* [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600
@ 2024-12-13 12:33 Chris Rankin
2024-12-16 18:36 ` Alex Deucher
0 siblings, 1 reply; 4+ messages in thread
From: Chris Rankin @ 2024-12-13 12:33 UTC (permalink / raw)
To: LKML, amd-gfx
Hi,
I've just noticed this warning in my dmesg log. This is a vanilla
6.12.4 kernel, with a Radeon RX6600 graphics card.
Cheers,
Chris
[ 4624.741148] ------------[ cut here ]------------
[ 4624.744474] 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]
[ 4624.744942] WARNING: CPU: 2 PID: 9069 at kernel/workqueue.c:3704
check_flush_dependency+0xbe/0xd0
[ 4624.765285] Modules linked in: snd_seq_dummy rpcrdma rdma_cm iw_cm
ib_cm ib_core af_packet 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
ebtable_nat ebtable_broute ip6table_nat ip6table_mangle ip6table_raw
ip6table_security iptable_nat iptable_mangle iptable_raw
iptable_security nf_nat_ftp nf_conntrack_ftp nf_nat nf_conntrack
nf_defrag_ipv6 nf_defrag_ipv4 nf_tables libcrc32c ebtable_filter
ebtables ip6table_filter ip6_tables iptable_filter ip_tables x_tables
bnep it87 hwmon_vid binfmt_misc snd_hda_codec_realtek
snd_hda_codec_generic snd_hda_codec_hdmi snd_hda_scodec_component
snd_hda_intel uvcvideo btusb uvc videobuf2_vmalloc btintel
videobuf2_memops videobuf2_v4l2 videodev btbcm snd_usb_audio bluetooth
snd_intel_dspcfg intel_powerclamp snd_hda_codec videobuf2_common
coretemp snd_virtuoso snd_usbmidi_lib snd_oxygen_lib snd_ctl_led
kvm_intel input_leds mc snd_hwdep led_class snd_mpu401_uart
[ 4624.765400] snd_hda_core joydev snd_rawmidi rfkill kvm snd_seq
snd_seq_device gpio_ich snd_pcm pktcdvd iTCO_wdt snd_hrtimer r8169
snd_timer intel_cstate realtek snd mdio_devres intel_uncore libphy
i2c_i801 soundcore lpc_ich tiny_power_button mxm_wmi i7core_edac
acpi_cpufreq i2c_smbus pcspkr button nfsd auth_rpcgss nfs_acl lockd
grace dm_mod fuse sunrpc loop configfs dax nfnetlink zram zsmalloc
ext4 crc32c_generic mbcache jbd2 amdgpu video amdxcp i2c_algo_bit
mfd_core drm_ttm_helper ttm drm_exec gpu_sched hid_microsoft
drm_suballoc_helper drm_buddy drm_display_helper drm_kms_helper usbhid
sr_mod sd_mod cdrom drm pata_jmicron ahci libahci uhci_hcd xhci_pci
libata ehci_pci ehci_hcd xhci_hcd scsi_mod firewire_ohci psmouse
firewire_core usbcore crc32c_intel sha512_ssse3 sha256_ssse3 bsg
serio_raw sha1_ssse3 drm_panel_orientation_quirks scsi_common crc16
usb_common crc_itu_t wmi msr gf128mul crypto_simd cryptd
[ 4624.932496] CPU: 2 UID: 0 PID: 9069 Comm: kworker/u32:3 Tainted: G
I 6.12.4 #1
[ 4624.939803] Tainted: [I]=FIRMWARE_WORKAROUND
[ 4624.942773] Hardware name: Gigabyte Technology Co., Ltd.
EX58-UD3R/EX58-UD3R, BIOS FB 05/04/2009
[ 4624.950340] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
[ 4624.954967] RIP: 0010:check_flush_dependency+0xbe/0xd0
[ 4624.958806] Code: 75 2a 48 8b 55 18 48 8d 8b c8 00 00 00 4d 89 e0
48 81 c6 c8 00 00 00 48 c7 c7 1b d6 e9 81 c6 05 a3 5f 56 01 01 e8 32
30 fe ff <0f> 0b 5b 5d 41 5c c3 cc cc cc cc 0f 1f 80 00 00 00 00 90 90
90 90
[ 4624.976253] RSP: 0018:ffffc9000bec7c88 EFLAGS: 00010086
[ 4624.980177] RAX: 0000000000000000 RBX: ffff888100118000 RCX: 0000000000000027
[ 4624.986003] RDX: 0000000000000003 RSI: ffffffff81eab2b9 RDI: 00000000ffffffff
[ 4624.991835] RBP: ffff888155daa900 R08: 0000000000000000 R09: 7067646d61006600
[ 4624.997668] R10: 0000000000000091 R11: fefefefefefefeff R12: ffffffffa05ec880
[ 4625.003501] R13: 0000000000000001 R14: ffff88810011c600 R15: ffff888163600000
[ 4625.009334] FS: 0000000000000000(0000) GS:ffff888343c80000(0000)
knlGS:0000000000000000
[ 4625.016118] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 4625.020555] CR2: 0000000099837000 CR3: 0000000144e4c000 CR4: 00000000000026f0
[ 4625.026381] Call Trace:
[ 4625.027525] <TASK>
[ 4625.028323] ? __warn+0x90/0x120
[ 4625.030255] ? report_bug+0xe2/0x160
[ 4625.032532] ? check_flush_dependency+0xbe/0xd0
[ 4625.035768] ? handle_bug+0x53/0x80
[ 4625.037959] ? exc_invalid_op+0x13/0x60
[ 4625.040499] ? asm_exc_invalid_op+0x16/0x20
[ 4625.043384] ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu]
[ 4625.049366] ? check_flush_dependency+0xbe/0xd0
[ 4625.052598] ? check_flush_dependency+0xbe/0xd0
[ 4625.055830] __flush_work+0xb2/0x1f0
[ 4625.058109] ? work_grab_pending+0x3f/0x120
[ 4625.060996] ? set_work_pool_and_clear_pending+0x14/0x20
[ 4625.065008] ? __cancel_work+0x89/0xc0
[ 4625.067460] __cancel_work_sync+0x4a/0x70
[ 4625.070173] amdgpu_gfx_off_ctrl+0xa6/0x100 [amdgpu]
[ 4625.074231] amdgpu_ring_alloc+0x52/0x60 [amdgpu]
[ 4625.077974] amdgpu_ib_schedule+0x155/0x640 [amdgpu]
[ 4625.081988] amdgpu_job_run+0xda/0x140 [amdgpu]
[ 4625.085663] drm_sched_run_job_work+0x246/0x310 [gpu_sched]
[ 4625.089935] process_scheduled_works+0x19c/0x2c0
[ 4625.093252] worker_thread+0x13b/0x1c0
[ 4625.095706] ? __pfx_worker_thread+0x10/0x10
[ 4625.098678] kthread+0xef/0x100
[ 4625.100523] ? __pfx_kthread+0x10/0x10
[ 4625.102976] ret_from_fork+0x24/0x40
[ 4625.105256] ? __pfx_kthread+0x10/0x10
[ 4625.107709] ret_from_fork_asm+0x1a/0x30
[ 4625.110338] </TASK>
[ 4625.111228] ---[ end trace 0000000000000000 ]---
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600
2024-12-13 12:33 [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600 Chris Rankin
@ 2024-12-16 18:36 ` Alex Deucher
2024-12-17 0:34 ` Matthew Brost
0 siblings, 1 reply; 4+ messages in thread
From: Alex Deucher @ 2024-12-16 18:36 UTC (permalink / raw)
To: Chris Rankin, Christian Koenig, Tvrtko Ursulin, Matthew Brost, Tejun Heo
Cc: LKML, amd-gfx
On Fri, Dec 13, 2024 at 7:53 AM Chris Rankin <rankincj@gmail.com> wrote:
>
> Hi,
>
> I've just noticed this warning in my dmesg log. This is a vanilla
> 6.12.4 kernel, with a Radeon RX6600 graphics card.
That was caused by this commit:
commit 746ae46c11137ba21f0c0c68f082a9d8c1222c78
Author: Matthew Brost <matthew.brost@intel.com>
Date: Wed Oct 23 16:59:17 2024 -0700
drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM
drm_gpu_scheduler.submit_wq is used to submit jobs, jobs are in the path
of dma-fences, and dma-fences are in the path of reclaim. Mark scheduler
work queue with WQ_MEM_RECLAIM to ensure forward progress during
reclaim; without WQ_MEM_RECLAIM, work queues cannot make forward
progress during reclaim.
However, after further discussion, I think the warning is actually a
false positive. See this discussion:
https://lists.freedesktop.org/archives/amd-gfx/2024-November/117349.html
From the thread:
"Question is - does check_flush_dependency() need to skip the
!WQ_MEM_RECLAIM flushing WQ_MEM_RECLAIM warning *if* the work is already
running *and* it was called from cancel_delayed_work_sync()?"
Thanks,
Alex
>
> Cheers,
> Chris
>
> [ 4624.741148] ------------[ cut here ]------------
> [ 4624.744474] 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]
> [ 4624.744942] WARNING: CPU: 2 PID: 9069 at kernel/workqueue.c:3704
> check_flush_dependency+0xbe/0xd0
> [ 4624.765285] Modules linked in: snd_seq_dummy rpcrdma rdma_cm iw_cm
> ib_cm ib_core af_packet 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
> ebtable_nat ebtable_broute ip6table_nat ip6table_mangle ip6table_raw
> ip6table_security iptable_nat iptable_mangle iptable_raw
> iptable_security nf_nat_ftp nf_conntrack_ftp nf_nat nf_conntrack
> nf_defrag_ipv6 nf_defrag_ipv4 nf_tables libcrc32c ebtable_filter
> ebtables ip6table_filter ip6_tables iptable_filter ip_tables x_tables
> bnep it87 hwmon_vid binfmt_misc snd_hda_codec_realtek
> snd_hda_codec_generic snd_hda_codec_hdmi snd_hda_scodec_component
> snd_hda_intel uvcvideo btusb uvc videobuf2_vmalloc btintel
> videobuf2_memops videobuf2_v4l2 videodev btbcm snd_usb_audio bluetooth
> snd_intel_dspcfg intel_powerclamp snd_hda_codec videobuf2_common
> coretemp snd_virtuoso snd_usbmidi_lib snd_oxygen_lib snd_ctl_led
> kvm_intel input_leds mc snd_hwdep led_class snd_mpu401_uart
> [ 4624.765400] snd_hda_core joydev snd_rawmidi rfkill kvm snd_seq
> snd_seq_device gpio_ich snd_pcm pktcdvd iTCO_wdt snd_hrtimer r8169
> snd_timer intel_cstate realtek snd mdio_devres intel_uncore libphy
> i2c_i801 soundcore lpc_ich tiny_power_button mxm_wmi i7core_edac
> acpi_cpufreq i2c_smbus pcspkr button nfsd auth_rpcgss nfs_acl lockd
> grace dm_mod fuse sunrpc loop configfs dax nfnetlink zram zsmalloc
> ext4 crc32c_generic mbcache jbd2 amdgpu video amdxcp i2c_algo_bit
> mfd_core drm_ttm_helper ttm drm_exec gpu_sched hid_microsoft
> drm_suballoc_helper drm_buddy drm_display_helper drm_kms_helper usbhid
> sr_mod sd_mod cdrom drm pata_jmicron ahci libahci uhci_hcd xhci_pci
> libata ehci_pci ehci_hcd xhci_hcd scsi_mod firewire_ohci psmouse
> firewire_core usbcore crc32c_intel sha512_ssse3 sha256_ssse3 bsg
> serio_raw sha1_ssse3 drm_panel_orientation_quirks scsi_common crc16
> usb_common crc_itu_t wmi msr gf128mul crypto_simd cryptd
> [ 4624.932496] CPU: 2 UID: 0 PID: 9069 Comm: kworker/u32:3 Tainted: G
> I 6.12.4 #1
> [ 4624.939803] Tainted: [I]=FIRMWARE_WORKAROUND
> [ 4624.942773] Hardware name: Gigabyte Technology Co., Ltd.
> EX58-UD3R/EX58-UD3R, BIOS FB 05/04/2009
> [ 4624.950340] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
> [ 4624.954967] RIP: 0010:check_flush_dependency+0xbe/0xd0
> [ 4624.958806] Code: 75 2a 48 8b 55 18 48 8d 8b c8 00 00 00 4d 89 e0
> 48 81 c6 c8 00 00 00 48 c7 c7 1b d6 e9 81 c6 05 a3 5f 56 01 01 e8 32
> 30 fe ff <0f> 0b 5b 5d 41 5c c3 cc cc cc cc 0f 1f 80 00 00 00 00 90 90
> 90 90
> [ 4624.976253] RSP: 0018:ffffc9000bec7c88 EFLAGS: 00010086
> [ 4624.980177] RAX: 0000000000000000 RBX: ffff888100118000 RCX: 0000000000000027
> [ 4624.986003] RDX: 0000000000000003 RSI: ffffffff81eab2b9 RDI: 00000000ffffffff
> [ 4624.991835] RBP: ffff888155daa900 R08: 0000000000000000 R09: 7067646d61006600
> [ 4624.997668] R10: 0000000000000091 R11: fefefefefefefeff R12: ffffffffa05ec880
> [ 4625.003501] R13: 0000000000000001 R14: ffff88810011c600 R15: ffff888163600000
> [ 4625.009334] FS: 0000000000000000(0000) GS:ffff888343c80000(0000)
> knlGS:0000000000000000
> [ 4625.016118] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 4625.020555] CR2: 0000000099837000 CR3: 0000000144e4c000 CR4: 00000000000026f0
> [ 4625.026381] Call Trace:
> [ 4625.027525] <TASK>
> [ 4625.028323] ? __warn+0x90/0x120
> [ 4625.030255] ? report_bug+0xe2/0x160
> [ 4625.032532] ? check_flush_dependency+0xbe/0xd0
> [ 4625.035768] ? handle_bug+0x53/0x80
> [ 4625.037959] ? exc_invalid_op+0x13/0x60
> [ 4625.040499] ? asm_exc_invalid_op+0x16/0x20
> [ 4625.043384] ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu]
> [ 4625.049366] ? check_flush_dependency+0xbe/0xd0
> [ 4625.052598] ? check_flush_dependency+0xbe/0xd0
> [ 4625.055830] __flush_work+0xb2/0x1f0
> [ 4625.058109] ? work_grab_pending+0x3f/0x120
> [ 4625.060996] ? set_work_pool_and_clear_pending+0x14/0x20
> [ 4625.065008] ? __cancel_work+0x89/0xc0
> [ 4625.067460] __cancel_work_sync+0x4a/0x70
> [ 4625.070173] amdgpu_gfx_off_ctrl+0xa6/0x100 [amdgpu]
> [ 4625.074231] amdgpu_ring_alloc+0x52/0x60 [amdgpu]
> [ 4625.077974] amdgpu_ib_schedule+0x155/0x640 [amdgpu]
> [ 4625.081988] amdgpu_job_run+0xda/0x140 [amdgpu]
> [ 4625.085663] drm_sched_run_job_work+0x246/0x310 [gpu_sched]
> [ 4625.089935] process_scheduled_works+0x19c/0x2c0
> [ 4625.093252] worker_thread+0x13b/0x1c0
> [ 4625.095706] ? __pfx_worker_thread+0x10/0x10
> [ 4625.098678] kthread+0xef/0x100
> [ 4625.100523] ? __pfx_kthread+0x10/0x10
> [ 4625.102976] ret_from_fork+0x24/0x40
> [ 4625.105256] ? __pfx_kthread+0x10/0x10
> [ 4625.107709] ret_from_fork_asm+0x1a/0x30
> [ 4625.110338] </TASK>
> [ 4625.111228] ---[ end trace 0000000000000000 ]---
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600
2024-12-16 18:36 ` Alex Deucher
@ 2024-12-17 0:34 ` Matthew Brost
2024-12-18 18:08 ` Tejun Heo
0 siblings, 1 reply; 4+ messages in thread
From: Matthew Brost @ 2024-12-17 0:34 UTC (permalink / raw)
To: Alex Deucher
Cc: Chris Rankin, Christian Koenig, Tvrtko Ursulin, Tejun Heo, LKML, amd-gfx
On Mon, Dec 16, 2024 at 01:36:29PM -0500, Alex Deucher wrote:
> On Fri, Dec 13, 2024 at 7:53 AM Chris Rankin <rankincj@gmail.com> wrote:
> >
> > Hi,
> >
> > I've just noticed this warning in my dmesg log. This is a vanilla
> > 6.12.4 kernel, with a Radeon RX6600 graphics card.
>
> That was caused by this commit:
>
> commit 746ae46c11137ba21f0c0c68f082a9d8c1222c78
> Author: Matthew Brost <matthew.brost@intel.com>
> Date: Wed Oct 23 16:59:17 2024 -0700
>
> drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM
>
> drm_gpu_scheduler.submit_wq is used to submit jobs, jobs are in the path
> of dma-fences, and dma-fences are in the path of reclaim. Mark scheduler
> work queue with WQ_MEM_RECLAIM to ensure forward progress during
> reclaim; without WQ_MEM_RECLAIM, work queues cannot make forward
> progress during reclaim.
>
> However, after further discussion, I think the warning is actually a
> false positive. See this discussion:
> https://lists.freedesktop.org/archives/amd-gfx/2024-November/117349.html
>
> From the thread:
> "Question is - does check_flush_dependency() need to skip the
> !WQ_MEM_RECLAIM flushing WQ_MEM_RECLAIM warning *if* the work is already
> running *and* it was called from cancel_delayed_work_sync()?"
>
See my reply just now [1] — I’m going to have to disagree with AMD's
assessment, but I’m not certain.
Again, I believe Tejun is the authority here.
Matt
[1] https://lore.kernel.org/all/154641d9-be2a-4018-af5e-a57dbffb45d5@igalia.com/T/#ma1ed4a99d9ad1a05f8d4648dd979d7c9d93591ff
> Thanks,
>
> Alex
>
>
> >
> > Cheers,
> > Chris
> >
> > [ 4624.741148] ------------[ cut here ]------------
> > [ 4624.744474] 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]
> > [ 4624.744942] WARNING: CPU: 2 PID: 9069 at kernel/workqueue.c:3704
> > check_flush_dependency+0xbe/0xd0
> > [ 4624.765285] Modules linked in: snd_seq_dummy rpcrdma rdma_cm iw_cm
> > ib_cm ib_core af_packet 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
> > ebtable_nat ebtable_broute ip6table_nat ip6table_mangle ip6table_raw
> > ip6table_security iptable_nat iptable_mangle iptable_raw
> > iptable_security nf_nat_ftp nf_conntrack_ftp nf_nat nf_conntrack
> > nf_defrag_ipv6 nf_defrag_ipv4 nf_tables libcrc32c ebtable_filter
> > ebtables ip6table_filter ip6_tables iptable_filter ip_tables x_tables
> > bnep it87 hwmon_vid binfmt_misc snd_hda_codec_realtek
> > snd_hda_codec_generic snd_hda_codec_hdmi snd_hda_scodec_component
> > snd_hda_intel uvcvideo btusb uvc videobuf2_vmalloc btintel
> > videobuf2_memops videobuf2_v4l2 videodev btbcm snd_usb_audio bluetooth
> > snd_intel_dspcfg intel_powerclamp snd_hda_codec videobuf2_common
> > coretemp snd_virtuoso snd_usbmidi_lib snd_oxygen_lib snd_ctl_led
> > kvm_intel input_leds mc snd_hwdep led_class snd_mpu401_uart
> > [ 4624.765400] snd_hda_core joydev snd_rawmidi rfkill kvm snd_seq
> > snd_seq_device gpio_ich snd_pcm pktcdvd iTCO_wdt snd_hrtimer r8169
> > snd_timer intel_cstate realtek snd mdio_devres intel_uncore libphy
> > i2c_i801 soundcore lpc_ich tiny_power_button mxm_wmi i7core_edac
> > acpi_cpufreq i2c_smbus pcspkr button nfsd auth_rpcgss nfs_acl lockd
> > grace dm_mod fuse sunrpc loop configfs dax nfnetlink zram zsmalloc
> > ext4 crc32c_generic mbcache jbd2 amdgpu video amdxcp i2c_algo_bit
> > mfd_core drm_ttm_helper ttm drm_exec gpu_sched hid_microsoft
> > drm_suballoc_helper drm_buddy drm_display_helper drm_kms_helper usbhid
> > sr_mod sd_mod cdrom drm pata_jmicron ahci libahci uhci_hcd xhci_pci
> > libata ehci_pci ehci_hcd xhci_hcd scsi_mod firewire_ohci psmouse
> > firewire_core usbcore crc32c_intel sha512_ssse3 sha256_ssse3 bsg
> > serio_raw sha1_ssse3 drm_panel_orientation_quirks scsi_common crc16
> > usb_common crc_itu_t wmi msr gf128mul crypto_simd cryptd
> > [ 4624.932496] CPU: 2 UID: 0 PID: 9069 Comm: kworker/u32:3 Tainted: G
> > I 6.12.4 #1
> > [ 4624.939803] Tainted: [I]=FIRMWARE_WORKAROUND
> > [ 4624.942773] Hardware name: Gigabyte Technology Co., Ltd.
> > EX58-UD3R/EX58-UD3R, BIOS FB 05/04/2009
> > [ 4624.950340] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
> > [ 4624.954967] RIP: 0010:check_flush_dependency+0xbe/0xd0
> > [ 4624.958806] Code: 75 2a 48 8b 55 18 48 8d 8b c8 00 00 00 4d 89 e0
> > 48 81 c6 c8 00 00 00 48 c7 c7 1b d6 e9 81 c6 05 a3 5f 56 01 01 e8 32
> > 30 fe ff <0f> 0b 5b 5d 41 5c c3 cc cc cc cc 0f 1f 80 00 00 00 00 90 90
> > 90 90
> > [ 4624.976253] RSP: 0018:ffffc9000bec7c88 EFLAGS: 00010086
> > [ 4624.980177] RAX: 0000000000000000 RBX: ffff888100118000 RCX: 0000000000000027
> > [ 4624.986003] RDX: 0000000000000003 RSI: ffffffff81eab2b9 RDI: 00000000ffffffff
> > [ 4624.991835] RBP: ffff888155daa900 R08: 0000000000000000 R09: 7067646d61006600
> > [ 4624.997668] R10: 0000000000000091 R11: fefefefefefefeff R12: ffffffffa05ec880
> > [ 4625.003501] R13: 0000000000000001 R14: ffff88810011c600 R15: ffff888163600000
> > [ 4625.009334] FS: 0000000000000000(0000) GS:ffff888343c80000(0000)
> > knlGS:0000000000000000
> > [ 4625.016118] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [ 4625.020555] CR2: 0000000099837000 CR3: 0000000144e4c000 CR4: 00000000000026f0
> > [ 4625.026381] Call Trace:
> > [ 4625.027525] <TASK>
> > [ 4625.028323] ? __warn+0x90/0x120
> > [ 4625.030255] ? report_bug+0xe2/0x160
> > [ 4625.032532] ? check_flush_dependency+0xbe/0xd0
> > [ 4625.035768] ? handle_bug+0x53/0x80
> > [ 4625.037959] ? exc_invalid_op+0x13/0x60
> > [ 4625.040499] ? asm_exc_invalid_op+0x16/0x20
> > [ 4625.043384] ? __pfx_amdgpu_device_delay_enable_gfx_off+0x10/0x10 [amdgpu]
> > [ 4625.049366] ? check_flush_dependency+0xbe/0xd0
> > [ 4625.052598] ? check_flush_dependency+0xbe/0xd0
> > [ 4625.055830] __flush_work+0xb2/0x1f0
> > [ 4625.058109] ? work_grab_pending+0x3f/0x120
> > [ 4625.060996] ? set_work_pool_and_clear_pending+0x14/0x20
> > [ 4625.065008] ? __cancel_work+0x89/0xc0
> > [ 4625.067460] __cancel_work_sync+0x4a/0x70
> > [ 4625.070173] amdgpu_gfx_off_ctrl+0xa6/0x100 [amdgpu]
> > [ 4625.074231] amdgpu_ring_alloc+0x52/0x60 [amdgpu]
> > [ 4625.077974] amdgpu_ib_schedule+0x155/0x640 [amdgpu]
> > [ 4625.081988] amdgpu_job_run+0xda/0x140 [amdgpu]
> > [ 4625.085663] drm_sched_run_job_work+0x246/0x310 [gpu_sched]
> > [ 4625.089935] process_scheduled_works+0x19c/0x2c0
> > [ 4625.093252] worker_thread+0x13b/0x1c0
> > [ 4625.095706] ? __pfx_worker_thread+0x10/0x10
> > [ 4625.098678] kthread+0xef/0x100
> > [ 4625.100523] ? __pfx_kthread+0x10/0x10
> > [ 4625.102976] ret_from_fork+0x24/0x40
> > [ 4625.105256] ? __pfx_kthread+0x10/0x10
> > [ 4625.107709] ret_from_fork_asm+0x1a/0x30
> > [ 4625.110338] </TASK>
> > [ 4625.111228] ---[ end trace 0000000000000000 ]---
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600
2024-12-17 0:34 ` Matthew Brost
@ 2024-12-18 18:08 ` Tejun Heo
0 siblings, 0 replies; 4+ messages in thread
From: Tejun Heo @ 2024-12-18 18:08 UTC (permalink / raw)
To: Matthew Brost
Cc: Alex Deucher, Chris Rankin, Christian Koenig, Tvrtko Ursulin,
LKML, amd-gfx
Hello, sorry about the delay.
On Mon, Dec 16, 2024 at 04:34:00PM -0800, Matthew Brost wrote:
> > However, after further discussion, I think the warning is actually a
> > false positive. See this discussion:
> > https://lists.freedesktop.org/archives/amd-gfx/2024-November/117349.html
> >
> > From the thread:
> > "Question is - does check_flush_dependency() need to skip the
> > !WQ_MEM_RECLAIM flushing WQ_MEM_RECLAIM warning *if* the work is already
> > running *and* it was called from cancel_delayed_work_sync()?"
> >
>
> See my reply just now [1] — I’m going to have to disagree with AMD's
> assessment, but I’m not certain.
>
> Again, I believe Tejun is the authority here.
I think we can skip the warning if the flushing is coming from
cancel*_work_sync() as flush takes place iff the work item already has a
worker running - ie. it can't be blocked from lack of memory. Tvrtko, can
you write up a patch to exclude the condition from check_flush_dependency()?
I think it can just skip check_flush_dependency() when @from_cancel is set.
Taking a step back, if an actual dependency develops in the future - memory
reclaim actually blocking on gpu work items, one way to handle that would be
adding subsystem-wide workqueues so that the rescuer can be shared across
GPU drivers / devices. As long as they don't depend on each other for making
forward progress, which they most likely wouldn't, sharing a rescuer across
them is completely fine.
Thanks.
--
tejun
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2024-12-18 18:09 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-12-13 12:33 [WARNING][AMDGPU] WQ_MEM_RECLAIM with Radeon RX 6600 Chris Rankin
2024-12-16 18:36 ` Alex Deucher
2024-12-17 0:34 ` Matthew Brost
2024-12-18 18:08 ` Tejun Heo
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®