Bug#1122106: linux-image-amd64: UBSAN create a systematic trace in DRM driver
Control: tags -1 + moreinfo
Hi,
On Sun, Dec 07, 2025 at 12:00:51PM +0100, Eric Valette wrote:
> Package: linux-image-amd64
> Version: 6.17.10-1
> Severity: normal
>
> It is a very old miniPC I use for selfhosted Web site but still.
> It runs unsable /07/12/2025...
>
> radeon 0000:00:01.0: vgaarb: deactivate vga console
> [ 25.183441] [drm] initializing kernel modesetting (PALM 0x1002:0x9804 0x1022:0x1511 0x00).
> [ 25.183775] ATOM BIOS: AMD
> [ 25.183905] radeon 0000:00:01.0: VRAM: 384M 0x0000000000000000 - 0x0000000017FFFFFF (384M used)
> [ 25.183924] radeon 0000:00:01.0: GTT: 1024M 0x0000000018000000 - 0x0000000057FFFFFF
> [ 25.183945] [drm] Detected VRAM RAM=384M, BAR=256M
> [ 25.183952] [drm] RAM width 32bits DDR
> [ 25.192896] [drm] radeon: 384M of VRAM memory ready
> [ 25.192916] [drm] radeon: 1024M of GTT memory ready.
> [ 25.193009] [drm] Loading PALM Microcode
> [ 25.212937] [drm] Internal thermal controller without fan control
> Internal thermal controller without fan control
> [ 25.212966] ------------[ cut here ]------------
> [ 25.212971] UBSAN: array-index-out-of-bounds in /build/reproducible-path/linux-6.17.10/drivers/gpu/drm/radeon/radeon_atombios.c:2720:34
> [ 25.213005] index 16 is out of range for type 'UCHAR [*]'
> [ 25.213025] CPU: 1 UID: 0 PID: 347 Comm: (udev-worker) Not tainted 6.17.10+deb14-amd64 #1 PREEMPT(lazy) Debian 6.17.10-1
> [ 25.213038] Hardware name: /Inagua CRB, BIOS 4.6.4 12/05/2011
> [ 25.213044] Call Trace:
> [ 25.213053] <TASK>
> [ 25.213063] dump_stack_lvl+0x5d/0x80
> [ 25.213084] ubsan_epilogue+0x5/0x2b
> [ 25.213099] __ubsan_handle_out_of_bounds.cold+0x54/0x59
> [ 25.213121] radeon_atombios_get_power_modes+0x8d6/0x900 [radeon]
> [ 25.213494] radeon_pm_init+0x142/0x760 [radeon]
> [ 25.213834] evergreen_init+0x154/0x2b0 [radeon]
> [ 25.214177] radeon_device_init+0x563/0xba0 [radeon]
> [ 25.214455] ? pci_find_capability+0x79/0xb0
> [ 25.214470] radeon_driver_load_kms+0xa1/0x260 [radeon]
> [ 25.214751] radeon_pci_probe+0x177/0x230 [radeon]
> [ 25.215028] local_pci_probe+0x42/0x90
> [ 25.215044] pci_device_probe+0xda/0x2b0
> [ 25.215057] ? sysfs_do_create_link_sd+0x6d/0xd0
> [ 25.215074] really_probe+0xde/0x340
> [ 25.215086] ? pm_runtime_barrier+0x55/0x90
> [ 25.215100] __driver_probe_device+0x78/0x140
> [ 25.215112] driver_probe_device+0x1f/0xa0
> [ 25.215123] ? __pfx___driver_attach+0x10/0x10
> [ 25.215133] __driver_attach+0xcb/0x1e0
> [ 25.215145] bus_for_each_dev+0x85/0xd0
> [ 25.215156] bus_add_driver+0x10b/0x1f0
> [ 25.215170] ? __pfx_radeon_module_init+0x10/0x10 [radeon]
> [ 25.215443] driver_register+0x75/0xe0
> [ 25.215456] ? radeon_register_atpx_handler+0xe/0x30 [radeon]
> [ 25.215752] do_one_initcall+0x5b/0x300
> [ 25.215766] do_init_module+0x62/0x250
> [ 25.215779] ? init_module_from_file+0x8a/0xe0
> [ 25.215791] init_module_from_file+0x8a/0xe0
> [ 25.215809] idempotent_init_module+0x114/0x310
> [ 25.215825] __x64_sys_finit_module+0x6d/0xd0
> [ 25.215836] ? syscall_trace_enter+0x8d/0x1d0
> [ 25.215845] do_syscall_64+0x82/0x320
> [ 25.215859] ? vfs_read+0x165/0x390
> [ 25.215872] ? vfs_read+0x165/0x390
> [ 25.215882] ? __rseq_handle_notify_resume+0xa6/0x4c0
> [ 25.215897] ? switch_fpu_return+0x5b/0xe0
> [ 25.215907] ? do_syscall_64+0x200/0x320
> [ 25.215917] ? vfs_fstatat+0x75/0xa0
> [ 25.215927] ? __do_sys_newfstatat+0x3b/0x80
> [ 25.215944] ? do_syscall_64+0xbb/0x320
> [ 25.215953] ? __x64_sys_openat+0x61/0xa0
> [ 25.215963] ? do_syscall_64+0xbb/0x320
> [ 25.215971] ? do_syscall_64+0xbb/0x320
> [ 25.215979] ? exc_page_fault+0x74/0x180
> [ 25.215988] entry_SYSCALL_64_after_hwframe+0x76/0x7e
> [ 25.215999] RIP: 0033:0x7f0ee2f0f219
> [ 25.216040] Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d c7 5b 10 00 f7 d8 64 89 01 48
> [ 25.216049] RSP: 002b:00007ffc54f73758 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
> [ 25.216061] RAX: ffffffffffffffda RBX: 000056319126a170 RCX: 00007f0ee2f0f219
> [ 25.216068] RDX: 0000000000000004 RSI: 000056319126b5b0 RDI: 0000000000000044
> [ 25.216073] RBP: 0000000000000004 R08: 0000000000000000 R09: 0000563190fa73e0
> [ 25.216079] R10: 0000000000000000 R11: 0000000000000246 R12: 000056319126b5b0
> [ 25.216084] R13: 0000000000020000 R14: 00005631912654d0 R15: 0000000000000000
> [ 25.216095] </TASK>
> [ 25.222786] ---[ end trace ]---
> [ 25.223016] [drm] Found smc ucode version: 0x00010601
> [ 25.223142] [drm] radeon: dpm initialized
> [ 25.247831] [drm] GART: num cpu pages 262144, num gpu pages 262144
> [ 25.322212] [drm] PCIE GART of 1024M enabled (table at 0x0000000000162000).
> [ 25.322514] radeon 0000:00:01.0: WB enabled
> [ 25.322531] radeon 0000:00:01.0: fence driver on ring 0 uses gpu addr 0x0000000018000c00
> [ 25.322542] radeon 0000:00:01.0: fence driver on ring 3 uses gpu addr 0x0000000018000c0c
> [ 25.323122] radeon 0000:00:01.0: fence driver on ring 5 uses gpu addr 0x0000000000072118
> [ 25.345406] radeon 0000:00:01.0: radeon: MSI limited to 32-bit
> [ 25.345630] radeon 0000:00:01.0: radeon: using MSI.
> [ 25.345714] [drm] radeon: irq initialized.
> [ 25.385815] [drm] ring test on 0 succeeded in 1 usecs
> [ 25.385847] [drm] ring test on 3 succeeded in 2 usecs
> [ 25.495066] [drm] ring test on 5 succeeded in 1 usecs
> [ 25.528183] [drm] UVD initialized successfully.
> [ 25.529107] snd_hda_intel 0000:00:01.1: bound 0000:00:01.0 (ops radeon_audio_component_bind_ops [radeon])
> [ 25.529694] [drm] ib test on ring 0 succeeded in 0 usecs
> [ 25.529760] [drm] ib test on ring 3 succeeded in 0 usecs
> [ 25.551633] [drm] ib test on ring 5 succeeded
> [ 25.590120] [drm] Radeon Display Connectors
> [ 25.590140] [drm] Connector 0:
There is an upstream report which seems to match:
https://gitlab.freedesktop.org/drm/amd/-/issues/4374
Alex Deucher proposed there two changes, can you test those please and
report back if they work? If so can you please directly confirm in the
upstream issue which fixes it and report back here as well?
Regards,
Salvatore
Reply to: