Skip to content

Commit

Permalink
arm64: kernel: restrict /dev/mem read() calls to linear region
Browse files Browse the repository at this point in the history
When running lscpu on an AArch64 system that has SMBIOS version 2.0
tables, it will segfault in the following way:

  Unable to handle kernel paging request at virtual address ffff8000bfff0000
  pgd = ffff8000f9615000
  [ffff8000bfff0000] *pgd=0000000000000000
  Internal error: Oops: 96000007 [#1] PREEMPT SMP
  Modules linked in:
  CPU: 0 PID: 1284 Comm: lscpu Not tainted 4.11.0-rc3+ torvalds#103
  Hardware name: QEMU QEMU Virtual Machine, BIOS 0.0.0 02/06/2015
  task: ffff8000fa78e800 task.stack: ffff8000f9780000
  PC is at __arch_copy_to_user+0x90/0x220
  LR is at read_mem+0xcc/0x140

This is caused by the fact that lspci issues a read() on /dev/mem at the
offset where it expects to find the SMBIOS structure array. However, this
region is classified as EFI_RUNTIME_SERVICE_DATA (as per the UEFI spec),
and so it is omitted from the linear mapping.

So let's restrict /dev/mem read/write access to those areas that are
covered by the linear region.

Reported-by: Alexander Graf <agraf@suse.de>
Fixes: 4dffbfc ("arm64/efi: mark UEFI reserved regions as MEMBLOCK_NOMAP")
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
(cherry picked from commit 1151f83)
  • Loading branch information
Ard Biesheuvel authored and jchandra-cavm committed Jul 13, 2017
1 parent 6f7da29 commit 6871924
Showing 1 changed file with 13 additions and 6 deletions.
19 changes: 13 additions & 6 deletions arch/arm64/mm/mmap.c
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@

#include <linux/elf.h>
#include <linux/fs.h>
#include <linux/memblock.h>
#include <linux/mm.h>
#include <linux/mman.h>
#include <linux/export.h>
Expand Down Expand Up @@ -103,12 +104,18 @@ void arch_pick_mmap_layout(struct mm_struct *mm)
*/
int valid_phys_addr_range(phys_addr_t addr, size_t size)
{
if (addr < PHYS_OFFSET)
return 0;
if (addr + size > __pa(high_memory - 1) + 1)
return 0;

return 1;
/*
* Check whether addr is covered by a memory region without the
* MEMBLOCK_NOMAP attribute, and whether that region covers the
* entire range. In theory, this could lead to false negatives
* if the range is covered by distinct but adjacent memory regions
* that only differ in other attributes. However, few of such
* attributes have been defined, and it is debatable whether it
* follows that /dev/mem read() calls should be able traverse
* such boundaries.
*/
return memblock_is_region_memory(addr, size) &&
memblock_is_map_memory(addr);
}

/*
Expand Down

0 comments on commit 6871924

Please sign in to comment.