linux/drivers/gpu, branch v6.4

drm: use mgr->dev in drm_dbg_kms in drm_dp_add_payload_part2

2023-06-20T20:00:09Z

I've been experiencing some intermittent crashes down in the display driver code. The symptoms are ususally a line like this in dmesg: amdgpu 0000:30:00.0: [drm] Failed to create MST payload for port 000000006d3a3885: -5 ...followed by an Oops due to a NULL pointer dereference. Switch to using mgr->dev instead of state->dev since "state" can be NULL in some cases. Link: https://bugzilla.redhat.com/show_bug.cgi?id=2184855 Suggested-by: Jani Nikula Signed-off-by: Jeff Layton Reviewed-by: Jani Nikula Reviewed-by: Lyude Paul Signed-off-by: Lyude Paul Link: https://patchwork.freedesktop.org/patch/msgid/20230419112447.18471-1-jlayton@kernel.org

Merge tag 'drm-misc-fixes-2023-06-16' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

2023-06-16T20:11:59Z

drm-misc-fixes maybe in time for v6.4-rc7: - qaic leak and null deref fix. - Fix runtime pm in nouveau. - Fix array overflow in ti-sn65dsi86 pwm chip handling. - Assorted null check fixes in nouveau. Signed-off-by: Dave Airlie From: Maarten Lankhorst Link: https://patchwork.freedesktop.org/patch/msgid/641eb8a8-fbd7-90ad-0805-310b7fec9344@lankhorst.se

nouveau: fix client work fence deletion race

2023-06-16T01:19:34Z

This seems to have existed for ever but is now more apparant after commit 9bff18d13473 ("drm/ttm: use per BO cleanup workers") My analysis: two threads are running, one in the irq signalling the fence, in dma_fence_signal_timestamp_locked, it has done the DMA_FENCE_FLAG_SIGNALLED_BIT setting, but hasn't yet reached the callbacks. The second thread in nouveau_cli_work_ready, where it sees the fence is signalled, so then puts the fence, cleanups the object and frees the work item, which contains the callback. Thread one goes again and tries to call the callback and causes the use-after-free. Proposed fix: lock the fence signalled check in nouveau_cli_work_ready, so either the callbacks are done or the memory is freed. Reviewed-by: Karol Herbst Fixes: 11e451e74050 ("drm/nouveau: remove fence wait code from deferred client work handler") Cc: stable@vger.kernel.org Signed-off-by: Dave Airlie Link: https://lore.kernel.org/dri-devel/20230615024008.1600281-1-airlied@gmail.com/

drm/amd/display: limit DPIA link rate to HBR3

2023-06-13T21:16:35Z

[Why] DPIA doesn't support UHBR, driver should not enable UHBR for dp tunneling [How] limit DPIA link rate to HBR3 Cc: Mario Limonciello Cc: Alex Deucher Cc: stable@vger.kernel.org Acked-by: Stylon Wang Signed-off-by: Peichen Huang Reviewed-by: Mustapha Ghaddar Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher

drm/amd/display: fix the system hang while disable PSR

2023-06-13T21:14:52Z

[Why] When the PSR enabled. If you try to adjust the timing parameters, it may cause system hang. Because the timing mismatch with the DMCUB settings. [How] Disable the PSR before adjusting timing parameters. Cc: Mario Limonciello Cc: Alex Deucher Cc: stable@vger.kernel.org Acked-by: Stylon Wang Signed-off-by: Tom Chung Reviewed-by: Wayne Lin Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher

drm/amd/display: edp do not add non-edid timings

2023-06-13T21:14:09Z

[Why] most edp support only timings from edid. applying non-edid timings, especially those timings out of edp bandwidth, may damage edp. [How] do not add non-edid timings for edp. Cc: Mario Limonciello Cc: Alex Deucher Cc: stable@vger.kernel.org Acked-by: Stylon Wang Signed-off-by: Hersen Wu Reviewed-by: Roman Li Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher

Revert "drm/amdgpu: remove TOPDOWN flags when allocating VRAM in large bar system"

2023-06-13T21:12:41Z

This reverts commit c105518679b6e87232874ffc989ec403bee59664. This patch disables the TOPDOWN flag for APU and few dGPU cards which has the VRAM size equal to the BAR size. When we enable the TOPDOWN flag, we get the free blocks at the highest available memory region and we don't split the lower order blocks. This change is required to keep off the fragmentation related issues particularly in ASIC which has VRAM space <= 500MiB Hence, we are reverting this patch. Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2270 Signed-off-by: Arunpravin Paneer Selvam Reviewed-by: Christian König Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org

drm/amdgpu: vcn_4_0 set instance 0 init sched score to 1

2023-06-13T21:08:42Z

Only vcn0 can process AV1 codecx. In order to use both vcn0 and vcn1 in h264/265 transcode to AV1 cases, set vcn0 sched score to 1 at initialization time. Signed-off-by: Sonny Jiang Reviewed-by: Leo Liu Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.1.x

drm/radeon: Disable outputs when releasing fbdev client

2023-06-13T21:08:11Z

Disable the modesetting pipeline before release the radeon's fbdev client. Fixes the following error: [ 17.217408] WARNING: CPU: 5 PID: 1464 at drivers/gpu/drm/ttm/ttm_bo.c:326 ttm_bo_release+0x27e/0x2d0 [ttm] [ 17.217418] Modules linked in: edac_mce_amd radeon(+) drm_ttm_helper ttm video drm_suballoc_helper drm_display_helper kvm irqbypass drm_kms_helper syscopyarea crc32_pclmul sysfillrect sha512_ssse3 sysimgblt sha512_generic cfbfillrect cfbimgblt wmi_bmof aesni_intel cfbcopyarea crypto_simd cryptd k10temp acpi_cpufreq wmi dm_mod [ 17.217432] CPU: 5 PID: 1464 Comm: systemd-udevd Not tainted 6.4.0-rc4+ #1 [ 17.217436] Hardware name: Micro-Star International Co., Ltd. MS-7A38/B450M PRO-VDH MAX (MS-7A38), BIOS B.G0 07/26/2022 [ 17.217438] RIP: 0010:ttm_bo_release+0x27e/0x2d0 [ttm] [ 17.217444] Code: 48 89 43 38 48 89 43 40 48 8b 5c 24 30 48 8b b5 40 08 00 00 48 8b 6c 24 38 48 83 c4 58 e9 7a 49 f7 e0 48 89 ef e9 6c fe ff ff <0f> 0b 48 83 7b 20 00 0f 84 b7 fd ff ff 0f 0b 0f 1f 00 e9 ad fd ff [ 17.217448] RSP: 0018:ffffc9000095fbb0 EFLAGS: 00010202 [ 17.217451] RAX: 0000000000000001 RBX: ffff8881052c8de0 RCX: 0000000000000000 [ 17.217453] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff8881052c8de0 [ 17.217455] RBP: ffff888104a66e00 R08: ffff8881052c8de0 R09: ffff888104a7cf08 [ 17.217457] R10: ffffc9000095fbe0 R11: ffffc9000095fbe8 R12: ffff8881052c8c78 [ 17.217458] R13: ffff8881052c8c78 R14: dead000000000100 R15: ffff88810528b108 [ 17.217460] FS: 00007f319fcbb8c0(0000) GS:ffff88881a540000(0000) knlGS:0000000000000000 [ 17.217463] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 17.217464] CR2: 000055dc8b0224a0 CR3: 000000010373d000 CR4: 0000000000750ee0 [ 17.217466] PKRU: 55555554 [ 17.217468] Call Trace: [ 17.217470] [ 17.217472] ? __warn+0x97/0x160 [ 17.217476] ? ttm_bo_release+0x27e/0x2d0 [ttm] [ 17.217481] ? report_bug+0x1ec/0x200 [ 17.217487] ? handle_bug+0x3c/0x70 [ 17.217490] ? exc_invalid_op+0x1f/0x90 [ 17.217493] ? preempt_count_sub+0xb5/0x100 [ 17.217496] ? asm_exc_invalid_op+0x16/0x20 [ 17.217500] ? ttm_bo_release+0x27e/0x2d0 [ttm] [ 17.217505] ? ttm_resource_move_to_lru_tail+0x1ab/0x1d0 [ttm] [ 17.217511] radeon_bo_unref+0x1a/0x30 [radeon] [ 17.217547] radeon_gem_object_free+0x20/0x30 [radeon] [ 17.217579] radeon_fbdev_fb_destroy+0x57/0x90 [radeon] [ 17.217616] unregister_framebuffer+0x72/0x110 [ 17.217620] drm_client_dev_unregister+0x6d/0xe0 [ 17.217623] drm_dev_unregister+0x2e/0x90 [ 17.217626] drm_put_dev+0x26/0x90 [ 17.217628] pci_device_remove+0x44/0xc0 [ 17.217631] really_probe+0x257/0x340 [ 17.217635] __driver_probe_device+0x73/0x120 [ 17.217638] driver_probe_device+0x2c/0xb0 [ 17.217641] __driver_attach+0xa0/0x150 [ 17.217643] ? __pfx___driver_attach+0x10/0x10 [ 17.217646] bus_for_each_dev+0x67/0xa0 [ 17.217649] bus_add_driver+0x10e/0x210 [ 17.217651] driver_register+0x5c/0x120 [ 17.217653] ? __pfx_radeon_module_init+0x10/0x10 [radeon] [ 17.217681] do_one_initcall+0x44/0x220 [ 17.217684] ? kmalloc_trace+0x37/0xc0 [ 17.217688] do_init_module+0x64/0x240 [ 17.217691] __do_sys_finit_module+0xb2/0x100 [ 17.217694] do_syscall_64+0x3b/0x90 [ 17.217697] entry_SYSCALL_64_after_hwframe+0x72/0xdc [ 17.217700] RIP: 0033:0x7f319feaa5a9 [ 17.217702] Code: 08 89 e8 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 90 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 27 08 0d 00 f7 d8 64 89 01 48 [ 17.217706] RSP: 002b:00007ffc6bf3e7f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000139 [ 17.217709] RAX: ffffffffffffffda RBX: 00005607204f3170 RCX: 00007f319feaa5a9 [ 17.217710] RDX: 0000000000000000 RSI: 00007f31a002eefd RDI: 0000000000000018 [ 17.217712] RBP: 00007f31a002eefd R08: 0000000000000000 R09: 00005607204f1860 [ 17.217714] R10: 0000000000000018 R11: 0000000000000246 R12: 0000000000020000 [ 17.217716] R13: 0000000000000000 R14: 0000560720522450 R15: 0000560720255899 [ 17.217718] [ 17.217719] ---[ end trace 0000000000000000 ]--- The buffer object backing the fbdev emulation got pinned twice: by the fb_probe helper radeon_fbdev_create_pinned_object() and the modesetting code when the framebuffer got displayed. It only got unpinned once by the fbdev helper radeon_fbdev_destroy_pinned_object(). Hence TTM's BO- release function complains about the pin counter. Forcing the outputs off also undoes the modesettings pin increment. Tested-by: Borislav Petkov (AMD) Reported-by: Borislav Petkov Closes: https://lore.kernel.org/dri-devel/20230603174814.GCZHt83pN+wNjf63sC@fat_crate.local/ Signed-off-by: Thomas Zimmermann Fixes: e317a69fe891 ("drm/radeon: Implement client-based fbdev emulation") Cc: Alex Deucher Cc: Thomas Zimmermann Cc: "Christian König" Cc: "Pan, Xinhui" Cc: amd-gfx@lists.freedesktop.org Signed-off-by: Alex Deucher

drm/amd/pm: workaround for compute workload type on some skus

2023-06-13T21:05:49Z

On smu 13.0.0, the compute workload type cannot be set on all the skus due to some other problems. This workaround is to make sure compute workload type can also run on some specific skus. v2: keep the variable consistent Signed-off-by: Kenneth Feng Acked-by: Lijo Lazar Reviewed-by: Feifei Xu Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.1.x