Ard Biesheuvel f75a91c82d arm64: efi: Execute runtime services from a dedicated stack
commit ff7a167961d1b97e0e205f245f806e564d3505e7 upstream.

With the introduction of PRMT in the ACPI subsystem, the EFI rts
workqueue is no longer the only caller of efi_call_virt_pointer() in the
kernel. This means the EFI runtime services lock is no longer sufficient
to manage concurrent calls into firmware, but also that firmware calls
may occur that are not marshalled via the workqueue mechanism, but
originate directly from the caller context.

For added robustness, and to ensure that the runtime services have 8 KiB
of stack space available as per the EFI spec, introduce a spinlock
protected EFI runtime stack of 8 KiB, where the spinlock also ensures
serialization between the EFI rts workqueue (which itself serializes EFI
runtime calls) and other callers of efi_call_virt_pointer().

While at it, use the stack pivot to avoid reloading the shadow call
stack pointer from the ordinary stack, as doing so could produce a
gadget to defeat it.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Cc: Lee Jones <lee@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2023-01-24 07:24:41 +01:00

174 lines
4.8 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
/*
* Extensible Firmware Interface
*
* Based on Extensible Firmware Interface Specification version 2.4
*
* Copyright (C) 2013, 2014 Linaro Ltd.
*/
#include <linux/efi.h>
#include <linux/init.h>
#include <asm/efi.h>
static bool region_is_misaligned(const efi_memory_desc_t *md)
{
if (PAGE_SIZE == EFI_PAGE_SIZE)
return false;
return !PAGE_ALIGNED(md->phys_addr) ||
!PAGE_ALIGNED(md->num_pages << EFI_PAGE_SHIFT);
}
/*
* Only regions of type EFI_RUNTIME_SERVICES_CODE need to be
* executable, everything else can be mapped with the XN bits
* set. Also take the new (optional) RO/XP bits into account.
*/
static __init pteval_t create_mapping_protection(efi_memory_desc_t *md)
{
u64 attr = md->attribute;
u32 type = md->type;
if (type == EFI_MEMORY_MAPPED_IO)
return PROT_DEVICE_nGnRE;
if (region_is_misaligned(md)) {
static bool __initdata code_is_misaligned;
/*
* Regions that are not aligned to the OS page size cannot be
* mapped with strict permissions, as those might interfere
* with the permissions that are needed by the adjacent
* region's mapping. However, if we haven't encountered any
* misaligned runtime code regions so far, we can safely use
* non-executable permissions for non-code regions.
*/
code_is_misaligned |= (type == EFI_RUNTIME_SERVICES_CODE);
return code_is_misaligned ? pgprot_val(PAGE_KERNEL_EXEC)
: pgprot_val(PAGE_KERNEL);
}
/* R-- */
if ((attr & (EFI_MEMORY_XP | EFI_MEMORY_RO)) ==
(EFI_MEMORY_XP | EFI_MEMORY_RO))
return pgprot_val(PAGE_KERNEL_RO);
/* R-X */
if (attr & EFI_MEMORY_RO)
return pgprot_val(PAGE_KERNEL_ROX);
/* RW- */
if (((attr & (EFI_MEMORY_RP | EFI_MEMORY_WP | EFI_MEMORY_XP)) ==
EFI_MEMORY_XP) ||
type != EFI_RUNTIME_SERVICES_CODE)
return pgprot_val(PAGE_KERNEL);
/* RWX */
return pgprot_val(PAGE_KERNEL_EXEC);
}
/* we will fill this structure from the stub, so don't put it in .bss */
struct screen_info screen_info __section(".data");
EXPORT_SYMBOL(screen_info);
int __init efi_create_mapping(struct mm_struct *mm, efi_memory_desc_t *md)
{
pteval_t prot_val = create_mapping_protection(md);
bool page_mappings_only = (md->type == EFI_RUNTIME_SERVICES_CODE ||
md->type == EFI_RUNTIME_SERVICES_DATA);
/*
* If this region is not aligned to the page size used by the OS, the
* mapping will be rounded outwards, and may end up sharing a page
* frame with an adjacent runtime memory region. Given that the page
* table descriptor covering the shared page will be rewritten when the
* adjacent region gets mapped, we must avoid block mappings here so we
* don't have to worry about splitting them when that happens.
*/
if (region_is_misaligned(md))
page_mappings_only = true;
create_pgd_mapping(mm, md->phys_addr, md->virt_addr,
md->num_pages << EFI_PAGE_SHIFT,
__pgprot(prot_val | PTE_NG), page_mappings_only);
return 0;
}
static int __init set_permissions(pte_t *ptep, unsigned long addr, void *data)
{
efi_memory_desc_t *md = data;
pte_t pte = READ_ONCE(*ptep);
if (md->attribute & EFI_MEMORY_RO)
pte = set_pte_bit(pte, __pgprot(PTE_RDONLY));
if (md->attribute & EFI_MEMORY_XP)
pte = set_pte_bit(pte, __pgprot(PTE_PXN));
set_pte(ptep, pte);
return 0;
}
int __init efi_set_mapping_permissions(struct mm_struct *mm,
efi_memory_desc_t *md)
{
BUG_ON(md->type != EFI_RUNTIME_SERVICES_CODE &&
md->type != EFI_RUNTIME_SERVICES_DATA);
if (region_is_misaligned(md))
return 0;
/*
* Calling apply_to_page_range() is only safe on regions that are
* guaranteed to be mapped down to pages. Since we are only called
* for regions that have been mapped using efi_create_mapping() above
* (and this is checked by the generic Memory Attributes table parsing
* routines), there is no need to check that again here.
*/
return apply_to_page_range(mm, md->virt_addr,
md->num_pages << EFI_PAGE_SHIFT,
set_permissions, md);
}
/*
* UpdateCapsule() depends on the system being shutdown via
* ResetSystem().
*/
bool efi_poweroff_required(void)
{
return efi_enabled(EFI_RUNTIME_SERVICES);
}
asmlinkage efi_status_t efi_handle_corrupted_x18(efi_status_t s, const char *f)
{
pr_err_ratelimited(FW_BUG "register x18 corrupted by EFI %s\n", f);
return s;
}
DEFINE_SPINLOCK(efi_rt_lock);
asmlinkage u64 *efi_rt_stack_top __ro_after_init;
/* EFI requires 8 KiB of stack space for runtime services */
static_assert(THREAD_SIZE >= SZ_8K);
static int __init arm64_efi_rt_init(void)
{
void *p;
if (!efi_enabled(EFI_RUNTIME_SERVICES))
return 0;
p = __vmalloc_node(THREAD_SIZE, THREAD_ALIGN, GFP_KERNEL,
NUMA_NO_NODE, &&l);
l: if (!p) {
pr_warn("Failed to allocate EFI runtime stack\n");
clear_bit(EFI_RUNTIME_SERVICES, &efi.flags);
return -ENOMEM;
}
efi_rt_stack_top = p + THREAD_SIZE;
return 0;
}
core_initcall(arm64_efi_rt_init);