linux-kernel - AW: [PATCH] drm/amdgpu: fix potential integer overflow on shift of a int

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <b9d63b56e0f849f1a5a2def73c899047@bfs.de>
Date:   Mon, 8 Feb 2021 15:55:16 +0000
From:   Walter Harms <wharms@....de>
To:     Christian König <christian.koenig@....com>,
        Colin King <colin.king@...onical.com>,
        Alex Deucher <alexander.deucher@....com>,
        "David Airlie" <airlied@...ux.ie>, Daniel Vetter <daniel@...ll.ch>,
        Huang Rui <ray.huang@....com>,
        Junwei Zhang <Jerry.Zhang@....com>,
        "amd-gfx@...ts.freedesktop.org" <amd-gfx@...ts.freedesktop.org>,
        "dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>
CC:     "kernel-janitors@...r.kernel.org" <kernel-janitors@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: AW: [PATCH] drm/amdgpu: fix potential integer overflow on shift of a
 int

thx for info
________________________________________
Von: Christian König <christian.koenig@....com>
Gesendet: Montag, 8. Februar 2021 13:14:49
An: Walter Harms; Colin King; Alex Deucher; David Airlie; Daniel Vetter; Huang Rui; Junwei Zhang; amd-gfx@...ts.freedesktop.org; dri-devel@...ts.freedesktop.org
Cc: kernel-janitors@...r.kernel.org; linux-kernel@...r.kernel.org
Betreff: Re: [PATCH] drm/amdgpu: fix potential integer overflow on shift of a int

For start and end?  The hardware has 48 bit address space and that won't
fit into 32bits.

Only the fragment handling can't do more than 2GB at the same time.

Christian.

Am 08.02.21 um 12:05 schrieb Walter Harms:
> i am curious:
> what is the win to have a unsigned 64 bit integer in the first
> place ?
>
> re,
>   wh
> ________________________________________
> Von: Christian König <christian.koenig@....com>
> Gesendet: Montag, 8. Februar 2021 10:17:42
> An: Colin King; Alex Deucher; David Airlie; Daniel Vetter; Huang Rui; Junwei Zhang; amd-gfx@...ts.freedesktop.org; dri-devel@...ts.freedesktop.org
> Cc: kernel-janitors@...r.kernel.org; linux-kernel@...r.kernel.org
> Betreff: Re: [PATCH] drm/amdgpu: fix potential integer overflow on shift of a int
>
> Am 08.02.21 um 00:07 schrieb Colin King:
>> From: Colin Ian King <colin.king@...onical.com>
>>
>> The left shift of int 32 bit integer constant 1 is evaluated using 32
>> bit arithmetic and then assigned to an unsigned 64 bit integer. In the
>> case where *frag is 32 or more this can lead to an oveflow.  Avoid this
>> by shifting 1ULL.
> Well that can't happen. Take a look at the code in that function:
>
>>                  max_frag = 31;
> ...
>>          if (*frag >= max_frag) {
>>                  *frag = max_frag;
>>                  *frag_end = end & ~((1ULL << max_frag) - 1);
>>          } else {
>>                  *frag_end = start + (1 << *frag);
>>          }
> But I'm fine with applying the patch if it silences your warning.
>
> Regards,
> Christian.
>
>> Addresses-Coverity: ("Unintentional integer overflow")
>> Fixes: dfcd99f6273e ("drm/amdgpu: meld together VM fragment and huge page handling")
>> Signed-off-by: Colin Ian King <colin.king@...onical.com>
>> ---
>>    drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 2 +-
>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>> index 9d19078246c8..53a925600510 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>> @@ -1412,7 +1412,7 @@ static void amdgpu_vm_fragment(struct amdgpu_vm_update_params *params,
>>                *frag = max_frag;
>>                *frag_end = end & ~((1ULL << max_frag) - 1);
>>        } else {
>> -             *frag_end = start + (1 << *frag);
>> +             *frag_end = start + (1ULL << *frag);
>>        }
>>    }
>>