linux-kernel - Re: [PATCH 1/4] drm/ttm: Fix ttm_lru_bulk_move_pos

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

Message-ID: <b9eb015b-6ba4-f980-17f1-db4921cefccc@amd.com>
Date:   Thu, 22 Jun 2023 14:26:15 +0200
From:   Christian König <christian.koenig@....com>
To:     Thomas Hellström 
        <thomas.hellstrom@...ux.intel.com>, intel-xe@...ts.freedesktop.org
Cc:     Christian König <ckoenig.leichtzumerken@...il.com>,
        Daniel Vetter <daniel.vetter@...ll.ch>,
        dri-devel@...ts.freedesktop.org, stable@...r.kernel.org,
        intel-gfx@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/4] drm/ttm: Fix ttm_lru_bulk_move_pos_tail()

Teddy already stumbled over this as well, but this patch here looks better.

Going to review and push it to drm-misc-fixes asap.

Thanks,
Christian.

Am 22.06.23 um 12:14 schrieb Thomas Hellström:
> The value of pos->first was not updated when the first resource of the
> range was moved. This could lead to errors like the one below.
> Fix this by updating pos->first when needed.
>
> <3> [218.963342] BUG: KASAN: null-ptr-deref in ttm_lru_bulk_move_del+0xc5/0x180 [ttm]
> <3> [218.963456] Read of size 8 at addr 0000000000000038 by task xe_evict/1529
> <3> [218.963546]
> <3> [218.963566] CPU: 0 PID: 1529 Comm: xe_evict Not tainted 6.3.0-xe #1
> <3> [218.963664] Hardware name: Intel Corporation Tiger Lake Client Platform/TigerLake H DDR4 SODIMM RVP, BIOS TGLSFWI1.R00.4064.A00.2102041619 02/04/2021
> <3> [218.963841] Call Trace:
> <3> [218.963881]  <TASK>
> <3> [218.963915]  dump_stack_lvl+0x64/0xb0
> <3> [218.963976]  print_report+0x3e5/0x600
> <3> [218.964036]  ? ttm_lru_bulk_move_del+0xc5/0x180 [ttm]
> <3> [218.964127]  kasan_report+0x96/0xc0
> <3> [218.964183]  ? ttm_lru_bulk_move_del+0xc5/0x180 [ttm]
> <3> [218.964276]  ttm_lru_bulk_move_del+0xc5/0x180 [ttm]
> <3> [218.964365]  ttm_bo_set_bulk_move+0x92/0x140 [ttm]
> <3> [218.964454]  xe_gem_object_close+0xc8/0x120 [xe]
> <3> [218.964675]  ? __pfx_xe_gem_object_close+0x10/0x10 [xe]
> <3> [218.964908]  ? drm_gem_object_handle_put_unlocked+0xc7/0x170 [drm]
> <3> [218.965071]  drm_gem_object_release_handle+0x45/0x80 [drm]
> <3> [218.965220]  ? __pfx_drm_gem_object_release_handle+0x10/0x10 [drm]
> <3> [218.965381]  idr_for_each+0xc9/0x180
> <3> [218.965437]  ? __pfx_idr_for_each+0x10/0x10
> <3> [218.965504]  drm_gem_release+0x20/0x30 [drm]
> <3> [218.965637]  drm_file_free.part.0+0x4cb/0x4f0 [drm]
> <3> [218.965778]  ? drm_close_helper.isra.0+0xb7/0xe0 [drm]
> <3> [218.965921]  drm_release_noglobal+0x49/0x90 [drm]
> <3> [218.966061]  __fput+0x122/0x450
> <3> [218.966115]  task_work_run+0xfe/0x190
> <3> [218.966175]  ? __pfx_task_work_run+0x10/0x10
> <3> [218.966239]  ? do_raw_spin_unlock+0xa7/0x140
> <3> [218.966308]  do_exit+0x55f/0x1430
> <3> [218.966364]  ? __pfx_lock_release+0x10/0x10
> <3> [218.966431]  ? do_raw_spin_lock+0x11d/0x1e0
> <3> [218.966498]  ? __pfx_do_exit+0x10/0x10
> <3> [218.966554]  ? __pfx_do_raw_spin_lock+0x10/0x10
> <3> [218.966625]  ? mark_held_locks+0x24/0x90
> <3> [218.966688]  ? lockdep_hardirqs_on_prepare+0x136/0x210
> <3> [218.966768]  do_group_exit+0x68/0x110
> <3> [218.966828]  __x64_sys_exit_group+0x2c/0x30
> <3> [218.966896]  do_syscall_64+0x3c/0x90
> <3> [218.966955]  entry_SYSCALL_64_after_hwframe+0x72/0xdc
> <3> [218.967035] RIP: 0033:0x7f77b194f146
> <3> [218.967094] Code: Unable to access opcode bytes at 0x7f77b194f11c.
> <3> [218.967174] RSP: 002b:00007ffc64791188 EFLAGS: 00000246 ORIG_RAX: 00000000000000e7
> <3> [218.967271] RAX: ffffffffffffffda RBX: 00007f77b1a548a0 RCX: 00007f77b194f146
> <3> [218.967364] RDX: 0000000000000062 RSI: 000000000000003c RDI: 0000000000000062
> <3> [218.967458] RBP: 0000000000000062 R08: 00000000000000e7 R09: ffffffffffffff78
> <3> [218.967553] R10: 0000000000000058 R11: 0000000000000246 R12: 00007f77b1a548a0
> <3> [218.967648] R13: 0000000000000003 R14: 00007f77b1a5d2e8 R15: 0000000000000000
> <3> [218.967745]  </TASK>
>
> Fixes: fee2ede15542 ("drm/ttm: rework bulk move handling v5")
> Cc: "Christian König" <christian.koenig@....com>
> Cc: "Christian König" <ckoenig.leichtzumerken@...il.com>
> Cc: Daniel Vetter <daniel.vetter@...ll.ch>
> Cc: dri-devel@...ts.freedesktop.org
> Cc: <stable@...r.kernel.org> # v5.19+
> Link: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/411
> Signed-off-by: Thomas Hellström <thomas.hellstrom@...ux.intel.com>
> ---
>   drivers/gpu/drm/ttm/ttm_resource.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
> index 7333f7a87a2f..cb05e0a36576 100644
> --- a/drivers/gpu/drm/ttm/ttm_resource.c
> +++ b/drivers/gpu/drm/ttm/ttm_resource.c
> @@ -86,6 +86,8 @@ static void ttm_lru_bulk_move_pos_tail(struct ttm_lru_bulk_move_pos *pos,
>   				       struct ttm_resource *res)
>   {
>   	if (pos->last != res) {
> +		if (pos->first == res)
> +			pos->first = list_next_entry(res, lru);
>   		list_move(&res->lru, &pos->last->lru);
>   		pos->last = res;
>   	}