(p)) {                                   \
+			WARN_ON(IS_ENABLED(DEBUG) && (len) >= 128);            \
+			stack[(len)++] = rcu_dereference_raw(p);               \
+		}                                                              \
+		true;                                                          \
+	})

This also looks like it could be a function.

Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+static void root_free_rcu(struct rcu_head *rcu)
+{
+	struct allowedips_node *node, *stack[128] = {
+		container_of(rcu, struct allowedips_node, rcu) };

drivers/net/wireguard/allowedips.c line 111

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+	unsigned int len = 1;
+
+	while (len > 0 && (node = stack[--len]) &&
+	       push_rcu(stack, node->bit[0], len) &&
+	       push_rcu(stack, node->bit[1], len))
+		kfree(node);
+}
+
+#define ref(p) rcu_access_pointer(p)
+#define deref(p) rcu_dereference_protected(*(p), lockdep_is_held(lock))

Macros should be uppercase, or better still, functions.

drivers/net/wireguard/allowedips.c line 115

Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+#define push(p) ({                                                             \
+		WARN_ON(IS_ENABLED(DEBUG) && len >= 128);                      \
+		stack[len++] = p;                                              \
+	})

This one definitely should be upper case, to warn readers it has unexpected side effects.

Quoted 41 lines
drivers/net/wireguard/allowedips.c new file +41 -0
+
+static void walk_remove_by_peer(struct allowedips_node __rcu **top,
+				struct wg_peer *peer, struct mutex *lock)
+{
+	struct allowedips_node __rcu **stack[128], **nptr;
+	struct allowedips_node *node, *prev;
+	unsigned int len;
+
+	if (unlikely(!peer || !ref(*top)))
+		return;
+
+	for (prev = NULL, len = 0, push(top); len > 0; prev = node) {
+		nptr = stack[len - 1];
+		node = deref(nptr);
+		if (!node) {
+			--len;
+			continue;
+		}
+		if (!prev || ref(prev->bit[0]) == node ||
+		    ref(prev->bit[1]) == node) {
+			if (ref(node->bit[0]))
+				push(&node->bit[0]);
+			else if (ref(node->bit[1]))
+				push(&node->bit[1]);
+		} else if (ref(node->bit[0]) == prev) {
+			if (ref(node->bit[1]))
+				push(&node->bit[1]);
+		} else {
+			if (rcu_dereference_protected(node->peer,
+				lockdep_is_held(lock)) == peer) {
+				RCU_INIT_POINTER(node->peer, NULL);
+				if (!node->bit[0] || !node->bit[1]) {
+					rcu_assign_pointer(*nptr,
+					deref(&node->bit[!ref(node->bit[0])]));
+					call_rcu_bh(&node->rcu, node_free_rcu);
+					node = deref(nptr);
+				}
+			}
+			--len;
+		}
+	}

drivers/net/wireguard/allowedips.c line 166

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+}
+
+#undef ref
+#undef deref
+#undef push
+
+static __always_inline unsigned int fls128(u64 a, u64 b)
+{
+	return a ? fls64(a) + 64U : fls64(b);
+}

Does the compiler actually get this wrong? Not inline it? There was an interesting LWN post about this recently:

https://lwn.net/Articles/767884/

But in general, inline of any form should be avoided in .c files.

Quoted 11 lines
drivers/net/wireguard/allowedips.c new file +11 -0
+
+static __always_inline u8 common_bits(const struct allowedips_node *node,
+				      const u8 *key, u8 bits)
+{
+	if (bits == 32)
+		return 32U - fls(*(const u32 *)node->bits ^ *(const u32 *)key);
+	else if (bits == 128)
+		return 128U - fls128(
+			*(const u64 *)&node->bits[0] ^ *(const u64 *)&key[0],
+			*(const u64 *)&node->bits[8] ^ *(const u64 *)&key[8]);
+	return 0;

drivers/net/wireguard/allowedips.c line 187

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+}
+
+/* This could be much faster if it actually just compared the common bits
+ * properly, by precomputing a mask bswap(~0 << (32 - cidr)), and the rest, but
+ * it turns out that common_bits is already super fast on modern processors,
+ * even taking into account the unfortunate bswap. So, we just inline it like
+ * this instead.
+ */
+#define prefix_matches(node, key, bits)                                        \
+	(common_bits(node, key, bits) >= (node)->cidr)

Could be a function.

Quoted 11 lines
drivers/net/wireguard/allowedips.c new file +11 -0
+
+static __always_inline struct allowedips_node *
+find_node(struct allowedips_node *trie, u8 bits, const u8 *key)
+{
+	struct allowedips_node *node = trie, *found = NULL;
+
+	while (node && prefix_matches(node, key, bits)) {
+		if (rcu_access_pointer(node->peer))
+			found = node;
+		if (node->cidr == bits)
+			break;

drivers/net/wireguard/allowedips.c line 208

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+		node = rcu_dereference_bh(choose_node(node, key));
+	}
+	return found;
+}
+
+/* Returns a strong reference to a peer */
+static __always_inline struct wg_peer *
+lookup(struct allowedips_node __rcu *root, u8 bits, const void *be_ip)
+{
+	u8 ip[16] __aligned(__alignof(u64));

You virtually never see aligned stack variables. This needs some sort of comment.

Quoted 15 lines
drivers/net/wireguard/allowedips.c new file +15 -0
+	struct allowedips_node *node;
+	struct wg_peer *peer = NULL;
+
+	swap_endian(ip, be_ip, bits);
+
+	rcu_read_lock_bh();
+retry:
+	node = find_node(rcu_dereference_bh(root), bits, ip);
+	if (node) {
+		peer = wg_peer_get_maybe_zero(rcu_dereference_bh(node->peer));
+		if (!peer)
+			goto retry;
+	}
+	rcu_read_unlock_bh();
+	return peer;

drivers/net/wireguard/allowedips.c line 233

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+}
+
+__attribute__((nonnull(1))) static bool
+node_placement(struct allowedips_node __rcu *trie, const u8 *key, u8 cidr,
+	       u8 bits, struct allowedips_node **rnode, struct mutex *lock)
+{
+	struct allowedips_node *node = rcu_dereference_protected(trie,
+						lockdep_is_held(lock));
+	struct allowedips_node *parent = NULL;
+	bool exact = false;

Should there be a WARN_ON(!key) here, since the attribute will only detect problems at compile time, and maybe some runtime cases will get passed it?

Quoted 37 lines
(quoted diff) +37 -0
+
+	while (node && node->cidr <= cidr && prefix_matches(node, key, bits)) {
+		parent = node;
+		if (parent->cidr == cidr) {
+			exact = true;
+			break;
+		}
+		node = rcu_dereference_protected(choose_node(parent, key),
+						 lockdep_is_held(lock));
+	}
+	*rnode = parent;
+	return exact;
+}
+void wg_cookie_message_consume(struct message_handshake_cookie *src,
+			       struct wg_device *wg)
+{
+	struct wg_peer *peer = NULL;
+	u8 cookie[COOKIE_LEN];
+	bool ret;
+
+	if (unlikely(!wg_index_hashtable_lookup(&wg->index_hashtable,
+					     INDEX_HASHTABLE_HANDSHAKE |
+					     INDEX_HASHTABLE_KEYPAIR,
+					     src->receiver_index, &peer)))
+		return;
+
+	down_read(&peer->latest_cookie.lock);
+	if (unlikely(!peer->latest_cookie.have_sent_mac1)) {
+		up_read(&peer->latest_cookie.lock);
+		goto out;
+	}
+	ret = xchacha20poly1305_decrypt(
+		cookie, src->encrypted_cookie, sizeof(src->encrypted_cookie),
+		peer->latest_cookie.last_mac1_sent, COOKIE_LEN, src->nonce,
+		peer->latest_cookie.cookie_decryption_key);
+	up_read(&peer->latest_cookie.lock);
+

Review

Quoted 10 lines
(quoted diff) +10 -0
+	if (ret) {
+		down_write(&peer->latest_cookie.lock);
+		memcpy(peer->latest_cookie.cookie, cookie, COOKIE_LEN);
+		peer->latest_cookie.birthdate = ktime_get_boot_fast_ns();
+		peer->latest_cookie.is_valid = true;
+		peer->latest_cookie.have_sent_mac1 = false;
+		up_write(&peer->latest_cookie.lock);
+	} else {
+		net_dbg_ratelimited("%s: Could not decrypt invalid cookie response\n",
+				    wg->dev->name);

It might be worth adding a netdev_dbg_ratelimited(), which takes a netdev as its first parameter, just line netdev_dbg().

Quoted 23 lines
drivers/net/wireguard/device.c new file +23 -0
+static int wg_open(struct net_device *dev)
+{
+	struct in_device *dev_v4 = __in_dev_get_rtnl(dev);
+	struct inet6_dev *dev_v6 = __in6_dev_get(dev);
+	struct wg_device *wg = netdev_priv(dev);
+	struct wg_peer *peer;
+	int ret;
+
+	if (dev_v4) {
+		/* At some point we might put this check near the ip_rt_send_
+		 * redirect call of ip_forward in net/ipv4/ip_forward.c, similar
+		 * to the current secpath check.
+		 */
+		IN_DEV_CONF_SET(dev_v4, SEND_REDIRECTS, false);
+		IPV4_DEVCONF_ALL(dev_net(dev), SEND_REDIRECTS) = false;
+	}
+	if (dev_v6)
+		dev_v6->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_NONE;
+
+	ret = wg_socket_init(wg, wg->incoming_port);
+	if (ret < 0)
+		return ret;
+	mutex_lock(&wg->device_update_lock);

drivers/net/wireguard/device.c line 61

Quoted 10 lines
drivers/net/wireguard/device.c new file +10 -0
+	list_for_each_entry(peer, &wg->peer_list, peer_list) {
+		wg_packet_send_staged_packets(peer);
+		if (peer->persistent_keepalive_interval)
+			wg_packet_send_keepalive(peer);
+	}
+	mutex_unlock(&wg->device_update_lock);
+	return 0;
+}
+
+#if defined(CONFIG_PM_SLEEP) && !defined(CONFIG_ANDROID)

I don't see any other code which uses this combination. Why is this needed?

Quoted 66 lines
drivers/net/wireguard/device.c new file +66 -0
+static netdev_tx_t wg_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct wg_device *wg = netdev_priv(dev);
+	struct sk_buff_head packets;
+	struct wg_peer *peer;
+	struct sk_buff *next;
+	sa_family_t family;
+	u32 mtu;
+	int ret;
+
+	if (unlikely(wg_skb_examine_untrusted_ip_hdr(skb) != skb->protocol)) {
+		ret = -EPROTONOSUPPORT;
+		net_dbg_ratelimited("%s: Invalid IP packet\n", dev->name);
+		goto err;
+	}
+
+	peer = wg_allowedips_lookup_dst(&wg->peer_allowedips, skb);
+	if (unlikely(!peer)) {
+		ret = -ENOKEY;
+		if (skb->protocol == htons(ETH_P_IP))
+			net_dbg_ratelimited("%s: No peer has allowed IPs matching %pI4\n",
+					    dev->name, &ip_hdr(skb)->daddr);
+		else if (skb->protocol == htons(ETH_P_IPV6))
+			net_dbg_ratelimited("%s: No peer has allowed IPs matching %pI6\n",
+					    dev->name, &ipv6_hdr(skb)->daddr);
+		goto err;
+	}
+
+	family = READ_ONCE(peer->endpoint.addr.sa_family);
+	if (unlikely(family != AF_INET && family != AF_INET6)) {
+		ret = -EDESTADDRREQ;
+		net_dbg_ratelimited("%s: No valid endpoint has been configured or discovered for peer %llu\n",
+				    dev->name, peer->internal_id);
+		goto err_peer;
+	}
+
+	mtu = skb_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
+
+	__skb_queue_head_init(&packets);
+	if (!skb_is_gso(skb)) {
+		skb->next = NULL;
+	} else {
+		struct sk_buff *segs = skb_gso_segment(skb, 0);
+
+		if (unlikely(IS_ERR(segs))) {
+			ret = PTR_ERR(segs);
+			goto err_peer;
+		}
+		dev_kfree_skb(skb);
+		skb = segs;
+	}
+	do {
+		next = skb->next;
+		skb->next = skb->prev = NULL;
+
+		skb = skb_share_check(skb, GFP_ATOMIC);
+		if (unlikely(!skb))
+			continue;
+
+		/* We only need to keep the original dst around for icmp,
+		 * so at this point we're in a position to drop it.
+		 */
+		skb_dst_drop(skb);
+
+		PACKET_CB(skb)->mtu = mtu;
+

drivers/net/wireguard/device.c line 186

Quoted 10 lines
drivers/net/wireguard/device.c new file +10 -0
+		__skb_queue_tail(&packets, skb);
+	} while ((skb = next) != NULL);
+
+	spin_lock_bh(&peer->staged_packet_queue.lock);
+	/* If the queue is getting too big, we start removing the oldest packets
+	 * until it's small again. We do this before adding the new packet, so
+	 * we don't remove GSO segments that are in excess.
+	 */
+	while (skb_queue_len(&peer->staged_packet_queue) > MAX_STAGED_PACKETS)
+		dev_kfree_skb(__skb_dequeue(&peer->staged_packet_queue));

It would be good to have some stats counters in here. Maybe the standard interface statistics will cover it, otherwise ethtool -S.

You should also get this code looked at by some of the queueing people. Rather than discarding, you might want to be applying back pressure to slow down the sender application?

Quoted 4 lines
drivers/net/wireguard/noise.c new file +4 -0
+static void kdf(u8 *first_dst, u8 *second_dst, u8 *third_dst, const u8 *data,
+		size_t first_len, size_t second_len, size_t third_len,
+		size_t data_len, const u8 chaining_key[NOISE_HASH_LEN])
+{

drivers/net/wireguard/noise.c line 300

Quoted 10 lines
drivers/net/wireguard/noise.c new file +10 -0
+	u8 output[BLAKE2S_HASH_SIZE + 1];
+	u8 secret[BLAKE2S_HASH_SIZE];
+
+	WARN_ON(IS_ENABLED(DEBUG) &&
+		(first_len > BLAKE2S_HASH_SIZE ||
+		 second_len > BLAKE2S_HASH_SIZE ||
+		 third_len > BLAKE2S_HASH_SIZE ||
+		 ((second_len || second_dst || third_len || third_dst) &&
+		  (!first_len || !first_dst)) ||
+		 ((third_len || third_dst) && (!second_len || !second_dst))));

Maybe split this up into a number of WARN_ON()s. At the moment, if it fires, you have little idea why, there are so many comparisons. Also, is this on the hot path? I guess not, since this is keys, not packets. Do you need to care about DEBUG here?

Andrew

author maintainer
#30 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Hi Andrew,

Thanks for the review. Comments and fix links are inline below.

Quoted 4 lines · Andrew Lunn

drivers/net/wireguard/allowedips.c line 44

Quoted 3 lines
drivers/net/wireguard/allowedips.c new file +2 -0
+#define choose_node(parent, key)                                               \
+     parent->bit[(key[parent->bit_at_a] >> parent->bit_at_b) & 1]

This should be a function, not a macro.

That prevents it from being used as an lvalue, which is mostly where it's used, with the rcu-protected structure, so this will remain a macro. I'll make it upper-case though.

Quoted 7 lines
Quoted 7 lines
drivers/net/wireguard/allowedips.c new file +7 -0
+
+static void node_free_rcu(struct rcu_head *rcu)
+{
+     kfree(container_of(rcu, struct allowedips_node, rcu));
+}
+
+#define push_rcu(stack, p, len) ({                                             \
Quoted 8 lines

drivers/net/wireguard/allowedips.c line 57

Quoted 6 lines
drivers/net/wireguard/allowedips.c new file +6 -0
+             if (rcu_access_pointer(p)) {                                   \
+                     WARN_ON(IS_ENABLED(DEBUG) && (len) >= 128);            \
+                     stack[(len)++] = rcu_dereference_raw(p);               \
+             }                                                              \
+             true;                                                          \
+     })

This also looks like it could be a function.

You're right. I've changed this now. That produces slightly worse code on gcc 8.2, but that's not really hot path anyway, so it doesn't matter.

Quoted 17 lines
Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+static void root_free_rcu(struct rcu_head *rcu)
+{
+     struct allowedips_node *node, *stack[128] = {
+             container_of(rcu, struct allowedips_node, rcu) };

drivers/net/wireguard/allowedips.c line 111

Quoted 10 lines
drivers/net/wireguard/allowedips.c new file +10 -0
+     unsigned int len = 1;
+
+     while (len > 0 && (node = stack[--len]) &&
+            push_rcu(stack, node->bit[0], len) &&
+            push_rcu(stack, node->bit[1], len))
+             kfree(node);
+}
+
+#define ref(p) rcu_access_pointer(p)
+#define deref(p) rcu_dereference_protected(*(p), lockdep_is_held(lock))

Macros should be uppercase, or better still, functions.

Quoted 7 lines

drivers/net/wireguard/allowedips.c line 115

Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+#define push(p) ({                                                             \
+             WARN_ON(IS_ENABLED(DEBUG) && len >= 128);                      \
+             stack[len++] = p;                                              \
+     })

This one definitely should be upper case, to warn readers it has unexpected side effects.

I've made these little helper macros uppercase.

Quoted 46 lines
Quoted 46 lines
drivers/net/wireguard/allowedips.c new file +45 -0
+
+static void walk_remove_by_peer(struct allowedips_node __rcu **top,
+                             struct wg_peer *peer, struct mutex *lock)
+{
+     struct allowedips_node __rcu **stack[128], **nptr;
+     struct allowedips_node *node, *prev;
+     unsigned int len;
+
+     if (unlikely(!peer || !ref(*top)))
+             return;
+
+     for (prev = NULL, len = 0, push(top); len > 0; prev = node) {
+             nptr = stack[len - 1];
+             node = deref(nptr);
+             if (!node) {
+                     --len;
+                     continue;
+             }
+             if (!prev || ref(prev->bit[0]) == node ||
+                 ref(prev->bit[1]) == node) {
+                     if (ref(node->bit[0]))
+                             push(&node->bit[0]);
+                     else if (ref(node->bit[1]))
+                             push(&node->bit[1]);
+             } else if (ref(node->bit[0]) == prev) {
+                     if (ref(node->bit[1]))
+                             push(&node->bit[1]);
+             } else {
+                     if (rcu_dereference_protected(node->peer,
+                             lockdep_is_held(lock)) == peer) {
+                             RCU_INIT_POINTER(node->peer, NULL);
+                             if (!node->bit[0] || !node->bit[1]) {
+                                     rcu_assign_pointer(*nptr,
+                                     deref(&node->bit[!ref(node->bit[0])]));
+                                     call_rcu_bh(&node->rcu, node_free_rcu);
+                                     node = deref(nptr);
+                             }
+                     }
+                     --len;
+             }
+     }
+}
+
+#undef ref
+#undef deref
Quoted 8 lines

drivers/net/wireguard/allowedips.c line 166

Quoted 6 lines
drivers/net/wireguard/allowedips.c new file +6 -0
+#undef push
+
+static __always_inline unsigned int fls128(u64 a, u64 b)
+{
+     return a ? fls64(a) + 64U : fls64(b);
+}

Does the compiler actually get this wrong? Not inline it?

Actually for this function (in contrast with all of the other ones marked as such, which really must have the annotation), recent gcc gets it right, so I've removed the annotation. Nice catch.

Quoted 15 lines
Quoted 15 lines
drivers/net/wireguard/allowedips.c new file +15 -0
+
+static __always_inline u8 common_bits(const struct allowedips_node *node,
+                                   const u8 *key, u8 bits)
+{
+     if (bits == 32)
+             return 32U - fls(*(const u32 *)node->bits ^ *(const u32 *)key);
+     else if (bits == 128)
+             return 128U - fls128(
+                     *(const u64 *)&node->bits[0] ^ *(const u64 *)&key[0],
+                     *(const u64 *)&node->bits[8] ^ *(const u64 *)&key[8]);
+     return 0;
+}
+
+/* This could be much faster if it actually just compared the common bits
+ * properly, by precomputing a mask bswap(~0 << (32 - cidr)), and the rest, but
Quoted 8 lines

drivers/net/wireguard/allowedips.c line 187

Quoted 6 lines

+ * it turns out that common_bits is already super fast on modern processors,
+ * even taking into account the unfortunate bswap. So, we just inline it like
+ * this instead.
+ */

drivers/net/wireguard/allowedips.c new file +2 -0
+#define prefix_matches(node, key, bits)                                        \
+     (common_bits(node, key, bits) >= (node)->cidr)

Could be a function.

I've made this a function now and confirmed codegen did not change significantly.

Quoted 8 lines

drivers/net/wireguard/allowedips.c line 208

Quoted 5 lines
drivers/net/wireguard/allowedips.c new file +5 -0
+/* Returns a strong reference to a peer */
+static __always_inline struct wg_peer *
+lookup(struct allowedips_node __rcu *root, u8 bits, const void *be_ip)
+{
+     u8 ip[16] __aligned(__alignof(u64));

You virtually never see aligned stack variables. This needs some sort of comment.

I've added a comment. The reason, if you're curious, is so that we can pass it to fls64 on all platforms.

Quoted 4 lines
Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+__attribute__((nonnull(1))) static bool
+node_placement(struct allowedips_node __rcu *trie, const u8 *key, u8 cidr,
+            u8 bits, struct allowedips_node **rnode, struct mutex *lock)
+{
Quoted 8 lines

drivers/net/wireguard/allowedips.c line 233

Quoted 4 lines
drivers/net/wireguard/allowedips.c new file +4 -0
+     struct allowedips_node *node = rcu_dereference_protected(trie,
+                                             lockdep_is_held(lock));
+     struct allowedips_node *parent = NULL;
+     bool exact = false;

Should there be a WARN_ON(!key) here, since the attribute will only detect problems at compile time, and maybe some runtime cases will get passed it?

Actually no, it's only used in one place, and that function is pretty clear about checking beforehand, and even this function checks implicitly. Looking at my commit log, I had added the annotation to make clang-analyzer happy, because it couldn't (at the time) deduce things correctly from rcu_access_pointer(p) checks. It looks like recent clang-analyzer has fixed that, and besides I'm not sure the kernel is in the business of adding annotations to work around static-analyzer-bug-of-the-week. So I've gotten rid of the annotation. Thanks for bringing my attention to it.

Quoted 5 lines

drivers/net/wireguard/cookie.c line 231

Quoted 2 lines
drivers/net/wireguard/cookie.c new file +2 -0
+             net_dbg_ratelimited("%s: Could not decrypt invalid cookie response\n",
+                                 wg->dev->name);

It might be worth adding a netdev_dbg_ratelimited(), which takes a netdev as its first parameter, just line netdev_dbg().

That sounds like it could be useful. Though, I'm trying hard to make the first patch submission _not_ need to touch any of the rest of the networking stack. I've actually got a small list of interesting networking stack changes that might be useful for WireGuard, but I think I'd prefer to submit these after this is all merged, and each one separately for a separate discussion, if that's okay with you.

Quoted 5 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+#if defined(CONFIG_PM_SLEEP) && !defined(CONFIG_ANDROID)

I don't see any other code which uses this combination. Why is this needed?

WireGuard clears private key material before going to sleep, so that ephemeral keys never live longer in ram than their expiration date. This works well for mostly everything, except Android devices where crazy out-of-tree wireless drivers (such as qcacld) will actually wake and sleep the phone constantly, for a few milliseconds each wake, to handle incoming packets. In this case, we must not zero out ephemeral keys, so that these packets can actually be decrypted. There are no other systems that have this characteristic, as far as I can tell, and I'm certainly not willing to add a runtime configuration nob for that, and it turns out that matching on CONFIG_ANDROID as such does the right thing all of the time. If the landscape becomes more complicated, we can visit it, probably in the form of passing reasons through to pm_notification and related scaffolding. But in lieu of this infrastructure someday existing, I prefer to do the simple-and-always-works-anyway solution.

Quoted 5 lines

drivers/net/wireguard/device.c line 186

Quoted 2 lines
drivers/net/wireguard/device.c new file +2 -0
+     while (skb_queue_len(&peer->staged_packet_queue) > MAX_STAGED_PACKETS)
+             dev_kfree_skb(__skb_dequeue(&peer->staged_packet_queue));

It would be good to have some stats counters in here. Maybe the standard interface statistics will cover it, otherwise ethtool -S.

Good point, I need to increment the drop counters there. I'll add that; good catch.

Willy and I actually discussed around a year ago adding a series of wireguard-specific error counters for all the various unusual things that can happen. This is something I'll probably investigate in earnest post-merge, as it's a bit of a "bell and whistle" thing, and the existing counters now are already sufficient for most purposes.

Quoted 3 lines

You should also get this code looked at by some of the queueing people. Rather than discarding, you might want to be applying back pressure to slow down the sender application?

Actually Toke, DaveT, and I have agonized quite a bit over the queueing mechanisms in WireGuard, as well as two GSoC summers looking at it with students. There's been lots of flent and rrul and tweaking involved, and the entire queueing system has gone through probably 15 different revisions of trying things out and experimenting. At this point, I think the way we're doing queueing and when we're dropping, and those various factors, are taken care of pretty well. Future work still involves considering shoehorning some fq_codel in a similar way that wireless does, and a few other tweeks, but I think the current system is good and stable and very much sufficient.

The particular instance you're pointing out though, with staged_packet_queue, is not the relevant queue and isn't what you have in mind with regards to backpressure. This is simply a very short lived thing for dealing with packets before a handshake is made, and for handling xmit being used from multiple cores, but it's not _the_ queueing system that's relevant in the actual buffering and latency calculations.

A little bit of the queueing stuff is discussed in the paper I presented at Netdev 2.2, if you're interested:
https://www.wireguard.com/papers/wireguard-netdev22.pdf Some of that is outdated by now, but should give you some idea.

Quoted 11 lines
Quoted 11 lines
drivers/net/wireguard/noise.c new file +11 -0
+static void kdf(u8 *first_dst, u8 *second_dst, u8 *third_dst, const u8 *data,
+             size_t first_len, size_t second_len, size_t third_len,
+             size_t data_len, const u8 chaining_key[NOISE_HASH_LEN])
+{
+     u8 output[BLAKE2S_HASH_SIZE + 1];
+     u8 secret[BLAKE2S_HASH_SIZE];
+
+     WARN_ON(IS_ENABLED(DEBUG) &&
+             (first_len > BLAKE2S_HASH_SIZE ||
+              second_len > BLAKE2S_HASH_SIZE ||
+              third_len > BLAKE2S_HASH_SIZE ||
Quoted 8 lines

drivers/net/wireguard/noise.c line 300

Quoted 3 lines
drivers/net/wireguard/noise.c new file +3 -0
+              ((second_len || second_dst || third_len || third_dst) &&
+               (!first_len || !first_dst)) ||
+              ((third_len || third_dst) && (!second_len || !second_dst))));

Maybe split this up into a number of WARN_ON()s. At the moment, if it fires, you have little idea why, there are so many comparisons. Also, is this on the hot path? I guess not, since this is keys, not packets. Do you need to care about DEBUG here?

This is on the hot path, actually. Well, it's not on path of data packets, but I do consider handshake packets to be fairly "warm". I'm not especially keen on splitting that up into multiple warn_ons, mostly because if that is ever hint, something is seriously wrong, and the whole flow would likely then need auditing anyway.

Regards,
Jason

#31 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Quoted 6 lines
Quoted 5 lines

drivers/net/wireguard/cookie.c line 231

Quoted 2 lines
drivers/net/wireguard/cookie.c new file +2 -0
+             net_dbg_ratelimited("%s: Could not decrypt invalid cookie response\n",
+                                 wg->dev->name);

It might be worth adding a netdev_dbg_ratelimited(), which takes a netdev as its first parameter, just line netdev_dbg().

Quoted 6 lines

That sounds like it could be useful. Though, I'm trying hard to make the first patch submission _not_ need to touch any of the rest of the networking stack. I've actually got a small list of interesting networking stack changes that might be useful for WireGuard, but I think I'd prefer to submit these after this is all merged, and each one separately for a separate discussion, if that's okay with you.

Hi Jason

I can understand that. But on the flip side, CAKE reached something like version 19 before it got merged. Wireguard is looking similar. An addition like the above, is not controversial. You could submit such a single patch in a weeks time when net-next reopens, and probably version 1 or 2 will get accepted.

Andrew

author maintainer
#32 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Hi Andrew,

Quoted 5 lines · Andrew Lunn

I can understand that. But on the flip side, CAKE reached something like version 19 before it got merged. Wireguard is looking similar. An addition like the above, is not controversial. You could submit such a single patch in a weeks time when net-next reopens, and probably version 1 or 2 will get accepted.

Alright, maybe I'll give it a try that way. Previously, every time I wanted to send things upstream for WireGuard, I'd be sure to find other users in the tree of what I was adding, since WireGuard wasn't ready for patchsets yet. But I guess now that some series have begun, it's more acceptable to add APIs here and there as you described.

Jason

#33 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

In reply to Jason A. Donenfeld
Quoted 10 lines
Quoted 4 lines
Quoted 1 line

+#if defined(CONFIG_PM_SLEEP) && !defined(CONFIG_ANDROID)

I don't see any other code which uses this combination. Why is this needed?

WireGuard clears private key material before going to sleep, so that ephemeral keys never live longer in ram than their expiration date. This works well for mostly everything, except Android devices where crazy out-of-tree wireless drivers (such as qcacld) will actually wake and sleep the phone constantly

Hi Jason

Out of tree is important here. To some degree, mainline does not care about out of tree drivers. Putting in a bandaid for them does not help get them fixed.

I would drop this bandaid. If the Android community decides to move wireguard from mainline into their fork, they can put the bandaid back in again, or get the driver fixed.

Andrew

author maintainer
#34 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Hi Andrew,

Quoted 7 lines · Andrew Lunn

Out of tree is important here. To some degree, mainline does not care about out of tree drivers. Putting in a bandaid for them does not help get them fixed.

I would drop this bandaid. If the Android community decides to move wireguard from mainline into their fork, they can put the bandaid back in again, or get the driver fixed.

No out-of-tree is really not important here, and sorry if mentioning that in my explanation was distracting and caused you to misunderstand. The issue isn't the crazy wireless driver; the issue is that suspend on Android happens all the time, and the entire Android way of using Linux is centered around non-stop suspending, with triggers to wake up (wifi drivers, for example, like the out-of-tree qcacld one, but presumably in tree ones too), and ways of controlling when it goes to sleep (screen blanking, wakelocks, etc). The Android model of Linux revolves around this, and hence the suspend semantics for WireGuard respect this model and adjust accordingly, using the appropriate CONFIG_ANDROID to determine which model we're operating under. This is not a bandaid, and it doesn't have to do with forks of the Linux kernel.

Jason

#35 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Quoted 6 lines · Jason A. Donenfeld

when it goes to sleep (screen blanking, wakelocks, etc). The Android model of Linux revolves around this, and hence the suspend semantics for WireGuard respect this model and adjust accordingly, using the appropriate CONFIG_ANDROID to determine which model we're operating under. This is not a bandaid, and it doesn't have to do with forks of the Linux kernel.

If that's what you are trying to conditionalize, why don't use CONFIG_PM_AUTOSLEEP? That way if there are other systems that want to use the Android wakelocks style of suspend management, your code will DTRT, as opposed to depending on CONFIG_ANDROID.

- Ted

author maintainer
#36 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Hey Ted,

Quoted 4 lines · Theodore Y. Ts'o

If that's what you are trying to conditionalize, why don't use CONFIG_PM_AUTOSLEEP? That way if there are other systems that want to use the Android wakelocks style of suspend management, your code will DTRT, as opposed to depending on CONFIG_ANDROID.

That's a terrific idea; thanks for the suggestion. I'll make that change for the next patchset version.

Jason

#37 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

In reply to Jason A. Donenfeld
Quoted 20 lines
Quoted 11 lines
Quoted 11 lines
drivers/net/wireguard/noise.c new file +11 -0
+static void kdf(u8 *first_dst, u8 *second_dst, u8 *third_dst, const u8 *data,
+             size_t first_len, size_t second_len, size_t third_len,
+             size_t data_len, const u8 chaining_key[NOISE_HASH_LEN])
+{
+     u8 output[BLAKE2S_HASH_SIZE + 1];
+     u8 secret[BLAKE2S_HASH_SIZE];
+
+     WARN_ON(IS_ENABLED(DEBUG) &&
+             (first_len > BLAKE2S_HASH_SIZE ||
+              second_len > BLAKE2S_HASH_SIZE ||
+              third_len > BLAKE2S_HASH_SIZE ||
Quoted 8 lines

drivers/net/wireguard/noise.c line 300

Quoted 3 lines
drivers/net/wireguard/noise.c new file +3 -0
+              ((second_len || second_dst || third_len || third_dst) &&
+               (!first_len || !first_dst)) ||
+              ((third_len || third_dst) && (!second_len || !second_dst))));

Maybe split this up into a number of WARN_ON()s. At the moment, if it fires, you have little idea why, there are so many comparisons. Also, is this on the hot path? I guess not, since this is keys, not packets. Do you need to care about DEBUG here?

Quoted 2 lines

This is on the hot path, actually. Well, it's not on path of data packets, but I do consider handshake packets to be fairly "warm".

Hi Jason

So for me, hot path is something called 10 million timers per second. How often do handshake packets happen?

Quoted 3 lines

I'm not especially keen on splitting that up into multiple warn_ons, mostly because if that is ever hint, something is seriously wrong, and the whole flow would likely then need auditing anyway.

On the flip side, if you think this is so unlikely, it probably means it is hard to reproduce. And then you will wish you knew exactly which triggered. If you can get the disassembly and stack trace you might be able to figure it out, but have a print could be a lot easier.

Anyway, it was just a suggestion. No worries.

Andrew

author maintainer
#38 lore raw

Re: [PATCH net-next v8 28/28] net: WireGuard secure network tunnel

Hi Andrew,

Quoted 5 lines · Andrew Lunn
Quoted 2 lines

This is on the hot path, actually. Well, it's not on path of data packets, but I do consider handshake packets to be fairly "warm".

So for me, hot path is something called 10 million timers per second. How often do handshake packets happen?

I can actually imagine situations in which this rate might happen. Handshake performance is important.

Jason