Graphical glitches on 6.12

Brief summary of the problem:

Hello!

Displays show graphical glitches which are briefly visible (e.g. GDM Login, GNOME-Shell Overview, gnome-terminal, Steam). They don't appear in recorded videos. Therefore, I link to this report with a similar picture. In my specific case glitches cover approximately two or three "lines", which is probably the updated area.

Graphic glitches are not visible in Counter-Strike 1 or Counter-Strike2.

Hardware description:

• CPU: AMD Ryzen™ 7 PRO 6850U
• GPU: Rembrandt Radeon 680M
• System Memory: 32 GB
• Display(s): built-in 2560x1600 (Lenovo ThinkPad X13 Gen 3)
• Type of Display Connection: eDP

I'm using amdgpu.dcdebugmask=0x200, which works fine for me the past weeks with Linux 6.11.

./psr.py 
DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: Parade
○ resv_40f: 01
○ ID String: 06-96
○ PSR Status: 00-00-00

System information:

• Distro name and Version: Archlinux
• Kernel version: 6.12.1

How to reproduce the issue:

• Login to GNOME
• Using GNOME-Shell -> Overview -> Type in application name
• Using the gnome-terminal -> autocompletion, Enter, resize
• Using Steam -> Clicking stuff in menus

Workaround

Force PSR OFF with amdgpu.dcdebugmask=0x10.

Related: Archlinux forum thread and #3742 #3658 (closed).

Thank you

Update 2025-01-01

Enabling PSR-SU also works fine in my case. Check if you can use PSR-SU with Linux 6.12. Usually enabling is only possible if you added yourself amdgpu.dcdebugmask with an argument to your boot command, you need to remove it.

changed the description

Workaround:

Downgrade to Linux 6.11 series. I prefer this, so I can keep PSR ON.

changed the description

Also seeing this on my Yoga 7 Gen 7, a regression from 6.11.10 to 6.12.3. Graphical corruption on changed elements, all over the screen. In dmesg, I can see a backtrace and a different backtrace after an error about REG_TIMEOUT. I also attached the full dmesg log, and copied the interesting bits I saw. I've attached a video of the issue, I also see it on elements of text and anything that changes on screen, not just at the bottom like in the clip.

• CPU: AMD Ryzen 5 6600U
• GPU: Radeon 660M
• Kernel 6.12.3

sudo ./psr.py:

DRI device 0 DMCUB F/W version: 0x04000044
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: Parade
○ resv_40f: 01
○ ID String: 08-01
○ PSR Status: 00-00-02

Backtrace 1:

[   14.593023] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu]
[   14.593442] WARNING: CPU: 8 PID: 96 at kernel/workqueue.c:3704 check_flush_dependency+0xfc/0x120
[   14.593455] Modules linked in: snd_seq_dummy snd_hrtimer snd_seq snd_seq_device rfcomm cmac algif_hash algif_skcipher af_alg qrtr bnep ip6t_REJECT nf_reject_ipv6 xt_hl ip6t_rt ipt_REJECT nf_reject_ipv4 xt_LOG nf_log_syslog nft_limit xt_limit xt_addrtype xt_tcpudp xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nft_compat binfmt_misc nf_tables nls_ascii nls_cp437 vfat fat amd_atl intel_rapl_msr intel_rapl_common amdgpu lenovo_ymc mt7921e mt7921_common edac_mce_amd mt792x_lib btusb snd_acp6x_pdm_dma snd_soc_acp6x_mach snd_soc_dmic mt76_connac_lib btrtl amdxcp snd_sof_amd_rembrandt snd_hda_codec_realtek drm_exec btintel uvcvideo snd_sof_amd_acp gpu_sched btbcm kvm_amd mt76 snd_hda_codec_generic videobuf2_vmalloc btmtk drm_buddy snd_sof_pci uvc videobuf2_memops snd_hda_scodec_component snd_sof_xtensa_dsp drm_suballoc_helper videobuf2_v4l2 bluetooth hid_sensor_accel_3d mac80211 snd_hda_codec_hdmi drm_display_helper snd_sof kvm hid_sensor_trigger videodev hid_sensor_iio_common snd_sof_utils cec libarc4
[   14.593646]  industrialio_triggered_buffer videobuf2_common snd_pci_ps rc_core snd_hda_intel kfifo_buf snd_rpl_pci_acp6x drm_ttm_helper mc industrialio rapl snd_intel_dspcfg pcspkr snd_pci_acp6x wmi_bmof snd_pci_acp5x ttm snd_intel_sdw_acpi ucsi_acpi snd_rn_pci_acp3x cfg80211 drm_kms_helper snd_hda_codec typec_ucsi ideapad_laptop snd_acp_config snd_hda_scodec_tas2781_i2c snd_soc_acpi sparse_keymap sp5100_tco typec snd_hda_core platform_profile i2c_algo_bit k10temp watchdog ccp snd_pci_acp3x snd_hwdep roles rfkill button snd_soc_core snd_compress snd_pcm_dmaengine snd_soc_tas2781_fmwlib snd_soc_tas2781_comlib snd_pcm snd_timer snd soundcore crc8 amd_pmc ac acpi_tad joydev evdev serio_raw msr parport_pc ppdev lp parport nvme_fabrics efi_pstore configfs nfnetlink ip_tables x_tables autofs4 ext4 crc16 mbcache jbd2 dm_crypt dm_mod efivarfs raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor raid6_pq libcrc32c crc32c_generic raid1 raid0 md_mod hid_sensor_custom wacom xhci_pci usbhid hid_sensor_hub
[   14.593863]  hid_multitouch crct10dif_pclmul sdhci_pci crc32_pclmul crc32c_intel xhci_hcd hid_generic cqhci ghash_clmulni_intel usbcore sha512_ssse3 sdhci i2c_hid_acpi nvme i2c_piix4 sha256_ssse3 i2c_hid drm thunderbolt sha1_ssse3 mmc_core nvme_core usb_common i2c_smbus video battery hid wmi aesni_intel gf128mul crypto_simd cryptd
[   14.593937] CPU: 8 UID: 0 PID: 96 Comm: kworker/u48:7 Not tainted 6.12.3-amd64 #1  Debian 6.12.3-1
[   14.593945] Hardware name: LENOVO 82QF/LNVNB161216, BIOS K5CN44WW 01/22/2024
[   14.593949] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
[   14.593962] RIP: 0010:check_flush_dependency+0xfc/0x120
[   14.593970] Code: 8b 45 18 48 8d b2 c0 00 00 00 49 89 e8 48 8d 8b c0 00 00 00 48 c7 c7 80 1e 14 a4 c6 05 28 ff d3 01 01 48 89 c2 e8 64 89 fd ff <0f> 0b e9 1f ff ff ff 80 3d 13 ff d3 01 00 75 93 e9 4a ff ff ff 66
[   14.593973] RSP: 0018:ffffb6cec04abc60 EFLAGS: 00010082
[   14.593978] RAX: 0000000000000000 RBX: ffff992140050800 RCX: 0000000000000027
[   14.593982] RDX: ffff992430021788 RSI: 0000000000000001 RDI: ffff992430021780
[   14.593984] RBP: ffffffffc1896a60 R08: 0000000000000000 R09: ffffb6cec04abae0
[   14.593987] R10: ffffffffa48b42c8 R11: 0000000000000003 R12: ffff992140d019c0
[   14.593990] R13: ffff992140c79300 R14: ffffb6cec04abc90 R15: 0000000000000001
[   14.593993] FS:  0000000000000000(0000) GS:ffff992430000000(0000) knlGS:0000000000000000
[   14.593996] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   14.593999] CR2: 00007ff93c016048 CR3: 00000003a8422000 CR4: 0000000000f50ef0
[   14.594003] PKRU: 55555554
[   14.594005] Call Trace:
[   14.594009]  <TASK>
[   14.594012]  ? check_flush_dependency+0xfc/0x120
[   14.594017]  ? __warn.cold+0x93/0xf6
[   14.594022]  ? check_flush_dependency+0xfc/0x120
[   14.594031]  ? report_bug+0xff/0x140
[   14.594038]  ? handle_bug+0x58/0x90
[   14.594045]  ? exc_invalid_op+0x17/0x70
[   14.594050]  ? asm_exc_invalid_op+0x1a/0x20
[   14.594058]  ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu]
[   14.594329]  ? check_flush_dependency+0xfc/0x120
[   14.594333]  __flush_work+0x110/0x2c0
[   14.594337]  cancel_delayed_work_sync+0x5e/0x80
[   14.594342]  amdgpu_gfx_off_ctrl+0xad/0x140 [amdgpu]
[   14.594503]  amdgpu_ring_alloc+0x43/0x60 [amdgpu]
[   14.594640]  amdgpu_ib_schedule+0xf0/0x730 [amdgpu]
[   14.594774]  amdgpu_job_run+0x8c/0x170 [amdgpu]
[   14.594943]  ? mod_delayed_work_on+0xa4/0xb0
[   14.594947]  drm_sched_run_job_work+0x26e/0x400 [gpu_sched]
[   14.594952]  process_one_work+0x177/0x330
[   14.594955]  worker_thread+0x252/0x390
[   14.594958]  ? __pfx_worker_thread+0x10/0x10
[   14.594961]  kthread+0xd2/0x100
[   14.594965]  ? __pfx_kthread+0x10/0x10
[   14.594968]  ret_from_fork+0x34/0x50
[   14.594972]  ? __pfx_kthread+0x10/0x10
[   14.594976]  ret_from_fork_asm+0x1a/0x30
[   14.594982]  </TASK>

Backtrace 2:

[ 4133.989986] amdgpu 0000:33:00.0: [drm] REG_WAIT timeout 10us * 331 tries - get_channel_status line:350
[ 4133.990044] ------------[ cut here ]------------
[ 4133.990046] WARNING: CPU: 2 PID: 6891 at drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dce_aux.c:393 dce_aux_transfer_raw+0x750/0x780 [amdgpu]
[ 4133.990499] Modules linked in: ccm snd_seq_dummy snd_hrtimer snd_seq snd_seq_device rfcomm cmac algif_hash algif_skcipher af_alg qrtr bnep ip6t_REJECT nf_reject_ipv6 xt_hl ip6t_rt ipt_REJECT nf_reject_ipv4 xt_LOG nf_log_syslog nft_limit xt_limit xt_addrtype xt_tcpudp xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nft_compat binfmt_misc nf_tables nls_ascii nls_cp437 vfat fat amd_atl intel_rapl_msr intel_rapl_common amdgpu lenovo_ymc mt7921e mt7921_common edac_mce_amd mt792x_lib btusb snd_acp6x_pdm_dma snd_soc_acp6x_mach snd_soc_dmic mt76_connac_lib btrtl amdxcp snd_sof_amd_rembrandt snd_hda_codec_realtek drm_exec btintel uvcvideo snd_sof_amd_acp gpu_sched btbcm kvm_amd mt76 snd_hda_codec_generic videobuf2_vmalloc btmtk drm_buddy snd_sof_pci uvc videobuf2_memops snd_hda_scodec_component snd_sof_xtensa_dsp drm_suballoc_helper videobuf2_v4l2 bluetooth hid_sensor_accel_3d mac80211 snd_hda_codec_hdmi drm_display_helper snd_sof kvm hid_sensor_trigger videodev hid_sensor_iio_common snd_sof_utils cec libarc4
[ 4133.990636]  industrialio_triggered_buffer videobuf2_common snd_pci_ps rc_core snd_hda_intel kfifo_buf snd_rpl_pci_acp6x drm_ttm_helper mc industrialio rapl snd_intel_dspcfg pcspkr snd_pci_acp6x wmi_bmof snd_pci_acp5x ttm snd_intel_sdw_acpi ucsi_acpi snd_rn_pci_acp3x cfg80211 drm_kms_helper snd_hda_codec typec_ucsi ideapad_laptop snd_acp_config snd_hda_scodec_tas2781_i2c snd_soc_acpi sparse_keymap sp5100_tco typec snd_hda_core platform_profile i2c_algo_bit k10temp watchdog ccp snd_pci_acp3x snd_hwdep roles rfkill button snd_soc_core snd_compress snd_pcm_dmaengine snd_soc_tas2781_fmwlib snd_soc_tas2781_comlib snd_pcm snd_timer snd soundcore crc8 amd_pmc ac acpi_tad joydev evdev serio_raw msr parport_pc ppdev lp parport nvme_fabrics efi_pstore configfs nfnetlink ip_tables x_tables autofs4 ext4 crc16 mbcache jbd2 dm_crypt dm_mod efivarfs raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor raid6_pq libcrc32c crc32c_generic raid1 raid0 md_mod hid_sensor_custom wacom xhci_pci usbhid hid_sensor_hub
[ 4133.990796]  hid_multitouch crct10dif_pclmul sdhci_pci crc32_pclmul crc32c_intel xhci_hcd hid_generic cqhci ghash_clmulni_intel usbcore sha512_ssse3 sdhci i2c_hid_acpi nvme i2c_piix4 sha256_ssse3 i2c_hid drm thunderbolt sha1_ssse3 mmc_core nvme_core usb_common i2c_smbus video battery hid wmi aesni_intel gf128mul crypto_simd cryptd
[ 4133.990855] CPU: 2 UID: 0 PID: 6891 Comm: psr.py Tainted: G        W          6.12.3-amd64 #1  Debian 6.12.3-1
[ 4133.990861] Tainted: [W]=WARN
[ 4133.990864] Hardware name: LENOVO 82QF/LNVNB161216, BIOS K5CN44WW 01/22/2024
[ 4133.990867] RIP: 0010:dce_aux_transfer_raw+0x750/0x780 [amdgpu]
[ 4133.991237] Code: b8 04 00 00 00 41 89 04 24 41 bd ff ff ff ff e9 88 fd ff ff b8 02 00 00 00 eb ea 3c 01 19 c0 83 e0 c0 83 c0 50 e9 57 f9 ff ff <0f> 0b b8 03 00 00 00 eb d2 b8 03 00 00 00 eb cb 41 c7 04 24 05 00
[ 4133.991240] RSP: 0018:ffffb6ceca18fc10 EFLAGS: 00010246
[ 4133.991245] RAX: 0000000011000000 RBX: ffff992152243080 RCX: ffffffffc24736b3
[ 4133.991248] RDX: 0000000000000000 RSI: 0000000000005414 RDI: ffff992142c80000
[ 4133.991250] RBP: ffffb6ceca18fc80 R08: ffffffffc23ae540 R09: 00000000000001dc
[ 4133.991253] R10: ffffffffa48b42c8 R11: 0000000000000003 R12: ffffb6ceca18fc7c
[ 4133.991255] R13: 0000000000000004 R14: 0000000000000000 R15: ffff992149e88d50
[ 4133.991258] FS:  00007f079d865080(0000) GS:ffff99242fd00000(0000) knlGS:0000000000000000
[ 4133.991261] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 4133.991263] CR2: 000055a87556bb08 CR3: 0000000224ece000 CR4: 0000000000f50ef0
[ 4133.991267] PKRU: 55555554
[ 4133.991269] Call Trace:
[ 4133.991275]  <TASK>
[ 4133.991278]  ? dce_aux_transfer_raw+0x750/0x780 [amdgpu]
[ 4133.991638]  ? __warn.cold+0x93/0xf6
[ 4133.991645]  ? dce_aux_transfer_raw+0x750/0x780 [amdgpu]
[ 4133.992017]  ? report_bug+0xff/0x140
[ 4133.992025]  ? handle_bug+0x58/0x90
[ 4133.992032]  ? exc_invalid_op+0x17/0x70
[ 4133.992036]  ? asm_exc_invalid_op+0x1a/0x20
[ 4133.992046]  ? dce_aux_transfer_raw+0x750/0x780 [amdgpu]
[ 4133.992463]  dm_dp_aux_transfer+0xda/0x1a0 [amdgpu]
[ 4133.992923]  drm_dp_dpcd_access+0xbb/0x150 [drm_display_helper]
[ 4133.992943]  ? drm_dp_dpcd_probe+0x8b/0xf0 [drm_display_helper]
[ 4133.992959]  drm_dp_dpcd_read+0xef/0x110 [drm_display_helper]
[ 4133.992975]  auxdev_read_iter+0x9a/0x1a0 [drm_display_helper]
[ 4133.992993]  vfs_read+0x29c/0x370
[ 4133.993004]  ksys_read+0x6d/0xf0
[ 4133.993011]  do_syscall_64+0x82/0x190
[ 4133.993020]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993029]  ? syscall_exit_to_user_mode+0x4d/0x210
[ 4133.993035]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993040]  ? do_syscall_64+0x8e/0x190
[ 4133.993044]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993049]  ? handle_irq_event+0x50/0x90
[ 4133.993058]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993063]  ? handle_edge_irq+0x8b/0x230
[ 4133.993069]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993074]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993079]  ? __irq_exit_rcu+0x37/0xb0
[ 4133.993084]  ? srso_alias_return_thunk+0x5/0xfbef5
[ 4133.993090]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[ 4133.993096] RIP: 0033:0x7f079d96945d
[ 4133.993162] Code: c9 39 0e 00 f7 d8 64 89 02 b8 ff ff ff ff eb bb e8 88 a0 01 00 0f 1f 84 00 00 00 00 00 80 3d 09 bc 0e 00 00 74 17 31 c0 0f 05 <48> 3d 00 f0 ff ff 77 5b c3 66 2e 0f 1f 84 00 00 00 00 00 48 83 ec
[ 4133.993165] RSP: 002b:00007ffd3fbcb238 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
[ 4133.993170] RAX: ffffffffffffffda RBX: 00007f079d864ff8 RCX: 00007f079d96945d
[ 4133.993173] RDX: 0000000000001000 RSI: 000000000c8c2330 RDI: 0000000000000003
[ 4133.993176] RBP: 0000000000001000 R08: 0000000000000000 R09: 0000000000000002
[ 4133.993179] R10: 0000000000000003 R11: 0000000000000246 R12: 0000000000000003
[ 4133.993181] R13: 000000000c8c2330 R14: 0000000000bb05f0 R15: 00007f079d2655c0
[ 4133.993190]  </TASK>

dmesg.log

VID_20241209_150150_2

I have no idea why the video is showing as an audio clip, but downloading it works

added DC Rembrandt labels

Update

I upgraded to Linux 6.12.4 and removed amdgpu.dcdebugmask entirely. Neither passing 0x200 (only disable PSR-SU) nor 0x10 (disable PSR completely). Therefore PSR-SU shall be in use? The graphical glitches are gone!
I'm regularly using amdgpu.dcdebugmask=0x200 because my display is likely not compatible with PSR-SU and turns off the display backlight. Using 0x200 worked well for me with 6.10 and 6.11.

Test

I need to test in long term if I can now use PSR-SU instead of PSR. If PSR-SU still switches off my display backlight occasionally, I'm in trouble. If not, I'm lucky.

Idea

Some people reverted the commit use a more lax vblank enable policy for older ASICs . But the mentioned change to vblank motivated me to test this again without passing any argument to amdgpu.dcdebugmask. Is it possible the 6.12 shuffled around the compatibility between various ASICs, displays and PSR-Modes {OFF, PSR-SU, PSR}?

I'm just poking around.

I've clarified the last section from the post above.

I used the laptop now with PSR-SU enabled for some others and I didn't recognized any graphical glitches. I'm now rather sure about this.

The display backlight also remained on surprisingly. Here I need keep testing at least week or better two weeks, to be sure.

I've noticed this doesn't happen while the display is continuously changing. If I start scrolling through a block of text, I get corruption where the text was as soon as I start scrolling, but this disappears and no more appears while I'm scrolling, until I stop and start again.

I'll try a bisect, any ideas which commit may have caused this to save a bit of time?

Bisecting done, it comes back to 58a261bfc96763a851cb48b203ed57da37e157b8 drm/amd/display: use a more lax vblank enable policy for older ASICs

Bisected with amdgpu.dcdebugmask=0x200, since that's enabled by default for my panel in newer kernels, and the issue only occurs with PSR-SU off.

This fit with the observations?

PSR-SU ON  -> Good # depending on display support and that backlight remains on
PSR    ON  -> Problematic # for some
PSR    OFF -> Good # higher power consumption

I'm still testing if my display backlight remains on with PSR-SU ON. Hadn't much time for testing but hopefully the Christmas days allow me to get some confidence.

Without amdgpu.dcdebugmask

# ./psr.py 
DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: Parade
○ resv_40f: 01
○ ID String: 06-96
○ PSR Status: 00-00-02

Question at developers: Is their a documentation about the meaning of 00-00-02. I guess this means PSR-SU ON.

Yeah, with PSR-SU on I don't get this issue, but my panel isn't happy with PSR-SU and I get other, unrelated issues. With PSR-SU off, I get this issue. I haven't tested with PSR off entirely.

Thanks for testing Stuart.

added bisected label

/cc @effective-light

On my laptop (7735HS) I don't have graphical glitches, instead everything freezes for a second and then I see a similar backtrace in dmesg:

backtrace

``` [ 1314.283432] ------------[ cut here ]------------ [ 1314.283443] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu] [ 1314.283999] WARNING: CPU: 13 PID: 2113 at kernel/workqueue.c:3704 check_flush_dependency+0xfc/0x120 [ 1314.284014] Modules linked in: nft_chain_nat xt_CT xt_tcpudp xt_helper nf_conntrack_ftp nft_limit nft_compat iptable_nat nf_nat xt_TCPMSS xt_LOG nf_log_syslog ipt_REJECT nf_reject_ipv4 iptable_raw iptable_mangle iptable_filter xt_multiport xt_state xt_limit xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nf_tables libcrc32c snd_seq_dummy snd_hrtimer rfcomm snd_seq snd_seq_device cmac algif_hash algif_skcipher af_alg bnep snd_soc_dmic snd_soc_acp6x_mach snd_acp6x_pdm_dma snd_sof_amd_acp70 snd_sof_amd_acp63 snd_soc_acpi_amd_match snd_sof_amd_vangogh snd_sof_amd_rembrandt snd_sof_amd_renoir snd_sof_amd_acp snd_sof_pci snd_sof_xtensa_dsp snd_sof snd_sof_utils snd_pci_ps snd_amd_sdw_acpi amd_atl intel_rapl_msr vfat soundwire_amd fat intel_rapl_common snd_hda_codec_realtek iwlmvm soundwire_generic_allocation amdgpu snd_hda_codec_generic soundwire_bus mac80211 snd_hda_scodec_component nouveau snd_hda_codec_hdmi snd_soc_core libarc4 snd_hda_intel snd_compress ptp uvcvideo ac97_bus pps_core snd_intel_dspcfg [ 1314.284117] snd_pcm_dmaengine snd_intel_sdw_acpi videobuf2_vmalloc mxm_wmi snd_rpl_pci_acp6x uvc cdc_ether snd_hda_codec drm_gpuvm amdxcp videobuf2_memops btusb usbnet snd_acp_pci drm_exec iwlwifi btrtl drm_buddy videobuf2_v4l2 snd_hda_core gpu_sched snd_acp_legacy_common kvm_amd btintel drm_suballoc_helper videobuf2_common snd_pci_acp6x snd_hwdep i2c_algo_bit btbcm drm_ttm_helper btmtk spd5118 kvm snd_pci_acp5x asus_wmi videodev snd_pcm r8169 ttm r8152 snd_rn_pci_acp3x platform_profile cfg80211 mc bluetooth rapl realtek snd_timer sparse_keymap snd_acp_config mii wmi_bmof drm_display_helper mdio_devres snd snd_soc_acpi i2c_piix4 cec rfkill soundcore k10temp snd_pci_acp3x libphy i2c_smbus joydev amd_pmc mousedev acpi_tad mac_hid loop nfnetlink ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 dm_crypt cbc encrypted_keys trusted asn1_encoder tee usbhid ccp crct10dif_pclmul crc32_pclmul crc32c_intel polyval_clmulni polyval_generic ghash_clmulni_intel sha512_ssse3 serio_raw sha256_ssse3 sha1_ssse3 atkbd nvme [ 1314.284241] aesni_intel libps2 vivaldi_fmap hid_multitouch gf128mul nvme_core hid_generic i8042 crypto_simd cryptd sp5100_tco nvme_auth video i2c_hid_acpi serio wmi i2c_hid dm_mod pkcs8_key_parser crypto_user [ 1314.284281] CPU: 13 UID: 0 PID: 2113 Comm: kworker/u64:64 Not tainted 6.12.6-arch1-1 #1 (closed) be8168881006593767299fff7299891c69c41600 [ 1314.284286] Hardware name: PCSpecialist GM6BG5Q/GM6BG5Q, BIOS N.1.12A01 04/03/2023 [ 1314.284288] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched] [ 1314.284294] RIP: 0010:check_flush_dependency+0xfc/0x120 [ 1314.284297] Code: 8b 45 18 48 8d b2 c0 00 00 00 49 89 e8 48 8d 8b c0 00 00 00 48 c7 c7 40 a3 ae b3 c6 05 e9 04 16 02 01 48 89 c2 e8 04 8e fd ff <0f> 0b e9 1f ff ff ff 80 3d d4 04 16 02 00 75 93 e9 4a ff ff ff 66 [ 1314.284300] RSP: 0018:ffffbb8987fb3c60 EFLAGS: 00010082 [ 1314.284303] RAX: 0000000000000000 RBX: ffff99a840050800 RCX: 0000000000000027 [ 1314.284305] RDX: ffff99af1eaa18c8 RSI: 0000000000000001 RDI: ffff99af1eaa18c0 [ 1314.284307] RBP: ffffffffc1b43b00 R08: 0000000000000000 R09: ffffbb8987fb3ae0 [ 1314.284309] R10: ffffffffb42b5528 R11: 0000000000000003 R12: ffff99a8944d5100 [ 1314.284311] R13: ffff99a8420600c0 R14: ffffbb8987fb3c90 R15: 0000000000000001 [ 1314.284313] FS: 0000000000000000(0000) GS:ffff99af1ea80000(0000) knlGS:0000000000000000 [ 1314.284315] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 1314.284317] CR2: 000074206e619000 CR3: 00000003d1e22000 CR4: 0000000000f50ef0 [ 1314.284320] PKRU: 55555554 [ 1314.284321] Call Trace: [ 1314.284323] [ 1314.284325] ? check_flush_dependency+0xfc/0x120 [ 1314.284328] ? __warn.cold+0x93/0xf6 [ 1314.284332] ? check_flush_dependency+0xfc/0x120 [ 1314.284336] ? report_bug+0xff/0x140 [ 1314.284342] ? handle_bug+0x58/0x90 [ 1314.284345] ? exc_invalid_op+0x17/0x70 [ 1314.284348] ? asm_exc_invalid_op+0x1a/0x20 [ 1314.284352] ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu c91cd080e1155e0debec7c4b8ca50ccbb8b029cb] [ 1314.284552] ? check_flush_dependency+0xfc/0x120 [ 1314.284557] __flush_work+0x110/0x2c0 [ 1314.284563] cancel_delayed_work_sync+0x5e/0x80 [ 1314.284568] amdgpu_gfx_off_ctrl+0xad/0x140 [amdgpu c91cd080e1155e0debec7c4b8ca50ccbb8b029cb] [ 1314.284812] amdgpu_ring_alloc+0x43/0x60 [amdgpu c91cd080e1155e0debec7c4b8ca50ccbb8b029cb] [ 1314.285020] amdgpu_ib_schedule+0xf0/0x730 [amdgpu c91cd080e1155e0debec7c4b8ca50ccbb8b029cb] [ 1314.285235] amdgpu_job_run+0x8c/0x170 [amdgpu c91cd080e1155e0debec7c4b8ca50ccbb8b029cb] [ 1314.285508] ? mod_delayed_work_on+0xa4/0xb0 [ 1314.285513] drm_sched_run_job_work+0x25c/0x3f0 [gpu_sched 49204b5052764fd0640351d5fd3aad21461d24e1] [ 1314.285520] process_one_work+0x17e/0x330 [ 1314.285524] worker_thread+0x2ce/0x3f0 [ 1314.285528] ? __pfx_worker_thread+0x10/0x10 [ 1314.285532] kthread+0xd2/0x100 [ 1314.285536] ? __pfx_kthread+0x10/0x10 [ 1314.285540] ret_from_fork+0x34/0x50 [ 1314.285544] ? __pfx_kthread+0x10/0x10 [ 1314.285548] ret_from_fork_asm+0x1a/0x30 [ 1314.285556] [ 1314.285558] ---[ end trace 0000000000000000 ]--- ```

mentioned in issue #3787

I have a 9950X with iGPU enabled and I believe that I am experiencing this bug BUT it persists even when booting with amdgpu.dcdebugmask=0x10. Any advice is appreciated.

Linux version 6.12.7-arch1-1 (linux@archlinux) (gcc (GCC) 14.2.1 20240910, GNU ld (GNU Binutils) 2.43.0) #1 SMP PREEMPT_DYNAMIC Fri, 27 Dec 2024 14:24:37 +0000
Command line: BOOT_IMAGE=/vmlinuz-linux root=UUID=08e3e8f3-f050-43fc-8b43-e9ed21443418 rw loglevel=3 quiet audit=0 amdgpu.dcdebugmask=0x10
...
------------[ cut here ]------------
workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu]
WARNING: CPU: 23 PID: 993 at kernel/workqueue.c:3704 check_flush_dependency+0xfc/0x120
Modules linked in: overlay amd_atl intel_rapl_msr intel_rapl_common snd_hda_codec_realtek kvm_amd snd_hda_codec_generic snd_hda_scodec_component snd_hda_codec_hdmi kvm snd_hda_intel crct10dif_pclmul snd_intel_dspcfg crc32_pclmul snd_intel_sdw_acpi polyval_clmulni polyval_generic snd_hda_codec ghash_clmulni_intel sha512_ssse3 snd_hda_core sha256_ssse3 sha1_ssse3 snd_hwdep spd5118 ip6t_REJECT aesni_intel sp5100_tco snd_pcm nf_reject_ipv6 r8169 gf128mul crypto_simd snd_timer realtek cryptd i2c_piix4 snd mdio_devres xt_hl wmi_bmof rapl pcspkr ccp i2c_smbus soundcore libphy ip6t_rt gpio_amdpt gpio_generic cfg80211 mousedev joydev ipt_REJECT nf_reject_ipv4 mac_hid xt_multiport xt_comment rfkill xt_limit xt_addrtype xt_tcpudp xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c ip6table_filter ip6_tables iptable_filter xt_iprange xt_mark xt_NFQUEUE k10temp nct6683 dm_mod loop nfnetlink ip_tables x_tables ext4 crc32c_generic mbcache jbd2 hid_microsoft ff_memless hid_generic nvme nvme_core crc32c_intel
 usbhid nvme_auth amdgpu video wmi amdxcp i2c_algo_bit drm_ttm_helper ttm drm_exec gpu_sched drm_suballoc_helper drm_buddy drm_display_helper cec crc16
CPU: 23 UID: 0 PID: 993 Comm: kworker/u128:2 Not tainted 6.12.7-arch1-1 #1 9e77c5d99557be92f482a3ac6317d887bb3ffaf9
Hardware name: Micro-Star International Co., Ltd. MS-7E16/X670E GAMING PLUS WIFI (MS-7E16), BIOS 1.93 12/02/2024
Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
RIP: 0010:check_flush_dependency+0xfc/0x120
Code: 8b 45 18 48 8d b2 c0 00 00 00 49 89 e8 48 8d 8b c0 00 00 00 48 c7 c7 10 32 0f 8f c6 05 c9 3a 16 02 01 48 89 c2 e8 04 8e fd ff <0f> 0b e9 1f ff ff ff 80 3d b4 3a 16 02 00 75 93 e9 4a ff ff ff 66
RSP: 0018:ffffad3d05bebc68 EFLAGS: 00010086
RAX: 0000000000000000 RBX: ffffa0b940050c00 RCX: 0000000000000027
RDX: ffffa0c44e3a18c8 RSI: 0000000000000001 RDI: ffffa0c44e3a18c0
RBP: ffffffffc0370b00 R08: 0000000000000000 R09: ffffad3d05bebae8
R10: ffffa0c47dcd1768 R11: 0000000000000003 R12: ffffa0b94907a140
R13: ffffa0b94ac1d2c0 R14: ffffad3d05bebc98 R15: 0000000000000001
FS:  0000000000000000(0000) GS:ffffa0c44e380000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007ca7028fe1dc CR3: 0000000c05a22000 CR4: 0000000000f50ef0
PKRU: 55555554
Call Trace:
 <TASK>
 ? check_flush_dependency+0xfc/0x120
 ? __warn.cold+0x93/0xf6
 ? check_flush_dependency+0xfc/0x120
 ? report_bug+0xff/0x140
 ? console_unlock+0x9d/0x140
 ? handle_bug+0x58/0x90
 ? exc_invalid_op+0x17/0x70
 ? asm_exc_invalid_op+0x1a/0x20
 ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu e86c125fc0d1d107466a775e7b7301e5c757afc6]
 ? check_flush_dependency+0xfc/0x120
 ? check_flush_dependency+0xfc/0x120
 __flush_work+0x110/0x2c0
 cancel_delayed_work_sync+0x5e/0x80
 amdgpu_gfx_off_ctrl+0xad/0x140 [amdgpu e86c125fc0d1d107466a775e7b7301e5c757afc6]
 amdgpu_ring_alloc+0x40/0x60 [amdgpu e86c125fc0d1d107466a775e7b7301e5c757afc6]
 amdgpu_ib_schedule+0xf0/0x730 [amdgpu e86c125fc0d1d107466a775e7b7301e5c757afc6]
 amdgpu_job_run+0x8e/0x1f0 [amdgpu e86c125fc0d1d107466a775e7b7301e5c757afc6]
 drm_sched_run_job_work+0x259/0x3f0 [gpu_sched 29c1ee69cf658188cc04c71397c34eee6e156b12]
 process_one_work+0x17b/0x330
 worker_thread+0x2ce/0x3f0
 ? __pfx_worker_thread+0x10/0x10
 kthread+0xcf/0x100
 ? __pfx_kthread+0x10/0x10
 ret_from_fork+0x31/0x50
 ? __pfx_kthread+0x10/0x10
 ret_from_fork_asm+0x1a/0x30
 </TASK>
---[ end trace 0000000000000000 ]---

mentioned in issue #2970 (closed)

I kept testing the last days and the results seem positive

• Linux 6.12
• PSR-SU ON
• Display doesn't show graphical glitches
• Display doesn't turn off randomly

If you have the possibility to turn PSR-SU ON, do it.

This requires a display capable of PSR-SU and removing amdgpu.dcdebugmask=0x200 or amdgpu.dcdebugmask=0x10. This is not a solution for people, which displays don't support PSR-SU. If if you still use one of these options because you had previous problems with PSR-SU, please give it a try with Linux 6.12. Even when other problems persist in your case.

# ./psr.py 
DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: Parade
○ resv_40f: 01
○ ID String: 06-96
○ PSR Status: 00-00-02

@Developers

Is it possible that there is a handling difference between PSR-SU and PSR? Stuarts testing also supports this. If you try to fix the graphical glitches, please keep PSR-SU working reliable. I am happy with my the current state

changed the description

I can confirm this also on a Lenovo ThinkPad T14s Gen 3 (AMD), running 6.12.7. Not using any debugflags or other special boot parameters.

33:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Rembrandt [Radeon 680M] (rev d1)

[14943.804762] ------------[ cut here ]------------
[14943.804770] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu]
[14943.805399] WARNING: CPU: 12 PID: 16991 at kernel/workqueue.c:3704 check_flush_dependency+0xfc/0x120
[14943.805418] Modules linked in: michael_mic uinput rfcomm snd_seq_dummy snd_hrtimer nf_conntrack_netbios_ns nf_conntrack_broadcast nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 ip_set nf_tables sunrpc bnep qrtr_mhi vfat fat intel_rapl_msr amd_atl intel_rapl_common edac_mce_amd snd_soc_acp6x_mach snd_soc_dmic snd_acp6x_pdm_dma snd_sof_amd_acp70 snd_sof_amd_acp63 snd_soc_acpi_amd_match snd_sof_amd_vangogh snd_sof_amd_rembrandt snd_sof_amd_renoir snd_sof_amd_acp snd_sof_pci snd_sof_xtensa_dsp snd_sof qrtr snd_sof_utils snd_pci_ps snd_amd_sdw_acpi snd_hda_codec_realtek soundwire_amd soundwire_generic_allocation snd_hda_codec_generic soundwire_bus ath11k_pci snd_hda_scodec_component ath11k snd_soc_core snd_hda_codec_hdmi snd_compress ac97_bus snd_pcm_dmaengine snd_hda_intel qmi_helpers snd_rpl_pci_acp6x snd_intel_dspcfg snd_acp_pci kvm_amd btusb uvcvideo btrtl snd_intel_sdw_acpi
[14943.805614]  snd_acp_legacy_common mac80211 uvc snd_hda_codec videobuf2_vmalloc btintel videobuf2_memops snd_ctl_led libarc4 snd_hda_core kvm snd_pci_acp6x videobuf2_v4l2 snd_hwdep btbcm snd_seq spd5118 videobuf2_common btmtk rapl snd_seq_device cfg80211 bluetooth videodev think_lmi wmi_bmof snd_pcm thinkpad_acpi mc firmware_attributes_class sparse_keymap snd_pci_acp5x platform_profile snd_rn_pci_acp3x snd_acp_config thunderbolt snd_timer rfkill snd_soc_acpi i2c_piix4 k10temp snd_pci_acp3x mhi i2c_smbus snd soundcore joydev amd_pmc acpi_tad loop nfnetlink zram amdgpu hid_logitech_hidpp amdxcp i2c_algo_bit drm_ttm_helper ttm drm_exec gpu_sched drm_suballoc_helper drm_buddy drm_display_helper hid_logitech_dj cec nvme nvme_core crct10dif_pclmul crc32_pclmul crc32c_intel polyval_clmulni video hid_multitouch polyval_generic ucsi_acpi ghash_clmulni_intel sha512_ssse3 sha256_ssse3 sha1_ssse3 typec_ucsi sp5100_tco typec nvme_auth i2c_hid_acpi wmi i2c_hid serio_raw fuse i2c_dev
[14943.805816] CPU: 12 UID: 0 PID: 16991 Comm: kworker/u64:3 Not tainted 6.12.7-200.fc41.x86_64 #1
[14943.805821] Hardware name: LENOVO 21CQCTO1WW/21CQCTO1WW, BIOS R22ET74W (1.44 ) 10/28/2024
[14943.805824] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
[14943.805830] RIP: 0010:check_flush_dependency+0xfc/0x120
[14943.805834] Code: 8b 45 18 48 8d b2 c0 00 00 00 49 89 e8 48 8d 8b c0 00 00 00 48 c7 c7 98 24 e5 b8 c6 05 ce 99 ad 02 01 48 89 c2 e8 d4 8b fd ff <0f> 0b e9 1f ff ff ff 80 3d b9 99 ad 02 00 75 93 e9 4a ff ff ff 66
[14943.805837] RSP: 0018:ffffaf6b0fcabc58 EFLAGS: 00010082
[14943.805840] RAX: 0000000000000000 RBX: ffff97d5801be600 RCX: 0000000000000027
[14943.805843] RDX: ffff97dc9f021908 RSI: 0000000000000001 RDI: ffff97dc9f021900
[14943.805845] RBP: ffffffffc0672470 R08: 0000000000000000 R09: 0000000000000000
[14943.805847] R10: ffffffffffffffff R11: 0000000000000000 R12: ffff97d58c635400
[14943.805848] R13: ffff97d584ff9200 R14: ffffaf6b0fcabc98 R15: 0000000000000001
[14943.805852] FS:  0000000000000000(0000) GS:ffff97dc9f000000(0000) knlGS:0000000000000000
[14943.805854] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[14943.805856] CR2: 00007f6059195000 CR3: 00000005b082a000 CR4: 0000000000f50ef0
[14943.805859] PKRU: 55555554
[14943.805861] Call Trace:
[14943.805870]  <TASK>
[14943.805873]  ? check_flush_dependency+0xfc/0x120
[14943.805877]  ? __warn.cold+0x93/0xfa
[14943.805886]  ? check_flush_dependency+0xfc/0x120
[14943.805897]  ? report_bug+0xff/0x140
[14943.805903]  ? console_unlock+0x9d/0x140
[14943.805912]  ? handle_bug+0x58/0x90
[14943.805920]  ? exc_invalid_op+0x17/0x70
[14943.805924]  ? asm_exc_invalid_op+0x1a/0x20
[14943.805932]  ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu]
[14943.806173]  ? check_flush_dependency+0xfc/0x120
[14943.806177]  ? check_flush_dependency+0xfc/0x120
[14943.806181]  __flush_work+0x110/0x310
[14943.806188]  cancel_delayed_work_sync+0x5e/0x80
[14943.806193]  amdgpu_gfx_off_ctrl+0xad/0x140 [amdgpu]
[14943.806477]  amdgpu_ring_alloc+0x43/0x60 [amdgpu]
[14943.806722]  amdgpu_ib_schedule+0xf0/0x730 [amdgpu]
[14943.806974]  amdgpu_job_run+0x8e/0x1f0 [amdgpu]
[14943.807281]  drm_sched_run_job_work+0x26e/0x400 [gpu_sched]
[14943.807289]  process_one_work+0x179/0x330
[14943.807295]  worker_thread+0x252/0x390
[14943.807299]  ? __pfx_worker_thread+0x10/0x10
[14943.807303]  kthread+0xd2/0x100
[14943.807307]  ? __pfx_kthread+0x10/0x10
[14943.807310]  ret_from_fork+0x34/0x50
[14943.807315]  ? __pfx_kthread+0x10/0x10
[14943.807318]  ret_from_fork_asm+0x1a/0x30
[14943.807326]  </TASK>
[14943.807328] ---[ end trace 0000000000000000 ]---

$ sudo ./psr.py
DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: Parade
○ resv_40f: 01
○ ID String: 08-03
○ PSR Status: 00-00-02

Running Fedora 41 (KDE).

Did any of those workarounds work for you? I use a ThinkPad T14 Gen 3 (AMD).

For me setting the adaptive sync to 'always' seems to have stopped the glitching. I'm not entirely sure what command line parameter is being set (should investigate), it's an option in the KDE Display Settings.

I am using GNOME. I enabled the experimental feature Adaptive Sync in GNOME, but I am not able to activate it for my display.

Right now I am using Kernel 6.6.72, which basically means that PSU is disabled (I guess). psu.py generates the following output:

DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: 00-22-b9
○ resv_40f: 01
○ ID String: 65-43
○ PSR Status: 00-00-00

With Kernel 6.12.10:

DRI device 1 DMCUB F/W version: 0x04000045
○ PSR 2 with Y coordinates (eDP 1.4a) [3]
○ Sink OUI: 00-22-b9
○ resv_40f: 01
○ ID String: 65-43
○ PSR Status: 00-00-02

If I set amdgpu.dcdebugmask=0x10 for Kernel 6.12.10, I get the same output as with Kernel 6.6.72. Glitches are gone. So disabling PSU for 6.12 seems to work.

I am now wondering if there is any difference between using 6.6 without any parameters or 6.12 with activated dcdebugmask.

Seeing this on an Asus ROG 14 (2022) with 680M graphics on 6.12 (tried 6.12.4 and 6.12.9) and a KDE Plasma Wayland session. I see some flashing extra-brief corruption appearing on an horizontal stripe through the display, more often in the bottom half. Quite evident when looking at web pages. Possibly more frequent when the machine is unloaded. Leaving a web page on and doing absolutely nothing the thing becomes almost periodic with a some-seconds period. The perception may be subjective, but I find it extra disturbing, to the point of being unable to work on the machine. Fortunately, setting "Adaptive vsync" to "Always" in the KDE display settings seems to radically improve the situation. Leaving it to "Automatic" or setting it to "Never" used to be fine on pre 6.12 kernels, though.

Interesting. This seems to work here, too.

I still have have these issues as well, even with kernel 6.12.10. Instead of disabling PSR completely (which resolves the issue but increases battery usage) I use the lts kernel for now. My OLED panel doesn't offer the adaptive sync option in KDE Plasma.

Operating System: CachyOS Linux
KDE Plasma Version: 6.2.5
KDE Frameworks Version: 6.10.0
Qt Version: 6.8.1
Kernel Version: 6.6.72-1-lts (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 PRO 6860Z with Radeon Graphics
Memory: 27,2 GiB of RAM
Graphics Processor: AMD Radeon Graphics
Manufacturer: LENOVO
Product Name: 21D2CTO1WW
System Version: ThinkPad Z13 Gen 1

This seems to be fixed for me with 6.12.11, can anyone else confirm?

It's working for me as well.

Fixed for me using 6.13.0.

I have a LENOVO 21D2CTO1WW (ThinkPad Z13 Gen 1) running BIOS 1.70 (N3GET70W (1.70 )). I rebooted into 6.12.11-200.fc41.x86_64 this morning, removing amdgpu.dcdebugmask=0x10 from the kernel cmd line.
I haven't seen any glitches so far.

If you have vrr capable display it's likely due to commit 67edb81d6e9a ("drm/amd/display: Disable replay and psr while VRR is enabled"). For everyone else it's probably due to commit ff2e4d874726 ("drm/amd/display: Do not wait for PSR disable on vbl enable") (both of which are in 6.12.11).

I think this is fixed.

btw.
With 6.13.x the backlight turns unconditionally off again with PSR-SU.

closed