Re: [PATCH] mm/pagewalk.c: Fix walk_page_range access wrong PTEs

[Naoya Horiguchi]

Hello, 

On Fri, Oct 18, 2013 at 04:15:17PM +0800, linx.z.chen@xxxxxxxxx wrote:
> From: Chen LinX <linx.z.chen@xxxxxxxxx>
> 
> When walk_page_range walk a memory map's page tables, it'll skip VM_PFNMAP area,
> then variable 'next' will to assign to vma->vm_end, it maybe larger than 'end'.
> In next loop, 'addr' will be larger than 'next'. Then in /proc/XXXX/pagemap file
> reading procedure, the 'addr' will growing forever in pagemap_pte_range,
> pte_to_pagemap_entry will access wrong pte.
> 
> [   93.387934] BUG: Bad page map in process procrank  pte:8437526f pmd:785de067
> [   93.387936] addr:9108d000 vm_flags:00200073 anon_vma:f0d99020 mapping:  (null) index:9108d
> [   93.387938] CPU: 1 PID: 4974 Comm: procrank Tainted: G    B   W  O 3.10.1+ #1
> [   93.387942]  f0d983c8 f0d983c8 e8271e48 c281ef4b e8271e84 c20ee794 c2a3f150 9108d000
> [   93.387946]  00200073 f0d99020 00000000 0009108d e8342d90 8437526f 0009108d 00000000
> [   93.387950]  00000000 00084375 fffa4234 e8271e98 c20ef836 00000000 e8271f28 f0d983c8
> [   93.387951] Call Trace:
> [   93.387953]  [<c281ef4b>] dump_stack+0x16/0x18
> [   93.387956]  [<c20ee794>] print_bad_pte+0x114/0x1b0
> [   93.387959]  [<c20ef836>] vm_normal_page+0x56/0x60
> [   93.387961]  [<c21594ba>] pagemap_pte_range+0x17a/0x1d0
> [   93.387963]  [<c2159340>] ? m_next+0x70/0x70
> [   93.387966]  [<c20fc60e>] walk_page_range+0x19e/0x2c0
> [   93.387969]  [<c20432c4>] ? ptrace_may_access+0x24/0x40
> [   93.387971]  [<c2159a4e>] pagemap_read+0x16e/0x200
> [   93.387973]  [<c2159340>] ? m_next+0x70/0x70
> [   93.387976]  [<c2158be0>] ? quota_send_warning+0x1f0/0x1f0
> [   93.387979]  [<c2158c30>] ? pagemap_pte_hole+0x50/0x50
> [   93.387981]  [<c21598e0>] ? clear_refs_pte_range+0xc0/0xc0
> [   93.387984]  [<c210c534>] vfs_read+0x84/0x150
> [   93.387986]  [<c21598e0>] ? clear_refs_pte_range+0xc0/0xc0
> [   93.387988]  [<c210c77a>] SyS_read+0x4a/0x80
> [   93.387991]  [<c28267c8>] syscall_call+0x7/0xb
> [   93.387996]  [<c2820000>] ? e1000_regdump+0x6f/0x37c
> [  102.866190]  [<c259fcd3>] ? cpuidle_idle_call+0x93/0x1b0
> [  102.872124]  [<c28297fa>] ? atomic_notifier_call_chain+0x1a/0x20
> [  102.878834]  [<c2008478>] ? arch_cpu_idle+0x8/0x20
> [  102.884183]  [<c207950d>] ? cpu_startup_entry+0x4d/0x1b0
> [  102.890116]  [<c280ecbd>] ? rest_init+0x5d/0x60
> [  102.895174]  [<c2b989f8>] ? start_kernel+0x31b/0x321
> [  102.900717]  [<c2b98512>] ? repair_env_string+0x51/0x51
> [  102.906551]  [<c2b98356>] ? i386_start_kernel+0x12c/0x12f
> 
> Signed-off-by: Liu ShuoX <shuox.liu@xxxxxxxxx>
> Signed-off-by: Chen LinX <linx.z.chen@xxxxxxxxx>
> ---
>  mm/pagewalk.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/pagewalk.c b/mm/pagewalk.c
> index 5da2cbc..2beeabf 100644
> --- a/mm/pagewalk.c
> +++ b/mm/pagewalk.c
> @@ -242,7 +242,7 @@ int walk_page_range(unsigned long addr, unsigned long end,
>  		if (err)
>  			break;
>  		pgd++;
> -	} while (addr = next, addr != end);
> +	} while (addr = next, addr < end);
>  
>  	return err;
>  }

I agree to this change.

Reviewed-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>

BTW, walk_(pte|pmd|pud|hugetlb)_range also have the same condition
(addr != end), so I think that it's safer to change them too.

And I think that it's worth sending to -stable.
The VM_PFNMAP check is introduced around v3.10-rc2,
so 3.10.X and 3.11.X should want this fix.

Thanks,
Naoya Horiguchi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/