efi/x86: Avoid corrupted config tables under Xen
Subsystems (3)
Ard Biesheuvel
maintainerv2 cover efi/x86: Avoid corrupted config tables under Xen
This is an alternate approach to addressing the issue that Demi Marie is attempting to fix in [0] (i.e., ESRT config table exposed to a x86 dom0 is corrupted because it resides in boot services memory as per the EFI spec, where it gets corrupted by Xen). My main objection to that approach is that it needs Xen-specific fixes in multiple different places, but we still end up only fixing the ESRT case specifically.
So instead, I am proposing this series as a more generic way to handle configuration tables that reside in boot services memory, and confining the Xen specific logic to the Xen EFI glue code.
Given that EFI boot without a memory map is only permitted on x86 and only when doing Xen boot, let's clear up some inconsistencies there first so we can set the EFI_PARAVIRT flag on all architectures that do pseudo-EFI boot straight into the core kernel (i.e., without going through the stub). This moves a good chunk of EFI memory map manipulation code into the x86 arch tree, where it arguably belongs as no other architectures rely on it. This is implemented in patches 1 - 3.
Patch #4 refactors the ESRT sanity checks on the memory descriptor, by moving them into the efi_mem_desc_lookup() helper, which should not return corrupted descriptors in the first place.
Patch #5 adds a Xen hypercall fallback to efi_mem_desc_lookup() when running under Xen without a EFI memory map, so that, e.g., the existing ESRT code will perform its validation against the Xen provided descriptor if no memory map is available.
Patch #6 updates the config table traversal code so that the Xen glue code can force them to be disregarded, which happens when the table in question points into a memory region that is not of a type that Xen automatically reserves. Future changes can refine this logic if needed.
Changes since v1:
- add patch #4
- move Xen descriptor lookup into efi_mem_desc_lookup()
- drop allowlist for ACPI and SMBIOS tables
[0] https://lore.kernel.org/all/[email protected]/
Ard Biesheuvel (6):
efi: Move EFI fake memmap support into x86 arch tree
efi: memmap: Move manipulation routines into x86 arch tree
efi: xen: Set EFI_PARAVIRT for Xen dom0 boot on all architectures
efi: memmap: Disregard bogus entries instead of returning them
efi: xen: Implement memory descriptor lookup based on hypercall
efi: Apply allowlist to EFI configuration tables when running under
Xen
arch/x86/Kconfig | 20 ++
arch/x86/include/asm/efi.h | 16 ++
arch/x86/kernel/setup.c | 1 +
arch/x86/platform/efi/Makefile | 4 +-
arch/x86/platform/efi/efi.c | 8 +-
{drivers/firmware => arch/x86/platform}/efi/fake_mem.c | 79 ++++++-
arch/x86/platform/efi/memmap.c | 238 ++++++++++++++++++++
drivers/firmware/efi/Kconfig | 22 --
drivers/firmware/efi/Makefile | 4 -
drivers/firmware/efi/efi.c | 25 +-
drivers/firmware/efi/esrt.c | 18 +-
drivers/firmware/efi/fake_mem.h | 10 -
drivers/firmware/efi/fdtparams.c | 4 +
drivers/firmware/efi/memmap.c | 224 +-----------------
drivers/firmware/efi/x86_fake_mem.c | 75 ------
drivers/xen/efi.c | 58 +++++
include/linux/efi.h | 19 +-
17 files changed, 446 insertions(+), 379 deletions(-)
rename {drivers/firmware => arch/x86/platform}/efi/fake_mem.c (58%)
create mode 100644 arch/x86/platform/efi/memmap.c
delete mode 100644 drivers/firmware/efi/fake_mem.h
delete mode 100644 drivers/firmware/efi/x86_fake_mem.cSignature · 1 line
2.35.1
Ard Biesheuvel
maintainerPATCH v2 1/6 efi: Move EFI fake memmap support into x86 arch tree
The EFI fake memmap support is specific to x86, which manipulates the EFI memory map in various different ways after receiving it from the EFI stub. On other architectures, we have manages to push back on this, and the EFI memory map is kept pristine.
So let's move the fake memmap code into the x86 arch tree, where it arguably belongs.
---
arch/x86/Kconfig | 20 +++++
arch/x86/include/asm/efi.h | 5 ++
arch/x86/kernel/setup.c | 1 +
arch/x86/platform/efi/Makefile | 1 +
{drivers/firmware => arch/x86/platform}/efi/fake_mem.c | 79 +++++++++++++++++++-
drivers/firmware/efi/Kconfig | 22 ------
drivers/firmware/efi/Makefile | 4 -
drivers/firmware/efi/fake_mem.h | 10 ---
drivers/firmware/efi/x86_fake_mem.c | 75 -------------------
include/linux/efi.h | 6 --
10 files changed, 103 insertions(+), 120 deletions(-)arch/x86/Kconfig +20 -0
@@ -1978,6 +1978,26 @@ config EFI_MIXED If unsure, say N. +config EFI_FAKE_MEMMAP + bool "Enable EFI fake memory map" + depends on EFI + help + Saying Y here will enable "efi_fake_mem" boot option. By specifying + this parameter, you can add arbitrary attribute to specific memory + range by updating original (firmware provided) EFI memmap. This is + useful for debugging of EFI memmap related feature, e.g., Address + Range Mirroring feature. + +config EFI_MAX_FAKE_MEM + int "maximum allowable number of ranges in efi_fake_mem boot option" + depends on EFI_FAKE_MEMMAP + range 1 128 + default 8 + help + Maximum allowable number of ranges in efi_fake_mem boot option. + Ranges can be set up to this value using comma-separated list. + The default value is 8. + source "kernel/Kconfig.hz" config KEXEC
arch/x86/include/asm/efi.h +5 -0
@@ -404,10 +404,15 @@ static inline void efi_reserve_boot_services(void)#ifdef CONFIG_EFI_FAKE_MEMMAP extern void __init efi_fake_memmap_early (void ); +extern void __init efi_fake_memmap (void );#else static inline void efi_fake_memmap_early (void ) { } + +static inline void efi_fake_memmap (void ) +{ +}#endif #define arch_ima_efi_boot_mode \
arch/x86/kernel/setup.c +1 -0
@@ -31,6 +31,7 @@#include <xen/xen.h> #include <asm/apic.h> +#include <asm/efi.h> #include <asm/numa.h> #include <asm/bios_ebda.h> #include <asm/bugs.h>
arch/x86/platform/efi/Makefile +1 -0
@@ -4,3 +4,4 @@ GCOV_PROFILE := n obj-$(CONFIG_EFI) += quirks.o efi.o efi_$(BITS).o efi_stub_$(BITS).o obj-$(CONFIG_EFI_MIXED) += efi_thunk_$(BITS).o +obj-$(CONFIG_EFI_FAKE_MEMMAP) += fake_mem.o
arch/x86/platform/efi/fake_mem.c renamed from drivers/firmware/efi/fake_mem.c +76 -3
89 lines not on this page · raw message
drivers/firmware/efi/Kconfig +0 -22
@@ -37,28 +37,6 @@ config EFI_RUNTIME_MAP See also Documentation/ABI/testing/sysfs-firmware-efi-runtime-map. -config EFI_FAKE_MEMMAP - bool "Enable EFI fake memory map" - depends on EFI && X86 - default n - help - Saying Y here will enable "efi_fake_mem" boot option. - By specifying this parameter, you can add arbitrary attribute - to specific memory range by updating original (firmware provided) - EFI memmap. - This is useful for debugging of EFI memmap related feature. - e.g. Address Range Mirroring feature. - -config EFI_MAX_FAKE_MEM - int "maximum allowable number of ranges in efi_fake_mem boot option" - depends on EFI_FAKE_MEMMAP - range 1 128 - default 8 - help - Maximum allowable number of ranges in efi_fake_mem boot option. - Ranges can be set up to this value using comma-separated list. - The default value is 8. - config EFI_SOFT_RESERVE bool "Reserve EFI Specific Purpose Memory" depends on EFI && EFI_STUB && ACPI_HMAT
drivers/firmware/efi/Makefile +0 -4
@@ -23,7 +23,6 @@ obj-$(CONFIG_UEFI_CPER) += cper.o obj-$(CONFIG_EFI_RUNTIME_MAP) += runtime-map.o obj-$(CONFIG_EFI_RUNTIME_WRAPPERS) += runtime-wrappers.o subdir-$(CONFIG_EFI_STUB) += libstub -obj-$(CONFIG_EFI_FAKE_MEMMAP) += fake_map.o obj-$(CONFIG_EFI_BOOTLOADER_CONTROL) += efibc.o obj-$(CONFIG_EFI_TEST) += test/ obj-$(CONFIG_EFI_DEV_PATH_PARSER) += dev-path-parser.o
@@ -32,9 +31,6 @@ obj-$(CONFIG_EFI_RCI2_TABLE) += rci2-table.o obj-$(CONFIG_EFI_EMBEDDED_FIRMWARE) += embedded-firmware.o obj-$(CONFIG_LOAD_UEFI_KEYS) += mokvar-table.o -fake_map-y += fake_mem.o -fake_map-$(CONFIG_X86) += x86_fake_mem.o - obj-$(CONFIG_SYSFB) += sysfb_efi.o arm-obj-$(CONFIG_EFI) := efi-init.o arm-runtime.o
drivers/firmware/efi/fake_mem.h deleted +0 -10
@@ -1,10 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -#ifndef __EFI_FAKE_MEM_H__ -#define __EFI_FAKE_MEM_H__ -#include <asm/efi.h> - -#define EFI_MAX_FAKEMEM CONFIG_EFI_MAX_FAKE_MEM - -extern struct efi_mem_range efi_fake_mems[EFI_MAX_FAKEMEM ]; -extern int nr_fake_mem; -#endif /* __EFI_FAKE_MEM_H__ */
drivers/firmware/efi/x86_fake_mem.c deleted +0 -75
75 lines not on this page · raw message
include/linux/efi.h +0 -6
@@ -749,12 +749,6 @@ extern struct kobject *efi_kobj;extern int efi_reboot_quirk_mode;extern bool efi_poweroff_required (void ); -#ifdef CONFIG_EFI_FAKE_MEMMAP -extern void __init efi_fake_memmap (void ); -#else -static inline void efi_fake_memmap (void ) { } -#endif -extern unsigned long efi_mem_attr_table;/*
Signature · 1 line
2.35.1
Ard Biesheuvel
maintainerPATCH v2 2/6 efi: memmap: Move manipulation routines into x86 arch tree
In reply to Ard BiesheuvelThe EFI memory map is a description of the memory layout as provided by the firmware, and only x86 manipulates it in various different ways for its own memory bookkeeping. So let's move the memmap routines that are only used by x86 into the x86 arch tree.
--- arch/x86/include/asm/efi.h | 11 + arch/x86/platform/efi/Makefile | 3 +- arch/x86/platform/efi/memmap.c | 235 ++++++++++++++++++++ drivers/firmware/efi/memmap.c | 221 +----------------- include/linux/efi.h | 10 +- 5 files changed, 251 insertions(+), 229 deletions(-)
arch/x86/include/asm/efi.h +11 -0
@@ -415,6 +415,17 @@ static inline void efi_fake_memmap(void) }#endif +extern int __init efi_memmap_alloc (unsigned int num_entries, +struct efi_memory_map_data *data); +extern void __efi_memmap_free (u64 phys,unsigned long size, +unsigned long flags); + +extern int __init efi_memmap_install (struct efi_memory_map_data *data); +extern int __init efi_memmap_split_count (efi_memory_desc_t *md, +struct range *range); +extern void __init efi_memmap_insert (struct efi_memory_map *old_memmap, +void *buf,struct efi_mem_range *mem); +#define arch_ima_efi_boot_mode \ ({ extern struct boot_params boot_params; boot_params.secure_boot; })
arch/x86/platform/efi/Makefile +2 -1
@@ -2,6 +2,7 @@ KASAN_SANITIZE := n GCOV_PROFILE := n -obj-$(CONFIG_EFI) += quirks.o efi.o efi_$(BITS).o efi_stub_$(BITS).o +obj-$(CONFIG_EFI) += memmap.o quirks.o efi.o efi_$(BITS).o \ + efi_stub_$(BITS).o obj-$(CONFIG_EFI_MIXED) += efi_thunk_$(BITS).o obj-$(CONFIG_EFI_FAKE_MEMMAP) += fake_mem.o
arch/x86/platform/efi/memmap.c new file +235 -0
235 lines not on this page · raw message
drivers/firmware/efi/memmap.c +2 -219
243 lines not on this page · raw message
include/linux/efi.h +1 -9
@@ -707,18 +707,10 @@ static inline efi_status_t efi_query_variable_store(u32 attributes,#endif extern void __iomem *efi_lookup_mapped_addr (u64 phys_addr); -extern int __init efi_memmap_alloc (unsigned int num_entries, -struct efi_memory_map_data *data); -extern void __efi_memmap_free (u64 phys,unsigned long size, -unsigned long flags); +extern int __init __efi_memmap_init (struct efi_memory_map_data *data);extern int __init efi_memmap_init_early (struct efi_memory_map_data *data);extern int __init efi_memmap_init_late (phys_addr_t addr,unsigned long size);extern void __init efi_memmap_unmap (void ); -extern int __init efi_memmap_install (struct efi_memory_map_data *data); -extern int __init efi_memmap_split_count (efi_memory_desc_t *md, -struct range *range); -extern void __init efi_memmap_insert (struct efi_memory_map *old_memmap, -void *buf,struct efi_mem_range *mem);#ifdef CONFIG_EFI_ESRT extern void __init efi_esrt_init (void );
Signature · 1 line
2.35.1
Ard Biesheuvel
maintainerPATCH v2 3/6 efi: xen: Set EFI_PARAVIRT for Xen dom0 boot on all architectures
In reply to Ard BiesheuvelCurrently, the EFI_PARAVIRT flag is only used by x86, even though other architectures also support pseudo-EFI boot, where the core kernel is invoked directly and provided with a set of data tables that resemble the ones constructed by the EFI stub, which never actually runs in that case.
Let's fix this inconsistency, and always set this flag when booting dom0 via the EFI boot path. Note that Xen on x86 does not provide the EFI memory map in this case, whereas other architectures do, so move the associated EFI_PARAVIRT check into the x86 platform code.
--- arch/x86/platform/efi/efi.c | 8 +++++--- arch/x86/platform/efi/memmap.c | 3 +++ drivers/firmware/efi/fdtparams.c | 4 ++++ drivers/firmware/efi/memmap.c | 3 --- 4 files changed, 12 insertions(+), 6 deletions(-)
arch/x86/platform/efi/efi.c +5 -3
@@ -214,9 +214,11 @@ int __init efi_memblock_x86_reserve_range(void) data.desc_size = e->efi_memdesc_size; data.desc_version = e->efi_memdesc_version; - rv =efi_memmap_init_early (&data); -if (rv) -return rv; +if (!efi_enabled (EFI_PARAVIRT )) { + rv =efi_memmap_init_early (&data); +if (rv) +return rv; + }if (add_efi_memmap ||do_efi_soft_reserve ())do_add_efi_memmap ();
arch/x86/platform/efi/memmap.c +3 -0
@@ -93,6 +93,9 @@ int __init efi_memmap_install(struct efi_memory_map_data *data) {efi_memmap_unmap (); +if (efi_enabled (EFI_PARAVIRT )) +return 0 ; +return __efi_memmap_init (data); }
drivers/firmware/efi/fdtparams.c +4 -0
@@ -30,11 +30,13 @@ static __initconst const char name[][22] = {static __initconst const struct {const char path[17 ]; +u8 paravirt;const char params[PARAMCOUNT ][26 ]; } dt_params[] = { { #ifdef CONFIG_XEN // <-------17------> .path ="/hypervisor/uefi" , + .paravirt =1 , .params = { [SYSTAB ] ="xen,uefi-system-table" , [MMBASE ] ="xen,uefi-mmap-start" ,
@@ -121,6 +123,8 @@ u64 __init efi_get_fdt_params(struct efi_memory_map_data *mm)pr_err ("Can't find property '%s' in DT!\n" , pname);return 0 ; } +if (dt_params[i].paravirt) +set_bit (EFI_PARAVIRT , &efi.flags);return systab; } notfound:
drivers/firmware/efi/memmap.c +0 -3
@@ -44,9 +44,6 @@ int __init __efi_memmap_init(struct efi_memory_map_data *data)struct efi_memory_map map;phys_addr_t phys_map; -if (efi_enabled (EFI_PARAVIRT )) -return 0 ; - phys_map = data->phys_map;if (data->flags &EFI_MEMMAP_LATE )
Signature · 1 line
2.35.1
Ard Biesheuvel
maintainerPATCH v2 4/6 efi: memmap: Disregard bogus entries instead of returning them
In reply to Ard BiesheuvelThe ESRT code currently contains some sanity checks on the memory descriptor it obtains, but these can only trigger when the descriptor is invalid (if at all).
So let's drop these checks, and instead, disregard descriptors entirely if the start address is misaligned, or the number of pages reaches beyond the end of the address space. Note that the memory map as a whole could still be inconsistent, i.e., multiple entries might cover the same area, or the address could be outside of the addressable VA space, but validating that goes beyond the scope of these helpers.
--- drivers/firmware/efi/efi.c | 13 +++++++------ drivers/firmware/efi/esrt.c | 18 +----------------- 2 files changed, 8 insertions(+), 23 deletions(-)
drivers/firmware/efi/efi.c +7 -6
@@ -461,19 +461,20 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)efi_memory_desc_t *md;if (!efi_enabled (EFI_MEMMAP )) { -pr_err_once ("EFI_MEMMAP is not enabled.\n" ); +pr_warn_once ("EFI_MEMMAP is not enabled.\n" );return -EINVAL ; } -if (!out_md) { -pr_err_once ("out_md is null.\n" ); -return -EINVAL ; - } -for_each_efi_memory_desc (md) {u64 size;u64 end; +/* skip bogus entries */ +if ((md->phys_addr & (EFI_PAGE_SIZE -1 )) || + (md->phys_addr >0 && + (md->num_pages > (U64_MAX - md->phys_addr +1 ) >>EFI_PAGE_SHIFT ))) +continue ; + size = md->num_pages <<EFI_PAGE_SHIFT ; end = md->phys_addr + size;if (phys_addr >= md->phys_addr && phys_addr < end) {
drivers/firmware/efi/esrt.c +1 -17
@@ -247,9 +247,6 @@ void __init efi_esrt_init(void)int rc;phys_addr_t end; -if (!efi_enabled (EFI_MEMMAP )) -return ; -pr_debug ("esrt-init: loading.\n" );if (!esrt_table_exists ())return ;
@@ -263,21 +260,8 @@ void __init efi_esrt_init(void)return ; } - max =efi_mem_desc_end (&md); -if (max < efi.esrt) { -pr_err ("EFI memory descriptor is invalid. (esrt: %p max: %p)\n" , - (void *)efi.esrt, (void *)max); -return ; - } - + max =efi_mem_desc_end (&md) - efi.esrt; size =sizeof (*esrt); - max -= efi.esrt; - -if (max < size) { -pr_err ("ESRT header doesn't fit on single memory map entry. (size: %zu max: %zu)\n" , - size, max); -return ; - } va =early_memremap (efi.esrt, size);if (!va) {
Signature · 1 line
2.35.1
Re: [PATCH v2 4/6] efi: memmap: Disregard bogus entries instead of returning them
Quoted 25 lines · Ard Biesheuvel
The ESRT code currently contains some sanity checks on the memory descriptor it obtains, but these can only trigger when the descriptor is invalid (if at all).
So let's drop these checks, and instead, disregard descriptors entirely if the start address is misaligned, or the number of pages reaches beyond the end of the address space. Note that the memory map as a whole could still be inconsistent, i.e., multiple entries might cover the same area, or the address could be outside of the addressable VA space, but validating that goes beyond the scope of these helpers.
Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 13 +++++++------ drivers/firmware/efi/esrt.c | 18 +----------------- 2 files changed, 8 insertions(+), 23 deletions(-)drivers/firmware/efi/efi.c +0
-0@@ -461,19 +461,20 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)efi_memory_desc_t *md;if (!efi_enabled (EFI_MEMMAP )) {
drivers/firmware/efi/efi.c int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) · line 472
Quoted 10 lines
drivers/firmware/efi/efi.c +1
-6-pr_err_once ("EFI_MEMMAP is not enabled.\n" ); +pr_warn_once ("EFI_MEMMAP is not enabled.\n" );return -EINVAL ; } -if (!out_md) { -pr_err_once ("out_md is null.\n" ); -return -EINVAL ; - } -
Nit: this seems unrelated.
drivers/firmware/efi/efi.c int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) · line 476
Quoted 9 lines
drivers/firmware/efi/efi.c +5
-0for_each_efi_memory_desc (md) {u64 size;u64 end; +/* skip bogus entries */ +if ((md->phys_addr & (EFI_PAGE_SIZE -1 )) || + (md->phys_addr >0 && + (md->num_pages > (U64_MAX - md->phys_addr +1 ) >>EFI_PAGE_SHIFT ))) +continue ;
Should this also check if md->num_pages is 0? Also, should this check be part of for_each_efi_memory_desc()?
Quoted 28 lines
drivers/firmware/efi/efi.c +1
-0+ size = md->num_pages <<EFI_PAGE_SHIFT ; end = md->phys_addr + size;if (phys_addr >= md->phys_addr && phys_addr < end) {drivers/firmware/efi/esrt.c +0
-9@@ -247,9 +247,6 @@ void __init efi_esrt_init(void)int rc;phys_addr_t end; -if (!efi_enabled (EFI_MEMMAP )) -return ; -pr_debug ("esrt-init: loading.\n" );if (!esrt_table_exists ())return ;@@ -263,21 +260,8 @@ void __init efi_esrt_init(void)return ; } - max =efi_mem_desc_end (&md); -if (max < efi.esrt) { -pr_err ("EFI memory descriptor is invalid. (esrt: %p max: %p)\n" , - (void *)efi.esrt, (void *)max); -return ; - }
drivers/firmware/efi/esrt.c void __init efi_esrt_init(void) · line 280
Quoted 10 lines
drivers/firmware/efi/esrt.c +1
-8- + max =efi_mem_desc_end (&md) - efi.esrt; size =sizeof (*esrt); - max -= efi.esrt; - -if (max < size) { -pr_err ("ESRT header doesn't fit on single memory map entry. (size: %zu max: %zu)\n" , - size, max); -return ; - }
This can still happen if the ESRT pointer is very very close to the end of a memory map entry, unless there is another check that handles such cases.
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Ard Biesheuvel
maintainerRe: [PATCH v2 4/6] efi: memmap: Disregard bogus entries instead of returning them
Quoted 44 lines · Demi Marie Obenour
Quoted 25 lines · Ard Biesheuvel
The ESRT code currently contains some sanity checks on the memory descriptor it obtains, but these can only trigger when the descriptor is invalid (if at all).
So let's drop these checks, and instead, disregard descriptors entirely if the start address is misaligned, or the number of pages reaches beyond the end of the address space. Note that the memory map as a whole could still be inconsistent, i.e., multiple entries might cover the same area, or the address could be outside of the addressable VA space, but validating that goes beyond the scope of these helpers.
Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 13 +++++++------ drivers/firmware/efi/esrt.c | 18 +----------------- 2 files changed, 8 insertions(+), 23 deletions(-)drivers/firmware/efi/efi.c +0
-0@@ -461,19 +461,20 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)efi_memory_desc_t *md;if (!efi_enabled (EFI_MEMMAP )) {drivers/firmware/efi/efi.c int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) · line 472
Quoted 10 lines
drivers/firmware/efi/efi.c +1
-6-pr_err_once ("EFI_MEMMAP is not enabled.\n" ); +pr_warn_once ("EFI_MEMMAP is not enabled.\n" );return -EINVAL ; } -if (!out_md) { -pr_err_once ("out_md is null.\n" ); -return -EINVAL ; - } -Nit: this seems unrelated.
Quoted 4 lines
drivers/firmware/efi/efi.c +0
-0for_each_efi_memory_desc (md) {u64 size;u64 end;
Quoted 7 lines
drivers/firmware/efi/efi.c int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) · line 476
Quoted 5 lines
drivers/firmware/efi/efi.c +5
-0+/* skip bogus entries */ +if ((md->phys_addr & (EFI_PAGE_SIZE -1 )) || + (md->phys_addr >0 && + (md->num_pages > (U64_MAX - md->phys_addr +1 ) >>EFI_PAGE_SHIFT ))) +continue ;Should this also check if
md->num_pagesis 0?
Yes, probably.
Quoted 3 lines
Also, should this check
be part offor_each_efi_memory_desc()?
No, I don't think so. The for_each_xxx() helpers we have throughout the kernel usually don't incorporate such checks, and I'd prefer to adhere to the principle of least surprise here.
Quoted 34 lines
Quoted 34 lines
drivers/firmware/efi/efi.c +1
-0+ size = md->num_pages <<EFI_PAGE_SHIFT ; end = md->phys_addr + size;if (phys_addr >= md->phys_addr && phys_addr < end) {drivers/firmware/efi/esrt.c +1
-13@@ -247,9 +247,6 @@ void __init efi_esrt_init(void)int rc;phys_addr_t end; -if (!efi_enabled (EFI_MEMMAP )) -return ; -pr_debug ("esrt-init: loading.\n" );if (!esrt_table_exists ())return ;@@ -263,21 +260,8 @@ void __init efi_esrt_init(void)return ; } - max =efi_mem_desc_end (&md); -if (max < efi.esrt) { -pr_err ("EFI memory descriptor is invalid. (esrt: %p max: %p)\n" , - (void *)efi.esrt, (void *)max); -return ; - } - + max =efi_mem_desc_end (&md) - efi.esrt; size =sizeof (*esrt); - max -= efi.esrt; - -if (max < size) {
Quoted 8 lines
drivers/firmware/efi/esrt.c void __init efi_esrt_init(void) · line 280
Quoted 4 lines
drivers/firmware/efi/esrt.c +0
-4-pr_err ("ESRT header doesn't fit on single memory map entry. (size: %zu max: %zu)\n" , - size, max); -return ; - }This can still happen if the ESRT pointer is very very close to the end of a memory map entry, unless there is another check that handles such cases.
You're right - I missed that.
Ard Biesheuvel
maintainerPATCH v2 5/6 efi: xen: Implement memory descriptor lookup based on hypercall
In reply to Ard BiesheuvelXen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)
drivers/firmware/efi/efi.c +4 -1
@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;
@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */
drivers/xen/efi.c +34 -0
@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h>
@@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; + +if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr; + out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ; + out_md->type = info->mem.type; + out_md->attribute = info->mem.attr; + +return 0 ; +}
include/linux/efi.h +1 -0
@@ -731,6 +731,7 @@ extern u64 efi_mem_attribute (unsigned long phys_addr, unsigned long size);extern int __init efi_uart_console_only (void );extern u64 efi_mem_desc_end (efi_memory_desc_t *md);extern int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md); +extern int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md);extern void efi_mem_reserve (phys_addr_t addr,u64 size);extern int efi_mem_reserve_persistent (phys_addr_t addr,u64 size); extern void efi_initialize_iomem_resources(struct resource *code_resource,
Signature · 1 line
2.35.1
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +
drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;
This will be equal to phys_addr, not the actual start of the memory region.
drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;
Similarly, this will be the number of bytes in the memory region after phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the original efi_mem_desc_lookup().
I am not sure if this matters in practice, but I thought you would want to be aware of it.
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Ard Biesheuvel
maintainerRe: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 89 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;
Quoted 7 lines
Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
In any case, it means we'll need to round down phys_addr to page size at the very least.
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 16 lines · Ard Biesheuvel
Quoted 14 lines · Demi Marie Obenour
Quoted 1 line · Ard Biesheuvel
+
out_md->phys_addr=info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.Quoted 1 line
+
out_md->num_pages=info->mem.size>>EFI_PAGE_SHIFT;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().
Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
I don't think so, unfortunately. That said, with the below adjustment, I think that's okay for the _current_ users of efi_mem_desc_lookup().
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.
Signature · 3 lines
Best Regards, Marek Marczykowski-Górecki Invisible Things Lab
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
In reply to Ard BiesheuvelQuoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().
Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.
That makes sense. Are there any callers that will be broken even with this rounding?
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Ard Biesheuvel
maintainerRe: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from efi_arch_mem_reserve() this way, but as far as I can tell, that should be fine too.
Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 115 lines · Ard Biesheuvel
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
Quoted 7 lines
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from
efi_arch_mem_reserve()this way, but as far as I can tell, that should be fine too.Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Demi, if you open a PR with this at
https://github.com/QubesOS/qubes-linux-kernel/pulls, I can run it through our CI - (at least) one of the machines has ESRT table. AFAIR your test laptop has it too.
Signature · 3 lines
Best Regards, Marek Marczykowski-Górecki Invisible Things Lab
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 121 lines · Marek Marczykowski-Górecki
Quoted 120 lines · Ard Biesheuvel
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from
efi_arch_mem_reserve()this way, but as far as I can tell, that should be fine too.
Quoted 7 lines
Quoted 2 lines
Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Demi, if you open a PR with this at
https://github.com/QubesOS/qubes-linux-kernel/pulls, I can run it through our CI - (at least) one of the machines has ESRT table. AFAIR your test laptop has it too.
Just this patch or the whole series?
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 125 lines · Demi Marie Obenour
Quoted 124 lines · Marek Marczykowski-Górecki
Quoted 122 lines · Ard Biesheuvel
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from
efi_arch_mem_reserve()this way, but as far as I can tell, that should be fine too.Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Quoted 6 lines
Quoted 4 lines
Demi, if you open a PR with this at
https://github.com/QubesOS/qubes-linux-kernel/pulls, I can run it through our CI - (at least) one of the machines has ESRT table. AFAIR your test laptop has it too.Just this patch or the whole series?
Whole series.
Signature · 3 lines
Best Regards, Marek Marczykowski-Górecki Invisible Things Lab
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
Quoted 126 lines · Marek Marczykowski-Górecki
Quoted 125 lines · Demi Marie Obenour
Quoted 124 lines · Marek Marczykowski-Górecki
Quoted 122 lines · Ard Biesheuvel
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from
efi_arch_mem_reserve()this way, but as far as I can tell, that should be fine too.Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Quoted 8 lines
Quoted 6 lines
Quoted 4 lines
Demi, if you open a PR with this at
https://github.com/QubesOS/qubes-linux-kernel/pulls, I can run it through our CI - (at least) one of the machines has ESRT table. AFAIR your test laptop has it too.Just this patch or the whole series?
Whole series.
I have tested the series as in
https://github.com/QubesOS/qubes-linux-kernel/pull/681 and it seems to work great.
Note the series there differs from this thread, and is marked as "v3" - I assume (but haven't verified) it has changes requested in this thread applied. Demi, can you confirm? If so, you can probably send this v3, and feel free to include my Tested-by (unless you make significant changes, ofc).
Signature · 3 lines
Best Regards, Marek Marczykowski-Górecki Invisible Things Lab
Re: [PATCH v2 5/6] efi: xen: Implement memory descriptor lookup based on hypercall
In reply to Marek Marczykowski-GóreckiQuoted 121 lines · Marek Marczykowski-Górecki
Quoted 120 lines · Ard Biesheuvel
Quoted 104 lines · Demi Marie Obenour
Quoted 96 lines · Ard Biesheuvel
Quoted 94 lines · Demi Marie Obenour
Quoted 71 lines · Ard Biesheuvel
Xen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some sanity checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 5 ++- drivers/xen/efi.c | 34 ++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 39 insertions(+), 1 deletion(-)drivers/firmware/efi/efi.c +4
-1@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;@@ -485,6 +485,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */ drivers/xen/efi.c +18
-0@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h> @@ -292,3 +293,36 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op = { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + } + }; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; +drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 322
Quoted 10 lines
drivers/xen/efi.c +10
-0+if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr;This will be equal to
phys_addr, not the actual start of the memory region.drivers/xen/efi.c void __init xen_efi_runtime_setup(void) · line 323
Quoted 1 line
drivers/xen/efi.c +1
-0+ out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ;Similarly, this will be the number of bytes in the memory region after
phys_addr, not the total number of bytes in the region. These two differences mean that this function is not strictly equivalent to the originalefi_mem_desc_lookup().Quoted 5 lines
Quoted 2 lines
I am not sure if this matters in practice, but I thought you would want to be aware of it.
This is a bit disappointing. Is there no way to obtain this information via a Xen hypercall?
Quoted 8 lines
It is possible, but doing so is very complex (it essentially requires a binary search). This really should be fixed on the Xen side.
Quoted 2 lines
In any case, it means we'll need to round down
phys_addrto page size at the very least.That makes sense. Are there any callers that will be broken even with this rounding?
As far as I can tell, it should work fine. The only thing to double check is whether we are not creating spurious error messages from
efi_arch_mem_reserve()this way, but as far as I can tell, that should be fine too.
Quoted 6 lines
Quoted 2 lines
Is there anyone at your end that can give this a spin on an actual Xen/x86 system?
Demi, if you open a PR with this at
https://github.com/QubesOS/qubes-linux-kernel/pulls, I can run it through our CI - (at least) one of the machines has ESRT table.
Done: https://github.com/QubesOS/qubes-linux-kernel/pull/681
Quoted 1 line
AFAIR your test laptop has it too.
It does; I plan to test a version that has the needed rounding.
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Ard Biesheuvel
maintainerPATCH v2 6/6 efi: Apply allowlist to EFI configuration tables when running under Xen
In reply to Ard BiesheuvelAs it turns out, Xen does not guarantee that EFI bootservices data regions in memory are preserved, which means that EFI configuration tables pointing into such memory regions may be corrupted before the dom0 OS has had a chance to inspect them.
Demi Marie reports that this is causing problems for Qubes OS when it attempts to perform system firmware updates, which requires that the contents of the ESRT configuration table are valid when the fwupd user space program runs.
However, other configuration tables such as the memory attributes table or the runtime properties table are equally affected, and so we need a comprehensive workaround that works for any table type.
So when running under Xen, check the EFI memory descriptor covering the start of the table, and disregard the table if it does not reside in memory that is preserved by Xen.
--- drivers/firmware/efi/efi.c | 7 ++++++ drivers/xen/efi.c | 24 ++++++++++++++++++++ include/linux/efi.h | 2 ++ 3 files changed, 33 insertions(+)
drivers/firmware/efi/efi.c +7 -0
@@ -560,6 +560,13 @@ static __init int match_config_table(const efi_guid_t *guid,for (i =0 ;efi_guidcmp (table_types[i].guid,NULL_GUID ); i++) {if (!efi_guidcmp (*guid, table_types[i].guid)) { +if (IS_ENABLED (CONFIG_XEN_EFI ) && + !xen_efi_config_table_is_usable (guid, table)) { +if (table_types[i].name[0 ]) +pr_cont ("(%s=0x%lx) " , + table_types[i].name, table); +return 1 ; + } *(table_types[i].ptr) = table;if (table_types[i].name[0 ]) pr_cont("%s=0x%lx ",
drivers/xen/efi.c +24 -0
@@ -326,3 +326,27 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return 0 ; } + +bool __init xen_efi_config_table_is_usable (const efi_guid_t *guid, +unsigned long table) +{ +efi_memory_desc_t md; +int rc; + +if (!efi_enabled (EFI_PARAVIRT )) +return true; + + rc =efi_mem_desc_lookup (table, &md); +if (rc) +return false; + +switch (md.type) { +case EFI_RUNTIME_SERVICES_CODE : +case EFI_RUNTIME_SERVICES_DATA : +case EFI_ACPI_RECLAIM_MEMORY : +case EFI_RESERVED_TYPE : +return true; + } + +return false; +}
include/linux/efi.h +2 -0
@@ -1352,4 +1352,6 @@ struct linux_efi_initrd {/* Header of a populated EFI secret area */ #define EFI_SECRET_TABLE_HEADER_GUID EFI_GUID(0x1e74f542, 0x71dd, 0x4d66, 0x96, 0x3e, 0xef, 0x42, 0x87, 0xff, 0x17, 0x3b) +bool xen_efi_config_table_is_usable (const efi_guid_t *,unsigned long table); +#endif /* _LINUX_EFI_H */
Signature · 1 line
2.35.1
Re: [PATCH v2 6/6] efi: Apply allowlist to EFI configuration tables when running under Xen
Quoted 62 lines · Ard Biesheuvel
As it turns out, Xen does not guarantee that EFI bootservices data regions in memory are preserved, which means that EFI configuration tables pointing into such memory regions may be corrupted before the dom0 OS has had a chance to inspect them.
Demi Marie reports that this is causing problems for Qubes OS when it attempts to perform system firmware updates, which requires that the contents of the ESRT configuration table are valid when the fwupd user space program runs.
However, other configuration tables such as the memory attributes table or the runtime properties table are equally affected, and so we need a comprehensive workaround that works for any table type.
So when running under Xen, check the EFI memory descriptor covering the start of the table, and disregard the table if it does not reside in memory that is preserved by Xen.
Co-developed-by Demi Marie Obenour <[email protected]>Signed-off-by Demi Marie Obenour <[email protected]>Signed-off-by Ard Biesheuvel <[email protected]>--- drivers/firmware/efi/efi.c | 7 ++++++ drivers/xen/efi.c | 24 ++++++++++++++++++++ include/linux/efi.h | 2 ++ 3 files changed, 33 insertions(+)drivers/firmware/efi/efi.c +7
-0@@ -560,6 +560,13 @@ static __init int match_config_table(const efi_guid_t *guid,for (i =0 ;efi_guidcmp (table_types[i].guid,NULL_GUID ); i++) {if (!efi_guidcmp (*guid, table_types[i].guid)) { +if (IS_ENABLED (CONFIG_XEN_EFI ) && + !xen_efi_config_table_is_usable (guid, table)) { +if (table_types[i].name[0 ]) +pr_cont ("(%s=0x%lx) " , + table_types[i].name, table); +return 1 ; + } *(table_types[i].ptr) = table;if (table_types[i].name[0 ]) pr_cont("%s=0x%lx ",drivers/xen/efi.c +9
-0@@ -326,3 +326,27 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return 0 ; } + +bool __init xen_efi_config_table_is_usable (const efi_guid_t *guid, +unsigned long table) +{ +efi_memory_desc_t md; +int rc; + +if (!efi_enabled (EFI_PARAVIRT )) +return true;
drivers/xen/efi.c int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) · line 347
Quoted 10 lines
drivers/xen/efi.c +10
-0+ + rc =efi_mem_desc_lookup (table, &md); +if (rc) +return false; + +switch (md.type) { +case EFI_RUNTIME_SERVICES_CODE : +case EFI_RUNTIME_SERVICES_DATA : +case EFI_ACPI_RECLAIM_MEMORY : +case EFI_RESERVED_TYPE :
Some firmware uses EFI_ACPI_MEMORY_NVS to store ACPI tables, so this needs to be added to the allowlist. Otherwise affected systems will not boot. Xen treats EFI_ACPI_MEMORY_NVS the way it treats EFI_ACPI_RECLAIM_MEMORY, so this is safe.
Quoted 18 lines
drivers/xen/efi.c +5
-0+return true; + } + +return false; +}include/linux/efi.h +2
-0@@ -1352,4 +1352,6 @@ struct linux_efi_initrd {/* Header of a populated EFI secret area */ #define EFI_SECRET_TABLE_HEADER_GUID EFI_GUID(0x1e74f542, 0x71dd, 0x4d66, 0x96, 0x3e, 0xef, 0x42, 0x87, 0xff, 0x17, 0x3b) +bool xen_efi_config_table_is_usable (const efi_guid_t *,unsigned long table); +#endif /* _LINUX_EFI_H */ Signature · 1 line
2.35.1
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
v3 cover efi: Support ESRT under Xen
In reply to Ard BiesheuvelThis patch series fixes handling of EFI tables when running under Xen. These fixes allow the ESRT to be loaded when running paravirtualized in dom0, making the use of EFI capsule updates possible.
Demi Marie Obenour (5):
efi: memmap: Disregard bogus entries instead of returning them
efi: xen: Implement memory descriptor lookup based on hypercall
efi: Apply allowlist to EFI configuration tables when running under
Xen
efi: Actually enable the ESRT under Xen
efi: Warn if trying to reserve memory under Xen
drivers/firmware/efi/efi.c | 22 ++++++++++++- drivers/firmware/efi/esrt.c | 15 +++------ drivers/xen/efi.c | 61 +++++++++++++++++++++++++++++++++++++ include/linux/efi.h | 3 ++ 4 files changed, 90 insertions(+), 11 deletions(-)
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
PATCH v3 1/5 efi: memmap: Disregard bogus entries instead of returning them
The ESRT code currently contains two consistency checks on the memory descriptor it obtains, but one of them is both incomplete and can only trigger on invalid descriptors.
So let's drop these checks, and instead disregard descriptors entirely if the start address is misaligned, or if the number of pages reaches to or beyond the end of the address space. Note that the memory map as a whole could still be inconsistent: multiple entries might cover the same area, or the address could be outside of the addressable PA space, but validating that goes beyond the scope of these helpers. Also note that since the physical address space is never 64-bits wide, a descriptor that includes the last page of memory is not valid. This is fortunate, since it means that a valid physical address will never be an error pointer and that the length of a memory descriptor in bytes will fit in a 64-bit unsigned integer.
--- drivers/firmware/efi/efi.c | 6 ++++++ drivers/firmware/efi/esrt.c | 9 +-------- 2 files changed, 7 insertions(+), 8 deletions(-)
drivers/firmware/efi/efi.c +6 -0
@@ -474,6 +474,12 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)u64 size;u64 end; +/* skip bogus entries (including empty ones) */ +if ((md->phys_addr & (EFI_PAGE_SIZE -1 )) || + (md->num_pages <=0 ) || + (md->num_pages > (U64_MAX - md->phys_addr) >>EFI_PAGE_SHIFT )) +continue ; + size = md->num_pages <<EFI_PAGE_SHIFT ; end = md->phys_addr + size;if (phys_addr >= md->phys_addr && phys_addr < end) {
drivers/firmware/efi/esrt.c +1 -8
@@ -263,15 +263,8 @@ void __init efi_esrt_init(void)return ; } - max =efi_mem_desc_end (&md); -if (max < efi.esrt) { -pr_err ("EFI memory descriptor is invalid. (esrt: %p max: %p)\n" , - (void *)efi.esrt, (void *)max); -return ; - } - + max =efi_mem_desc_end (&md) - efi.esrt; size =sizeof (*esrt); - max -= efi.esrt;if (max < size) { pr_err("ESRT header doesn't fit on single memory map entry. (size: %zu max: %zu)\n",
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
PATCH v3 2/5 efi: xen: Implement memory descriptor lookup based on hypercall
In reply to Demi Marie ObenourXen on x86 boots dom0 in EFI mode but without providing a memory map. This means that some consistency checks we would like to perform on configuration tables or other data structures in memory are not currently possible. Xen does, however, expose EFI memory descriptor info via a Xen hypercall, so let's wire that up instead. It turns out that the returned information is not identical to what Linux's efi_mem_desc_lookup would return: the address returned is the address passed to the hypercall, and the size returned is the number of bytes remaining in the configuration table. However, none of the callers of efi_mem_desc_lookup() currently care about this. In the future, Xen may gain a hypercall that returns the actual start address, which can be used instead.
--- drivers/firmware/efi/efi.c | 5 ++++- drivers/xen/efi.c | 36 ++++++++++++++++++++++++++++++++++++ include/linux/efi.h | 1 + 3 files changed, 41 insertions(+), 1 deletion(-)
drivers/firmware/efi/efi.c +4 -1
@@ -456,7 +456,7 @@ void __init efi_find_mirror(void)* and if so, populate the supplied memory descriptor with the appropriate * data. */ -int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) {efi_memory_desc_t *md;
@@ -490,6 +490,9 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return -ENOENT ; } +extern int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md) + __weak __alias(__efi_mem_desc_lookup); +/* * Calculate the highest address of an efi memory descriptor. */
drivers/xen/efi.c +36 -0
@@ -26,6 +26,7 @@#include <xen/interface/xen.h> #include <xen/interface/platform.h> +#include <xen/page.h> #include <xen/xen.h> #include <xen/xen-ops.h>
@@ -292,3 +293,38 @@ void __init xen_efi_runtime_setup(void) efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count; efi.reset_system = xen_efi_reset_system; } + +int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md) +{ +static_assert (XEN_PAGE_SHIFT ==EFI_PAGE_SHIFT , +"Mismatch between EFI_PAGE_SHIFT and XEN_PAGE_SHIFT" ); +struct xen_platform_op op; +union xenpf_efi_info *info = &op.u.firmware_info.u.efi_info; +int rc; + +if (!efi_enabled (EFI_PARAVIRT ) ||efi_enabled (EFI_MEMMAP )) +return __efi_mem_desc_lookup (phys_addr, out_md); + phys_addr &= ~(u64)(EFI_PAGE_SIZE -1 ); + op = (struct xen_platform_op ) { + .cmd = XENPF_firmware_info, + .u.firmware_info = { + .type =XEN_FW_EFI_INFO , + .index =XEN_FW_EFI_MEM_INFO , + .u.efi_info.mem.addr = phys_addr, + .u.efi_info.mem.size =U64_MAX - phys_addr, + }, + }; + + rc =HYPERVISOR_platform_op (&op); +if (rc) { +pr_warn ("Failed to lookup header 0x%llx in Xen memory map: error %d\n" , + phys_addr, rc); + } + + out_md->phys_addr = info->mem.addr; + out_md->num_pages = info->mem.size >>EFI_PAGE_SHIFT ; + out_md->type = info->mem.type; + out_md->attribute = info->mem.attr; + +return 0 ; +}
include/linux/efi.h +1 -0
@@ -724,6 +724,7 @@ extern u64 efi_mem_attribute (unsigned long phys_addr, unsigned long size);extern int __init efi_uart_console_only (void );extern u64 efi_mem_desc_end (efi_memory_desc_t *md);extern int efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md); +extern int __efi_mem_desc_lookup (u64 phys_addr,efi_memory_desc_t *out_md);extern void efi_mem_reserve (phys_addr_t addr,u64 size);extern int efi_mem_reserve_persistent (phys_addr_t addr,u64 size); extern void efi_initialize_iomem_resources(struct resource *code_resource,
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
PATCH v3 3/5 efi: Apply allowlist to EFI configuration tables when running under Xen
In reply to Demi Marie ObenourAs it turns out, Xen does not guarantee that EFI boot services data regions in memory are preserved, which means that EFI configuration tables pointing into such memory regions may be corrupted before the dom0 OS has had a chance to inspect them.
This is causing problems for Qubes OS when it attempts to perform system firmware updates, which requires that the contents of the EFI System Resource Table are valid when the fwupd userspace program runs.
However, other configuration tables such as the memory attributes table or the runtime properties table are equally affected, and so we need a comprehensive workaround that works for any table type.
So when running under Xen, check the EFI memory descriptor covering the start of the table, and disregard the table if it does not reside in memory that is preserved by Xen.
--- drivers/firmware/efi/efi.c | 7 +++++++ drivers/xen/efi.c | 25 +++++++++++++++++++++++++ include/linux/efi.h | 2 ++ 3 files changed, 34 insertions(+)
drivers/firmware/efi/efi.c +7 -0
@@ -564,6 +564,13 @@ static __init int match_config_table(const efi_guid_t *guid,for (i =0 ;efi_guidcmp (table_types[i].guid,NULL_GUID ); i++) {if (!efi_guidcmp (*guid, table_types[i].guid)) { +if (IS_ENABLED (CONFIG_XEN_EFI ) && + !xen_efi_config_table_is_usable (guid, table)) { +if (table_types[i].name[0 ]) +pr_cont ("(%s=0x%lx) may have been clobbered by Xen " , + table_types[i].name, table); +return 1 ; + } *(table_types[i].ptr) = table;if (table_types[i].name[0 ]) pr_cont("%s=0x%lx ",
drivers/xen/efi.c +25 -0
@@ -328,3 +328,28 @@ int efi_mem_desc_lookup(u64 phys_addr, efi_memory_desc_t *out_md)return 0 ; } + +bool __init xen_efi_config_table_is_usable (const efi_guid_t *guid, +unsigned long table) +{ +efi_memory_desc_t md; +int rc; + +if (!efi_enabled (EFI_PARAVIRT )) +return true; + + rc =efi_mem_desc_lookup (table, &md); +if (rc) +return false; + +switch (md.type) { +case EFI_RUNTIME_SERVICES_CODE : +case EFI_RUNTIME_SERVICES_DATA : +case EFI_ACPI_RECLAIM_MEMORY : +case EFI_ACPI_MEMORY_NVS : +case EFI_RESERVED_TYPE : +return true; +default : +return false; + } +}
include/linux/efi.h +2 -0
@@ -1322,4 +1322,6 @@ struct linux_efi_coco_secret_area {/* Header of a populated EFI secret area */ #define EFI_SECRET_TABLE_HEADER_GUID EFI_GUID(0x1e74f542, 0x71dd, 0x4d66, 0x96, 0x3e, 0xef, 0x42, 0x87, 0xff, 0x17, 0x3b) +bool xen_efi_config_table_is_usable (const efi_guid_t *guid,unsigned long table); +#endif /* _LINUX_EFI_H */
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
PATCH v3 4/5 efi: Actually enable the ESRT under Xen
In reply to Demi Marie ObenourThe ESRT can be parsed if EFI_PARAVIRT is enabled, even if EFI_MEMMAP is not. Also allow the ESRT to be in reclaimable memory, as that is where future Xen versions will put it.
--- drivers/firmware/efi/esrt.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/firmware/efi/esrt.c +4 -2
@@ -247,7 +247,7 @@ void __init efi_esrt_init(void)int rc;phys_addr_t end; -if (!efi_enabled (EFI_MEMMAP )) +if (!efi_enabled (EFI_MEMMAP ) && !efi_enabled (EFI_PARAVIRT ))return ;pr_debug ("esrt-init: loading.\n" );
@@ -258,7 +258,9 @@ void __init efi_esrt_init(void)if (rc <0 || (!(md.attribute &EFI_MEMORY_RUNTIME ) && md.type !=EFI_BOOT_SERVICES_DATA && - md.type !=EFI_RUNTIME_SERVICES_DATA )) { + md.type !=EFI_RUNTIME_SERVICES_DATA && + md.type !=EFI_ACPI_RECLAIM_MEMORY && + md.type !=EFI_ACPI_MEMORY_NVS )) {pr_warn ("ESRT header is not in the memory map.\n" );return ; }
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
PATCH v3 5/5 efi: Warn if trying to reserve memory under Xen
In reply to Demi Marie ObenourDoing so cannot work and should never happen.
--- drivers/firmware/efi/efi.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/firmware/efi/efi.c +4 -0
@@ -519,6 +519,10 @@ void __init __weak efi_arch_mem_reserve(phys_addr_t addr, u64 size) {}*/ void __init efi_mem_reserve (phys_addr_t addr,u64 size) { +/* efi_mem_reserve() does not work under Xen */ +if (WARN_ON_ONCE (efi_enabled (EFI_PARAVIRT ))) +return ; +if (!memblock_is_region_reserved (addr, size))memblock_reserve (addr, size);
Signature · 3 lines
Sincerely, Demi Marie Obenour (she/her/hers) Invisible Things Lab
Ard Biesheuvel
maintainerRe: [PATCH v3 0/5] efi: Support ESRT under Xen
In reply to Demi Marie ObenourQuoted 5 lines · Demi Marie Obenour
This patch series fixes handling of EFI tables when running under Xen. These fixes allow the ESRT to be loaded when running paravirtualized in dom0, making the use of EFI capsule updates possible.
Quoted 8 lines
Demi Marie Obenour (5):
efi: memmap: Disregard bogus entries instead of returning them
efi: xen: Implement memory descriptor lookup based on hypercall
efi: Apply allowlist to EFI configuration tables when running under
Xen
efi: Actually enable the ESRT under Xen
efi: Warn if trying to reserve memory under Xen
I have given these a spin on a system with a dodgy ESRT (the region in question is not covered by the memory map at all), and things are exactly as broken before, which is good.
I have queued these up in efi/next now, they should appear in -next tomorrow.
Quoted 10 lines
drivers/firmware/efi/efi.c | 22 ++++++++++++- drivers/firmware/efi/esrt.c | 15 +++------ drivers/xen/efi.c | 61 +++++++++++++++++++++++++++++++++++++ include/linux/efi.h | 3 ++ 4 files changed, 90 insertions(+), 11 deletions(-)--
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab