Skip to contentSkip to conversation

Cleanups and improvements

PATCH Superseded Mark Rutland 5 participants 43 messages Jan 14, 2022 – Jan 25, 2022 Kbuild

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover Cleanups and improvements

This series reworks the aarch64 boot-wrapper, moving a fair amount of initialization code to C. This has a few benefits:

1) This makes the code more legible and easier to maintain.

Current feature detection and system register configuration is written in assembly, requiring runs of *almost* identical blocks of assembly to read ID registers and conditionally initialize register values. This requires a bunch of labels (which are all named numerically), and all the magic numbers are hard coded, so this gets pretty painful to read:

| mrs x1, id_aa64isar0_el1
| ubfx x1, x1, #24, #4
| cbz x1, 1f
|
| orr x0, x0, #(1 << 34) // TME enable
|
| 1:

In C, it's much easier to add helpers which use mnemonics, which makes the code much easier to read, and avoids the need to manually allocate registers, etc:

| if (mrs_field(ID_AA64ISAR0_EL1, TME))
| scr |= SCR_EL3_TME;

This should make it easier to handle new architectural extensions (and/or architecture variants such as ARMv8-R) in future.

2) This allows for better diagnostics.

Currently a mis-configured boot-wrapper is rather unforgiving, and provides no indication as to what might have gone wrong. By moving initialization to C, we can make use to the UART code to log diagnostic information, and we can more easily add additional error handling and conditional logic.

This series adds diagnostic information and error handling that can be used to identify problems such as the boot-wrapper being launched at the wrong exception level:

| Boot-wrapper v0.2
| Entered at EL2
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper | [000000008000fff8..0000000080010000] => mbox | [0000000080200000..00000000822af200] => kernel | [0000000088000000..0000000088002857] => dtb |
| WARNING: PSCI could not be initialized. Boot may fail

Originally I had planned for this to be a more expansive set of changes, unifying some early control-flow, fixing some latent bugs, and making the boot-wrapper dynamically handle being entered at any of EL{3,2,1} with automated selection of a suitable DT. As the series has already become pretty long, I'd like to get this preparatory cleanup out of the way first, and handle those cases with subsequent patches.

If there are no objections, I intend to apply these by the end of the day.

I've pushed the series to the cleanup branch in the boot-wrapper repo:

https://git.kernel.org/pub/scm/linux/kernel/git/mark/boot-wrapper-aarch64.git/
git://git.kernel.org/pub/scm/linux/kernel/git/mark/boot-wrapper-aarch64.git

... and it should apply cleanup atop the master branch.

Since v1 [1]:
* Add comments for bit-field macros
* Use BIT() for *_RES1 definitions
* Complete start_el3/start_no_el3 cleanup
* Remove extraneous macro definition

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Mark.

Mark Rutland (13):
Document entry requirements
Add bit-field macros
aarch64: add system register accessors
aarch32: add coprocessor accessors
aarch64: add mov_64 macro
aarch64: initialize SCTLR_ELx for the boot-wrapper Rework common init C code
Announce boot-wrapper mode / exception level
aarch64: move the bulk of EL3 initialization to C
aarch32: move the bulk of Secure PL1 initialization to C Announce locations of memory objects
Rework bootmethod initialization
Unify start_el3 & start_no_el3

 Makefile.am                       |   6 +-
 arch/aarch32/boot.S               |  33 +++---
 arch/aarch32/include/asm/cpu.h    |  62 ++++++++---
 arch/aarch32/include/asm/gic-v3.h |   6 +-
 arch/aarch32/include/asm/psci.h   |  28 +++++
 arch/aarch32/init.c               |  42 ++++++++
 arch/aarch32/psci.S               |  16 +--
 arch/aarch32/utils.S              |   9 --
 arch/aarch64/boot.S               | 169 +++++++++++++-----------------
 arch/aarch64/common.S             |  10 +-
 arch/aarch64/include/asm/cpu.h    | 102 +++++++++++++++---
 arch/aarch64/include/asm/gic-v3.h |  10 +-
 arch/aarch64/include/asm/psci.h   |  28 +++++
 arch/aarch64/init.c               |  88 ++++++++++++++++
 arch/aarch64/psci.S               |  22 +---
 arch/aarch64/spin.S               |   6 +-
 arch/aarch64/utils.S              |   9 --
 common/boot.c                     |   4 -
 common/init.c                     |  60 +++++++++++
 common/platform.c                 |  45 +++++---
 common/psci.c                     |  22 +++-
 include/bits.h                    |  79 ++++++++++++++
 include/boot.h                    |   2 +
 include/platform.h                |  20 ++++
 model.lds.S                       |  20 ++--
 25 files changed, 668 insertions(+), 230 deletions(-)
 create mode 100644 arch/aarch32/include/asm/psci.h
 create mode 100644 arch/aarch32/init.c
 create mode 100644 arch/aarch64/include/asm/psci.h
 create mode 100644 arch/aarch64/init.c
 create mode 100644 common/init.c
 create mode 100644 include/bits.h
 create mode 100644 include/platform.h
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#1 lore raw

PATCH v2 1/13 Document entry requirements

Currently the boot-wrapper only supports some combinations of exception levels, with other combinations not being supported.

While we generally expect the boot-wrapper to be entered at the highest implemented exception level, the AArch32 boot-wrapper has a comment implying it supports being entered with something else owning EL3. As this would require such EL3 firmware to always be in sync with the boot-wrapper's requirements, which change over time, we don't actually support such a configuration.

Some CPU state (such as CNTFRQ/CNTFRQ_EL0) needs to be initialized at the highest implemented exception level, but today the boot-wrapper only does so when entered at EL3 / Secure-PL1. Thus, today the only completely supported configurations are EL3 / Secure-PL1, and entering in other configurations is not entirely supported.

The aarch64 jump_kernel function always writes to SCTLR_EL2, which is UNDEFINED at EL1. Hence, the aarch64 boot-wrapper does not support being entered at EL1.

The aarch32 code assumes that any non-hyp mode is Secure PL1, and attempt to switch to monitor mode. If entered on a system without the security extensions, where the highest privileged mode is Non-secure PL1, this will not work. Hence the aarch32 boot-wrapper does not support being entered at Non-secure PL1.

Actually supporting all of these configurations requires restructuring much of the boot-wrapper. For now, document the supported configurations in each architecture's boot.S, and remove the misleading comment from arch/aarch32/boot.S. Subsequent patches will improve the support and add support for additional configurations.

There should be no functional change as a result of this patch.

---
 arch/aarch32/boot.S | 14 ++++++++++++++
 arch/aarch32/psci.S |  5 -----
 arch/aarch64/boot.S | 13 +++++++++++++
 3 files changed, 27 insertions(+), 5 deletions(-)
arch/aarch32/boot.S +14 -0
index 4add338..00c432d 100644
@@ -16,6 +16,20 @@
 	.arch_extension virt
 
 	.section .init
+
+	/*
+	 * The boot-wrapper must be entered from the reset vector at the
+	 * highest implemented exception level. The boot-wrapper only supports
+	 * being entered in the following modes:
+	 *
+	 * - PL1 / EL3 (Secure) Supervisor mode
+	 *   Entering in this mode is strongly recommended.
+	 *   PL2 must be implemented.
+	 *
+	 * - PL2 / EL2 (Non-secure) Hypervisor mode
+	 *   Entering in this mode is partially supported.
+	 *   PSCI is not supported when entered in this mode.
+	 */
 ASM_FUNC(_start)
 	/* Stack initialisation */
 	cpuid	r0, r1
arch/aarch32/psci.S +0 -5
index dc7aeb7..e0d2972 100644
@@ -44,11 +44,6 @@ ASM_FUNC(start_el3)
 	/* pass through */
 
 ASM_FUNC(start_no_el3)
-	/*
-	 * For no-el3, we assume that firmware launched the boot-wrapper in
-	 * non-secure EL2 or EL1. We assume it has its own PSCI implementation
-	 * sitting at EL3, and that this path is only taken by primary CPU.
-	 */
 	cpuid	r0, r1
 	blx	find_logical_id
 	b	psci_first_spin
arch/aarch64/boot.S +13 -0
index 27ba449..900b9f8 100644
@@ -12,6 +12,19 @@
 
 	.section .init
 
+	/*
+	 * The boot-wrapper must be entered from the reset vector at the
+	 * highest implemented exception level. The boot-wrapper only supports
+	 * being entered at the following exception levels:
+	 *
+	 * - EL3 (Secure)
+	 *   Entering at EL3 is strongly recommended.
+	 *   EL2 must be implemented.
+	 *
+	 * - EL2 (Non-secure)
+	 *   Entering at EL2 is partially supported.
+	 *   PSCI is not supported when entered in this exception level.
+	 */
 ASM_FUNC(_start)
 	cpuid	x0, x1
 	bl	find_logical_id
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#2 lore raw

PATCH v2 2/13 Add bit-field macros

In reply to Mark Rutland

Arm architectural documentation typically defines bit-fields as [msb,lsb] and single-bit fields as [bit]. For clarity it would be helpful if we could define fields in the same way.

Add helpers so that we can do so, along with helper to extract/insert bit-field values.

There should be no functional change as a result of this patch.

---
 include/bits.h | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 79 insertions(+)
 create mode 100644 include/bits.h
include/bits.h new file +79 -0
new file mode 100644
index 0000000..0bf2c67
@@ -0,0 +1,79 @@
+/*
+ * include/bits.h - helpers for bit-field definitions.
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __BITS_H
+#define __BITS_H
+
+#ifdef __ASSEMBLY__
+#define UL(x)	x
+#define ULL(x)	x
+#else
+#define UL(x)	x##UL
+#define ULL(x)	x##ULL
+#endif
+
+/*
+ * Define a contiguous mask of bits with `msb` as the most significant bit and
+ * `lsb` as the least significant bit. The `msb` value must be greater than or
+ * equal to `lsb`.
+ *
+ * For example:
+ * - BITS(63, 63) is 0x8000000000000000
+ * - BITS(63, 0)  is 0xFFFFFFFFFFFFFFFF
+ * - BITS(0, 0)   is 0x0000000000000001
+ * - BITS(49, 17) is 0x0001FFFFFFFE0000
+ */
+#define BITS(msb, lsb) \
+	((~ULL(0) >> (63 - msb)) & (~ULL(0) << lsb))
+
+/*
+ * Define a mask of a single set bit `b`.
+ *
+ * For example:
+ * - BIT(63) is 0x8000000000000000
+ * - BIT(0)  is 0x0000000000000001
+ * - BIT(32) is 0x0000000100000000
+ */
+#define BIT(b)	BITS(b, b)
+
+/*
+ * Find the least significant set bit in the contiguous set of bits in `mask`.
+ *
+ * For example:
+ * - BITS_LSB(0x0000000000000001) is 0
+ * - BITS_LSB(0x000000000000ff00) is 8
+ * - BITS_LSB(0x8000000000000000) is 63
+ */
+#define BITS_LSB(mask)	(__builtin_ffsll(mask) - 1)
+
+/*
+ * Extract a bit-field out of `val` described by the contiguous set of bits in
+ * `mask`.
+ *
+ * For example:
+ * - BITS_EXTRACT(0xABCD, BITS(15, 12)) is 0xA
+ * - BITS_EXTRACT(0xABCD, BITS(11, 8))  is 0xB
+ * - BITS_EXTRACT(0xABCD, BIT(7))       is 0x1
+ */
+#define BITS_EXTRACT(val, mask) \
+	(((val) & (mask)) >> BITS_LSB(mask))
+
+/*
+ * Insert the least significant bits of `val` into the bit-field described by
+ * the contiguous set of bits in `mask`.
+ *
+ * For example:
+ * - BITS_INSERT(BITS(3, 0), 0xA)   is 0x000A
+ * - BITS_INSERT(BITS(15, 12), 0xA) is 0xA000
+ * - BITS_INSERT(BIT(15), 0xF)      is 0x1000
+ *
+ */
+#define BITS_INSERT(mask, val) \
+	(((val) << BITS_LSB(mask)) & (mask))
+
+#endif
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#3 lore raw

Re: [bootwrapper PATCH v2 02/13] Add bit-field macros

Quoted 84 lines · Mark Rutland

Arm architectural documentation typically defines bit-fields as [msb,lsb] and single-bit fields as [bit]. For clarity it would be helpful if we could define fields in the same way.

Add helpers so that we can do so, along with helper to extract/insert bit-field values.

There should be no functional change as a result of this patch.

---
 include/bits.h | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 79 insertions(+)
 create mode 100644 include/bits.h
include/bits.h new file +63 -0
new file mode 100644
index 0000000..0bf2c67
@@ -0,0 +1,79 @@
+/*
+ * include/bits.h - helpers for bit-field definitions.
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __BITS_H
+#define __BITS_H
+
+#ifdef __ASSEMBLY__
+#define UL(x)	x
+#define ULL(x)	x
+#else
+#define UL(x)	x##UL
+#define ULL(x)	x##ULL
+#endif
+
+/*
+ * Define a contiguous mask of bits with `msb` as the most significant bit and
+ * `lsb` as the least significant bit. The `msb` value must be greater than or
+ * equal to `lsb`.
+ *
+ * For example:
+ * - BITS(63, 63) is 0x8000000000000000
+ * - BITS(63, 0)  is 0xFFFFFFFFFFFFFFFF
+ * - BITS(0, 0)   is 0x0000000000000001
+ * - BITS(49, 17) is 0x0001FFFFFFFE0000
+ */
+#define BITS(msb, lsb) \
+	((~ULL(0) >> (63 - msb)) & (~ULL(0) << lsb))
+
+/*
+ * Define a mask of a single set bit `b`.
+ *
+ * For example:
+ * - BIT(63) is 0x8000000000000000
+ * - BIT(0)  is 0x0000000000000001
+ * - BIT(32) is 0x0000000100000000
+ */
+#define BIT(b)	BITS(b, b)
+
+/*
+ * Find the least significant set bit in the contiguous set of bits in `mask`.
+ *
+ * For example:
+ * - BITS_LSB(0x0000000000000001) is 0
+ * - BITS_LSB(0x000000000000ff00) is 8
+ * - BITS_LSB(0x8000000000000000) is 63
+ */
+#define BITS_LSB(mask)	(__builtin_ffsll(mask) - 1)
+
+/*
+ * Extract a bit-field out of `val` described by the contiguous set of bits in
+ * `mask`.
+ *
+ * For example:
+ * - BITS_EXTRACT(0xABCD, BITS(15, 12)) is 0xA
+ * - BITS_EXTRACT(0xABCD, BITS(11, 8))  is 0xB
+ * - BITS_EXTRACT(0xABCD, BIT(7))       is 0x1
+ */
+#define BITS_EXTRACT(val, mask) \

include/bits.h line 73

Quoted 10 lines
include/bits.h new file +10 -0
+	(((val) & (mask)) >> BITS_LSB(mask))
+
+/*
+ * Insert the least significant bits of `val` into the bit-field described by
+ * the contiguous set of bits in `mask`.
+ *
+ * For example:
+ * - BITS_INSERT(BITS(3, 0), 0xA)   is 0x000A
+ * - BITS_INSERT(BITS(15, 12), 0xA) is 0xA000
+ * - BITS_INSERT(BIT(15), 0xF)      is 0x1000

* - BITS_INSERT(BIT(35), 0x1) is 0x0

... which might be surprising!

of course BITS_INSERT(BIT(35), ULL(0x1)) works, but I feel either the description above needs to signpost this gotcha, or a cast is needed below.

AFAICT this macro isn't actually used in your series.

Steve

include/bits.h line 79

Quoted 7 lines
include/bits.h new file +6 -0
+ *
+ */
+#define BITS_INSERT(mask, val) \
+	(((val) << BITS_LSB(mask)) & (mask))
+
+#endif

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#4 lore raw

Re: [bootwrapper PATCH v2 02/13] Add bit-field macros

Quoted 10 lines · Steven Price
Quoted 8 lines · Mark Rutland

+/*
+ * Insert the least significant bits of val into the bit-field described by
+ * the contiguous set of bits in mask.
+ *
+ * For example:
+ * - BITS_INSERT(BITS(3, 0), 0xA) is 0x000A
+ * - BITS_INSERT(BITS(15, 12), 0xA) is 0xA000
+ * - BITS_INSERT(BIT(15), 0xF) is 0x1000

Quoted 8 lines

* - BITS_INSERT(BIT(35), 0x1) is 0x0

... which might be surprising!

of course BITS_INSERT(BIT(35), ULL(0x1)) works, but I feel either the description above needs to signpost this gotcha, or a cast is needed below.

AFAICT this macro isn't actually used in your series.

Good point; I'll either add the cast or drop that for now.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#5 lore raw

PATCH v2 3/13 aarch64: add system register accessors

In reply to Mark Rutland

We open code the use of mrs/msr for specific registers, which is somewhat tedious. Add macros to do this generically, along with a helper to extract a specific register field. Existing C usage is converted to the new helpers, and register definitions moved to a common location.

There should be no functional change as a result of this patch.

---
 arch/aarch64/include/asm/cpu.h    | 41 ++++++++++++++++++++++---------
 arch/aarch64/include/asm/gic-v3.h | 10 +++-----
 2 files changed, 32 insertions(+), 19 deletions(-)
arch/aarch64/include/asm/cpu.h +29 -12
index 63eb1c3..1053414 100644
@@ -9,10 +9,14 @@
 #ifndef __ASM_AARCH64_CPU_H
 #define __ASM_AARCH64_CPU_H
 
+#include <bits.h>
+
 #define MPIDR_ID_BITS		0xff00ffffff
 
 #define CURRENTEL_EL3		(3 << 2)
 
+#define ID_AA64PFR0_EL1_GIC	BITS(27, 24)
+
 /*
  * RES1 bits,  little-endian, caches and MMU off, no alignment checking,
  * no WXN.
@@ -29,6 +33,12 @@
 
 #define CPTR_EL3_EZ		(1 << 8)
 
+#define ICC_SRE_EL2		S3_4_C12_C9_5
+#define ICC_SRE_EL3		S3_6_C12_C12_5
+#define ICC_CTLR_EL1		S3_0_C12_C12_4
+#define ICC_CTLR_EL3		S3_6_C12_C12_4
+#define ICC_PMR_EL1		S3_0_C4_C6_0
+
 #define ZCR_EL3			s3_6_c1_c2_0
 #define ZCR_EL3_LEN_MASK	0x1ff
 
@@ -50,20 +60,27 @@
 
 #define sevl()		asm volatile ("sevl\n" : : : "memory")
 
-static inline unsigned long read_mpidr(void)
-{
-	unsigned long mpidr;
+#define __str(def)	#def
 
-	asm volatile ("mrs	%0, mpidr_el1\n" : "=r" (mpidr));
-	return mpidr & MPIDR_ID_BITS;
-}
+#define mrs(reg)							\
+({									\
+	unsigned long __mrs_val;					\
+	asm volatile("mrs %0, " __str(reg) : "=r" (__mrs_val));		\
+	__mrs_val;							\
+})
 
-static inline uint64_t read_id_aa64pfr0(void)
-{
-	uint64_t val;
+#define msr(reg, val)							\
+do {									\
+	unsigned long __msr_val = val;					\
+	asm volatile("msr " __str(reg) ", %0" : : "r" (__msr_val));	\
+} while (0)
+
+#define mrs_field(reg, field) \
+	BITS_EXTRACT(mrs(reg), (reg##_##field))
 
-	asm volatile ("mrs	%0, id_aa64pfr0_el1\n" : "=r" (val));
-	return val;
+static inline unsigned long read_mpidr(void)
+{
+	return mrs(mpidr_el1) & MPIDR_ID_BITS;
 }
 
 static inline void iciallu(void)
@@ -73,7 +90,7 @@ static inline void iciallu(void)
 
 static inline int has_gicv3_sysreg(void)
 {
-	return !!((read_id_aa64pfr0() >> 24) & 0xf);
+	return !!mrs_field(ID_AA64PFR0_EL1, GIC);
 }
 
 #endif /* !__ASSEMBLY__ */
arch/aarch64/include/asm/gic-v3.h +3 -7
index 5b32380..2447480 100644
@@ -9,20 +9,16 @@
 #ifndef __ASM_AARCH64_GICV3_H
 #define __ASM_AARCH64_GICV3_H
 
-#define ICC_SRE_EL2	"S3_4_C12_C9_5"
-#define ICC_SRE_EL3	"S3_6_C12_C12_5"
-#define ICC_CTLR_EL1	"S3_0_C12_C12_4"
-#define ICC_CTLR_EL3	"S3_6_C12_C12_4"
-#define ICC_PMR_EL1	"S3_0_C4_C6_0"
+#include <asm/cpu.h>
 
 static inline void gic_write_icc_sre(uint32_t val)
 {
-	asm volatile ("msr " ICC_SRE_EL3 ", %0" : : "r" (val));
+	msr(ICC_SRE_EL3, val);
 }
 
 static inline void gic_write_icc_ctlr(uint32_t val)
 {
-	asm volatile ("msr " ICC_CTLR_EL3 ", %0" : : "r" (val));
+	msr(ICC_CTLR_EL3, val);
 }
 
 #endif
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 3/13
#6 lore raw

Re: [bootwrapper PATCH v2 03/13] aarch64: add system register accessors

On Fri, 14 Jan 2022 10:56:43 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 8 lines

We open code the use of mrs/msr for specific registers, which is somewhat tedious. Add macros to do this generically, along with a helper to extract a specific register field. Existing C usage is converted to the new helpers, and register definitions moved to a common location.

There should be no functional change as a result of this patch.

Stared at and verified the architecture bits used in here:

Cheers,
Andre

Quoted 103 lines
---
 arch/aarch64/include/asm/cpu.h    | 41 ++++++++++++++++++++++---------
 arch/aarch64/include/asm/gic-v3.h | 10 +++-----
 2 files changed, 32 insertions(+), 19 deletions(-)
arch/aarch64/include/asm/cpu.h +29 -12
index 63eb1c3..1053414 100644
@@ -9,10 +9,14 @@
 #ifndef __ASM_AARCH64_CPU_H
 #define __ASM_AARCH64_CPU_H
 
+#include <bits.h>
+
 #define MPIDR_ID_BITS		0xff00ffffff
 
 #define CURRENTEL_EL3		(3 << 2)
 
+#define ID_AA64PFR0_EL1_GIC	BITS(27, 24)
+
 /*
  * RES1 bits,  little-endian, caches and MMU off, no alignment checking,
  * no WXN.
@@ -29,6 +33,12 @@
 
 #define CPTR_EL3_EZ		(1 << 8)
 
+#define ICC_SRE_EL2		S3_4_C12_C9_5
+#define ICC_SRE_EL3		S3_6_C12_C12_5
+#define ICC_CTLR_EL1		S3_0_C12_C12_4
+#define ICC_CTLR_EL3		S3_6_C12_C12_4
+#define ICC_PMR_EL1		S3_0_C4_C6_0
+
 #define ZCR_EL3			s3_6_c1_c2_0
 #define ZCR_EL3_LEN_MASK	0x1ff
 
@@ -50,20 +60,27 @@
 
 #define sevl()		asm volatile ("sevl\n" : : : "memory")
 
-static inline unsigned long read_mpidr(void)
-{
-	unsigned long mpidr;
+#define __str(def)	#def
 
-	asm volatile ("mrs	%0, mpidr_el1\n" : "=r" (mpidr));
-	return mpidr & MPIDR_ID_BITS;
-}
+#define mrs(reg)							\
+({									\
+	unsigned long __mrs_val;					\
+	asm volatile("mrs %0, " __str(reg) : "=r" (__mrs_val));		\
+	__mrs_val;							\
+})
 
-static inline uint64_t read_id_aa64pfr0(void)
-{
-	uint64_t val;
+#define msr(reg, val)							\
+do {									\
+	unsigned long __msr_val = val;					\
+	asm volatile("msr " __str(reg) ", %0" : : "r" (__msr_val));	\
+} while (0)
+
+#define mrs_field(reg, field) \
+	BITS_EXTRACT(mrs(reg), (reg##_##field))
 
-	asm volatile ("mrs	%0, id_aa64pfr0_el1\n" : "=r" (val));
-	return val;
+static inline unsigned long read_mpidr(void)
+{
+	return mrs(mpidr_el1) & MPIDR_ID_BITS;
 }
 
 static inline void iciallu(void)
@@ -73,7 +90,7 @@ static inline void iciallu(void)
 
 static inline int has_gicv3_sysreg(void)
 {
-	return !!((read_id_aa64pfr0() >> 24) & 0xf);
+	return !!mrs_field(ID_AA64PFR0_EL1, GIC);
 }
 
 #endif /* !__ASSEMBLY__ */
arch/aarch64/include/asm/gic-v3.h +1 -6
index 5b32380..2447480 100644
@@ -9,20 +9,16 @@
 #ifndef __ASM_AARCH64_GICV3_H
 #define __ASM_AARCH64_GICV3_H
 
-#define ICC_SRE_EL2	"S3_4_C12_C9_5"
-#define ICC_SRE_EL3	"S3_6_C12_C12_5"
-#define ICC_CTLR_EL1	"S3_0_C12_C12_4"
-#define ICC_CTLR_EL3	"S3_6_C12_C12_4"
-#define ICC_PMR_EL1	"S3_0_C4_C6_0"
+#include <asm/cpu.h>
 
 static inline void gic_write_icc_sre(uint32_t val)
 {
-	asm volatile ("msr " ICC_SRE_EL3 ", %0" : : "r" (val));

arch/aarch64/include/asm/gic-v3.h line 24

Quoted 10 lines
arch/aarch64/include/asm/gic-v3.h +2 -1
+	msr(ICC_SRE_EL3, val);
 }
 
 static inline void gic_write_icc_ctlr(uint32_t val)
 {
-	asm volatile ("msr " ICC_CTLR_EL3 ", %0" : : "r" (val));
+	msr(ICC_CTLR_EL3, val);
 }
 
 #endif

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#7 lore raw

PATCH v2 4/13 aarch32: add coprocessor accessors

In reply to Mark Rutland

We open code the use of mrc/mcr for specific registers, which is somewhat tedious. Add macros to do this generically, along with a helper to extract a specific register field. Existing C usage is converted to the new helpers, and register definitions moved to a common location.

There should be no functional change as a result of this patch.

---
 arch/aarch32/include/asm/cpu.h    | 45 ++++++++++++++++++++-----------
 arch/aarch32/include/asm/gic-v3.h |  6 +++--
 2 files changed, 34 insertions(+), 17 deletions(-)
arch/aarch32/include/asm/cpu.h +30 -15
index 105cae5..d691c7b 100644
@@ -9,9 +9,13 @@
 #ifndef __ASM_AARCH32_CPU_H
 #define __ASM_AARCH32_CPU_H
 
+#include <bits.h>
+
 #define MPIDR_ID_BITS		0x00ffffff
 #define MPIDR_INVALID		(-1)
 
+#define ID_PFR1_GIC		BITS(31, 28)
+
 /* Only RES1 bits and CP15 barriers for the kernel */
 #define HSCTLR_KERNEL		(3 << 28 | 3 << 22 | 1 << 18 | 1 << 16 | 1 << 11 | 3 << 4)
 #define SCTLR_KERNEL		(3 << 22 | 1 << 11 | 1 << 5 | 3 << 4)
@@ -40,32 +44,43 @@
 #define sevl()		asm volatile ("sev" : : : "memory")
 #endif
 
-static inline unsigned long read_mpidr(void)
-{
-	unsigned long mpidr;
+#define MPIDR		"p15, 0, %0, c0, c0, 5"
+#define ID_PFR1		"p15, 0, %0, c0, c1, 1"
+#define ICIALLU		"p15, 0, %0, c7, c5, 0"
 
-	asm volatile ("mrc	p15, 0, %0, c0, c0, 5\n" : "=r" (mpidr));
-	return mpidr & MPIDR_ID_BITS;
-}
+#define ICC_SRE		"p15, 6, %0, c12, c12, 5"
+#define ICC_CTLR	"p15, 6, %0, c12, c12, 4"
 
-static inline uint32_t read_id_pfr1(void)
-{
-	uint32_t val;
+#define mrc(reg)						\
+({								\
+	unsigned long __mrc_val;				\
+	asm volatile("mrc " reg : "=r" (__mrc_val));		\
+	__mrc_val;						\
+})
 
-	asm volatile ("mrc	p15, 0, %0, c0, c1, 1\n" : "=r" (val));
-	return val;
+#define mcr(reg, val)						\
+do {								\
+	unsigned long __mcr_val = val;				\
+	asm volatile("mcr " reg : : "r" (__mcr_val));		\
+} while (0)
+
+
+#define mrc_field(reg, field) \
+	BITS_EXTRACT(mrc(reg), (reg##_##field))
+
+static inline unsigned long read_mpidr(void)
+{
+	return mrc(MPIDR) & MPIDR_ID_BITS;
 }
 
 static inline void iciallu(void)
 {
-	uint32_t val = 0;
-
-	asm volatile ("mcr	p15, 0, %0, c7, c5, 0" : : "r" (val));
+	mcr(ICIALLU, 0);
 }
 
 static inline int has_gicv3_sysreg(void)
 {
-	return !!((read_id_pfr1() >> 28) & 0xf);
+	return !!mrc_field(ID_PFR1, GIC);
 }
 
 #endif /* __ASSEMBLY__ */
arch/aarch32/include/asm/gic-v3.h +4 -2
index 65f38de..b28136a 100644
@@ -9,14 +9,16 @@
 #ifndef __ASM_AARCH32_GICV3_H
 #define __ASM_AARCH32_GICV3_H
 
+#include <asm/cpu.h>
+
 static inline void gic_write_icc_sre(uint32_t val)
 {
-	asm volatile ("mcr p15, 6, %0, c12, c12, 5" : : "r" (val));
+	mcr(ICC_SRE, val);
 }
 
 static inline void gic_write_icc_ctlr(uint32_t val)
 {
-	asm volatile ("mcr p15, 6, %0, c12, c12, 4" : : "r" (val));
+	mcr(ICC_CTLR, val);
 }
 
 #endif
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#8 lore raw

PATCH v2 5/13 aarch64: add mov_64 macro

In reply to Mark Rutland

In subsequent patches we'll need to load 64-bit values into GPRs before the CPU is in a known endianness, where we cannot use literal pools.

In preparation for that, this patch adds a new mov_64 macro to load a 64-bit value into a GPR using a sequence of MOV and MOVKs, which will function the same regardless of the CPU's endianness.

At the same time, move the cpuid macro to use mov_64 internally.

---
 arch/aarch64/common.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
arch/aarch64/common.S +9 -1
index c7171a9..3279fa9 100644
@@ -9,9 +9,17 @@
 
 #include <cpu.h>
 
+	/* Load a 64-bit value using immediates */
+	.macro	mov_64 dest, val
+	mov	\dest, #(((\val) >>  0) & 0xffff)
+	movk	\dest, #(((\val) >> 16) & 0xffff), lsl #16
+	movk	\dest, #(((\val) >> 32) & 0xffff), lsl #32
+	movk	\dest, #(((\val) >> 48) & 0xffff), lsl #48
+	.endm
+
 	/* Put MPIDR into \dest, clobber \tmp and flags */
 	.macro cpuid dest, tmp
 	mrs	\dest, mpidr_el1
-	ldr	\tmp, =MPIDR_ID_BITS
+	mov_64	\tmp, MPIDR_ID_BITS
 	ands	\dest, \dest, \tmp
 	.endm
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 5/13
#9 lore raw

Re: [bootwrapper PATCH v2 05/13] aarch64: add mov_64 macro

On Fri, 14 Jan 2022 10:56:45 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 10 lines

In subsequent patches we'll need to load 64-bit values into GPRs before the CPU is in a known endianness, where we cannot use literal pools.

In preparation for that, this patch adds a new mov_64 macro to load a 64-bit value into a GPR using a sequence of MOV and MOVKs, which will function the same regardless of the CPU's endianness.

At the same time, move the cpuid macro to use mov_64 internally.

Not sure that's worth it, but there is a simpler version of that TF-A macro, along the lines of:

	.if ((\val) >> 16) & 0xffff
		movk    \dest, #(((\val) >> 16) & 0xffff), lsl #16
	.endif

It's a bit less optimal in a corner case than the TF-A version, but avoids the few pointless " movk x0, #0x0, lsl #..." I found in the code.

But that's an optimisation detail, so anyway:

Cheers,
Andre

Quoted 17 lines
---
 arch/aarch64/common.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
arch/aarch64/common.S +5 -0
index c7171a9..3279fa9 100644
@@ -9,9 +9,17 @@
 
 #include <cpu.h>
 
+	/* Load a 64-bit value using immediates */
+	.macro	mov_64 dest, val
+	mov	\dest, #(((\val) >>  0) & 0xffff)
+	movk	\dest, #(((\val) >> 16) & 0xffff), lsl #16
+	movk	\dest, #(((\val) >> 32) & 0xffff), lsl #32

arch/aarch64/common.S line 25

Quoted 10 lines
arch/aarch64/common.S +4 -1
+	movk	\dest, #(((\val) >> 48) & 0xffff), lsl #48
+	.endm
+
 	/* Put MPIDR into \dest, clobber \tmp and flags */
 	.macro cpuid dest, tmp
 	mrs	\dest, mpidr_el1
-	ldr	\tmp, =MPIDR_ID_BITS
+	mov_64	\tmp, MPIDR_ID_BITS
 	ands	\dest, \dest, \tmp
 	.endm

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#10 lore raw

PATCH v2 6/13 aarch64: initialize SCTLR_ELx for the boot-wrapper

In reply to Mark Rutland

The SCTLR_ELx registers contain fields which are UNKNOWN or IMPLEMENTATION DEFINED out of reset. This includes SCTLR_ELx.EE, which defines the endianness of memory accesses (e.g. reads from literal pools). Due to this, portions of boot-wrapper code are not guaranteed to work correctly.

Rework the startup code to explicitly initialize SCTLR_ELx for the exception level the boot-wrapper was entered at. When entered at EL2 it's necessary to first initialise HCR_EL2.E2H as this affects the RESx behaviour of bits in SCTLR_EL2, and also aliases SCTLR_EL1 to SCTLR_EL2, which would break the initialization performed in jump_kernel.

As we plan to eventually support the highest implemented EL being any of EL3/EL2/EL1, code is added to handle all of these exception levels, even though we do not currently support starting at EL1.

We'll initialize other registers in subsequent patches.

---
 arch/aarch64/boot.S            | 74 +++++++++++++++++++++++++++-------
 arch/aarch64/include/asm/cpu.h | 27 ++++++++++++-
 2 files changed, 85 insertions(+), 16 deletions(-)
arch/aarch64/boot.S +59 -15
index 900b9f8..45a0367 100644
@@ -26,26 +26,26 @@
 	 *   PSCI is not supported when entered in this exception level.
 	 */
 ASM_FUNC(_start)
-	cpuid	x0, x1
-	bl	find_logical_id
-	cmp	x0, #MPIDR_INVALID
-	beq	err_invalid_id
-	bl	setup_stack
-
-	/*
-	 * EL3 initialisation
-	 */
 	mrs	x0, CurrentEL
 	cmp	x0, #CURRENTEL_EL3
-	b.eq	1f
+	b.eq	reset_at_el3
+	cmp	x0, #CURRENTEL_EL2
+	b.eq	reset_at_el2
+	cmp	x0, #CURRENTEL_EL1
+	b.eq	reset_at_el1
 
-	mov	w0, #1
-	ldr	x1, =flag_no_el3
-	str	w0, [x1]
+	/* Booting at EL0 is not supported */
+	b	.
 
-	b	start_no_el3
+	/*
+	 * EL3 initialisation
+	 */
+reset_at_el3:
+	mov_64	x0, SCTLR_EL3_RESET
+	msr	sctlr_el3, x0
+	isb
 
-1:	mov	x0, #0x30			// RES1
+	mov	x0, #0x30			// RES1
 	orr	x0, x0, #(1 << 0)		// Non-secure EL1
 	orr	x0, x0, #(1 << 8)		// HVC enable
 
@@ -135,10 +135,54 @@ ASM_FUNC(_start)
 	ldr	x0, =COUNTER_FREQ
 	msr	cntfrq_el0, x0
 
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
 	bl	gic_secure_init
 
 	b	start_el3
 
+	/*
+	 * EL2 initialization
+	 */
+reset_at_el2:
+	// Ensure E2H is not in use
+	mov_64	x0, HCR_EL2_RESET
+	msr	hcr_el2, x0
+	isb
+
+	mov_64	x0, SCTLR_EL2_RESET
+	msr	sctlr_el2, x0
+	isb
+
+	b	reset_no_el3
+
+	/*
+	 * EL1 initialization
+	 */
+reset_at_el1:
+	mov_64	x0, SCTLR_EL1_RESET
+	msr	sctlr_el1, x0
+	isb
+
+	b	reset_no_el3
+
+reset_no_el3:
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
+	mov	w0, #1
+	ldr	x1, =flag_no_el3
+	str	w0, [x1]
+
+	b	start_no_el3
+
 err_invalid_id:
 	b	.
 
arch/aarch64/include/asm/cpu.h +26 -1
index 1053414..1e9141a 100644
@@ -14,6 +14,32 @@
 #define MPIDR_ID_BITS		0xff00ffffff
 
 #define CURRENTEL_EL3		(3 << 2)
+#define CURRENTEL_EL2		(2 << 2)
+#define CURRENTEL_EL1		(1 << 2)
+
+/*
+ * RES1 bit definitions definitions as of ARM DDI 0487G.b
+ *
+ * These includes bits which are RES1 in some configurations.
+ */
+#define SCTLR_EL3_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))
+
+#define SCTLR_EL2_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))
+
+#define HCR_EL2_RES1		(BIT(1))
+
+/*
+ * Initial register values required for the boot-wrapper to run out-of-reset.
+ */
+#define SCTLR_EL3_RESET		SCTLR_EL3_RES1
+#define SCTLR_EL2_RESET		SCTLR_EL2_RES1
+#define SCTLR_EL1_RESET		SCTLR_EL1_RES1
+#define HCR_EL2_RESET		HCR_EL2_RES1
 
 #define ID_AA64PFR0_EL1_GIC	BITS(27, 24)
 
@@ -43,7 +69,6 @@
 #define ZCR_EL3_LEN_MASK	0x1ff
 
 #define SCTLR_EL1_CP15BEN	(1 << 5)
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#11 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

On Fri, 14 Jan 2022 10:56:46 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 12 lines

The SCTLR_ELx registers contain fields which are UNKNOWN or IMPLEMENTATION DEFINED out of reset. This includes SCTLR_ELx.EE, which defines the endianness of memory accesses (e.g. reads from literal pools). Due to this, portions of boot-wrapper code are not guaranteed to work correctly.

Rework the startup code to explicitly initialize SCTLR_ELx for the exception level the boot-wrapper was entered at. When entered at EL2 it's necessary to first initialise HCR_EL2.E2H as this affects the RESx behaviour of bits in SCTLR_EL2, and also aliases SCTLR_EL1 to SCTLR_EL2, which would break the initialization performed in jump_kernel.

Quoted 5 lines

As we plan to eventually support the highest implemented EL being any of EL3/EL2/EL1, code is added to handle all of these exception levels, even though we do not currently support starting at EL1.

We'll initialize other registers in subsequent patches.

So the idea of initialising each EL and the respective code below looks good to me, however I have some questions about the SCTLR reset values below:

Quoted 119 lines
---
 arch/aarch64/boot.S            | 74 +++++++++++++++++++++++++++-------
 arch/aarch64/include/asm/cpu.h | 27 ++++++++++++-
 2 files changed, 85 insertions(+), 16 deletions(-)
arch/aarch64/boot.S +59 -15
index 900b9f8..45a0367 100644
@@ -26,26 +26,26 @@
 	 *   PSCI is not supported when entered in this exception level.
 	 */
 ASM_FUNC(_start)
-	cpuid	x0, x1
-	bl	find_logical_id
-	cmp	x0, #MPIDR_INVALID
-	beq	err_invalid_id
-	bl	setup_stack
-
-	/*
-	 * EL3 initialisation
-	 */
 	mrs	x0, CurrentEL
 	cmp	x0, #CURRENTEL_EL3
-	b.eq	1f
+	b.eq	reset_at_el3
+	cmp	x0, #CURRENTEL_EL2
+	b.eq	reset_at_el2
+	cmp	x0, #CURRENTEL_EL1
+	b.eq	reset_at_el1
 
-	mov	w0, #1
-	ldr	x1, =flag_no_el3
-	str	w0, [x1]
+	/* Booting at EL0 is not supported */
+	b	.
 
-	b	start_no_el3
+	/*
+	 * EL3 initialisation
+	 */
+reset_at_el3:
+	mov_64	x0, SCTLR_EL3_RESET
+	msr	sctlr_el3, x0
+	isb
 
-1:	mov	x0, #0x30			// RES1
+	mov	x0, #0x30			// RES1
 	orr	x0, x0, #(1 << 0)		// Non-secure EL1
 	orr	x0, x0, #(1 << 8)		// HVC enable
 
@@ -135,10 +135,54 @@ ASM_FUNC(_start)
 	ldr	x0, =COUNTER_FREQ
 	msr	cntfrq_el0, x0
 
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
 	bl	gic_secure_init
 
 	b	start_el3
 
+	/*
+	 * EL2 initialization
+	 */
+reset_at_el2:
+	// Ensure E2H is not in use
+	mov_64	x0, HCR_EL2_RESET
+	msr	hcr_el2, x0
+	isb
+
+	mov_64	x0, SCTLR_EL2_RESET
+	msr	sctlr_el2, x0
+	isb
+
+	b	reset_no_el3
+
+	/*
+	 * EL1 initialization
+	 */
+reset_at_el1:
+	mov_64	x0, SCTLR_EL1_RESET
+	msr	sctlr_el1, x0
+	isb
+
+	b	reset_no_el3
+
+reset_no_el3:
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
+	mov	w0, #1
+	ldr	x1, =flag_no_el3
+	str	w0, [x1]
+
+	b	start_no_el3
+
 err_invalid_id:
 	b	.
 
arch/aarch64/include/asm/cpu.h +3 -0
index 1053414..1e9141a 100644
@@ -14,6 +14,32 @@
 #define MPIDR_ID_BITS		0xff00ffffff
 
 #define CURRENTEL_EL3		(3 << 2)
+#define CURRENTEL_EL2		(2 << 2)
+#define CURRENTEL_EL1		(1 << 2)
+

arch/aarch64/include/asm/cpu.h line 29

Quoted 10 lines
arch/aarch64/include/asm/cpu.h +10 -0
+/*
+ * RES1 bit definitions definitions as of ARM DDI 0487G.b
+ *
+ * These includes bits which are RES1 in some configurations.
+ */
+#define SCTLR_EL3_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))
+
+#define SCTLR_EL2_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))

I compared all bits against the ARM ARM and the kernel version for EL2, that looks correct to me.

arch/aarch64/include/asm/cpu.h line 32

Quoted 3 lines
arch/aarch64/include/asm/cpu.h +3 -0
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))

- The kernel sets TSCXT(bit[20]), and the ARM ARM says that the value should be RES1 if FEAT_CSV2_* is not implemented. Should we copy this?
- The kernel clears ITD(bit[7]), and the ARM ARM says it's *Otherwise* RES1 (no AArch32 in EL0). I feel like we should not disable IT instructions in EL0 needlessly?
- I also feel like we should set CP15BEN(bit[5]), for similar reasons.

Granted those bits affect only EL0 execution, which we don't care about in the boot-wrapper, but I was wondering if we should change those anyway? At least bit 20?

arch/aarch64/include/asm/cpu.h line 34

Quoted 2 lines
arch/aarch64/include/asm/cpu.h +2 -0
+
+#define HCR_EL2_RES1		(BIT(1))

Should we set RW(bit[31]), just to be safe? Not sure this is explicitly mentioned somewhere, but is the boot flow when we are entered in EL2 to stay in EL2 and launch the kernel in there as well?

Quoted 19 lines
arch/aarch64/include/asm/cpu.h +8 -1
+
+/*
+ * Initial register values required for the boot-wrapper to run out-of-reset.
+ */
+#define SCTLR_EL3_RESET		SCTLR_EL3_RES1
+#define SCTLR_EL2_RESET		SCTLR_EL2_RES1
+#define SCTLR_EL1_RESET		SCTLR_EL1_RES1
+#define HCR_EL2_RESET		HCR_EL2_RES1
 
 #define ID_AA64PFR0_EL1_GIC	BITS(27, 24)
 
@@ -43,7 +69,6 @@
 #define ZCR_EL3_LEN_MASK	0x1ff
 
 #define SCTLR_EL1_CP15BEN	(1 << 5)
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */
Quoted 1 line

#define SCTLR_EL1_KERNEL (SCTLR_EL1_RES1 | SCTLR_EL1_CP15BEN)

So I wonder if this actually works? The ARMv7 version of SCTLR differs in some bits from both the ARMv8 AArch32 version and more importantly the AArch64 version. I had troubles the other day running the arm32 Linux kernel decompressor with some ARMv8 SCTLR_EL1 reset value. The decompressor code does only read-modify-write of SCTLR (probably to cover multiple architecture revisions), so some bits might stay wrong. In particular I think having bits 28 and 29 set caused problems. By looking at the ARMv7 ARM and with experimentation I came up with 0x00c00878 as a safe and working value.
Shall we have a separate reset value for 32bit?

Cheers,
Andre

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#12 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

Quoted 4 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:46 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Hi Andre,

Quoted 155 lines
Quoted 12 lines

The SCTLR_ELx registers contain fields which are UNKNOWN or IMPLEMENTATION DEFINED out of reset. This includes SCTLR_ELx.EE, which defines the endianness of memory accesses (e.g. reads from literal pools). Due to this, portions of boot-wrapper code are not guaranteed to work correctly.

Rework the startup code to explicitly initialize SCTLR_ELx for the exception level the boot-wrapper was entered at. When entered at EL2 it's necessary to first initialise HCR_EL2.E2H as this affects the RESx behaviour of bits in SCTLR_EL2, and also aliases SCTLR_EL1 to SCTLR_EL2, which would break the initialization performed in jump_kernel.

Quoted 5 lines

As we plan to eventually support the highest implemented EL being any of EL3/EL2/EL1, code is added to handle all of these exception levels, even though we do not currently support starting at EL1.

We'll initialize other registers in subsequent patches.

So the idea of initialising each EL and the respective code below looks good to me, however I have some questions about the SCTLR reset values below:

Quoted 119 lines
---
 arch/aarch64/boot.S            | 74 +++++++++++++++++++++++++++-------
 arch/aarch64/include/asm/cpu.h | 27 ++++++++++++-
 2 files changed, 85 insertions(+), 16 deletions(-)
arch/aarch64/boot.S +59 -15
index 900b9f8..45a0367 100644
@@ -26,26 +26,26 @@
 	 *   PSCI is not supported when entered in this exception level.
 	 */
 ASM_FUNC(_start)
-	cpuid	x0, x1
-	bl	find_logical_id
-	cmp	x0, #MPIDR_INVALID
-	beq	err_invalid_id
-	bl	setup_stack
-
-	/*
-	 * EL3 initialisation
-	 */
 	mrs	x0, CurrentEL
 	cmp	x0, #CURRENTEL_EL3
-	b.eq	1f
+	b.eq	reset_at_el3
+	cmp	x0, #CURRENTEL_EL2
+	b.eq	reset_at_el2
+	cmp	x0, #CURRENTEL_EL1
+	b.eq	reset_at_el1
 
-	mov	w0, #1
-	ldr	x1, =flag_no_el3
-	str	w0, [x1]
+	/* Booting at EL0 is not supported */
+	b	.
 
-	b	start_no_el3
+	/*
+	 * EL3 initialisation
+	 */
+reset_at_el3:
+	mov_64	x0, SCTLR_EL3_RESET
+	msr	sctlr_el3, x0
+	isb
 
-1:	mov	x0, #0x30			// RES1
+	mov	x0, #0x30			// RES1
 	orr	x0, x0, #(1 << 0)		// Non-secure EL1
 	orr	x0, x0, #(1 << 8)		// HVC enable
 
@@ -135,10 +135,54 @@ ASM_FUNC(_start)
 	ldr	x0, =COUNTER_FREQ
 	msr	cntfrq_el0, x0
 
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
 	bl	gic_secure_init
 
 	b	start_el3
 
+	/*
+	 * EL2 initialization
+	 */
+reset_at_el2:
+	// Ensure E2H is not in use
+	mov_64	x0, HCR_EL2_RESET
+	msr	hcr_el2, x0
+	isb
+
+	mov_64	x0, SCTLR_EL2_RESET
+	msr	sctlr_el2, x0
+	isb
+
+	b	reset_no_el3
+
+	/*
+	 * EL1 initialization
+	 */
+reset_at_el1:
+	mov_64	x0, SCTLR_EL1_RESET
+	msr	sctlr_el1, x0
+	isb
+
+	b	reset_no_el3
+
+reset_no_el3:
+	cpuid	x0, x1
+	bl	find_logical_id
+	cmp	x0, #MPIDR_INVALID
+	b.eq	err_invalid_id
+	bl	setup_stack
+
+	mov	w0, #1
+	ldr	x1, =flag_no_el3
+	str	w0, [x1]
+
+	b	start_no_el3
+
 err_invalid_id:
 	b	.
 
arch/aarch64/include/asm/cpu.h +3 -0
index 1053414..1e9141a 100644
@@ -14,6 +14,32 @@
 #define MPIDR_ID_BITS		0xff00ffffff
 
 #define CURRENTEL_EL3		(3 << 2)
+#define CURRENTEL_EL2		(2 << 2)
+#define CURRENTEL_EL1		(1 << 2)
+

arch/aarch64/include/asm/cpu.h line 29

Quoted 10 lines
arch/aarch64/include/asm/cpu.h +10 -0
+/*
+ * RES1 bit definitions definitions as of ARM DDI 0487G.b
+ *
+ * These includes bits which are RES1 in some configurations.
+ */
+#define SCTLR_EL3_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))
+
+#define SCTLR_EL2_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(18) | BIT(16) | BIT(11) | BIT(5) | BIT(4))

I compared all bits against the ARM ARM and the kernel version for EL2, that looks correct to me.

Quoted 6 lines

arch/aarch64/include/asm/cpu.h line 32

Quoted 3 lines
arch/aarch64/include/asm/cpu.h +3 -0
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))

- The kernel sets TSCXT(bit[20]), and the ARM ARM says that the value should be RES1 if FEAT_CSV2_* is not implemented. Should we copy this?

Yes, we should. I'll go and fold that in.

Quoted 3 lines

- The kernel clears ITD(bit[7]), and the ARM ARM says it's *Otherwise* RES1 (no AArch32 in EL0). I feel like we should not disable IT instructions in EL0 needlessly?

Per the ARM ARM the bit resets to an UNKNOWN value, and so per our usual policy the kernel must initialize that before it can depend upon it, and IIUC you say the kernel already does so.

So it shouldn't matter what the boot-wrapper does, and for consitency, I think the boot-wrapper should set this to 0b1.

Quoted 1 line

- I also feel like we should set CP15BEN(bit[5]), for similar reasons.

I agree.

Quoted 3 lines

Granted those bits affect only EL0 execution, which we don't care about in the boot-wrapper, but I was wondering if we should change those anyway? At least bit 20?

I'll set all three of those bits. as above.

Quoted 5 lines

arch/aarch64/include/asm/cpu.h line 34

Quoted 1 line
arch/aarch64/include/asm/cpu.h +1 -0
+#define HCR_EL2_RES1		(BIT(1))

Should we set RW(bit[31]), just to be safe? Not sure this is explicitly mentioned somewhere, but is the boot flow when we are entered in EL2 to stay in EL2 and launch the kernel in there as well?

I don't think we need to currently. HCR_EL2.RW is UNKNOWN out-of-reset, is RAO/WI rather than RES1, and only affects lower ELs, so per usual policy we leave this to the kernel to initialize. Of the configurations this could affect:

* BW @ EL3 + {psci,spin}
Kernel entered at EL2 on all PEs.

* BW @ EL2 + spin-table
Kernel entered at EL2 on all PEs.

* BW @ EL2 + PSCI
Kernel entered at EL2 on boot PE.
BROKEN due to PSCI conduit anyhow.

I'd like to clean up the EL + boot-method combinations, but this reqiures some more work which I had punted out of this series for now.

For now, I'd prefer to leave this as-is. I would like to fix this as part of a subsequent boot-method cleanup, configuring HCR_EL2.RW as part of the kernel handover logic later in the boot flow.

Quoted 18 lines
Quoted 18 lines
arch/aarch64/include/asm/cpu.h +8 -1
+
+/*
+ * Initial register values required for the boot-wrapper to run out-of-reset.
+ */
+#define SCTLR_EL3_RESET		SCTLR_EL3_RES1
+#define SCTLR_EL2_RESET		SCTLR_EL2_RES1
+#define SCTLR_EL1_RESET		SCTLR_EL1_RES1
+#define HCR_EL2_RESET		HCR_EL2_RES1
 
 #define ID_AA64PFR0_EL1_GIC	BITS(27, 24)
 
@@ -43,7 +69,6 @@
 #define ZCR_EL3_LEN_MASK	0x1ff
 
 #define SCTLR_EL1_CP15BEN	(1 << 5)
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
Quoted 7 lines
Quoted 3 lines
arch/aarch64/include/asm/cpu.h +0 -0
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */

#define SCTLR_EL1_KERNEL (SCTLR_EL1_RES1 | SCTLR_EL1_CP15BEN)

So I wonder if this actually works? The ARMv7 version of SCTLR differs in some bits from both the ARMv8 AArch32 version and more importantly the AArch64 version.

Hmm. SCTLR_EL1 is architecturally mapped to SCTLR, and for some reason I thought that applied field-wise, but IIUC you're saying that applies bit-wise, right?

The ARM ARM is delightfully unclear about this, which is reather unfortunate. Per the glossary in ARM DDI 0487G.b:

| Where this manual describes a register as being architecturally mapped
| to another register, this indicates that, in an implementation that
| supports both of the registers, the two registers access the same
| state.

That being the case, we'd need to do likewise for all the other registers accessible at kernel ELs, right?

Quoted 7 lines

I had troubles the other day running the
arm32 Linux kernel decompressor with some ARMv8 SCTLR_EL1 reset value. The decompressor code does only read-modify-write of SCTLR (probably to cover multiple architecture revisions), so some bits might stay wrong. In particular I think having bits 28 and 29 set caused problems. By looking at the ARMv7 ARM and with experimentation I came up with 0x00c00878 as a safe and working value.

Do we have any comments/documentation saying what the 32-bit kernel actually needs/wants here?

Quoted 1 line

Shall we have a separate reset value for 32bit?

Currently, the 32-bit side of the boot-wrapper uses:

	(3 << 22 | 1 << 11 | 1 << 5 | 3 << 4)

... noting that 1 << 5 and 3 << 4 overlap..

Which is:
* bit[22]: RES1
* bit[11]: RES1
* bit[5]: CP15BEN
* bit[4]: LSMAOE

Per the above, it sounds like you think that's wrong too?

Whatever we do, the 32-bit/64-bit boot-wrapper should program something functionally equivalent when booting a 32-bit kernel.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#13 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

Quoted 29 lines · Mark Rutland
Quoted 4 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:46 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Hi Andre,

Quoted 21 lines
Quoted 9 lines
arch/aarch64/include/asm/cpu.h +3 -1
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))
 
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */

#define SCTLR_EL1_KERNEL (SCTLR_EL1_RES1 | SCTLR_EL1_CP15BEN)

So I wonder if this actually works? The ARMv7 version of SCTLR differs in some bits from both the ARMv8 AArch32 version and more importantly the AArch64 version.

I had troubles the other day running the
arm32 Linux kernel decompressor with some ARMv8 SCTLR_EL1 reset value. The decompressor code does only read-modify-write of SCTLR (probably to cover multiple architecture revisions), so some bits might stay wrong. In particular I think having bits 28 and 29 set caused problems. By looking at the ARMv7 ARM and with experimentation I came up with 0x00c00878 as a safe and working value.

Having re-read all of this, I believe you're right; I'll rework the AArch32-kernel SCTLR_EL1_KERNEL to not use the AArch64 bit definitions, and I'll add some commentary explaining that we're writing to the AArch32 format.

Quoted 1 line
Quoted 1 line

Shall we have a separate reset value for 32bit?

Assuming you meant to alter the SCTLR_ELx_KERNEL definition, yes.

The idea of splitting the SCTLR_ELx_RESET and SCTLR_ELx_KERNEL definitions was that the former would be whatever the boot-wrapper needed to run at ELx, and the latter was whatever we must initialize for the kernel to run at ELx, so I don't want to put the AArch32 kernel bits into SCTLR_EL1_RESET.

My only remaining concern is exactly what we must initialize. If there's any documentation we can refer to, that'd be great, otherwise I'll dig through your prior suggestion.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#14 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

On Mon, 17 Jan 2022 13:05:54 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 46 lines · Mark Rutland
Quoted 29 lines · Andre Przywara
Quoted 4 lines

On Fri, 14 Jan 2022 10:56:46 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Hi Andre,

Quoted 21 lines
Quoted 9 lines
arch/aarch64/include/asm/cpu.h +3 -1
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))  
 
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */

#define SCTLR_EL1_KERNEL (SCTLR_EL1_RES1 | SCTLR_EL1_CP15BEN)

So I wonder if this actually works? The ARMv7 version of SCTLR differs in some bits from both the ARMv8 AArch32 version and more importantly the AArch64 version.

I had troubles the other day running the
arm32 Linux kernel decompressor with some ARMv8 SCTLR_EL1 reset value. The decompressor code does only read-modify-write of SCTLR (probably to cover multiple architecture revisions), so some bits might stay wrong. In particular I think having bits 28 and 29 set caused problems. By looking at the ARMv7 ARM and with experimentation I came up with 0x00c00878 as a safe and working value.

Having re-read all of this, I believe you're right; I'll rework the AArch32-kernel SCTLR_EL1_KERNEL to not use the AArch64 bit definitions, and I'll add some commentary explaining that we're writing to the AArch32 format.

Quoted 1 line
Quoted 1 line

Shall we have a separate reset value for 32bit?

Assuming you meant to alter the SCTLR_ELx_KERNEL definition, yes.

The idea of splitting the SCTLR_ELx_RESET and SCTLR_ELx_KERNEL definitions was that the former would be whatever the boot-wrapper needed to run at ELx, and the latter was whatever we must initialize for the kernel to run at ELx, so I don't want to put the AArch32 kernel bits into SCTLR_EL1_RESET.

Quoted 3 lines

My only remaining concern is exactly what we must initialize. If there's any documentation we can refer to, that'd be great, otherwise I'll dig through your prior suggestion.

I don't know of any recommendation what to initialise, though the ARMv7 ARM speaks a bit about reset values.
So the SCTLR_KERNEL value we use in arch/aarch32/include/asm/cpu.h seems to work (although it's not perfect), but we should use the same in the KERNEL_32 case in arch/aarch64/include/asm/cpu.h.

Going through the SCTLR description in the ARMv8 *and* ARMv7 ARM again, I think a safe and sane init value would be to set bits [23,22,18,16,11,5,4,3] to one. This differs slightly from the value I told you above, which I think was what I observed on an Cortex-A7 when dumping SCTLR in the decompressor code.

I became aware of the issue when I tried to start an ARM kernel on a Cortex-A53 board, with U-Boot dropping from AArch64-EL2 to AArch32-EL1. SCTLR_EL1 there got initialised according to the ARMv8 ARM (bits [29,28,23,22,20,11] set), but this made the kernel decompressor hang as soon as the MMU got enabled. I *think* I bisected it down to bits 28 and 29, which are RES1 in AArch64, but enable TEX remap and the Access Flag in v7, and both reset to 0 there. I haven't tried that on the model with the boot-wrapper, but I guess we see the same problem there.

In any case it seems to be already broken, so a fix or discussion doesn't need to block this series.

Cheers,
Andre

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#15 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

Quoted 55 lines · Andre Przywara

On Mon, 17 Jan 2022 13:05:54 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 46 lines · Mark Rutland
Quoted 29 lines · Andre Przywara
Quoted 4 lines

On Fri, 14 Jan 2022 10:56:46 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Hi Andre,

Quoted 21 lines
Quoted 9 lines
arch/aarch64/include/asm/cpu.h +3 -1
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))  
 
-#define SCTLR_EL1_RES1		(3 << 28 | 3 << 22 | 1 << 11)
 
 #ifdef KERNEL_32
 /* 32-bit kernel decompressor uses CP15 barriers */

#define SCTLR_EL1_KERNEL (SCTLR_EL1_RES1 | SCTLR_EL1_CP15BEN)

So I wonder if this actually works? The ARMv7 version of SCTLR differs in some bits from both the ARMv8 AArch32 version and more importantly the AArch64 version.

I had troubles the other day running the
arm32 Linux kernel decompressor with some ARMv8 SCTLR_EL1 reset value. The decompressor code does only read-modify-write of SCTLR (probably to cover multiple architecture revisions), so some bits might stay wrong. In particular I think having bits 28 and 29 set caused problems. By looking at the ARMv7 ARM and with experimentation I came up with 0x00c00878 as a safe and working value.

Having re-read all of this, I believe you're right; I'll rework the AArch32-kernel SCTLR_EL1_KERNEL to not use the AArch64 bit definitions, and I'll add some commentary explaining that we're writing to the AArch32 format.

Quoted 1 line
Quoted 1 line

Shall we have a separate reset value for 32bit?

Assuming you meant to alter the SCTLR_ELx_KERNEL definition, yes.

The idea of splitting the SCTLR_ELx_RESET and SCTLR_ELx_KERNEL definitions was that the former would be whatever the boot-wrapper needed to run at ELx, and the latter was whatever we must initialize for the kernel to run at ELx, so I don't want to put the AArch32 kernel bits into SCTLR_EL1_RESET.

Quoted 3 lines

My only remaining concern is exactly what we must initialize. If there's any documentation we can refer to, that'd be great, otherwise I'll dig through your prior suggestion.

Quoted 5 lines

I don't know of any recommendation what to initialise, though the ARMv7 ARM speaks a bit about reset values.
So the SCTLR_KERNEL value we use in arch/aarch32/include/asm/cpu.h seems to work (although it's not perfect), but we should use the same in the KERNEL_32 case in arch/aarch64/include/asm/cpu.h.

Ok; for now I'm going to copy the SCTLR_KERNEL definition as-is from the aarch32 header, and add a comment that we should restructure things to share a single definition in future.

The rest of the reply is mostly for future reference.

Quoted 5 lines

Going through the SCTLR description in the ARMv8 *and* ARMv7 ARM again, I think a safe and sane init value would be to set bits [23,22,18,16,11,5,4,3] to one. This differs slightly from the value I told you above, which I think was what I observed on an Cortex-A7 when dumping SCTLR in the decompressor code.

Having delved through the ARM ARM, those bits look right to me. Below is expanded detail for each bit to save having to delve into the ARM ARM again.

Looking at the latest ARM ARM (ARM DDI 0487G.b), section G8.2.126 "SCTLR, System Control Register" starting on page G8-6810:

* Bit 25 / EE
RES0/RES1 in depending on implemented endianness support. Currently the boot-wrapper doesn't handle missin support for either BE or LE, so I'm going to ignore this for now and just assume LE is present (so 0 is a legitimate value).

* Bit 23 / SPAN
RES1 where FEAT_PAN is not implemented.

* Bit 22
RES1 always.

* Bit 18 / nTWE
Not RESx, but warm-resets to 1.

* Bit 16 / nTWI
Not RESx, but warm-resets to 1.

* Bit 11
RES1 always.

* Bit 5 / CP15BEN
Not RESx, but warm-resets to 1.

* Bit 4 / LSMAOE
RES1 where FEAT_LSMAOC is not implemented. Warm-resets to 1 where FEAT_LSMAOC is implemented.

* Bit 3 / nTLSMD
RES1 where FEAT_LSMAOC is not implemented. Warm-resets to 1 where FEAT_LSMAOC is implemented.

... with all other bits being UNKNOWN, RES0, or warm-rest to 0.

Quoted 11 lines

I became aware of the issue when I tried to start an ARM kernel on a Cortex-A53 board, with U-Boot dropping from AArch64-EL2 to AArch32-EL1. SCTLR_EL1 there got initialised according to the ARMv8 ARM (bits [29,28,23,22,20,11] set), but this made the kernel decompressor hang as soon as the MMU got enabled. I *think* I bisected it down to bits 28 and 29, which are RES1 in AArch64, but enable TEX remap and the Access Flag in v7, and both reset to 0 there. I haven't tried that on the model with the boot-wrapper, but I guess we see the same problem there.

In any case it seems to be already broken, so a fix or discussion doesn't need to block this series.

Sure; I think I'll save the AArch32 cleanup for a follow-up and go for a simple change to jsut avoid making it worse in this patch.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#16 lore raw

Re: [bootwrapper PATCH v2 06/13] aarch64: initialize SCTLR_ELx for the boot-wrapper

In reply to Mark Rutland
Quoted 21 lines · Mark Rutland
Quoted 9 lines · Andre Przywara

arch/aarch64/include/asm/cpu.h line 32

Quoted 3 lines · On Fri, 14 Jan 2022 10:56:46 +0000 Mark Rutland <[email protected]> wrote:
arch/aarch64/include/asm/cpu.h +3 -0
+
+#define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
+				 BIT(11) | BIT(8) | BIT(7) | BIT(4))

- The kernel sets TSCXT(bit[20]), and the ARM ARM says that the value should be RES1 if FEAT_CSV2_* is not implemented. Should we copy this?

Yes, we should. I'll go and fold that in.

Quoted 3 lines

- The kernel clears ITD(bit[7]), and the ARM ARM says it's *Otherwise* RES1 (no AArch32 in EL0). I feel like we should not disable IT instructions in EL0 needlessly?

Per the ARM ARM the bit resets to an UNKNOWN value, and so per our usual policy the kernel must initialize that before it can depend upon it, and IIUC you say the kernel already does so.

Quoted 6 lines

So it shouldn't matter what the boot-wrapper does, and for consitency, I think the boot-wrapper should set this to 0b1.

Quoted 1 line

- I also feel like we should set CP15BEN(bit[5]), for similar reasons.

I agree.

I thought you meant it was also RES1 in the same case as ITD. It's actually RES0 when there's no AArch32 EL0 support, so I'm not going to add it to SCTLR_EL1_RES1.

I have added TSCXT (bit[20]) and ITD (bit[7]), as above, as those are RES1 in some contexts.

THanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#17 lore raw

PATCH v2 7/13 Rework common init C code

In reply to Mark Rutland

In init_platform() we initialize a UART and announce the presence of the bootwrapper to the world. We do this relatively late in the boot-flow, and prior to this will silently ignore errors (e.g. in gic_secure_init).

To make it possible to provide improved diagnostics, and to allow us to move more initialization into C, this patch reworks the init code to call a C function earlier, where we can announce the presence of the boot-wrapper and perform other initialization.

In subsequent patches this will be expanded with more CPU initialization.

---
 Makefile.am         |  2 +-
 arch/aarch32/boot.S |  5 +++++
 arch/aarch64/boot.S |  4 ++++
 common/boot.c       |  4 ----
 common/init.c       | 24 ++++++++++++++++++++++++
 common/platform.c   | 12 +++++++-----
 include/platform.h  | 17 +++++++++++++++++
 7 files changed, 58 insertions(+), 10 deletions(-)
 create mode 100644 common/init.c
 create mode 100644 include/platform.h
Makefile.am +1 -1
index f941b07..0651c38 100644
@@ -34,7 +34,7 @@ endif
 PSCI_CPU_OFF	:= 0x84000002
 
 COMMON_SRC	:= common/
-COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o
+COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o init.o
 
 ARCH_OBJ	:= boot.o stack.o utils.o
 
arch/aarch32/boot.S +5 -0
index 00c432d..ee073ea 100644
@@ -48,6 +48,9 @@ ASM_FUNC(_start)
 	mov	r0, #1
 	ldr	r1, =flag_no_el3
 	str	r0, [r1]
+
+	bl	cpu_init_bootwrapper
+
 	b	start_no_el3
 
 _switch_monitor:
@@ -71,6 +74,8 @@ _monitor:
 	ldr	r0, =COUNTER_FREQ
 	mcr	p15, 0, r0, c14, c0, 0		@ CNTFRQ
 
+	bl	cpu_init_bootwrapper
+
 	bl	gic_secure_init
 
 	/* Initialise boot method */
arch/aarch64/boot.S +4 -0
index 45a0367..17b4a75 100644
@@ -141,6 +141,8 @@ reset_at_el3:
 	b.eq	err_invalid_id
 	bl	setup_stack
 
+	bl	cpu_init_bootwrapper
+
 	bl	gic_secure_init
 
 	b	start_el3
@@ -181,6 +183,8 @@ reset_no_el3:
 	ldr	x1, =flag_no_el3
 	str	w0, [x1]
 
+	bl	cpu_init_bootwrapper
+
 	b	start_no_el3
 
 err_invalid_id:
common/boot.c +0 -4
index c74d34c..29d53a4 100644
@@ -12,8 +12,6 @@
 extern unsigned long entrypoint;
 extern unsigned long dtb;
 
-void init_platform(void);
-
 void __noreturn jump_kernel(unsigned long address,
 			    unsigned long a0,
 			    unsigned long a1,
@@ -62,8 +60,6 @@ void __noreturn first_spin(unsigned int cpu, unsigned long *mbox,
 			   unsigned long invalid)
 {
 	if (cpu == 0) {
-		init_platform();
-
 		*mbox = (unsigned long)&entrypoint;
 		sevl();
 		spin(mbox, invalid, 1);
common/init.c new file +24 -0
new file mode 100644
index 0000000..9c471c9
@@ -0,0 +1,24 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <cpu.h>
+#include <platform.h>
+
+static void announce_bootwrapper(void)
+{
+	print_string("Boot-wrapper v0.2\r\n\r\n");
+}
+
+void cpu_init_bootwrapper(void)
+{
+	if (this_cpu_logical_id() == 0) {
+		init_uart();
+		announce_bootwrapper();
+		init_platform();
+	}
+}
common/platform.c +7 -5
index d11f568..47bf547 100644
@@ -1,5 +1,5 @@
 /*
- * platform.c - code to initialise everything required when first booting.
+ * platform.c - Platform initialization and I/O.
  *
  * Copyright (C) 2015 ARM Limited. All rights reserved.
  *
@@ -7,6 +7,7 @@
  * found in the LICENSE.txt file.
  */
 
+#include <cpu.h>
 #include <stdint.h>
 
 #include <asm/io.h>
@@ -30,7 +31,7 @@
 #define V2M_SYS(reg)	((void *)SYSREGS_BASE + V2M_SYS_##reg)
 #endif
 
-static void print_string(const char *str)
+void print_string(const char *str)
 {
 	uint32_t flags;
 
@@ -47,7 +48,7 @@ static void print_string(const char *str)
 	}
 }
 
-void init_platform(void)
+void init_uart(void)
 {
 	/*
 	 * UART initialisation (38400 8N1)
@@ -58,9 +59,10 @@ void init_platform(void)
 	raw_writel(0x70,	PL011(UART_LCR_H));
 	/* Enable the UART, TXen and RXen */
 	raw_writel(0x301,	PL011(UARTCR));
+}
 
-	print_string("Boot-wrapper v0.2\r\n\r\n");
-
+void init_platform(void)
+{
 #ifdef SYSREGS_BASE
 	/*
 	 * CLCD output site MB
include/platform.h new file +17 -0
new file mode 100644
index 0000000..e5248e1
@@ -0,0 +1,17 @@
+/*
+ * include/platform.h - Platform initialization and I/O.
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __PLATFORM_H
+#define __PLATFORM_H
+
+void print_string(const char *str);
+void init_uart(void);
+
+void init_platform(void);
+
+#endif /* __PLATFORM_H */
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 7/13
#18 lore raw

Re: [bootwrapper PATCH v2 07/13] Rework common init C code

Quoted 9 lines · On Fri, 14 Jan 2022 10:56:47 +0000 Mark Rutland <[email protected]> wrote:

In init_platform() we initialize a UART and announce the presence of the bootwrapper to the world. We do this relatively late in the boot-flow, and prior to this will silently ignore errors (e.g. in gic_secure_init).

To make it possible to provide improved diagnostics, and to allow us to move more initialization into C, this patch reworks the init code to call a C function earlier, where we can announce the presence of the boot-wrapper and perform other initialization.

Quoted 4 lines

In subsequent patches this will be expanded with more CPU initialization.

Cheers,
Andre

Quoted 185 lines
---
 Makefile.am         |  2 +-
 arch/aarch32/boot.S |  5 +++++
 arch/aarch64/boot.S |  4 ++++
 common/boot.c       |  4 ----
 common/init.c       | 24 ++++++++++++++++++++++++
 common/platform.c   | 12 +++++++-----
 include/platform.h  | 17 +++++++++++++++++
 7 files changed, 58 insertions(+), 10 deletions(-)
 create mode 100644 common/init.c
 create mode 100644 include/platform.h
Makefile.am +1 -1
index f941b07..0651c38 100644
@@ -34,7 +34,7 @@ endif
 PSCI_CPU_OFF	:= 0x84000002
 
 COMMON_SRC	:= common/
-COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o
+COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o init.o
 
 ARCH_OBJ	:= boot.o stack.o utils.o
 
arch/aarch32/boot.S +5 -0
index 00c432d..ee073ea 100644
@@ -48,6 +48,9 @@ ASM_FUNC(_start)
 	mov	r0, #1
 	ldr	r1, =flag_no_el3
 	str	r0, [r1]
+
+	bl	cpu_init_bootwrapper
+
 	b	start_no_el3
 
 _switch_monitor:
@@ -71,6 +74,8 @@ _monitor:
 	ldr	r0, =COUNTER_FREQ
 	mcr	p15, 0, r0, c14, c0, 0		@ CNTFRQ
 
+	bl	cpu_init_bootwrapper
+
 	bl	gic_secure_init
 
 	/* Initialise boot method */
arch/aarch64/boot.S +4 -0
index 45a0367..17b4a75 100644
@@ -141,6 +141,8 @@ reset_at_el3:
 	b.eq	err_invalid_id
 	bl	setup_stack
 
+	bl	cpu_init_bootwrapper
+
 	bl	gic_secure_init
 
 	b	start_el3
@@ -181,6 +183,8 @@ reset_no_el3:
 	ldr	x1, =flag_no_el3
 	str	w0, [x1]
 
+	bl	cpu_init_bootwrapper
+
 	b	start_no_el3
 
 err_invalid_id:
common/boot.c +0 -4
index c74d34c..29d53a4 100644
@@ -12,8 +12,6 @@
 extern unsigned long entrypoint;
 extern unsigned long dtb;
 
-void init_platform(void);
-
 void __noreturn jump_kernel(unsigned long address,
 			    unsigned long a0,
 			    unsigned long a1,
@@ -62,8 +60,6 @@ void __noreturn first_spin(unsigned int cpu, unsigned long *mbox,
 			   unsigned long invalid)
 {
 	if (cpu == 0) {
-		init_platform();
-
 		*mbox = (unsigned long)&entrypoint;
 		sevl();
 		spin(mbox, invalid, 1);
common/init.c new file +24 -0
new file mode 100644
index 0000000..9c471c9
@@ -0,0 +1,24 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <cpu.h>
+#include <platform.h>
+
+static void announce_bootwrapper(void)
+{
+	print_string("Boot-wrapper v0.2\r\n\r\n");
+}
+
+void cpu_init_bootwrapper(void)
+{
+	if (this_cpu_logical_id() == 0) {
+		init_uart();
+		announce_bootwrapper();
+		init_platform();
+	}
+}
common/platform.c +7 -5
index d11f568..47bf547 100644
@@ -1,5 +1,5 @@
 /*
- * platform.c - code to initialise everything required when first booting.
+ * platform.c - Platform initialization and I/O.
  *
  * Copyright (C) 2015 ARM Limited. All rights reserved.
  *
@@ -7,6 +7,7 @@
  * found in the LICENSE.txt file.
  */
 
+#include <cpu.h>
 #include <stdint.h>
 
 #include <asm/io.h>
@@ -30,7 +31,7 @@
 #define V2M_SYS(reg)	((void *)SYSREGS_BASE + V2M_SYS_##reg)
 #endif
 
-static void print_string(const char *str)
+void print_string(const char *str)
 {
 	uint32_t flags;
 
@@ -47,7 +48,7 @@ static void print_string(const char *str)
 	}
 }
 
-void init_platform(void)
+void init_uart(void)
 {
 	/*
 	 * UART initialisation (38400 8N1)
@@ -58,9 +59,10 @@ void init_platform(void)
 	raw_writel(0x70,	PL011(UART_LCR_H));
 	/* Enable the UART, TXen and RXen */
 	raw_writel(0x301,	PL011(UARTCR));
+}
 
-	print_string("Boot-wrapper v0.2\r\n\r\n");
-
+void init_platform(void)
+{
 #ifdef SYSREGS_BASE
 	/*
 	 * CLCD output site MB
include/platform.h new file +7 -0
new file mode 100644
index 0000000..e5248e1
@@ -0,0 +1,17 @@
+/*
+ * include/platform.h - Platform initialization and I/O.
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.

include/platform.h line 17

Quoted 10 lines
include/platform.h new file +10 -0
+ */
+#ifndef __PLATFORM_H
+#define __PLATFORM_H
+
+void print_string(const char *str);
+void init_uart(void);
+
+void init_platform(void);
+
+#endif /* __PLATFORM_H */

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#19 lore raw

PATCH v2 8/13 Announce boot-wrapper mode / exception level

In reply to Mark Rutland

When something goes wrong within the boot-wrapper, it can be very helpful to know where we started from. Add an arch_announce() function to log this early in the boot process. More information can be added here in future.

This is logged ot the serial console as:

| Boot-wrapper v0.2
| Entered at EL3

---
 Makefile.am                    |  2 +-
 arch/aarch32/include/asm/cpu.h |  9 +++++++++
 arch/aarch32/init.c            | 30 ++++++++++++++++++++++++++++++
 arch/aarch64/init.c            | 19 +++++++++++++++++++
 common/init.c                  |  6 +++++-
 common/platform.c              | 24 ++++++++++++++----------
 include/platform.h             |  1 +
 7 files changed, 79 insertions(+), 12 deletions(-)
 create mode 100644 arch/aarch32/init.c
 create mode 100644 arch/aarch64/init.c
Makefile.am +1 -1
index 0651c38..885a77c 100644
@@ -36,7 +36,7 @@ PSCI_CPU_OFF	:= 0x84000002
 COMMON_SRC	:= common/
 COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o init.o
 
-ARCH_OBJ	:= boot.o stack.o utils.o
+ARCH_OBJ	:= boot.o stack.o utils.o init.o
 
 if BOOTWRAPPER_32
 CPPFLAGS	+= -DBOOTWRAPPER_32
arch/aarch32/include/asm/cpu.h +9 -0
index d691c7b..aa72204 100644
@@ -44,6 +44,15 @@
 #define sevl()		asm volatile ("sev" : : : "memory")
 #endif
 
+static inline unsigned long read_cpsr(void)
+{
+	unsigned long cpsr;
+	asm volatile ("mrs      %0, cpsr\n" : "=r" (cpsr));
+	return cpsr;
+}
+
+#define read_cpsr_mode()       (read_cpsr() & PSR_MODE_MASK)
+
 #define MPIDR		"p15, 0, %0, c0, c0, 5"
 #define ID_PFR1		"p15, 0, %0, c0, c1, 1"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
arch/aarch32/init.c new file +30 -0
new file mode 100644
index 0000000..b29ebb4
@@ -0,0 +1,30 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <platform.h>
+
+#include <asm/cpu.h>
+
+static const char *mode_string(void)
+{
+	switch (read_cpsr_mode()) {
+	case PSR_MON:
+		return "PL1";
+	case PSR_HYP:
+		return "PL2 (Non-secure)";
+	default:
+		return "<UNKNOWN MODE>";
+	}
+}
+
+void announce_arch(void)
+{
+	print_string("Entered at ");
+	print_string(mode_string());
+	print_string("\r\n");
+}
arch/aarch64/init.c new file +19 -0
new file mode 100644
index 0000000..82816e7
@@ -0,0 +1,19 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <cpu.h>
+#include <platform.h>
+
+void announce_arch(void)
+{
+	unsigned char el = mrs(CurrentEl) >> 2;
+
+	print_string("Entered at EL");
+	print_char('0' + el);
+	print_string("\r\n");
+}
common/init.c +5 -1
index 9c471c9..2600f73 100644
@@ -11,14 +11,18 @@
 
 static void announce_bootwrapper(void)
 {
-	print_string("Boot-wrapper v0.2\r\n\r\n");
+	print_string("Boot-wrapper v0.2\r\n");
 }
 
+void announce_arch(void);
+
 void cpu_init_bootwrapper(void)
 {
 	if (this_cpu_logical_id() == 0) {
 		init_uart();
 		announce_bootwrapper();
+		announce_arch();
+		print_string("\r\n");
 		init_platform();
 	}
 }
common/platform.c +14 -10
index 47bf547..80d0562 100644
@@ -31,21 +31,25 @@
 #define V2M_SYS(reg)	((void *)SYSREGS_BASE + V2M_SYS_##reg)
 #endif
 
-void print_string(const char *str)
+void print_char(char c)
 {
 	uint32_t flags;
 
-	while (*str) {
-		do
-			flags = raw_readl(PL011(UARTFR));
-		while (flags & PL011_UARTFR_FIFO_FULL);
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_FIFO_FULL);
+
+	raw_writel(c, PL011(UARTDR));
 
-		raw_writel(*str++, PL011(UARTDR));
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_BUSY);
+}
 
-		do
-			flags = raw_readl(PL011(UARTFR));
-		while (flags & PL011_UARTFR_BUSY);
-	}
+void print_string(const char *str)
+{
+	while (*str)
+		print_char(*str++);
 }
 
 void init_uart(void)
include/platform.h +1 -0
index e5248e1..237b481 100644
@@ -9,6 +9,7 @@
 #ifndef __PLATFORM_H
 #define __PLATFORM_H
 
+void print_char(char c);
 void print_string(const char *str);
 void init_uart(void);
 
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 8/13
#20 lore raw

Re: [bootwrapper PATCH v2 08/13] Announce boot-wrapper mode / exception level

On Fri, 14 Jan 2022 10:56:48 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 9 lines

When something goes wrong within the boot-wrapper, it can be very helpful to know where we started from. Add an arch_announce() function to log this early in the boot process. More information can be added here in future.

This is logged ot the serial console as:

| Boot-wrapper v0.2
| Entered at EL3

I like that one, and apart from the (already existing) UART issue below, this looks fine:

Quoted 1 line

Cheers,
Andre

Quoted 147 lines
---
 Makefile.am                    |  2 +-
 arch/aarch32/include/asm/cpu.h |  9 +++++++++
 arch/aarch32/init.c            | 30 ++++++++++++++++++++++++++++++
 arch/aarch64/init.c            | 19 +++++++++++++++++++
 common/init.c                  |  6 +++++-
 common/platform.c              | 24 ++++++++++++++----------
 include/platform.h             |  1 +
 7 files changed, 79 insertions(+), 12 deletions(-)
 create mode 100644 arch/aarch32/init.c
 create mode 100644 arch/aarch64/init.c
Makefile.am +1 -1
index 0651c38..885a77c 100644
@@ -36,7 +36,7 @@ PSCI_CPU_OFF	:= 0x84000002
 COMMON_SRC	:= common/
 COMMON_OBJ	:= boot.o bakery_lock.o platform.o lib.o init.o
 
-ARCH_OBJ	:= boot.o stack.o utils.o
+ARCH_OBJ	:= boot.o stack.o utils.o init.o
 
 if BOOTWRAPPER_32
 CPPFLAGS	+= -DBOOTWRAPPER_32
arch/aarch32/include/asm/cpu.h +9 -0
index d691c7b..aa72204 100644
@@ -44,6 +44,15 @@
 #define sevl()		asm volatile ("sev" : : : "memory")
 #endif
 
+static inline unsigned long read_cpsr(void)
+{
+	unsigned long cpsr;
+	asm volatile ("mrs      %0, cpsr\n" : "=r" (cpsr));
+	return cpsr;
+}
+
+#define read_cpsr_mode()       (read_cpsr() & PSR_MODE_MASK)
+
 #define MPIDR		"p15, 0, %0, c0, c0, 5"
 #define ID_PFR1		"p15, 0, %0, c0, c1, 1"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
arch/aarch32/init.c new file +30 -0
new file mode 100644
index 0000000..b29ebb4
@@ -0,0 +1,30 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <platform.h>
+
+#include <asm/cpu.h>
+
+static const char *mode_string(void)
+{
+	switch (read_cpsr_mode()) {
+	case PSR_MON:
+		return "PL1";
+	case PSR_HYP:
+		return "PL2 (Non-secure)";
+	default:
+		return "<UNKNOWN MODE>";
+	}
+}
+
+void announce_arch(void)
+{
+	print_string("Entered at ");
+	print_string(mode_string());
+	print_string("\r\n");
+}
arch/aarch64/init.c new file +19 -0
new file mode 100644
index 0000000..82816e7
@@ -0,0 +1,19 @@
+/*
+ * init.c - common boot-wrapper initialization
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#include <cpu.h>
+#include <platform.h>
+
+void announce_arch(void)
+{
+	unsigned char el = mrs(CurrentEl) >> 2;
+
+	print_string("Entered at EL");
+	print_char('0' + el);
+	print_string("\r\n");
+}
common/init.c +5 -1
index 9c471c9..2600f73 100644
@@ -11,14 +11,18 @@
 
 static void announce_bootwrapper(void)
 {
-	print_string("Boot-wrapper v0.2\r\n\r\n");
+	print_string("Boot-wrapper v0.2\r\n");
 }
 
+void announce_arch(void);
+
 void cpu_init_bootwrapper(void)
 {
 	if (this_cpu_logical_id() == 0) {
 		init_uart();
 		announce_bootwrapper();
+		announce_arch();
+		print_string("\r\n");
 		init_platform();
 	}
 }
common/platform.c +1 -5
index 47bf547..80d0562 100644
@@ -31,21 +31,25 @@
 #define V2M_SYS(reg)	((void *)SYSREGS_BASE + V2M_SYS_##reg)
 #endif
 
-void print_string(const char *str)
+void print_char(char c)
 {
 	uint32_t flags;
 
-	while (*str) {
-		do
-			flags = raw_readl(PL011(UARTFR));
-		while (flags & PL011_UARTFR_FIFO_FULL);

common/platform.c line 46

Quoted 10 lines
common/platform.c +8 -1
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_FIFO_FULL);
+
+	raw_writel(c, PL011(UARTDR));
 
-		raw_writel(*str++, PL011(UARTDR));
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_BUSY);

Apologies if that appears to be nitpicking over a totally pointless issue (given the nature of the *emulated* PL011 in the model), but:

I understand that this code has not changed, but this loop basically renders the FIFOs ineffective. Is this intended? I mean we explicitly enable the FIFOs in init_uart(), but then poll here until the transmit FIFO becomes empty, after *every* character pushed. I see that we probably want the output to be synchronous, for debug reasons, but I wonder if this should be achieved via an extra uart_flush() routine, once per line output? So call uart_flush() just in print_string(), but not in print_char(), for instance?

Cheers,
Andre

Quoted 18 lines
common/platform.c +5 -4
+}
 
-		do
-			flags = raw_readl(PL011(UARTFR));
-		while (flags & PL011_UARTFR_BUSY);
-	}
+void print_string(const char *str)
+{
+	while (*str)
+		print_char(*str++);
 }
 
 void init_uart(void)
include/platform.h +0 -0
index e5248e1..237b481 100644

include/platform.h line 15

Quoted 7 lines
include/platform.h +1 -0
@@ -9,6 +9,7 @@
 #ifndef __PLATFORM_H
 #define __PLATFORM_H
 
+void print_char(char c);
 void print_string(const char *str);
 void init_uart(void);
 

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#21 lore raw

Re: [bootwrapper PATCH v2 08/13] Announce boot-wrapper mode / exception level

Quoted 15 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:48 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 9 lines

When something goes wrong within the boot-wrapper, it can be very helpful to know where we started from. Add an arch_announce() function to log this early in the boot process. More information can be added here in future.

This is logged ot the serial console as:

| Boot-wrapper v0.2
| Entered at EL3

Quoted 6 lines

I like that one, and apart from the (already existing) UART issue below, this looks fine:

Quoted 1 line

Thanks!

Quoted 31 lines
Quoted 17 lines
common/platform.c +1 -5
index 47bf547..80d0562 100644
@@ -31,21 +31,25 @@
 #define V2M_SYS(reg)	((void *)SYSREGS_BASE + V2M_SYS_##reg)
 #endif
 
-void print_string(const char *str)
+void print_char(char c)
 {
 	uint32_t flags;
 
-	while (*str) {
-		do
-			flags = raw_readl(PL011(UARTFR));
-		while (flags & PL011_UARTFR_FIFO_FULL);

common/platform.c line 46

Quoted 10 lines
common/platform.c +8 -1
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_FIFO_FULL);
+
+	raw_writel(c, PL011(UARTDR));
 
-		raw_writel(*str++, PL011(UARTDR));
+	do {
+		flags = raw_readl(PL011(UARTFR));
+	} while (flags & PL011_UARTFR_BUSY);

Apologies if that appears to be nitpicking over a totally pointless issue (given the nature of the *emulated* PL011 in the model), but:

Quoted 2 lines

I understand that this code has not changed, but this loop basically renders the FIFOs ineffective. Is this intended?

Hmm... AFAICT, we started to enable the FIFOs in commit:

26a17ad59544f026 ("bootwrapper: improve UART initialisation")

... but back then the boot-wrapper didn't print anything to the UART, so that was purely about putting the UART into a sane state for the OS.

We only started to write to the UART in commit:

7ff3872adb33b068 ("Rewrite platform initialisation in C")

... where we said nothing about the FIFOs.

Therefore, I think there was no intent either way to use/avoid the FIFOs within the boot-wrapper.

Quoted 6 lines

I mean we explicitly enable the FIFOs in init_uart(), but then poll here until the transmit FIFO becomes empty, after *every* character pushed. I see that we probably want the output to be synchronous, for debug reasons, but I wonder if this should be achieved via an extra uart_flush() routine, once per line output? So call uart_flush() just in print_string(), but not in print_char(), for instance?

Hmm... we also use print_char() directly from the aarch64 init code when logging the exception level, so we'd need an explicit sync there, if we do want that output before other things can go wrong.

Otherwise, I don't have strong feelings either way, but I would like to have the output completed before print_*() return. We could add an internal print_char_nosync() that print_{char,string}() use to handle that.

I'm happy to take a follow-up patch for that, but for now I'm going to leave this as-is.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#22 lore raw

PATCH v2 9/13 aarch64: move the bulk of EL3 initialization to C

In reply to Mark Rutland

The majority of state that we initialize at EL3 is necessary for code at lower ELs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

This patch migrates the AArch64 EL3 initialization to C.

There should be no functional change as a result of this patch.

---
 Makefile.am                    |  2 +-
 arch/aarch64/boot.S            | 92 +---------------------------------
 arch/aarch64/include/asm/cpu.h | 36 ++++++++++++-
 arch/aarch64/init.c            | 69 +++++++++++++++++++++++++
 4 files changed, 107 insertions(+), 92 deletions(-)
Makefile.am +1 -1
index 885a77c..a598b95 100644
@@ -153,7 +153,7 @@ $(COMMON_SRC):
 %.o: %.S Makefile | $(ARCH_SRC)
 	$(CC) $(CPPFLAGS) -D__ASSEMBLY__ $(CFLAGS) $(DEFINES) -c -o $@ $<
 
-%.o: %.c Makefile | $(COMMON_SRC)
+%.o: %.c Makefile | $(ARCH_SRC) $(COMMON_SRC)
 	$(CC) $(CPPFLAGS) $(CFLAGS) $(DEFINES) -c -o $@ $<
 
 model.lds: $(LD_SCRIPT) Makefile
arch/aarch64/boot.S +2 -90

104 lines not on this page · raw message

arch/aarch64/include/asm/cpu.h +35 -1
index 1e9141a..e078ae4 100644
@@ -31,7 +31,41 @@
 #define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
 				 BIT(11) | BIT(8) | BIT(7) | BIT(4))
 
-#define HCR_EL2_RES1		(BIT(1))
+#define ZCR_EL3				s3_6_c1_c2_0
+#define ZCR_EL3_LEN			BITS(3, 1)
+
+#define MDCR_EL3_NSPB_NS_NOTRAP		(UL(3) << 12)
+#define MDCR_EL3_NSTB_NS_NOTRAP		(UL(3) << 24)
+#define MDCR_EL3_ENPMSN			BIT(36)
+
+#define SCR_EL3_RES1			BITS(5, 4)
+#define SCR_EL3_NS			BIT(0)
+#define SCR_EL3_HCE			BIT(8)
+#define SCR_EL3_RW			BIT(10)
+#define SCR_EL3_APK			BIT(16)
+#define SCR_EL3_API			BIT(17)
+#define SCR_EL3_ATA			BIT(26)
+#define SCR_EL3_FGTEN			BIT(27)
+#define SCR_EL3_ECVEN			BIT(28)
+#define SCR_EL3_TME			BIT(34)
+
+#define HCR_EL2_RES1			BIT(1)
+
+#define ID_AA64DFR0_EL1_PMSVER		BITS(35, 32)
+#define ID_AA64DFR0_EL1_TRACEBUFFER	BITS(47, 44)
+
+#define ID_AA64ISAR0_EL1_TME		BITS(27, 24)
+
+#define ID_AA64ISAR1_EL1_APA		BITS(7, 4)
+#define ID_AA64ISAR1_EL1_API		BITS(11, 8)
+#define ID_AA64ISAR1_EL1_GPA		BITS(27, 24)
+#define ID_AA64ISAR1_EL1_GPI		BITS(31, 28)
+
+#define ID_AA64MMFR0_EL1_FGT		BITS(59, 56)
+#define ID_AA64MMFR0_EL1_ECV		BITS(63, 60)
+
+#define ID_AA64PFR1_EL1_MTE		BITS(11, 8)
+#define ID_AA64PFR0_EL1_SVE		BITS(35, 32)
 
 /*
  * Initial register values required for the boot-wrapper to run out-of-reset.
arch/aarch64/init.c +69 -0

78 lines not on this page · raw message

Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#23 lore raw

Re: [bootwrapper PATCH v2 09/13] aarch64: move the bulk of EL3 initialization to C

On Fri, 14 Jan 2022 10:56:49 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 5 lines

The majority of state that we initialize at EL3 is necessary for code at lower ELs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

Ah, thanks, that looks much nicer and easier to read now, also is more robust, as keeping register values alive for more than a few assembly lines always scares me.

Quoted 3 lines

This patch migrates the AArch64 EL3 initialization to C.

There should be no functional change as a result of this patch.

I compared the removed assembly code against to added C code, and also checked the register bits against the ARM ARM.
Two (and a half) things stood out, see below:

Quoted 137 lines
---
 Makefile.am                    |  2 +-
 arch/aarch64/boot.S            | 92 +---------------------------------
 arch/aarch64/include/asm/cpu.h | 36 ++++++++++++-
 arch/aarch64/init.c            | 69 +++++++++++++++++++++++++
 4 files changed, 107 insertions(+), 92 deletions(-)
Makefile.am +1 -1
index 885a77c..a598b95 100644
@@ -153,7 +153,7 @@ $(COMMON_SRC):
 %.o: %.S Makefile | $(ARCH_SRC)
 	$(CC) $(CPPFLAGS) -D__ASSEMBLY__ $(CFLAGS) $(DEFINES) -c -o $@ $<
 
-%.o: %.c Makefile | $(COMMON_SRC)
+%.o: %.c Makefile | $(ARCH_SRC) $(COMMON_SRC)
 	$(CC) $(CPPFLAGS) $(CFLAGS) $(DEFINES) -c -o $@ $<
 
 model.lds: $(LD_SCRIPT) Makefile
arch/aarch64/boot.S +2 -90
index 17b4a75..c0ec518 100644
@@ -45,96 +45,6 @@ reset_at_el3:
 	msr	sctlr_el3, x0
 	isb
 
-	mov	x0, #0x30			// RES1
-	orr	x0, x0, #(1 << 0)		// Non-secure EL1
-	orr	x0, x0, #(1 << 8)		// HVC enable
-
-	/* Enable pointer authentication if present */
-	mrs	x1, id_aa64isar1_el1
-	/* We check for APA+API and GPA+GPI */
-	ldr	x2, =((0xff << 24) | (0xff << 4))
-	and	x1, x1, x2
-	cbz	x1, 1f
-
-	orr	x0, x0, #(1 << 16)		// AP key enable
-	orr	x0, x0, #(1 << 17)		// AP insn enable
-1:
-	/* Enable TME if present */
-	mrs	x1, id_aa64isar0_el1
-	ubfx	x1, x1, #24, #4
-	cbz	x1, 1f
-
-	orr	x0, x0, #(1 << 34)		// TME enable
-1:
-	/* Enable FGT if present */
-	mrs	x1, id_aa64mmfr0_el1
-	ubfx	x1, x1, #56, #4
-	cbz	x1, 1f
-
-	orr	x0, x0, #(1 << 27)		// FGT enable
-1:
-	/* Enable ECV2 if present (allows CNTPOFF_EL2) */
-	mrs	x1, id_aa64mmfr0_el1
-	ubfx	x1, x1, #60, #4
-	cmp	x1, #2
-	b.lt	1f
-
-	orr	x0, x0, #(1 << 28)		// ECV enable
-1:
-	/* Enable MTE if present */
-	mrs	x10, id_aa64pfr1_el1
-	ubfx	x10, x10, #8, #4
-	cmp	x10, #2
-	b.lt	1f
-
-	orr	x0, x0, #(1 << 26)		// ATA enable
-1:
-#ifndef KERNEL_32
-	orr	x0, x0, #(1 << 10)		// 64-bit EL2
-#endif
-	msr	scr_el3, x0
-
-	msr	cptr_el3, xzr			// Disable copro. traps to EL3
-
-	mov	x0, xzr
-	mrs	x1, id_aa64dfr0_el1
-	ubfx	x1, x1, #32, #4
-	cbz	x1, 1f
-
-	// Enable SPE for the non-secure world.
-	orr	x0, x0, #(0x3 << 12)
-
-	// Do not trap PMSNEVFR_EL1 if present
-	cmp	x1, #3
-	b.lt	1f
-	orr	x0, x0, #(1 << 36)
-
-1:	mrs	x1, id_aa64dfr0_el1
-	ubfx	x1, x1, #44, #4
-	cbz	x1, 1f
-
-	// Enable TRBE for the non-secure world.
-	ldr	x1, =(0x3 << 24)
-	orr	x0, x0, x1
-
-1:	msr	mdcr_el3, x0			// Disable traps to EL3
-
-	mrs	x0, id_aa64pfr0_el1
-	ubfx	x0, x0, #32, #4			// SVE present?
-	cbz	x0, 1f				// Skip SVE init if not
-
-	mrs	x0, cptr_el3
-	orr	x0, x0, #CPTR_EL3_EZ		// enable SVE
-	msr	cptr_el3, x0
-	isb
-
-	mov	x0, #ZCR_EL3_LEN_MASK		// SVE: Enable full vector len
-	msr	ZCR_EL3, x0			// for EL2.
-
-1:
-	ldr	x0, =COUNTER_FREQ
-	msr	cntfrq_el0, x0
-
 	cpuid	x0, x1
 	bl	find_logical_id
 	cmp	x0, #MPIDR_INVALID
@@ -143,6 +53,8 @@ reset_at_el3:
 
 	bl	cpu_init_bootwrapper
 
+	bl	cpu_init_el3
+
 	bl	gic_secure_init
 
 	b	start_el3
arch/aarch64/include/asm/cpu.h +0 -0
index 1e9141a..e078ae4 100644

arch/aarch64/include/asm/cpu.h line 35

Quoted 6 lines
arch/aarch64/include/asm/cpu.h +2 -1
@@ -31,7 +31,41 @@
 #define SCTLR_EL1_RES1		(BIT(29) | BIT(28) | BIT(23) | BIT(22) | \
 				 BIT(11) | BIT(8) | BIT(7) | BIT(4))
 
-#define HCR_EL2_RES1		(BIT(1))
+#define ZCR_EL3				s3_6_c1_c2_0
+#define ZCR_EL3_LEN			BITS(3, 1)

The (current) actual length field should be BITS(3, 0), no?

Quoted 81 lines
arch/aarch64/include/asm/cpu.h +33 -0
+
+#define MDCR_EL3_NSPB_NS_NOTRAP		(UL(3) << 12)
+#define MDCR_EL3_NSTB_NS_NOTRAP		(UL(3) << 24)
+#define MDCR_EL3_ENPMSN			BIT(36)
+
+#define SCR_EL3_RES1			BITS(5, 4)
+#define SCR_EL3_NS			BIT(0)
+#define SCR_EL3_HCE			BIT(8)
+#define SCR_EL3_RW			BIT(10)
+#define SCR_EL3_APK			BIT(16)
+#define SCR_EL3_API			BIT(17)
+#define SCR_EL3_ATA			BIT(26)
+#define SCR_EL3_FGTEN			BIT(27)
+#define SCR_EL3_ECVEN			BIT(28)
+#define SCR_EL3_TME			BIT(34)
+
+#define HCR_EL2_RES1			BIT(1)
+
+#define ID_AA64DFR0_EL1_PMSVER		BITS(35, 32)
+#define ID_AA64DFR0_EL1_TRACEBUFFER	BITS(47, 44)
+
+#define ID_AA64ISAR0_EL1_TME		BITS(27, 24)
+
+#define ID_AA64ISAR1_EL1_APA		BITS(7, 4)
+#define ID_AA64ISAR1_EL1_API		BITS(11, 8)
+#define ID_AA64ISAR1_EL1_GPA		BITS(27, 24)
+#define ID_AA64ISAR1_EL1_GPI		BITS(31, 28)
+
+#define ID_AA64MMFR0_EL1_FGT		BITS(59, 56)
+#define ID_AA64MMFR0_EL1_ECV		BITS(63, 60)
+
+#define ID_AA64PFR1_EL1_MTE		BITS(11, 8)
+#define ID_AA64PFR0_EL1_SVE		BITS(35, 32)
 
 /*
  * Initial register values required for the boot-wrapper to run out-of-reset.
arch/aarch64/init.c +30 -0
index 82816e7..ded9871 100644
@@ -8,6 +8,7 @@
  */
 #include <cpu.h>
 #include <platform.h>
+#include <stdbool.h>
 
 void announce_arch(void)
 {
@@ -17,3 +18,71 @@ void announce_arch(void)
 	print_char('0' + el);
 	print_string("\r\n");
 }
+
+static inline bool kernel_is_32bit(void)
+{
+#ifdef KERNEL_32
+	return true;
+#else
+	return false;
+#endif
+}
+
+static inline bool cpu_has_pauth(void)
+{
+	const unsigned long id_pauth = ID_AA64ISAR1_EL1_APA |
+				       ID_AA64ISAR1_EL1_API |
+				       ID_AA64ISAR1_EL1_GPA |
+				       ID_AA64ISAR1_EL1_GPI;
+
+	return mrs(ID_AA64ISAR1_EL1) & id_pauth;
+}
+
+void cpu_init_el3(void)
+{
+	unsigned long scr = SCR_EL3_RES1 | SCR_EL3_NS | SCR_EL3_HCE;
+	unsigned long mdcr = 0;
+	unsigned long cptr = 0;
+
+	if (cpu_has_pauth())
+		scr |= SCR_EL3_APK | SCR_EL3_API;
+

arch/aarch64/init.c void announce_arch(void) · line 59

Quoted 10 lines
arch/aarch64/init.c +10 -0
+	if (mrs_field(ID_AA64ISAR0_EL1, TME))
+		scr |= SCR_EL3_TME;
+
+	if (mrs_field(ID_AA64MMFR0_EL1, FGT))
+		scr |= SCR_EL3_FGTEN;
+
+	if (mrs_field(ID_AA64MMFR0_EL1, ECV) >= 2)
+		scr |= SCR_EL3_ECVEN;
+
+	if (mrs_field(ID_AA64PFR1_EL1, MTE))

The assembly code checked for >=2, which seems correct to me, as SCR_EL3_ATA is about MTE2?

Quoted 15 lines
arch/aarch64/init.c +15 -0
+		scr |= SCR_EL3_ATA;
+
+	if (!kernel_is_32bit())
+		scr |= SCR_EL3_RW;
+
+	msr(SCR_EL3, scr);
+
+	msr(CPTR_EL3, cptr);
+
+	if (mrs_field(ID_AA64DFR0_EL1, PMSVER))
+		mdcr |= MDCR_EL3_NSPB_NS_NOTRAP;
+
+	if (mrs_field(ID_AA64DFR0_EL1, PMSVER) >= 3)
+		mdcr |= MDCR_EL3_ENPMSN;
+

arch/aarch64/init.c void announce_arch(void) · line 84

Quoted 10 lines
arch/aarch64/init.c +10 -0
+	if (mrs_field(ID_AA64DFR0_EL1, TRACEBUFFER))
+		mdcr |= MDCR_EL3_NSTB_NS_NOTRAP;
+
+	msr(MDCR_EL3, mdcr);
+
+	if (mrs_field(ID_AA64PFR0_EL1, SVE)) {
+		cptr |= CPTR_EL3_EZ;
+		msr(CPTR_EL3, cptr);
+		isb();
+		msr(ZCR_EL3, ZCR_EL3_LEN);

So when comparing this to the other uses of XXX_EL3_YYY, they typically describe a mask, but here we seems to abuse this as a value? And apart from bit 0 missing from it (as noted above), the existing code writes 0x1ff into that register, presumable to cover future vector length extensions beyond 2048 bits (which those RAZ/WI fields in bits[8:4] seem to suggest). So shall we define ZCR_EL3_MAX_VEC_LEN to 0x1ff above, and then use that? Or ignore the crystal ball, and just stick with 2048 bits, by writing 0xf?

The rest passed the checks.

Cheers,
Andre

arch/aarch64/init.c void announce_arch(void) · line 88

Quoted 4 lines
arch/aarch64/init.c +4 -0
+	}
+
+	msr(CNTFRQ_EL0, COUNTER_FREQ);
+}

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#24 lore raw

Re: [bootwrapper PATCH v2 09/13] aarch64: move the bulk of EL3 initialization to C

Quoted 15 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:49 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 5 lines

The majority of state that we initialize at EL3 is necessary for code at lower ELs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

Ah, thanks, that looks much nicer and easier to read now, also is more robust, as keeping register values alive for more than a few assembly lines always scares me.

Quoted 7 lines
Quoted 3 lines

This patch migrates the AArch64 EL3 initialization to C.

There should be no functional change as a result of this patch.

I compared the removed assembly code against to added C code, and also checked the register bits against the ARM ARM.
Two (and a half) things stood out, see below:

Thanks for this! I've fixed those as noted below.

[...]

Quoted 5 lines

arch/aarch64/include/asm/cpu.h line 35

Quoted 3 lines
arch/aarch64/include/asm/cpu.h +2 -1
-#define HCR_EL2_RES1		(BIT(1))
+#define ZCR_EL3				s3_6_c1_c2_0
+#define ZCR_EL3_LEN			BITS(3, 1)

The (current) actual length field should be BITS(3, 0), no?

Yes, it should. I've corrected that to BITS(3, 0) now.

[...]

Quoted 14 lines
Quoted 14 lines
arch/aarch64/init.c +14 -0
+void cpu_init_el3(void)
+{
+	unsigned long scr = SCR_EL3_RES1 | SCR_EL3_NS | SCR_EL3_HCE;
+	unsigned long mdcr = 0;
+	unsigned long cptr = 0;
+
+	if (cpu_has_pauth())
+		scr |= SCR_EL3_APK | SCR_EL3_API;
+
+	if (mrs_field(ID_AA64ISAR0_EL1, TME))
+		scr |= SCR_EL3_TME;
+
+	if (mrs_field(ID_AA64MMFR0_EL1, FGT))
+		scr |= SCR_EL3_FGTEN;
Quoted 8 lines

arch/aarch64/init.c void announce_arch(void) · line 59

Quoted 5 lines
arch/aarch64/init.c +5 -0
+
+	if (mrs_field(ID_AA64MMFR0_EL1, ECV) >= 2)
+		scr |= SCR_EL3_ECVEN;
+
+	if (mrs_field(ID_AA64PFR1_EL1, MTE))

The assembly code checked for >=2, which seems correct to me, as SCR_EL3_ATA is about MTE2?

Yes; I botched that when converting to C. SCR_EL3.ATA is RES0 in the absence of MTE2. I've added >= 2 to the condition to match that.

Quoted 1 line
Quoted 1 line
arch/aarch64/init.c +1 -0
+		scr |= SCR_EL3_ATA;

[...]

Quoted 8 lines

arch/aarch64/init.c void announce_arch(void) · line 84

Quoted 5 lines
arch/aarch64/init.c +5 -0
+	if (mrs_field(ID_AA64PFR0_EL1, SVE)) {
+		cptr |= CPTR_EL3_EZ;
+		msr(CPTR_EL3, cptr);
+		isb();
+		msr(ZCR_EL3, ZCR_EL3_LEN);

So when comparing this to the other uses of XXX_EL3_YYY, they typically describe a mask, but here we seems to abuse this as a value?

True; I'll add a separate defintion for the value.

Quoted 4 lines

And apart from bit 0 missing from it (as noted above), the existing code writes 0x1ff into that register, presumable to cover future vector length extensions beyond 2048 bits (which those RAZ/WI fields in bits[8:4] seem to suggest).

Hmm... I went and found the SVE supplement and I can't see any rationale for what SW *should* do, nor can I find a description of the register (that seems to have been factored into some XML files I can't convince anything to load on my machine).

Quoted 2 lines

So shall we define ZCR_EL3_MAX_VEC_LEN to 0x1ff above, and then use that? Or ignore the crystal ball, and just stick with 2048 bits, by writing 0xf?

TBH, I'm not sure. In the absence of some documented guidance I'd prefer to go with 0xf, but given we already use 0x1ff, I want to dig into this a bit more.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#25 lore raw

Re: [bootwrapper PATCH v2 09/13] aarch64: move the bulk of EL3 initialization to C

On Mon, 17 Jan 2022 18:08:13 +0000
Mark Rutland <[email protected]> wrote:

Hi,

Quoted 89 lines · Andre Przywara
Quoted 15 lines

On Fri, 14 Jan 2022 10:56:49 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 5 lines

The majority of state that we initialize at EL3 is necessary for code at lower ELs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

Ah, thanks, that looks much nicer and easier to read now, also is more robust, as keeping register values alive for more than a few assembly lines always scares me.

Quoted 7 lines
Quoted 3 lines

This patch migrates the AArch64 EL3 initialization to C.

There should be no functional change as a result of this patch.

I compared the removed assembly code against to added C code, and also checked the register bits against the ARM ARM.
Two (and a half) things stood out, see below:

Thanks for this! I've fixed those as noted below.

[...]

Quoted 5 lines

arch/aarch64/include/asm/cpu.h line 35

Quoted 3 lines
arch/aarch64/include/asm/cpu.h +2 -1
-#define HCR_EL2_RES1		(BIT(1))
+#define ZCR_EL3				s3_6_c1_c2_0
+#define ZCR_EL3_LEN			BITS(3, 1)  

The (current) actual length field should be BITS(3, 0), no?

Yes, it should. I've corrected that to BITS(3, 0) now.

[...]

Quoted 14 lines
Quoted 14 lines
arch/aarch64/init.c +14 -0
+void cpu_init_el3(void)
+{
+	unsigned long scr = SCR_EL3_RES1 | SCR_EL3_NS | SCR_EL3_HCE;
+	unsigned long mdcr = 0;
+	unsigned long cptr = 0;
+
+	if (cpu_has_pauth())
+		scr |= SCR_EL3_APK | SCR_EL3_API;
+
+	if (mrs_field(ID_AA64ISAR0_EL1, TME))
+		scr |= SCR_EL3_TME;
+
+	if (mrs_field(ID_AA64MMFR0_EL1, FGT))
+		scr |= SCR_EL3_FGTEN;
Quoted 8 lines

arch/aarch64/init.c void announce_arch(void) · line 59

Quoted 5 lines
arch/aarch64/init.c +5 -0
+
+	if (mrs_field(ID_AA64MMFR0_EL1, ECV) >= 2)
+		scr |= SCR_EL3_ECVEN;
+
+	if (mrs_field(ID_AA64PFR1_EL1, MTE))  

The assembly code checked for >=2, which seems correct to me, as SCR_EL3_ATA is about MTE2?

Yes; I botched that when converting to C. SCR_EL3.ATA is RES0 in the absence of MTE2. I've added >= 2 to the condition to match that.

Quoted 1 line
Quoted 1 line
arch/aarch64/init.c +1 -0
+		scr |= SCR_EL3_ATA;  

[...]

Quoted 8 lines

arch/aarch64/init.c void announce_arch(void) · line 84

Quoted 5 lines
arch/aarch64/init.c +5 -0
+	if (mrs_field(ID_AA64PFR0_EL1, SVE)) {
+		cptr |= CPTR_EL3_EZ;
+		msr(CPTR_EL3, cptr);
+		isb();
+		msr(ZCR_EL3, ZCR_EL3_LEN);  

So when comparing this to the other uses of XXX_EL3_YYY, they typically describe a mask, but here we seems to abuse this as a value?

True; I'll add a separate defintion for the value.

Quoted 4 lines

And apart from bit 0 missing from it (as noted above), the existing code writes 0x1ff into that register, presumable to cover future vector length extensions beyond 2048 bits (which those RAZ/WI fields in bits[8:4] seem to suggest).

Hmm... I went and found the SVE supplement and I can't see any rationale for what SW *should* do, nor can I find a description of the register (that seems to have been factored into some XML files I can't convince anything to load on my machine).

Quoted 6 lines
Quoted 2 lines

So shall we define ZCR_EL3_MAX_VEC_LEN to 0x1ff above, and then use that? Or ignore the crystal ball, and just stick with 2048 bits, by writing 0xf?

TBH, I'm not sure. In the absence of some documented guidance I'd prefer to go with 0xf, but given we already use 0x1ff, I want to dig into this a bit more.

My impression was that this "[8:4] = RAZ/WI" compared to the "[63:9] = RES0" fields suggests this is for a potential extension, but I guess there would be more changes needed if SVE ever goes beyond 2048. So chances are high we need to adopt the code then anyway, and fixing the number then is the least of our problems.

So I feel we should stick to what's explicitly documented, and put 0xf in there.

Cheers,
Andre

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#26 lore raw

Re: [bootwrapper PATCH v2 09/13] aarch64: move the bulk of EL3 initialization to C

Quoted 4 lines · Andre Przywara
Quoted 3 lines · Mark Rutland
Quoted 2 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:49 +0000
Mark Rutland <[email protected]> wrote:

[Out of office this week so replies might be intermittent]

Quoted 9 lines
Quoted 9 lines
Quoted 4 lines

And apart from bit 0 missing from it (as noted above), the existing code writes 0x1ff into that register, presumable to cover future vector length extensions beyond 2048 bits (which those RAZ/WI fields in bits[8:4] seem to suggest).

Hmm... I went and found the SVE supplement and I can't see any rationale for what SW *should* do, nor can I find a description of the register (that seems to have been factored into some XML files I can't convince anything to load on my machine).

...

Quoted 3 lines
Quoted 3 lines

TBH, I'm not sure. In the absence of some documented guidance I'd prefer to go with 0xf, but given we already use 0x1ff, I want to dig into this a bit more.

I'm fairly sure I've seen some explicit discussion of this lurking somewhere, though I couldn't tell you where - DDI0584 A.i doesn't spell out the enumeration algorithm unfortunately AFAICT. The theory is that the extra bits are reserved for any future extension of the vector length if needed since it'd be inconvenient to have to split the vector length field up.

Quoted 8 lines

My impression was that this "[8:4] = RAZ/WI" compared to the "[63:9] = RES0" fields suggests this is for a potential extension, but I guess there would be more changes needed if SVE ever goes beyond 2048. So chances are high we need to adopt the code then anyway, and fixing the number then is the least of our problems.

So I feel we should stick to what's explicitly documented, and put 0xf in there.

We shouldn't need particularly many changes if the size of the field ever gets raised, with everything being dynamically sized already and the existing code starting off setting the WI bits to 1 the updates that are needed should just be on input validation.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#27 lore raw

Re: [bootwrapper PATCH v2 09/13] aarch64: move the bulk of EL3 initialization to C

Quoted 37 lines · Mark Brown
Quoted 4 lines · Andre Przywara
Quoted 3 lines · Mark Rutland
Quoted 2 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:49 +0000
Mark Rutland <[email protected]> wrote:

[Out of office this week so replies might be intermittent]

Quoted 9 lines
Quoted 9 lines
Quoted 4 lines

And apart from bit 0 missing from it (as noted above), the existing code writes 0x1ff into that register, presumable to cover future vector length extensions beyond 2048 bits (which those RAZ/WI fields in bits[8:4] seem to suggest).

Hmm... I went and found the SVE supplement and I can't see any rationale for what SW *should* do, nor can I find a description of the register (that seems to have been factored into some XML files I can't convince anything to load on my machine).

...

Quoted 3 lines
Quoted 3 lines

TBH, I'm not sure. In the absence of some documented guidance I'd prefer to go with 0xf, but given we already use 0x1ff, I want to dig into this a bit more.

I'm fairly sure I've seen some explicit discussion of this lurking somewhere, though I couldn't tell you where - DDI0584 A.i doesn't spell out the enumeration algorithm unfortunately AFAICT. The theory is that the extra bits are reserved for any future extension of the vector length if needed since it'd be inconvenient to have to split the vector length field up.

Quoted 5 lines

My impression was that this "[8:4] = RAZ/WI" compared to the "[63:9] = RES0" fields suggests this is for a potential extension, but I guess there would be more changes needed if SVE ever goes beyond 2048. So chances are high we need to adopt the code then anyway, and fixing the number then is the least of our problems.

Quoted 7 lines
Quoted 2 lines

So I feel we should stick to what's explicitly documented, and put 0xf in there.

We shouldn't need particularly many changes if the size of the field ever gets raised, with everything being dynamically sized already and the existing code starting off setting the WI bits to 1 the updates that are needed should just be on input validation.

Taking a look around, TF-A programs 0xf, and for consistency I think it'd be clearer to only program the currently-allocated LEN bits.

I'll spin a preparatory patch reducing ZCR_EL3_LEN_MASK to 0xf (and renaming that to ZCR_EL3_LEN_MAX).

If we need to grow it in future it should be trivial to do so.

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#28 lore raw

PATCH v2 10/13 aarch32: move the bulk of Secure PL1 initialization to C

In reply to Mark Rutland

The majority of state that we initialize at Secure PL1 is necessary for code at lower PLs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

This patch migrates the AArch32 Secure PL1 initialization to C.

There should be no functional change as a result of this patch.

---
 arch/aarch32/boot.S            | 11 +----------
 arch/aarch32/include/asm/cpu.h |  9 +++++++++
 arch/aarch32/init.c            | 12 ++++++++++++
 3 files changed, 22 insertions(+), 10 deletions(-)
arch/aarch32/boot.S +1 -10
index ee073ea..820957b 100644
@@ -63,16 +63,7 @@ _monitor:
 	/* Move the stack to Monitor mode*/
 	mrs	sp, sp_svc
 
-	/* Setup secure registers and devices */
-	mov	r0, #1				@ Non-secure lower level
-	orr	r0, #(1 << 8)			@ HVC enable
-	mcr	p15, 0, r0, c1, c1, 0		@ SCR
-
-	mov	r0, #(1 << 10 | 1 << 11)	@ Enable NS access to CPACR
-	mcr	p15, 0, r0, c1, c1, 2		@ NSACR
-
-	ldr	r0, =COUNTER_FREQ
-	mcr	p15, 0, r0, c14, c0, 0		@ CNTFRQ
+	bl	cpu_init_secure_pl1
 
 	bl	cpu_init_bootwrapper
 
arch/aarch32/include/asm/cpu.h +9 -0
index aa72204..c1bce9a 100644
@@ -30,6 +30,11 @@
 #define PSR_I			(1 << 7)
 #define PSR_A			(1 << 8)
 
+#define SCR_NS			BIT(0)
+#define SCR_HCE			BIT(8)
+
+#define NSACR_CP10		BIT(10)
+#define NSACR_CP11		BIT(11)
 
 #define SPSR_KERNEL		(PSR_A | PSR_I | PSR_F | PSR_HYP)
 
@@ -55,11 +60,15 @@ static inline unsigned long read_cpsr(void)
 
 #define MPIDR		"p15, 0, %0, c0, c0, 5"
 #define ID_PFR1		"p15, 0, %0, c0, c1, 1"
+#define SCR		"p15, 0, %0, c1, c1, 0"
+#define NSACR		"p15, 0, %0, c1, c1, 2"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
 
 #define ICC_SRE		"p15, 6, %0, c12, c12, 5"
 #define ICC_CTLR	"p15, 6, %0, c12, c12, 4"
 
+#define CNTFRQ		"p15, 0, %0, c14, c0, 0"
+
 #define mrc(reg)						\
 ({								\
 	unsigned long __mrc_val;				\
arch/aarch32/init.c +12 -0
index b29ebb4..5b69dcd 100644
@@ -28,3 +28,15 @@ void announce_arch(void)
 	print_string(mode_string());
 	print_string("\r\n");
 }
+
+void cpu_init_secure_pl1(void)
+{
+	unsigned long scr = SCR_NS | SCR_HCE;
+	unsigned long nsacr = NSACR_CP10 | NSACR_CP11;
+
+	mcr(SCR, scr);
+
+	mcr(NSACR, nsacr);
+
+	mcr(CNTFRQ, COUNTER_FREQ);
+}
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 10/13
#29 lore raw

Re: [bootwrapper PATCH v2 10/13] aarch32: move the bulk of Secure PL1 initialization to C

On Fri, 14 Jan 2022 10:56:50 +0000
Mark Rutland <[email protected]> wrote:

Hi,

Quoted 9 lines

The majority of state that we initialize at Secure PL1 is necessary for code at lower PLs to function, but isnt' necessary for the boot-wrapper itself. Given that, it would be better to write this in C where it can be written mode clearly, and where it will be possible to add logging/debug logic.

This patch migrates the AArch32 Secure PL1 initialization to C.

There should be no functional change as a result of this patch.

I compared the removed assembly code against to added C code, and also checked the register bits against the ARMv7 ARM.
Everything checks out, so:

Quoted 1 line

Cheers,
Andre

Quoted 70 lines
---
 arch/aarch32/boot.S            | 11 +----------
 arch/aarch32/include/asm/cpu.h |  9 +++++++++
 arch/aarch32/init.c            | 12 ++++++++++++
 3 files changed, 22 insertions(+), 10 deletions(-)
arch/aarch32/boot.S +1 -10
index ee073ea..820957b 100644
@@ -63,16 +63,7 @@ _monitor:
 	/* Move the stack to Monitor mode*/
 	mrs	sp, sp_svc
 
-	/* Setup secure registers and devices */
-	mov	r0, #1				@ Non-secure lower level
-	orr	r0, #(1 << 8)			@ HVC enable
-	mcr	p15, 0, r0, c1, c1, 0		@ SCR
-
-	mov	r0, #(1 << 10 | 1 << 11)	@ Enable NS access to CPACR
-	mcr	p15, 0, r0, c1, c1, 2		@ NSACR
-
-	ldr	r0, =COUNTER_FREQ
-	mcr	p15, 0, r0, c14, c0, 0		@ CNTFRQ
+	bl	cpu_init_secure_pl1
 
 	bl	cpu_init_bootwrapper
 
arch/aarch32/include/asm/cpu.h +9 -0
index aa72204..c1bce9a 100644
@@ -30,6 +30,11 @@
 #define PSR_I			(1 << 7)
 #define PSR_A			(1 << 8)
 
+#define SCR_NS			BIT(0)
+#define SCR_HCE			BIT(8)
+
+#define NSACR_CP10		BIT(10)
+#define NSACR_CP11		BIT(11)
 
 #define SPSR_KERNEL		(PSR_A | PSR_I | PSR_F | PSR_HYP)
 
@@ -55,11 +60,15 @@ static inline unsigned long read_cpsr(void)
 
 #define MPIDR		"p15, 0, %0, c0, c0, 5"
 #define ID_PFR1		"p15, 0, %0, c0, c1, 1"
+#define SCR		"p15, 0, %0, c1, c1, 0"
+#define NSACR		"p15, 0, %0, c1, c1, 2"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
 
 #define ICC_SRE		"p15, 6, %0, c12, c12, 5"
 #define ICC_CTLR	"p15, 6, %0, c12, c12, 4"
 
+#define CNTFRQ		"p15, 0, %0, c14, c0, 0"
+
 #define mrc(reg)						\
 ({								\
 	unsigned long __mrc_val;				\
arch/aarch32/init.c +2 -0
index b29ebb4..5b69dcd 100644
@@ -28,3 +28,15 @@ void announce_arch(void)
 	print_string(mode_string());
 	print_string("\r\n");
 }
+
+void cpu_init_secure_pl1(void)

arch/aarch32/init.c void announce_arch(void) · line 42

Quoted 10 lines
arch/aarch32/init.c +10 -0
+{
+	unsigned long scr = SCR_NS | SCR_HCE;
+	unsigned long nsacr = NSACR_CP10 | NSACR_CP11;
+
+	mcr(SCR, scr);
+
+	mcr(NSACR, nsacr);
+
+	mcr(CNTFRQ, COUNTER_FREQ);
+}

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#30 lore raw

PATCH v2 11/13 Announce locations of memory objects

In reply to Mark Rutland

To make it easier to debug boot failures, log the location of memory objects at boot time.

This is logged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL3
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
---
 common/init.c      | 27 +++++++++++++++++++++++++++
 common/platform.c  | 13 +++++++++++++
 include/platform.h |  2 ++
 model.lds.S        | 20 ++++++++++++++------
 4 files changed, 56 insertions(+), 6 deletions(-)
common/init.c +27 -0
index 2600f73..fc74b9e 100644
@@ -14,6 +14,32 @@ static void announce_bootwrapper(void)
 	print_string("Boot-wrapper v0.2\r\n");
 }
 
+#define announce_object(object, desc)				\
+do {								\
+	extern char object##__start[];				\
+	extern char object##__end[];				\
+	print_string("[");					\
+	print_ulong_hex((unsigned long)object##__start);	\
+	print_string("..");					\
+	print_ulong_hex((unsigned long)object##__end);		\
+	print_string("] => " desc "\r\n");			\
+} while (0)
+
+static void announce_objects(void)
+{
+	print_string("Memory layout:\r\n");
+	announce_object(text, "boot-wrapper");
+	announce_object(mbox, "mbox");
+	announce_object(kernel, "kernel");
+#ifdef XEN
+	announce_object(xen, "xen");
+#endif
+	announce_object(dtb, "dtb");
+#ifdef USE_INITRD
+	announce_object(filesystem, "initrd");
+#endif
+}
+
 void announce_arch(void);
 
 void cpu_init_bootwrapper(void)
@@ -22,6 +48,7 @@ void cpu_init_bootwrapper(void)
 		init_uart();
 		announce_bootwrapper();
 		announce_arch();
+		announce_objects();
 		print_string("\r\n");
 		init_platform();
 	}
common/platform.c +13 -0
index 80d0562..26e5944 100644
@@ -52,6 +52,19 @@ void print_string(const char *str)
 		print_char(*str++);
 }
 
+#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";
+	int i;
+
+	for (i = HEX_CHARS_PER_LONG - 1; i >= 0; i--) {
+		int v = (val >> (4 * i)) & 0xf;
+		print_char(hex_chars[v]);
+	}
+}
+
 void init_uart(void)
 {
 	/*
include/platform.h +2 -0
index 237b481..c88e124 100644
@@ -11,6 +11,8 @@
 
 void print_char(char c);
 void print_string(const char *str);
+void print_ulong_hex(unsigned long val);
+
 void init_uart(void);
 
 void init_platform(void);
model.lds.S +14 -6
index d4e7e13..dacaa25 100644
@@ -35,46 +35,54 @@ SECTIONS
 	 * the boot section's *(.data)
 	 */
 	.kernel (PHYS_OFFSET + KERNEL_OFFSET): {
-		kernel = .;
+		kernel__start = .;
 		KERNEL
+		kernel__end = .;
 	}
 
 #ifdef XEN
 	.xen (PHYS_OFFSET + XEN_OFFSET): {
-		xen = .;
+		xen__start = .;
 		XEN
+		xen__end = .;
 	}
 
-	entrypoint = xen;
+	entrypoint = xen__start;
 #else
-	entrypoint = kernel;
+	entrypoint = kernel__start;
 #endif
 
 	.dtb (PHYS_OFFSET + FDT_OFFSET): {
+		dtb__start = .;
 		dtb = .;
 		./fdt.dtb
+		dtb__end = .;
 	}
 
 #ifdef USE_INITRD
 	.filesystem (PHYS_OFFSET + FS_OFFSET): {
-		filesystem = .;
+		filesystem__start = .;
 		FILESYSTEM
-		fs_size = . - filesystem;
+		filesystem__end = .;
 	}
 #endif
 
 	.boot PHYS_OFFSET: {
+		text__start = .;
 		*(.init)
 		*(.text*)
 		*(.data* .rodata* .bss* COMMON)
 		*(.vectors)
 		*(.stack)
 		PROVIDE(etext = .);
+		text__end = .;
 	}
 
 	.mbox (PHYS_OFFSET + MBOX_OFFSET): {
+		mbox__start = .;
 		mbox = .;
 		QUAD(0x0)
+		mbox__end = .;
 	}
 
 	ASSERT(etext <= (PHYS_OFFSET + TEXT_LIMIT), ".text overflow!")
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#31 lore raw

Re: [bootwrapper PATCH v2 11/13] Announce locations of memory objects

On Fri, 14 Jan 2022 10:56:51 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 71 lines

To make it easier to debug boot failures, log the location of memory objects at boot time.

This is logged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL3
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
---
 common/init.c      | 27 +++++++++++++++++++++++++++
 common/platform.c  | 13 +++++++++++++
 include/platform.h |  2 ++
 model.lds.S        | 20 ++++++++++++++------
 4 files changed, 56 insertions(+), 6 deletions(-)
common/init.c +27 -0
index 2600f73..fc74b9e 100644
@@ -14,6 +14,32 @@ static void announce_bootwrapper(void)
 	print_string("Boot-wrapper v0.2\r\n");
 }
 
+#define announce_object(object, desc)				\
+do {								\
+	extern char object##__start[];				\
+	extern char object##__end[];				\
+	print_string("[");					\
+	print_ulong_hex((unsigned long)object##__start);	\
+	print_string("..");					\
+	print_ulong_hex((unsigned long)object##__end);		\
+	print_string("] => " desc "\r\n");			\
+} while (0)
+
+static void announce_objects(void)
+{
+	print_string("Memory layout:\r\n");
+	announce_object(text, "boot-wrapper");
+	announce_object(mbox, "mbox");
+	announce_object(kernel, "kernel");
+#ifdef XEN
+	announce_object(xen, "xen");
+#endif
+	announce_object(dtb, "dtb");
+#ifdef USE_INITRD
+	announce_object(filesystem, "initrd");
+#endif
+}
+
 void announce_arch(void);
 
 void cpu_init_bootwrapper(void)
@@ -22,6 +48,7 @@ void cpu_init_bootwrapper(void)
 		init_uart();
 		announce_bootwrapper();
 		announce_arch();
+		announce_objects();
 		print_string("\r\n");
 		init_platform();
 	}
common/platform.c +0 -0
index 80d0562..26e5944 100644

common/platform.c void print_string(const char *str) · line 59

Quoted 8 lines
common/platform.c +5 -0
@@ -52,6 +52,19 @@ void print_string(const char *str)
 		print_char(*str++);
 }
 
+#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";

This breaks the build for me (I guess because of the const?):
-------------------
ld --gc-sections arch/aarch64/boot.o arch/aarch64/stack.o arch/aarch64/utils.o arch/aarch64/init.o arch/aarch64/psci.o common/boot.o common/bakery_lock.o common/platform.o common/lib.o common/init.o common/psci.o common/gic-v3.o -o linux-system.axf --script=model.lds
ld: common/platform.o: in function `print_ulong_hex':
/src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:66: undefined reference to `__stack_chk_fail'
/src/boot-wrapper-aarch64.git/common/platform.c:66:(.text.print_ulong_hex+0xa0): relocation truncated to fit: R_AARCH64_CALL26 against undefined symbol `__stack_chk_fail'
make: *** [Makefile:684: linux-system.axf] Error 1
------------------

Adding -fno-stack-protector fixes it again.

I am still checking on each one of those compiler options, shall I send a smaller version of that patch 3/9 of mine, meanwhile, just carrying the uncontested -ffreestanding and -nostdlib, plus -fno-stack-protector?

Cheers,
Andre

Quoted 79 lines
common/platform.c +8 -0
+	int i;
+
+	for (i = HEX_CHARS_PER_LONG - 1; i >= 0; i--) {
+		int v = (val >> (4 * i)) & 0xf;
+		print_char(hex_chars[v]);
+	}
+}
+
 void init_uart(void)
 {
 	/*
include/platform.h +2 -0
index 237b481..c88e124 100644
@@ -11,6 +11,8 @@
 
 void print_char(char c);
 void print_string(const char *str);
+void print_ulong_hex(unsigned long val);
+
 void init_uart(void);
 
 void init_platform(void);
model.lds.S +12 -6
index d4e7e13..dacaa25 100644
@@ -35,46 +35,54 @@ SECTIONS
 	 * the boot section's *(.data)
 	 */
 	.kernel (PHYS_OFFSET + KERNEL_OFFSET): {
-		kernel = .;
+		kernel__start = .;
 		KERNEL
+		kernel__end = .;
 	}
 
 #ifdef XEN
 	.xen (PHYS_OFFSET + XEN_OFFSET): {
-		xen = .;
+		xen__start = .;
 		XEN
+		xen__end = .;
 	}
 
-	entrypoint = xen;
+	entrypoint = xen__start;
 #else
-	entrypoint = kernel;
+	entrypoint = kernel__start;
 #endif
 
 	.dtb (PHYS_OFFSET + FDT_OFFSET): {
+		dtb__start = .;
 		dtb = .;
 		./fdt.dtb
+		dtb__end = .;
 	}
 
 #ifdef USE_INITRD
 	.filesystem (PHYS_OFFSET + FS_OFFSET): {
-		filesystem = .;
+		filesystem__start = .;
 		FILESYSTEM
-		fs_size = . - filesystem;
+		filesystem__end = .;
 	}
 #endif
 
 	.boot PHYS_OFFSET: {
+		text__start = .;
 		*(.init)
 		*(.text*)
 		*(.data* .rodata* .bss* COMMON)
 		*(.vectors)
 		*(.stack)
 		PROVIDE(etext = .);
+		text__end = .;

model.lds.S SECTIONS · line 88

Quoted 10 lines
model.lds.S +2 -0
 	}
 
 	.mbox (PHYS_OFFSET + MBOX_OFFSET): {
+		mbox__start = .;
 		mbox = .;
 		QUAD(0x0)
+		mbox__end = .;
 	}
 
 	ASSERT(etext <= (PHYS_OFFSET + TEXT_LIMIT), ".text overflow!")

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#32 lore raw

Re: [bootwrapper PATCH v2 11/13] Announce locations of memory objects

Quoted 79 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:51 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 74 lines

To make it easier to debug boot failures, log the location of memory objects at boot time.

This is logged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL3
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
---
  common/init.c      | 27 +++++++++++++++++++++++++++
  common/platform.c  | 13 +++++++++++++
  include/platform.h |  2 ++
  model.lds.S        | 20 ++++++++++++++------
  4 files changed, 56 insertions(+), 6 deletions(-)
common/init.c +27 -0
index 2600f73..fc74b9e 100644
@@ -14,6 +14,32 @@ static void announce_bootwrapper(void)
  	print_string("Boot-wrapper v0.2\r\n");
  }
 
+#define announce_object(object, desc)				\
+do {								\
+	extern char object##__start[];				\
+	extern char object##__end[];				\
+	print_string("[");					\
+	print_ulong_hex((unsigned long)object##__start);	\
+	print_string("..");					\
+	print_ulong_hex((unsigned long)object##__end);		\
+	print_string("] => " desc "\r\n");			\
+} while (0)
+
+static void announce_objects(void)
+{
+	print_string("Memory layout:\r\n");
+	announce_object(text, "boot-wrapper");
+	announce_object(mbox, "mbox");
+	announce_object(kernel, "kernel");
+#ifdef XEN
+	announce_object(xen, "xen");
+#endif
+	announce_object(dtb, "dtb");
+#ifdef USE_INITRD
+	announce_object(filesystem, "initrd");
+#endif
+}
+
  void announce_arch(void);
 
  void cpu_init_bootwrapper(void)
@@ -22,6 +48,7 @@ void cpu_init_bootwrapper(void)
  		init_uart();
  		announce_bootwrapper();
  		announce_arch();
+		announce_objects();
  		print_string("\r\n");
  		init_platform();
  	}
common/platform.c +0 -0
index 80d0562..26e5944 100644
@@ -52,6 +52,19 @@ void print_string(const char *str)
  		print_char(*str++);
  }
 
Quoted 16 lines

common/platform.c void print_string(const char *str) · line 59

Quoted 5 lines
common/platform.c +5 -0
+#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";

This breaks the build for me (I guess because of the const?):
-------------------
ld --gc-sections arch/aarch64/boot.o arch/aarch64/stack.o arch/aarch64/utils.o arch/aarch64/init.o arch/aarch64/psci.o common/boot.o common/bakery_lock.o common/platform.o common/lib.o common/init.o common/psci.o common/gic-v3.o -o linux-system.axf --script=model.lds
ld: common/platform.o: in function `print_ulong_hex':
/src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:66: undefined reference to `__stack_chk_fail'
/src/boot-wrapper-aarch64.git/common/platform.c:66:(.text.print_ulong_hex+0xa0): relocation truncated to fit: R_AARCH64_CALL26 against undefined symbol `__stack_chk_fail'
make: *** [Makefile:684: linux-system.axf] Error 1
------------------

I got curious and tried a quick test compiling this function in isolation, and indeed for some reason GCC (9.3.0 at -O2) seems to end up allocating 80 bytes of stack, which is apparently large enough to trigger insertion of a stack check, and laboriously copying the string from static data into it. FWIW either making hex_chars static const, or simply '#define hex_chars "01234567abcdef"', not only avoids a stack check but also leads to notably nicer code.

Of course, something like "print_char(v + (v > 9 ? 'a' : '0'));" is best of all in terms of code/data, but risks venturing into "too clever" territory.

Robin.

Quoted 90 lines

Adding -fno-stack-protector fixes it again.

I am still checking on each one of those compiler options, shall I send a smaller version of that patch 3/9 of mine, meanwhile, just carrying the uncontested -ffreestanding and -nostdlib, plus -fno-stack-protector?

Cheers,
Andre

Quoted 81 lines
common/platform.c +8 -0
+	int i;
+
+	for (i = HEX_CHARS_PER_LONG - 1; i >= 0; i--) {
+		int v = (val >> (4 * i)) & 0xf;
+		print_char(hex_chars[v]);
+	}
+}
+
  void init_uart(void)
  {
  	/*
include/platform.h +2 -0
index 237b481..c88e124 100644
@@ -11,6 +11,8 @@
 
  void print_char(char c);
  void print_string(const char *str);
+void print_ulong_hex(unsigned long val);
+
  void init_uart(void);
 
  void init_platform(void);
model.lds.S +12 -6
index d4e7e13..dacaa25 100644
@@ -35,46 +35,54 @@ SECTIONS
  	 * the boot section's *(.data)
  	 */
  	.kernel (PHYS_OFFSET + KERNEL_OFFSET): {
-		kernel = .;
+		kernel__start = .;
  		KERNEL
+		kernel__end = .;
  	}
 
  #ifdef XEN
  	.xen (PHYS_OFFSET + XEN_OFFSET): {
-		xen = .;
+		xen__start = .;
  		XEN
+		xen__end = .;
  	}
 
-	entrypoint = xen;
+	entrypoint = xen__start;
  #else
-	entrypoint = kernel;
+	entrypoint = kernel__start;
  #endif
 
  	.dtb (PHYS_OFFSET + FDT_OFFSET): {
+		dtb__start = .;
  		dtb = .;
  		./fdt.dtb
+		dtb__end = .;
  	}
 
  #ifdef USE_INITRD
  	.filesystem (PHYS_OFFSET + FS_OFFSET): {
-		filesystem = .;
+		filesystem__start = .;
  		FILESYSTEM
-		fs_size = . - filesystem;
+		filesystem__end = .;
  	}
  #endif
 
  	.boot PHYS_OFFSET: {
+		text__start = .;
  		*(.init)
  		*(.text*)
  		*(.data* .rodata* .bss* COMMON)
  		*(.vectors)
  		*(.stack)
  		PROVIDE(etext = .);
+		text__end = .;
  	}
 
Quoted 9 lines

model.lds.S SECTIONS · line 88

Quoted 8 lines
model.lds.S +2 -0
  	.mbox (PHYS_OFFSET + MBOX_OFFSET): {
+		mbox__start = .;
  		mbox = .;
  		QUAD(0x0)
+		mbox__end = .;
  	}
 
  	ASSERT(etext <= (PHYS_OFFSET + TEXT_LIMIT), ".text overflow!")

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#33 lore raw

Re: [bootwrapper PATCH v2 11/13] Announce locations of memory objects

Quoted 18 lines · Robin Murphy
Quoted 16 lines · Andre Przywara

common/platform.c void print_string(const char *str) · line 59

Quoted 3 lines · On Fri, 14 Jan 2022 10:56:51 +0000 Mark Rutland <[email protected]> wrote:
common/platform.c +3 -0
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";

This breaks the build for me (I guess because of the const?):
-------------------
ld --gc-sections arch/aarch64/boot.o arch/aarch64/stack.o arch/aarch64/utils.o arch/aarch64/init.o arch/aarch64/psci.o common/boot.o common/bakery_lock.o common/platform.o common/lib.o common/init.o common/psci.o common/gic-v3.o -o linux-system.axf --script=model.lds
ld: common/platform.o: in function `print_ulong_hex':
/src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:66: undefined reference to `__stack_chk_fail'
/src/boot-wrapper-aarch64.git/common/platform.c:66:(.text.print_ulong_hex+0xa0): relocation truncated to fit: R_AARCH64_CALL26 against undefined symbol `__stack_chk_fail'
make: *** [Makefile:684: linux-system.axf] Error 1
------------------

Quoted 7 lines

I got curious and tried a quick test compiling this function in isolation, and indeed for some reason GCC (9.3.0 at -O2) seems to end up allocating 80 bytes of stack, which is apparently large enough to trigger insertion of a stack check, and laboriously copying the string from static data into it. FWIW either making hex_chars static const, or simply '#define hex_chars "01234567abcdef"', not only avoids a stack check but also leads to notably nicer code.

For the stack check, I think Andre's patch to add -fno-stack-protector is the correct fix, so I'd like to take that as a prerequisite.

To avoid using the stack unnecessarily and to make the code nicer, I'll also take your suggestion with the fixup below.

Thanks,
Mark.

---->8----
-#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+#define HEX_CHARS "0123456789abcdef"

void print_ulong_hex(unsigned long val)
{

-       const char hex_chars[16] = "0123456789abcdef";
        int i;
 
        for (i = HEX_CHARS_PER_LONG - 1; i >= 0; i--) {
                int v = (val >> (4 * i)) & 0xf;
-               print_char(hex_chars[v]);
+               print_char(HEX_CHARS[v]);
        }

}

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#34 lore raw

Re: [bootwrapper PATCH v2 11/13] Announce locations of memory objects

In reply to Andre Przywara
Quoted 33 lines · Andre Przywara
Quoted 16 lines · On Fri, 14 Jan 2022 10:56:51 +0000 Mark Rutland <[email protected]> wrote:

To make it easier to debug boot failures, log the location of memory objects at boot time.

This is logged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL3
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb

---

common/platform.c void print_string(const char *str) · line 59

Quoted 3 lines
common/platform.c +3 -0
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";

This breaks the build for me (I guess because of the const?):
-------------------
ld --gc-sections arch/aarch64/boot.o arch/aarch64/stack.o arch/aarch64/utils.o arch/aarch64/init.o arch/aarch64/psci.o common/boot.o common/bakery_lock.o common/platform.o common/lib.o common/init.o common/psci.o common/gic-v3.o -o linux-system.axf --script=model.lds
ld: common/platform.o: in function `print_ulong_hex':
/src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:58: undefined reference to `__stack_chk_guard'
ld: /src/boot-wrapper-aarch64.git/common/platform.c:66: undefined reference to `__stack_chk_fail'
/src/boot-wrapper-aarch64.git/common/platform.c:66:(.text.print_ulong_hex+0xa0): relocation truncated to fit: R_AARCH64_CALL26 against undefined symbol `__stack_chk_fail'
make: *** [Makefile:684: linux-system.axf] Error 1
------------------

Quoted 5 lines

Adding -fno-stack-protector fixes it again.

I am still checking on each one of those compiler options, shall I send a smaller version of that patch 3/9 of mine, meanwhile, just carrying the uncontested -ffreestanding and -nostdlib, plus -fno-stack-protector?

Yes please!

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Reviewed-by on v2 11/13
#35 lore raw

Re: [bootwrapper PATCH v2 11/13] Announce locations of memory objects

In reply to Mark Rutland

On Fri, 14 Jan 2022 10:56:51 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 5 lines

To make it easier to debug boot failures, log the location of memory objects at boot time.

This is logged to the serial console as:

Quoted 7 lines
| Boot-wrapper v0.2
| Entered at EL3
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb

That looks indeed very useful!

Quoted 1 line

Cheers,
Andre

Quoted 144 lines
---
 common/init.c      | 27 +++++++++++++++++++++++++++
 common/platform.c  | 13 +++++++++++++
 include/platform.h |  2 ++
 model.lds.S        | 20 ++++++++++++++------
 4 files changed, 56 insertions(+), 6 deletions(-)
common/init.c +27 -0
index 2600f73..fc74b9e 100644
@@ -14,6 +14,32 @@ static void announce_bootwrapper(void)
 	print_string("Boot-wrapper v0.2\r\n");
 }
 
+#define announce_object(object, desc)				\
+do {								\
+	extern char object##__start[];				\
+	extern char object##__end[];				\
+	print_string("[");					\
+	print_ulong_hex((unsigned long)object##__start);	\
+	print_string("..");					\
+	print_ulong_hex((unsigned long)object##__end);		\
+	print_string("] => " desc "\r\n");			\
+} while (0)
+
+static void announce_objects(void)
+{
+	print_string("Memory layout:\r\n");
+	announce_object(text, "boot-wrapper");
+	announce_object(mbox, "mbox");
+	announce_object(kernel, "kernel");
+#ifdef XEN
+	announce_object(xen, "xen");
+#endif
+	announce_object(dtb, "dtb");
+#ifdef USE_INITRD
+	announce_object(filesystem, "initrd");
+#endif
+}
+
 void announce_arch(void);
 
 void cpu_init_bootwrapper(void)
@@ -22,6 +48,7 @@ void cpu_init_bootwrapper(void)
 		init_uart();
 		announce_bootwrapper();
 		announce_arch();
+		announce_objects();
 		print_string("\r\n");
 		init_platform();
 	}
common/platform.c +13 -0
index 80d0562..26e5944 100644
@@ -52,6 +52,19 @@ void print_string(const char *str)
 		print_char(*str++);
 }
 
+#define HEX_CHARS_PER_LONG (2 * sizeof(long))
+
+void print_ulong_hex(unsigned long val)
+{
+	const char hex_chars[16] = "0123456789abcdef";
+	int i;
+
+	for (i = HEX_CHARS_PER_LONG - 1; i >= 0; i--) {
+		int v = (val >> (4 * i)) & 0xf;
+		print_char(hex_chars[v]);
+	}
+}
+
 void init_uart(void)
 {
 	/*
include/platform.h +2 -0
index 237b481..c88e124 100644
@@ -11,6 +11,8 @@
 
 void print_char(char c);
 void print_string(const char *str);
+void print_ulong_hex(unsigned long val);
+
 void init_uart(void);
 
 void init_platform(void);
model.lds.S +12 -6
index d4e7e13..dacaa25 100644
@@ -35,46 +35,54 @@ SECTIONS
 	 * the boot section's *(.data)
 	 */
 	.kernel (PHYS_OFFSET + KERNEL_OFFSET): {
-		kernel = .;
+		kernel__start = .;
 		KERNEL
+		kernel__end = .;
 	}
 
 #ifdef XEN
 	.xen (PHYS_OFFSET + XEN_OFFSET): {
-		xen = .;
+		xen__start = .;
 		XEN
+		xen__end = .;
 	}
 
-	entrypoint = xen;
+	entrypoint = xen__start;
 #else
-	entrypoint = kernel;
+	entrypoint = kernel__start;
 #endif
 
 	.dtb (PHYS_OFFSET + FDT_OFFSET): {
+		dtb__start = .;
 		dtb = .;
 		./fdt.dtb
+		dtb__end = .;
 	}
 
 #ifdef USE_INITRD
 	.filesystem (PHYS_OFFSET + FS_OFFSET): {
-		filesystem = .;
+		filesystem__start = .;
 		FILESYSTEM
-		fs_size = . - filesystem;
+		filesystem__end = .;
 	}
 #endif
 
 	.boot PHYS_OFFSET: {
+		text__start = .;
 		*(.init)
 		*(.text*)
 		*(.data* .rodata* .bss* COMMON)
 		*(.vectors)
 		*(.stack)
 		PROVIDE(etext = .);
+		text__end = .;

model.lds.S SECTIONS · line 88

Quoted 10 lines
model.lds.S +2 -0
 	}
 
 	.mbox (PHYS_OFFSET + MBOX_OFFSET): {
+		mbox__start = .;
 		mbox = .;
 		QUAD(0x0)
+		mbox__end = .;
 	}
 
 	ASSERT(etext <= (PHYS_OFFSET + TEXT_LIMIT), ".text overflow!")

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#36 lore raw

PATCH v2 12/13 Rework bootmethod initialization

In reply to Mark Rutland

We currently initialize the bootmethod late, in assembly code. This requires us to maintain the el3/no_el3 distintion late into the boot process, and means we cannot produce any helpful diagnostic when booted at an unexpected exception level.

Rework things so that we initialize the bootmethod early, with a warning when things are wrong. The el3/no_el3 distinction is now irrelevant to the bootmethod code, and can be removed in subsequent patches.

When a boot-wrapper configured for PSCI is entered at EL2, a warning is looged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL2
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
|
| WARNING: PSCI could not be initialized. Boot may fail
---
 arch/aarch32/include/asm/cpu.h  |  1 +
 arch/aarch32/include/asm/psci.h | 28 ++++++++++++++++++++++++++++
 arch/aarch32/psci.S             |  8 +-------
 arch/aarch32/utils.S            |  9 ---------
 arch/aarch64/include/asm/psci.h | 28 ++++++++++++++++++++++++++++
 arch/aarch64/psci.S             | 21 ++-------------------
 arch/aarch64/spin.S             |  3 +++
 arch/aarch64/utils.S            |  9 ---------
 common/init.c                   |  7 ++++++-
 common/psci.c                   | 22 +++++++++++++++++++---
 include/boot.h                  |  2 ++
 11 files changed, 90 insertions(+), 48 deletions(-)
 create mode 100644 arch/aarch32/include/asm/psci.h
 create mode 100644 arch/aarch64/include/asm/psci.h
arch/aarch32/include/asm/cpu.h +1 -0
index c1bce9a..3426075 100644
@@ -63,6 +63,7 @@ static inline unsigned long read_cpsr(void)
 #define SCR		"p15, 0, %0, c1, c1, 0"
 #define NSACR		"p15, 0, %0, c1, c1, 2"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
+#define MVBAR		"p15, 0, %0, c12, c0, 1"
 
 #define ICC_SRE		"p15, 6, %0, c12, c12, 5"
 #define ICC_CTLR	"p15, 6, %0, c12, c12, 4"
arch/aarch32/include/asm/psci.h new file +28 -0
new file mode 100644
index 0000000..50c7c70
@@ -0,0 +1,28 @@
+/*
+ * arch/aarch64/include/asm/psci.h
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __ASM_AARCH64_PSCI_H
+#define __ASM_AARCH64_PSCI_H
+
+#include <cpu.h>
+#include <stdbool.h>
+
+extern char psci_vectors[];
+
+static inline bool cpu_init_psci_arch(void)
+{
+	if (read_cpsr_mode() != PSR_MON)
+		return false;
+
+	mcr(MVBAR, (unsigned long)psci_vectors);
+	isb();
+
+	return true;
+}
+
+#endif
arch/aarch32/psci.S +1 -7
index e0d2972..cdc36b0 100644
@@ -15,7 +15,7 @@
 
 	.section .vectors
 	.align 6
-smc_vectors:
+ASM_DATA(psci_vectors)
 	b	err_exception			@ Reset
 	b	err_exception			@ Undef
 	b	handle_smc			@ SMC
@@ -39,11 +39,5 @@ handle_smc:
 	movs	pc, lr
 
 ASM_FUNC(start_el3)
-	ldr	r0, =smc_vectors
-	blx	setup_vector
-	/* pass through */
-
 ASM_FUNC(start_no_el3)
-	cpuid	r0, r1
-	blx	find_logical_id
 	b	psci_first_spin
arch/aarch32/utils.S +0 -9
index 5809f48..58279aa 100644
@@ -35,12 +35,3 @@ ASM_FUNC(find_logical_id)
 	bx	lr
 3:	mov	r0, #MPIDR_INVALID
 	bx	lr
-
-/*
- * Setup EL3 vectors.
- * r0: vector address
- */
-ASM_FUNC(setup_vector)
-	mcr	p15, 0, r0, c12, c0, 1	@ MVBAR
-	isb
-	bx	lr
arch/aarch64/include/asm/psci.h new file +28 -0
new file mode 100644
index 0000000..491e685
@@ -0,0 +1,28 @@
+/*
+ * arch/aarch64/include/asm/psci.h
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __ASM_AARCH64_PSCI_H
+#define __ASM_AARCH64_PSCI_H
+
+#include <cpu.h>
+#include <stdbool.h>
+
+extern char psci_vectors[];
+
+static inline bool cpu_init_psci_arch(void)
+{
+	if (mrs(CurrentEL) != CURRENTEL_EL3)
+		return false;
+
+	msr(VBAR_EL3, (unsigned long)psci_vectors);
+	isb();
+
+	return true;
+}
+
+#endif
arch/aarch64/psci.S +2 -19
index 8bd224b..d6ca2eb 100644
@@ -19,7 +19,7 @@
 	.section .vectors, "w"
 
 	.align 11
-vector:
+ASM_DATA(psci_vectors)
 	// current EL, SP_EL0
 	ventry	err_exception	// synchronous
 	ventry	err_exception	// IRQ
@@ -80,22 +80,5 @@ smc_exit:
 	eret
 
 ASM_FUNC(start_el3)
-	ldr	x0, =vector
-	bl	setup_vector
-
-	/* only boot the primary cpu (entry 0 in the table) */
-	cpuid	x0, x1
-	bl	find_logical_id
-	b	psci_first_spin
-
-/*
- * This PSCI implementation requires EL3. Without EL3 we'll only boot the
- * primary cpu, all others will be trapped in an infinite loop.
- */
 ASM_FUNC(start_no_el3)
-	cpuid	x0, x1
-	bl	find_logical_id
-	cbz	x0, psci_first_spin
-spin_dead:
-	wfe
-	b	spin_dead
+	b	psci_first_spin
arch/aarch64/spin.S +3 -0
index 1ea1c0b..764c532 100644
@@ -12,6 +12,9 @@
 
 	.text
 
+ASM_FUNC(cpu_init_bootmethod)
+	ret
+
 ASM_FUNC(start_el3)
 ASM_FUNC(start_no_el3)
 	cpuid	x0, x1
arch/aarch64/utils.S +0 -9
index 85c7f8a..32393cc 100644
@@ -32,12 +32,3 @@ ASM_FUNC(find_logical_id)
 	ret
 3:	mov	x0, #MPIDR_INVALID
 	ret
-
-/*
- * Setup EL3 vectors
- * x0: vector address
- */
-ASM_FUNC(setup_vector)
-	msr	VBAR_EL3, x0
-	isb
-	ret
common/init.c +6 -1
index fc74b9e..3c05ac3 100644
@@ -6,6 +6,7 @@
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE.txt file.
  */
+#include <boot.h>
 #include <cpu.h>
 #include <platform.h>
 
@@ -44,7 +45,9 @@ void announce_arch(void);
 
 void cpu_init_bootwrapper(void)
 {
-	if (this_cpu_logical_id() == 0) {
+	unsigned int cpu = this_cpu_logical_id();
+
+	if (cpu == 0) {
 		init_uart();
 		announce_bootwrapper();
 		announce_arch();
@@ -52,4 +55,6 @@ void cpu_init_bootwrapper(void)
 		print_string("\r\n");
 		init_platform();
 	}
+
+	cpu_init_bootmethod(cpu);
 }
common/psci.c +19 -3
index a0e8700..8af6870 100644
@@ -12,8 +12,11 @@
 #include <bakery_lock.h>
 #include <boot.h>
 #include <cpu.h>
+#include <platform.h>
 #include <psci.h>
 
+#include <asm/psci.h>
+
 #ifndef CPU_IDS
 #error "No MPIDRs provided"
 #endif
@@ -78,12 +81,25 @@ long psci_call(unsigned long fid, unsigned long arg1, unsigned long arg2)
 	}
 }
 
-void __noreturn psci_first_spin(unsigned int cpu)
+void __noreturn psci_first_spin(void)
 {
-	if (cpu == MPIDR_INVALID)
-		while (1);
+	unsigned int cpu = this_cpu_logical_id();
 
 	first_spin(cpu, branch_table + cpu, PSCI_ADDR_INVALID);
 
 	unreachable();
 }
+
+void cpu_init_bootmethod(unsigned int cpu)
+{
+	if (cpu_init_psci_arch())
+		return;
+
+	if (cpu == 0) {
+		print_string("WARNING: PSCI could not be initialized. Boot may fail\r\n\r\n");
+		return;
+	}
+
+	while (1)
+		wfe();
+}
include/boot.h +2 -0
index d75e013..10968f8 100644
@@ -16,4 +16,6 @@ void __noreturn spin(unsigned long *mbox, unsigned long invalid, int is_entry);
 void __noreturn first_spin(unsigned int cpu, unsigned long *mbox,
 			   unsigned long invalid_addr);
 
+void cpu_init_bootmethod(unsigned int cpu);
+
 #endif
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#37 lore raw

Re: [bootwrapper PATCH v2 12/13] Rework bootmethod initialization

On Fri, 14 Jan 2022 10:56:52 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 151 lines

We currently initialize the bootmethod late, in assembly code. This requires us to maintain the el3/no_el3 distintion late into the boot process, and means we cannot produce any helpful diagnostic when booted at an unexpected exception level.

Rework things so that we initialize the bootmethod early, with a warning when things are wrong. The el3/no_el3 distinction is now irrelevant to the bootmethod code, and can be removed in subsequent patches.

When a boot-wrapper configured for PSCI is entered at EL2, a warning is looged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL2
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
|
| WARNING: PSCI could not be initialized. Boot may fail
---
 arch/aarch32/include/asm/cpu.h  |  1 +
 arch/aarch32/include/asm/psci.h | 28 ++++++++++++++++++++++++++++
 arch/aarch32/psci.S             |  8 +-------
 arch/aarch32/utils.S            |  9 ---------
 arch/aarch64/include/asm/psci.h | 28 ++++++++++++++++++++++++++++
 arch/aarch64/psci.S             | 21 ++-------------------
 arch/aarch64/spin.S             |  3 +++
 arch/aarch64/utils.S            |  9 ---------
 common/init.c                   |  7 ++++++-
 common/psci.c                   | 22 +++++++++++++++++++---
 include/boot.h                  |  2 ++
 11 files changed, 90 insertions(+), 48 deletions(-)
 create mode 100644 arch/aarch32/include/asm/psci.h
 create mode 100644 arch/aarch64/include/asm/psci.h
arch/aarch32/include/asm/cpu.h +1 -0
index c1bce9a..3426075 100644
@@ -63,6 +63,7 @@ static inline unsigned long read_cpsr(void)
 #define SCR		"p15, 0, %0, c1, c1, 0"
 #define NSACR		"p15, 0, %0, c1, c1, 2"
 #define ICIALLU		"p15, 0, %0, c7, c5, 0"
+#define MVBAR		"p15, 0, %0, c12, c0, 1"
 
 #define ICC_SRE		"p15, 6, %0, c12, c12, 5"
 #define ICC_CTLR	"p15, 6, %0, c12, c12, 4"
arch/aarch32/include/asm/psci.h new file +28 -0
new file mode 100644
index 0000000..50c7c70
@@ -0,0 +1,28 @@
+/*
+ * arch/aarch64/include/asm/psci.h
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __ASM_AARCH64_PSCI_H
+#define __ASM_AARCH64_PSCI_H
+
+#include <cpu.h>
+#include <stdbool.h>
+
+extern char psci_vectors[];
+
+static inline bool cpu_init_psci_arch(void)
+{
+	if (read_cpsr_mode() != PSR_MON)
+		return false;
+
+	mcr(MVBAR, (unsigned long)psci_vectors);
+	isb();
+
+	return true;
+}
+
+#endif
arch/aarch32/psci.S +1 -7
index e0d2972..cdc36b0 100644
@@ -15,7 +15,7 @@
 
 	.section .vectors
 	.align 6
-smc_vectors:
+ASM_DATA(psci_vectors)
 	b	err_exception			@ Reset
 	b	err_exception			@ Undef
 	b	handle_smc			@ SMC
@@ -39,11 +39,5 @@ handle_smc:
 	movs	pc, lr
 
 ASM_FUNC(start_el3)
-	ldr	r0, =smc_vectors
-	blx	setup_vector
-	/* pass through */
-
 ASM_FUNC(start_no_el3)
-	cpuid	r0, r1
-	blx	find_logical_id
 	b	psci_first_spin
arch/aarch32/utils.S +0 -9
index 5809f48..58279aa 100644
@@ -35,12 +35,3 @@ ASM_FUNC(find_logical_id)
 	bx	lr
 3:	mov	r0, #MPIDR_INVALID
 	bx	lr
-
-/*
- * Setup EL3 vectors.
- * r0: vector address
- */
-ASM_FUNC(setup_vector)
-	mcr	p15, 0, r0, c12, c0, 1	@ MVBAR
-	isb
-	bx	lr
arch/aarch64/include/asm/psci.h new file +18 -0
new file mode 100644
index 0000000..491e685
@@ -0,0 +1,28 @@
+/*
+ * arch/aarch64/include/asm/psci.h
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __ASM_AARCH64_PSCI_H
+#define __ASM_AARCH64_PSCI_H
+
+#include <cpu.h>
+#include <stdbool.h>
+
+extern char psci_vectors[];
+
+static inline bool cpu_init_psci_arch(void)
+{

arch/aarch64/include/asm/psci.h line 28

Quoted 10 lines
arch/aarch64/include/asm/psci.h new file +10 -0
+	if (mrs(CurrentEL) != CURRENTEL_EL3)
+		return false;
+
+	msr(VBAR_EL3, (unsigned long)psci_vectors);
+	isb();
+
+	return true;
+}
+
+#endif

Is there any particular reason that needs to live as a static inline in a header file? Can't we have the prototype in, say include/boot.h, and then have this in a proper C file, for instance arch/aarch<xx>/init.c?

The rest looks alright, though the change is somewhat hard to digest.

Cheers,
Andre

Quoted 148 lines
arch/aarch64/psci.S +2 -19
index 8bd224b..d6ca2eb 100644
@@ -19,7 +19,7 @@
 	.section .vectors, "w"
 
 	.align 11
-vector:
+ASM_DATA(psci_vectors)
 	// current EL, SP_EL0
 	ventry	err_exception	// synchronous
 	ventry	err_exception	// IRQ
@@ -80,22 +80,5 @@ smc_exit:
 	eret
 
 ASM_FUNC(start_el3)
-	ldr	x0, =vector
-	bl	setup_vector
-
-	/* only boot the primary cpu (entry 0 in the table) */
-	cpuid	x0, x1
-	bl	find_logical_id
-	b	psci_first_spin
-
-/*
- * This PSCI implementation requires EL3. Without EL3 we'll only boot the
- * primary cpu, all others will be trapped in an infinite loop.
- */
 ASM_FUNC(start_no_el3)
-	cpuid	x0, x1
-	bl	find_logical_id
-	cbz	x0, psci_first_spin
-spin_dead:
-	wfe
-	b	spin_dead
+	b	psci_first_spin
arch/aarch64/spin.S +3 -0
index 1ea1c0b..764c532 100644
@@ -12,6 +12,9 @@
 
 	.text
 
+ASM_FUNC(cpu_init_bootmethod)
+	ret
+
 ASM_FUNC(start_el3)
 ASM_FUNC(start_no_el3)
 	cpuid	x0, x1
arch/aarch64/utils.S +0 -9
index 85c7f8a..32393cc 100644
@@ -32,12 +32,3 @@ ASM_FUNC(find_logical_id)
 	ret
 3:	mov	x0, #MPIDR_INVALID
 	ret
-
-/*
- * Setup EL3 vectors
- * x0: vector address
- */
-ASM_FUNC(setup_vector)
-	msr	VBAR_EL3, x0
-	isb
-	ret
common/init.c +6 -1
index fc74b9e..3c05ac3 100644
@@ -6,6 +6,7 @@
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE.txt file.
  */
+#include <boot.h>
 #include <cpu.h>
 #include <platform.h>
 
@@ -44,7 +45,9 @@ void announce_arch(void);
 
 void cpu_init_bootwrapper(void)
 {
-	if (this_cpu_logical_id() == 0) {
+	unsigned int cpu = this_cpu_logical_id();
+
+	if (cpu == 0) {
 		init_uart();
 		announce_bootwrapper();
 		announce_arch();
@@ -52,4 +55,6 @@ void cpu_init_bootwrapper(void)
 		print_string("\r\n");
 		init_platform();
 	}
+
+	cpu_init_bootmethod(cpu);
 }
common/psci.c +19 -3
index a0e8700..8af6870 100644
@@ -12,8 +12,11 @@
 #include <bakery_lock.h>
 #include <boot.h>
 #include <cpu.h>
+#include <platform.h>
 #include <psci.h>
 
+#include <asm/psci.h>
+
 #ifndef CPU_IDS
 #error "No MPIDRs provided"
 #endif
@@ -78,12 +81,25 @@ long psci_call(unsigned long fid, unsigned long arg1, unsigned long arg2)
 	}
 }
 
-void __noreturn psci_first_spin(unsigned int cpu)
+void __noreturn psci_first_spin(void)
 {
-	if (cpu == MPIDR_INVALID)
-		while (1);
+	unsigned int cpu = this_cpu_logical_id();
 
 	first_spin(cpu, branch_table + cpu, PSCI_ADDR_INVALID);
 
 	unreachable();
 }
+
+void cpu_init_bootmethod(unsigned int cpu)
+{
+	if (cpu_init_psci_arch())
+		return;
+
+	if (cpu == 0) {
+		print_string("WARNING: PSCI could not be initialized. Boot may fail\r\n\r\n");
+		return;
+	}
+
+	while (1)
+		wfe();
+}
include/boot.h +0 -0
index d75e013..10968f8 100644

include/boot.h void __noreturn spin(unsigned long *mbox, unsigned long invalid, int is_entry); · line 21

Quoted 6 lines
include/boot.h +2 -0
@@ -16,4 +16,6 @@ void __noreturn spin(unsigned long *mbox, unsigned long invalid, int is_entry);
 void __noreturn first_spin(unsigned int cpu, unsigned long *mbox,
 			   unsigned long invalid_addr);
 
+void cpu_init_bootmethod(unsigned int cpu);
+
 #endif

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#38 lore raw

Re: [bootwrapper PATCH v2 12/13] Rework bootmethod initialization

Quoted 59 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:52 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 54 lines

We currently initialize the bootmethod late, in assembly code. This requires us to maintain the el3/no_el3 distintion late into the boot process, and means we cannot produce any helpful diagnostic when booted at an unexpected exception level.

Rework things so that we initialize the bootmethod early, with a warning when things are wrong. The el3/no_el3 distinction is now irrelevant to the bootmethod code, and can be removed in subsequent patches.

When a boot-wrapper configured for PSCI is entered at EL2, a warning is looged to the serial console as:

| Boot-wrapper v0.2
| Entered at EL2
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper
| [000000008000fff8..0000000080010000] => mbox
| [0000000080200000..00000000822af200] => kernel
| [0000000088000000..0000000088002857] => dtb
|
| WARNING: PSCI could not be initialized. Boot may fail
arch/aarch64/include/asm/psci.h new file +24 -0
new file mode 100644
index 0000000..491e685
@@ -0,0 +1,28 @@
+/*
+ * arch/aarch64/include/asm/psci.h
+ *
+ * Copyright (C) 2021 ARM Limited. All rights reserved.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE.txt file.
+ */
+#ifndef __ASM_AARCH64_PSCI_H
+#define __ASM_AARCH64_PSCI_H
+
+#include <cpu.h>
+#include <stdbool.h>
+
+extern char psci_vectors[];
+
+static inline bool cpu_init_psci_arch(void)
+{
+	if (mrs(CurrentEL) != CURRENTEL_EL3)
+		return false;
+
+	msr(VBAR_EL3, (unsigned long)psci_vectors);
+	isb();
+
Quoted 8 lines

arch/aarch64/include/asm/psci.h line 28

Quoted 4 lines
arch/aarch64/include/asm/psci.h new file +4 -0
+	return true;
+}
+
+#endif

Is there any particular reason that needs to live as a static inline in a header file? Can't we have the prototype in, say include/boot.h, and then have this in a proper C file, for instance arch/aarch<xx>/init.c?

At the time I originally wrote it, I had thought that this was the simplest option, but you're right that it's cleaner to place this in the relevant init.c file -- done. :)

Thanks,
Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#39 lore raw

PATCH v2 13/13 Unify start_el3 & start_no_el3

In reply to Mark Rutland

Now that the start_el3 and start_no_el3 labels point at the same place, unify them into a start_bootmethod label and update callers.

There should be no functional change as a result of this patch.

---
 arch/aarch32/boot.S | 5 ++---
 arch/aarch32/psci.S | 3 +--
 arch/aarch64/boot.S | 4 ++--
 arch/aarch64/psci.S | 3 +--
 arch/aarch64/spin.S | 3 +--
 5 files changed, 7 insertions(+), 11 deletions(-)
arch/aarch32/boot.S +2 -3
index 820957b..4d16c9c 100644
@@ -51,7 +51,7 @@ ASM_FUNC(_start)
 
 	bl	cpu_init_bootwrapper
 
-	b	start_no_el3
+	b	start_bootmethod
 
 _switch_monitor:
 	adr	lr, _monitor
@@ -69,8 +69,7 @@ _monitor:
 
 	bl	gic_secure_init
 
-	/* Initialise boot method */
-	b	start_el3
+	b	start_bootmethod
 
 err_invalid_id:
 	b	.
arch/aarch32/psci.S +1 -2
index cdc36b0..6613b6f 100644
@@ -38,6 +38,5 @@ handle_smc:
 	pop	{r4 - r12, lr}
 	movs	pc, lr
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	b	psci_first_spin
arch/aarch64/boot.S +2 -2
index c0ec518..da5fa65 100644
@@ -57,7 +57,7 @@ reset_at_el3:
 
 	bl	gic_secure_init
 
-	b	start_el3
+	b	start_bootmethod
 
 	/*
 	 * EL2 initialization
@@ -97,7 +97,7 @@ reset_no_el3:
 
 	bl	cpu_init_bootwrapper
 
-	b	start_no_el3
+	b	start_bootmethod
 
 err_invalid_id:
 	b	.
arch/aarch64/psci.S +1 -2
index d6ca2eb..9709dbb 100644
@@ -79,6 +79,5 @@ smc_exit:
 	ldp	x18, x19, [sp], #16
 	eret
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	b	psci_first_spin
arch/aarch64/spin.S +1 -2
index 764c532..375f732 100644
@@ -15,8 +15,7 @@
 ASM_FUNC(cpu_init_bootmethod)
 	ret
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	cpuid	x0, x1
 	bl	find_logical_id
 
Signature · 7 lines
2.30.2


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Reviewed-by on v2 13/13
#40 lore raw

Re: [bootwrapper PATCH v2 13/13] Unify start_el3 & start_no_el3

Quoted 2 lines · On Fri, 14 Jan 2022 10:56:53 +0000 Mark Rutland <[email protected]> wrote:

Now that the start_el3 and start_no_el3 labels point at the same place, unify them into a start_bootmethod label and update callers.

Nice one.

Quoted 3 lines

There should be no functional change as a result of this patch.

Cheers,
Andre

Quoted 82 lines
---
 arch/aarch32/boot.S | 5 ++---
 arch/aarch32/psci.S | 3 +--
 arch/aarch64/boot.S | 4 ++--
 arch/aarch64/psci.S | 3 +--
 arch/aarch64/spin.S | 3 +--
 5 files changed, 7 insertions(+), 11 deletions(-)
arch/aarch32/boot.S +2 -3
index 820957b..4d16c9c 100644
@@ -51,7 +51,7 @@ ASM_FUNC(_start)
 
 	bl	cpu_init_bootwrapper
 
-	b	start_no_el3
+	b	start_bootmethod
 
 _switch_monitor:
 	adr	lr, _monitor
@@ -69,8 +69,7 @@ _monitor:
 
 	bl	gic_secure_init
 
-	/* Initialise boot method */
-	b	start_el3
+	b	start_bootmethod
 
 err_invalid_id:
 	b	.
arch/aarch32/psci.S +1 -2
index cdc36b0..6613b6f 100644
@@ -38,6 +38,5 @@ handle_smc:
 	pop	{r4 - r12, lr}
 	movs	pc, lr
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	b	psci_first_spin
arch/aarch64/boot.S +2 -2
index c0ec518..da5fa65 100644
@@ -57,7 +57,7 @@ reset_at_el3:
 
 	bl	gic_secure_init
 
-	b	start_el3
+	b	start_bootmethod
 
 	/*
 	 * EL2 initialization
@@ -97,7 +97,7 @@ reset_no_el3:
 
 	bl	cpu_init_bootwrapper
 
-	b	start_no_el3
+	b	start_bootmethod
 
 err_invalid_id:
 	b	.
arch/aarch64/psci.S +1 -2
index d6ca2eb..9709dbb 100644
@@ -79,6 +79,5 @@ smc_exit:
 	ldp	x18, x19, [sp], #16
 	eret
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	b	psci_first_spin
arch/aarch64/spin.S +0 -0
index 764c532..375f732 100644

arch/aarch64/spin.S line 21

Quoted 9 lines
arch/aarch64/spin.S +1 -2
@@ -15,8 +15,7 @@
 ASM_FUNC(cpu_init_bootmethod)
 	ret
 
-ASM_FUNC(start_el3)
-ASM_FUNC(start_no_el3)
+ASM_FUNC(start_bootmethod)
 	cpuid	x0, x1
 	bl	find_logical_id
 

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#41 lore raw
In reply to Mark Rutland

On Fri, 14 Jan 2022 10:56:40 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 59 lines

This series reworks the aarch64 boot-wrapper, moving a fair amount of initialization code to C. This has a few benefits:

1) This makes the code more legible and easier to maintain.

Current feature detection and system register configuration is written in assembly, requiring runs of *almost* identical blocks of assembly to read ID registers and conditionally initialize register values. This requires a bunch of labels (which are all named numerically), and all the magic numbers are hard coded, so this gets pretty painful to read:

| mrs x1, id_aa64isar0_el1
| ubfx x1, x1, #24, #4
| cbz x1, 1f
|
| orr x0, x0, #(1 << 34) // TME enable
|
| 1:

In C, it's much easier to add helpers which use mnemonics, which makes the code much easier to read, and avoids the need to manually allocate registers, etc:

| if (mrs_field(ID_AA64ISAR0_EL1, TME))
| scr |= SCR_EL3_TME;

This should make it easier to handle new architectural extensions (and/or architecture variants such as ARMv8-R) in future.

2) This allows for better diagnostics.

Currently a mis-configured boot-wrapper is rather unforgiving, and provides no indication as to what might have gone wrong. By moving initialization to C, we can make use to the UART code to log diagnostic information, and we can more easily add additional error handling and conditional logic.

This series adds diagnostic information and error handling that can be used to identify problems such as the boot-wrapper being launched at the wrong exception level:

| Boot-wrapper v0.2
| Entered at EL2
| Memory layout:
| [0000000080000000..0000000080001f90] => boot-wrapper | [000000008000fff8..0000000080010000] => mbox | [0000000080200000..00000000822af200] => kernel | [0000000088000000..0000000088002857] => dtb |
| WARNING: PSCI could not be initialized. Boot may fail

Originally I had planned for this to be a more expansive set of changes, unifying some early control-flow, fixing some latent bugs, and making the boot-wrapper dynamically handle being entered at any of EL{3,2,1} with automated selection of a suitable DT. As the series has already become pretty long, I'd like to get this preparatory cleanup out of the way first, and handle those cases with subsequent patches.

Quoted 2 lines

If there are no objections, I intend to apply these by the end of the day.

Can you hold back with that till the beginning of next week? I got stuck halfway in the series with my review, but am on it now as we speak.

Cheers,
Andre

Quoted 59 lines

I've pushed the series to the cleanup branch in the boot-wrapper repo:

https://git.kernel.org/pub/scm/linux/kernel/git/mark/boot-wrapper-aarch64.git/
git://git.kernel.org/pub/scm/linux/kernel/git/mark/boot-wrapper-aarch64.git

... and it should apply cleanup atop the master branch.

Since v1 [1]:
* Add comments for bit-field macros
* Use BIT() for *_RES1 definitions
* Complete start_el3/start_no_el3 cleanup
* Remove extraneous macro definition

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Mark.

Mark Rutland (13):
Document entry requirements
Add bit-field macros
aarch64: add system register accessors
aarch32: add coprocessor accessors
aarch64: add mov_64 macro
aarch64: initialize SCTLR_ELx for the boot-wrapper Rework common init C code
Announce boot-wrapper mode / exception level
aarch64: move the bulk of EL3 initialization to C
aarch32: move the bulk of Secure PL1 initialization to C Announce locations of memory objects
Rework bootmethod initialization
Unify start_el3 & start_no_el3

 Makefile.am                       |   6 +-
 arch/aarch32/boot.S               |  33 +++---
 arch/aarch32/include/asm/cpu.h    |  62 ++++++++---
 arch/aarch32/include/asm/gic-v3.h |   6 +-
 arch/aarch32/include/asm/psci.h   |  28 +++++
 arch/aarch32/init.c               |  42 ++++++++
 arch/aarch32/psci.S               |  16 +--
 arch/aarch32/utils.S              |   9 --
 arch/aarch64/boot.S               | 169 +++++++++++++-----------------
 arch/aarch64/common.S             |  10 +-
 arch/aarch64/include/asm/cpu.h    | 102 +++++++++++++++---
 arch/aarch64/include/asm/gic-v3.h |  10 +-
 arch/aarch64/include/asm/psci.h   |  28 +++++
 arch/aarch64/init.c               |  88 ++++++++++++++++
 arch/aarch64/psci.S               |  22 +---
 arch/aarch64/spin.S               |   6 +-
 arch/aarch64/utils.S              |   9 --
 common/boot.c                     |   4 -
 common/init.c                     |  60 +++++++++++
 common/platform.c                 |  45 +++++---
 common/psci.c                     |  22 +++-
 include/bits.h                    |  79 ++++++++++++++
 include/boot.h                    |   2 +
 include/platform.h                |  20 ++++
 model.lds.S                       |  20 ++--
Quoted 9 lines
 25 files changed, 668 insertions(+), 230 deletions(-)
 create mode 100644 arch/aarch32/include/asm/psci.h
 create mode 100644 arch/aarch32/init.c
 create mode 100644 arch/aarch64/include/asm/psci.h
 create mode 100644 arch/aarch64/init.c
 create mode 100644 common/init.c
 create mode 100644 include/bits.h
 create mode 100644 include/platform.h

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#42 lore raw
Quoted 10 lines · Andre Przywara

On Fri, 14 Jan 2022 10:56:40 +0000
Mark Rutland <[email protected]> wrote:

Hi Mark,

Quoted 2 lines

If there are no objections, I intend to apply these by the end of the day.

Can you hold back with that till the beginning of next week? I got stuck halfway in the series with my review, but am on it now as we speak.

Sure.

Thanks for taking a look. :)

Mark.

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel