Fedora Account System
Red Hat Associate
Red Hat Customer
Created attachment 1788039 [details] dmesg from kernel 5.12.7 showing panfrost GPU faults 1. Please describe the problem: On a Pinebook Pro machine, the following messages are repeated in the kernel log: May 29 22:49:51 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000015600000 Reason: TODO raw fault status: 0x660003C2 decoded fault status: SLAVE FAULT exception type 0xC2: TRANSLATION_FAULT_LEVEL2 access type 0x3: WRITE source id 0x6600 ... May 29 22:49:51 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x0000000008000000 Reason: TODO raw fault status: 0x660003C3 decoded fault status: SLAVE FAULT exception type 0xC3: TRANSLATION_FAULT_LEVEL3 access type 0x3: WRITE source id 0x6600 May 29 22:49:52 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=1, config=0x3300, status=0x8, head=0x53e7000, tail> May 29 22:49:52 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x00000000A315FD80 Reason: TODO raw fault status: 0xF002C1 decoded fault status: SLAVE FAULT exception type 0xC1: TRANSLATION_FAULT_LEVEL1 access type 0x2: READ source id 0xF0 May 29 22:49:52 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=1, config=0x3301, status=0x8, head=0x59d0500, tail> May 29 22:49:52 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x51e3680, tail> May 29 22:49:52 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x00000000A315FC00 Reason: TODO raw fault status: 0xF002C1 decoded fault status: SLAVE FAULT exception type 0xC1: TRANSLATION_FAULT_LEVEL1 access type 0x2: READ source id 0xF0 May 29 22:49:53 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x77af0c0, tail> May 29 22:49:53 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS1 at VA 0x0000000015E00000 Reason: TODO raw fault status: 0x660003C2 decoded fault status: SLAVE FAULT exception type 0xC2: TRANSLATION_FAULT_LEVEL2 access type 0x3: WRITE source id 0x6600 ... May 29 22:49:53 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=1, config=0x3301, status=0x8, head=0x77cd000, tail> May 29 22:49:53 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x00000000A315FC00 Reason: TODO raw fault status: 0xF002C1 decoded fault status: SLAVE FAULT exception type 0xC1: TRANSLATION_FAULT_LEVEL1 access type 0x2: READ source id 0xF0 May 29 22:49:54 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=1, config=0x3301, status=0x8, head=0x5993000, tail> May 29 22:49:54 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS0 at VA 0x00000000A315FC00 Reason: TODO raw fault status: 0xF002C1 decoded fault status: SLAVE FAULT exception type 0xC1: TRANSLATION_FAULT_LEVEL1 access type 0x2: READ source id 0xF0 May 29 22:49:54 kernel: panfrost ff9a0000.gpu: gpu sched timeout, js=0, config=0x3300, status=0x8, head=0x37d10c0, tail> May 29 22:49:54 kernel: panfrost ff9a0000.gpu: Unhandled Page fault in AS1 at VA 0x00000000A315FC00 Reason: TODO raw fault status: 0xF002C1 decoded fault status: SLAVE FAULT exception type 0xC1: TRANSLATION_FAULT_LEVEL1 access type 0x2: READ source id 0xF0 and similar. 2. What is the Version-Release number of the kernel: 5.12.7-300.fc34.aarch64 3. Did it work previously in Fedora? If so, what kernel version did the issue *first* appear? Old kernels are available for download at https://koji.fedoraproject.org/koji/packageinfo?packageID=8 : I think this started with Mesa upgrade to 21.1.1. 4. Can you reproduce this issue? If so, please provide the steps to reproduce the issue below: 1. Enable OpenGL compositing in Firefox, i.e. the following preferences: gfx.canvas.azure.accelerated true gfx.xrender.enabled true layers.accelerate-all true layers.acceleration.force-enabled true webgl.out-of-process true 2. Go to https://njumobile.pl/ 5. Does this problem occur with the latest Rawhide kernel? To install the Rawhide kernel, run ``sudo dnf install fedora-repos-rawhide`` followed by ``sudo dnf update --enablerepo=rawhide kernel``: I haven't tried. 6. Are you running any modules that not shipped with directly Fedora's kernel?: No. 7. Please attach the kernel logs. You can get the complete kernel log for a boot with ``journalctl --no-hostname -k > dmesg.txt``. If the issue occurred on a previous boot, use the journalctl ``-b`` flag.
Please use abrt or similar to get a full log with debug symbols
I'm pretty sure mesa-21.1.1 is the cause/trigger as downgrading to mesa-21.0.2 makes the issue stop. abrt is not catching anything, any hints on how to get a full log? Should I run a rawhide kernel? I got a kernel WARNING this time, too: May 31 12:30:53 kernel: ------------[ cut here ]------------ May 31 12:30:53 kernel: Memory manager not clean during takedown. May 31 12:30:53 kernel: WARNING: CPU: 4 PID: 7032 at drivers/gpu/drm/drm_mm.c:998 drm_mm_takedown+0x34/0x44 [drm] May 31 12:30:53 kernel: Modules linked in: rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace nfs_ssc fscache rfcomm snd_seq_dummy snd_hrtimer nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_f> May 31 12:30:53 kernel: videobuf2_memops industrialio_triggered_buffer snd_pcm brcmutil videobuf2_v4l2 kfifo_buf cfg80211 videobuf2_common nvmem_rockchip_efuse videodev mc snd_timer snd rfkill indus> May 31 12:30:53 kernel: CPU: 4 PID: 7032 Comm: JS Helper Tainted: G WC 5.12.7-300.fc34.aarch64 #1 May 31 12:30:53 kernel: Hardware name: /, BIOS 2021.04 04/28/2021 May 31 12:30:53 kernel: pstate: 40400005 (nZcv daif +PAN -UAO -TCO BTYPE=--) May 31 12:30:53 kernel: pc : drm_mm_takedown+0x34/0x44 [drm] May 31 12:30:53 kernel: lr : drm_mm_takedown+0x34/0x44 [drm] May 31 12:30:53 kernel: sp : ffff8000132cbb80 May 31 12:30:53 kernel: x29: ffff8000132cbb80 x28: ffff30dcbdf70000 May 31 12:30:53 kernel: x27: ffff30dc85b76ca8 x26: ffffa2abdc402000 May 31 12:30:53 kernel: x25: ffff30dc85b76cf0 x24: ffff30dc81d8c800 May 31 12:30:53 kernel: x23: 0000000000000000 x22: ffff30dc81d8c800 May 31 12:30:53 kernel: x21: ffff30dc85b76c00 x20: ffff30dc85b76cc8 May 31 12:30:53 kernel: x19: ffff30dcf0946c00 x18: 0000000000000000 May 31 12:30:53 kernel: x17: 0000000000000000 x16: ffffa2abdb7a0a54 May 31 12:30:53 kernel: x15: 0000000000000040 x14: 0000000000000000 May 31 12:30:53 kernel: x13: 0000000000000040 x12: ffff30dc91b756d0 May 31 12:30:53 kernel: x11: ffffa2abdc9fcc98 x10: 00000000ffffe000 May 31 12:30:53 kernel: x9 : ffffa2abda9e0d80 x8 : 00000000ffffdfff May 31 12:30:53 kernel: x7 : ffffa2abdc9fcc98 x6 : 0000000000000001 May 31 12:30:53 kernel: x5 : ffff30dd76ef3148 x4 : 0000000000000000 May 31 12:30:53 kernel: x3 : 0000000000000027 x2 : 0000000000000023 May 31 12:30:53 kernel: x1 : ffff30dd76ef3150 x0 : 0000000000000029 May 31 12:30:53 kernel: Call trace: May 31 12:30:53 kernel: drm_mm_takedown+0x34/0x44 [drm] May 31 12:30:53 kernel: panfrost_postclose+0x40/0x5c [panfrost] May 31 12:30:53 kernel: drm_file_free.part.0+0x1ac/0x250 [drm] May 31 12:30:53 kernel: drm_close_helper.isra.0+0x74/0x84 [drm] May 31 12:30:53 kernel: drm_release+0x78/0x154 [drm] May 31 12:30:53 kernel: __fput+0x88/0x244 May 31 12:30:53 kernel: ____fput+0x1c/0x30 May 31 12:30:53 kernel: task_work_run+0xcc/0x22c May 31 12:30:53 kernel: do_exit+0x1cc/0x460 May 31 12:30:53 kernel: do_group_exit+0x44/0xac May 31 12:30:53 kernel: get_signal+0x1e4/0x940 May 31 12:30:53 kernel: do_signal+0x84/0x270 May 31 12:30:53 kernel: do_notify_resume+0xe0/0x390 May 31 12:30:53 kernel: work_pending+0xc/0x498 May 31 12:30:53 kernel: ---[ end trace b5135138d9d2c413 ]--- May 31 12:31:28 kernel: panfrost_gem_shrinker_scan: 60 callbacks suppressed
It seems to be fixed in mesa main branch. I've just built 21.2.0-devel (git 234e1b7) and I'm unable to reproduce this anymore. Reassigning to mesa, then.
Upstream commit https://gitlab.freedesktop.org/mesa/mesa/-/commit/a89bc59980b3ea7b2f03d2994bae7dda689f637f looks relevant here. I'm going to try building 21.1.1 with that patch applied and report back.
Sadly, the patch depends on other patches not present in 21.1.1.
Upstream fix: https://gitlab.freedesktop.org/mesa/mesa/-/commit/fe9d37b0c6e89f11a5f25022a851da81d19dab73.patch . Scratch build with that patch fixes the issue: https://koji.fedoraproject.org/koji/taskinfo?taskID=69050219 . Upstream says fix will be included in the next 21.1.x release.
FEDORA-2021-0ec322843a has been submitted as an update to Fedora 34. https://bodhi.fedoraproject.org/updates/FEDORA-2021-0ec322843a
FEDORA-2021-0ec322843a has been pushed to the Fedora 34 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2021-0ec322843a` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2021-0ec322843a See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
FEDORA-2021-0ec322843a has been pushed to the Fedora 34 stable repository. If problem still persists, please make note of it in this bug report.