Project 'drm/intel' was moved to 'drm/i915/kernel'. Please update any links and bookmarks that may still have the old path.
[i915] GPU HANG: ecode 7:1:8edcfc79, in systemd-logind [411]
I'm using Archlinux. After upgrading to linux-5.7.2, the GPU hangs right after Xorg starts. The system becomes unresponsive and visual artifacts are visible (the screen starts blinking randomly).
The issue is not reproducible with linux-lts-5.4.46 (from which I'm currently writing this message). And was also not reproducible with linux-5.6.15, which makes me believe it's something about linux-5.7.x that is not compatible with i915. (??)
Apologies if this has been reported before.
Activity
It's reporting a hang from the wa_ctx.
47f8253d2b89 ("drm/i915/gen7: Clear all EU/L3 residual contexts")
which is new for 5.7, but should have been in the stable backports.
Edited by Chris WilsonMy bisect log:
git bisect start # bad: [cc328351e1e44c4373fabb6f98f0b5d636890225] drm/i915/gt: Wait for the wa batch to be pinned git bisect bad cc328351e1e44c4373fabb6f98f0b5d636890225 # good: [50148a25f841dc1e2cffe1669d4782dfa75d85d6] drm/i915/tgl: Move and restrict Wa_1408615072 git bisect good 50148a25f841dc1e2cffe1669d4782dfa75d85d6 # good: [c834d03cca61bf88aa8c2f9524360cce484686ac] drm/i915: Implement display w/a 1140 for glk/cnl git bisect good c834d03cca61bf88aa8c2f9524360cce484686ac # bad: [26fc4e4ba1605e365482a86ec8f33d91552db271] drm/i915: Always propagate the invocation to i915_schedule git bisect bad 26fc4e4ba1605e365482a86ec8f33d91552db271 # good: [ee2413eeed767d1fb794651e63e45c5e292c8180] drm/i915: Add mechanism to submit a context WA on ring submission git bisect good ee2413eeed767d1fb794651e63e45c5e292c8180 # bad: [64dc802aa5accfe36d431814ffdf15303b7f7b2b] drm/i915/phys: unconditionally call release_memory_region git bisect bad 64dc802aa5accfe36d431814ffdf15303b7f7b2b # bad: [47f8253d2b8947d79fd3196bf96c1959c0f25f20] drm/i915/gen7: Clear all EU/L3 residual contexts git bisect bad 47f8253d2b8947d79fd3196bf96c1959c0f25f20 # first bad commit: [47f8253d2b8947d79fd3196bf96c1959c0f25f20] drm/i915/gen7: Clear all EU/L3 residual contextsCherries required (in Arch):
dbe748cd3af4a7c264a94e3e7c56a084dbd0164d e33a814e772cdc36436c8c188d8c42d019fda639 78a5255ffb6a1af189a83e493d916ba1c54d8c75 df6d4f9db79c1a5d6f48b59db35ccd1e9ff9adfc a9a3ed1eff3601b63aea4fb462d8b3b92c7c1e7eEdited by Tom Yan
added GPU hang platform: HSW severity::critical labels
assigned to @baleboy
If you can, please try drm-tip as that's just landed an important fix for Haswell. If that is still failing, please try
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c index 68a08486fc87..51545897f1ec 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c @@ -1277,7 +1277,7 @@ int intel_ring_submission_setup(struct intel_engine_cs *engine) GEM_BUG_ON(timeline->hwsp_ggtt != engine->status_page.vma); - if (IS_HASWELL(engine->i915) && engine->class == RENDER_CLASS) { + if (IS_HASWELL(engine->i915) && engine->class == RENDER_CLASS && 0) { err = gen7_ctx_switch_bb_init(engine); if (err) goto err_ring_unpin;marked #2032 (closed) as a duplicate of this issue
I was getting the same error code in most 3D accelerated programs like Blender or even glxgears. Using drm-tip commit 999bc17a2471df17a3af3001d094cf6d5d4849b0 made the system freeze completely when launching those programs and I was unable to even switch to console to save a GPU crash dump. But recompiling with
&& 0seems to have prevented the errors.marked #2063 (closed) as a duplicate of this issue
Not sure if it helps, but here is the dmesg when booting without sddm:
[ 4.118714] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 4.126643] [drm] Initialized i915 1.6.0 20200313 for 0000:00:02.0 on minor 0 [ 4.128264] ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no) [ 4.128557] input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input15 [ 4.128691] snd_hda_intel 0000:00:03.0: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915]) [ 4.168197] fbcon: i915drmfb (fb0) is primary device [ 4.183571] usbcore: registered new interface driver btusb [ 4.188042] Bluetooth: hci0: RTL: examining hci_ver=06 hci_rev=000a lmp_ver=06 lmp_subver=8821 [ 4.189036] Bluetooth: hci0: RTL: rom_version status=0 version=1 [ 4.189037] Bluetooth: hci0: RTL: loading rtl_bt/rtl8821a_fw.bin [ 4.191683] Bluetooth: hci0: RTL: loading rtl_bt/rtl8821a_config.bin [ 4.191701] bluetooth hci0: Direct firmware load for rtl_bt/rtl8821a_config.bin failed with error -2 [ 4.191705] Bluetooth: hci0: RTL: cfg_sz -2, total sz 17428 [ 4.238748] Console: switching to colour frame buffer device 240x67 [ 4.242538] audit: type=1130 audit(1592506033.358:10): pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=systemd-rfkill comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' [ 4.258561] i915 0000:00:02.0: fb0: i915drmfb frame buffer deviceHere is the link for all of this dmesg.
I also use an haswell cpu (celeron G1820) and can't start X11 session of KDE Plasma on Arch Linux after upgrade to kernel 5.7.2. After I enter my password in login manager (SDDM), I only see a black screen with cursor during ~2 minutes, then my system goes back to login screen.
Edited by Strangiato
FWIW, I am also getting this issue on a Sandybridge CPU/GPU (Intel i5-2520M, Intel HD 3000). I'm on kernel 5.7.3 on Arch Linux, but I am not able to make the module crash (though it managed to reset once with 5.7.2). It just hangs the Xorg process to become an unkillable process (even with
-9).I also noticed this new release (5.7) has caused frequent artifacting when playing video through VA-API on MPV. Here's a video of the artifacting if that helps. I can reproduce the issue even with the compositor turned off, it just becomes less visible.
Edit: None of these are an issue with kernel 5.6.15
Edit2: I managed to track down the
ecodefrom the journal log. Here's what it showed:Jun 13 20:52:48 halifax kernel: [drm:fw_domains_get [i915]] *ERROR* render: timed out waiting for forcewake ack request. Jun 13 20:52:48 halifax kernel: ------------[ cut here ]------------ Jun 13 20:52:48 halifax kernel: i915 0000:00:02.0: GT thread status wait timed out Jun 13 20:52:48 halifax kernel: WARNING: CPU: 2 PID: 82943 at drivers/gpu/drm/i915/intel_uncore.c:327 __gen6_gt_wait_for_thread_c0+0x84/0x90 [i915] Jun 13 20:52:48 halifax kernel: Modules linked in: hid_generic usbhid hid vhost_net vhost tap vhost_iotlb nls_iso8859_1 nls_cp437 vfat fat fuse rfcomm acpi_call(OE) cmac algif_hash algif_skcipher af_alg bnep btusb btrtl btbcm btintel bluetooth uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_common ecdh_generic videodev ecc mc xt_CHECKSUM xt_MASQUERADE xt_conntrack ipt_REJECT nf_reject_ipv4 xt_tcpudp ip6table_mangle ip6table_nat iptable_mangle iptable_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter tun bridge stp llc intel_rapl_msr intel_rapl_common x86_pkg_temp_thermal intel_powerclamp kvm_intel kvm qmi_wwan cdc_wdm usbnet mii qcserial usb_wwan joydev mousedev snd_hda_codec_hdmi snd_hda_codec_conexant snd_hda_codec_generic irqbypass wl(POE) crct10dif_pclmul crc32_pclmul ghash_clmulni_intel i915 iTCO_wdt mei_wdt mei_hdcp aesni_intel iTCO_vendor_support snd_hda_intel wmi_bmof crypto_simd cryptd i2c_algo_bit Jun 13 20:52:48 halifax kernel: glue_helper intel_cstate snd_intel_dspcfg intel_uncore drm_kms_helper snd_hda_codec cec rc_core intel_rapl_perf psmouse input_leds e1000e snd_hda_core i2c_i801 intel_gtt cfg80211 mei_me syscopyarea sysfillrect sysimgblt mei fb_sys_fops lpc_ich snd_hwdep wmi thinkpad_acpi tpm_tis tpm_tis_core tpm nvram ledtrig_audio battery ac rfkill rng_core evdev mac_hid tp_smapi(OE) thinkpad_ec(OE) coretemp msr sg crypto_user snd_seq_dummy snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_oss snd_seq_midi_event snd_seq drm snd_seq_device snd_timer snd soundcore agpgart ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 serio_raw atkbd libps2 crc32c_intel sdhci_pci cqhci sdhci mmc_core ehci_pci ehci_hcd i8042 serio Jun 13 20:52:48 halifax kernel: CPU: 2 PID: 82943 Comm: kworker/2:1 Tainted: P W IOE 5.7.2-arch1-1 #1 Jun 13 20:52:48 halifax kernel: Hardware name: LENOVO 4290CTO/4290CTO, BIOS 8DET76WW (1.46 ) 06/21/2018 Jun 13 20:52:48 halifax kernel: Workqueue: events retire_work_handler [i915] Jun 13 20:52:48 halifax kernel: RIP: 0010:__gen6_gt_wait_for_thread_c0+0x84/0x90 [i915] Jun 13 20:52:48 halifax kernel: Code: 21 00 01 48 8b 78 18 4c 8b 67 50 4d 85 e4 75 03 4c 8b 27 e8 4e 8e 06 e0 4c 89 e2 48 c7 c7 e8 db d9 c0 48 89 c6 e8 e3 2e a5 df <0f> 0b eb bc 0f 1f 84 00 00 00 00 00 66 66 66 66 90 89 f0 41 54 48 Jun 13 20:52:48 halifax kernel: RSP: 0018:ffffafdbc0f6fc78 EFLAGS: 00010086 Jun 13 20:52:48 halifax kernel: RAX: 0000000000000000 RBX: ffffa251079407b8 RCX: 0000000000000000 Jun 13 20:52:48 halifax kernel: RDX: 0000000000000003 RSI: 0000000000000082 RDI: 00000000ffffffff Jun 13 20:52:48 halifax kernel: RBP: 00000e6faa521c0f R08: 0000000000000670 R09: 0000000000000001 Jun 13 20:52:48 halifax kernel: R10: 0000000000000000 R11: 0000000000000001 R12: ffffa25114674240 Jun 13 20:52:48 halifax kernel: R13: 0000000000000206 R14: ffffa251079407d0 R15: ffffa250e20de308 Jun 13 20:52:48 halifax kernel: FS: 0000000000000000(0000) GS:ffffa25116280000(0000) knlGS:0000000000000000 Jun 13 20:52:48 halifax kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Jun 13 20:52:48 halifax kernel: CR2: 00005626274fe4b8 CR3: 000000016020a006 CR4: 00000000000626e0 Jun 13 20:52:48 halifax kernel: Call Trace: Jun 13 20:52:48 halifax kernel: intel_uncore_forcewake_get+0x4c/0xd0 [i915] Jun 13 20:52:48 halifax kernel: gen6_bsd_submit_request+0x20/0x90 [i915] Jun 13 20:52:48 halifax kernel: submit_notify+0x5e/0x9c [i915] Jun 13 20:52:48 halifax kernel: __i915_sw_fence_complete+0x3c/0x1b0 [i915] Jun 13 20:52:48 halifax kernel: __engine_park+0xe5/0x1f0 [i915] Jun 13 20:52:48 halifax kernel: ____intel_wakeref_put_last+0x1b/0x50 [i915] Jun 13 20:52:48 halifax kernel: i915_active_acquire_barrier+0xf0/0x100 [i915] Jun 13 20:52:48 halifax kernel: intel_context_unpin+0x48/0xb0 [i915] Jun 13 20:52:48 halifax kernel: i915_request_retire+0x18a/0x330 [i915] Jun 13 20:52:48 halifax kernel: intel_gt_retire_requests_timeout+0x233/0x500 [i915] Jun 13 20:52:48 halifax kernel: process_one_work+0x1da/0x3d0 Jun 13 20:52:48 halifax kernel: worker_thread+0x4d/0x3e0 Jun 13 20:52:48 halifax kernel: ? rescuer_thread+0x3f0/0x3f0 Jun 13 20:52:48 halifax kernel: kthread+0x13e/0x160 Jun 13 20:52:48 halifax kernel: ? __kthread_bind_mask+0x60/0x60 Jun 13 20:52:48 halifax kernel: ret_from_fork+0x35/0x40 Jun 13 20:52:48 halifax kernel: ---[ end trace fddf503f57a5bc9e ]--- Jun 13 20:52:53 halifax kernel: [drm:fw_domains_get [i915]] *ERROR* render: timed out waiting for forcewake ack request. Jun 13 20:52:53 halifax kernel: i915 0000:00:02.0: GPU HANG: ecode 6:1:7a000fb9, in mpv/vo [84824] Jun 13 20:52:53 halifax kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jun 13 20:52:53 halifax kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jun 13 20:52:53 halifax kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jun 13 20:52:53 halifax kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jun 13 20:52:53 halifax kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jun 13 20:52:53 halifax kernel: GPU crash dump saved to /sys/class/drm/card0/error Jun 13 20:52:53 halifax kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0Edited by Angelo Geulinhaving same issue with kernel 5.7.2 on arch asus chromebox CN60 Intel Celeron 2955U gpu-error
One more log.
Operating System: Arch Linux KDE Plasma Version: 5.18.5 KDE Frameworks Version: 5.70.0 Qt Version: 5.15.0 Kernel Version (that does NOT hang): 5.6.15-arch1-1 OS Type: 64-bit Processors: 2 × Intel® Pentium® CPU G3258 @ 3.20GHz Memory: 15.3 GiB of RAM
/sys/class/drm/card0/errorreports:No error state collectedHere is excerpt (full journal linked also). Note on this install the HANG occurred in sddm-greeter:
Jun 23 17:10:09 archmain sddm[595]: Message received from greeter: Connect Jun 23 17:10:09 archmain sddm-greeter[647]: Loading file:///usr/share/sddm/themes/maldives/Main.qml... Jun 23 17:10:09 archmain sddm-greeter[647]: file:///usr/lib/qt/qml/SddmComponents/LayoutBox.qml:35:5: QML Connections: Implicitly defined onFoo properties in Connections are deprecated. Use this syntax instead: function onFoo(<arguments>) { ... } Jun 23 17:10:09 archmain sddm-greeter[647]: file:///usr/share/sddm/themes/maldives/Main.qml:40:5: QML Connections: Implicitly defined onFoo properties in Connections are deprecated. Use this syntax instead: function onFoo(<arguments>) { ... } Jun 23 17:10:09 archmain sddm-greeter[647]: Adding view for "HDMI1" QRect(0,0 1920x1080) Jun 23 17:10:09 archmain sddm-greeter[647]: Loading file:///usr/share/sddm/themes/maldives/Main.qml... Jun 23 17:10:09 archmain sddm-greeter[647]: file:///usr/lib/qt/qml/SddmComponents/LayoutBox.qml:35:5: QML Connections: Implicitly defined onFoo properties in Connections are deprecated. Use this syntax instead: function onFoo(<arguments>) { ... } Jun 23 17:10:09 archmain sddm-greeter[647]: file:///usr/share/sddm/themes/maldives/Main.qml:40:5: QML Connections: Implicitly defined onFoo properties in Connections are deprecated. Use this syntax instead: function onFoo(<arguments>) { ... } Jun 23 17:10:09 archmain sddm-greeter[647]: Adding view for "HDMI2" QRect(0,0 1920x1080) Jun 23 17:10:15 archmain kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in sddm-greeter [647] Jun 23 17:10:15 archmain kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jun 23 17:10:15 archmain kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jun 23 17:10:15 archmain kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jun 23 17:10:15 archmain kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jun 23 17:10:15 archmain kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jun 23 17:10:15 archmain kernel: GPU crash dump saved to /sys/class/drm/card0/error Jun 23 17:10:15 archmain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 17:10:15 archmain kernel: i915 0000:00:02.0: sddm-greeter[647] context reset due to GPU hang Jun 23 17:10:15 archmain sddm-greeter[647]: Message received from daemon: Capabilities Jun 23 17:10:15 archmain sddm-greeter[647]: Message received from daemon: HostName Jun 23 17:10:19 archmain dhcpcd[508]: enp2s0: no IPv6 Routers available Jun 23 17:10:21 archmain kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in sddm-greeter [647] Jun 23 17:10:21 archmain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 17:10:21 archmain kernel: i915 0000:00:02.0: sddm-greeter[647] context reset due to GPU hang Jun 23 17:10:21 archmain sddm[595]: Message received from greeter: PowerOff Jun 23 17:10:21 archmain systemd-logind[505]: System is powering down. Jun 23 17:10:21 archmain sddm[595]: Error from greeter session: "Process crashed" Jun 23 17:10:21 archmain sddm[595]: Auth: sddm-helper crashed (exit code 15) Jun 23 17:10:21 archmain sddm[595]: Error from greeter session: "Process crashed" Jun 23 17:10:21 archmain sddm[595]: Auth: sddm-helper exited with 15 Jun 23 17:10:21 archmain sddm[595]: Greeter stopped. Jun 23 17:10:21 archmain systemd[1]: Stopping Session c1 of user sddm.Edited by Harpo3I see that kernel 5.7.6 had some updates for i915, but I still have the GPU hang issue as reported in #2063 (closed)
Jun 25 22:20:18 archless kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in Xorg [551] Jun 25 22:20:18 archless kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jun 25 22:20:18 archless kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jun 25 22:20:18 archless kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jun 25 22:20:18 archless kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jun 25 22:20:18 archless kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jun 25 22:20:18 archless kernel: GPU crash dump saved to /sys/class/drm/card0/error Jun 25 22:20:18 archless kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 25 22:20:18 archless kernel: i915 0000:00:02.0: Xorg[551] context reset due to GPU hang Jun 25 22:20:21 archless kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in kscreenlocker_g [3310] Jun 25 22:20:21 archless kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0marked #2107 (closed) as a duplicate of this issue
I'm running a full Arch install on an Acer C720 with 4G RAM, & I'm having the same hang as the others. Everything ran perfectly through Kernel 5.6. I included the dmesg with Kernel 5.7.8-arch1-1. I can get more logs if you need them but would need to know what commands to run.dmesg_dump.txt
5.7.10-arch1-1 kernel, the same issue still on G3220 gpu_dump.txt
[ 22.545127] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in plasmashell [702] [ 22.545129] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 22.545129] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. [ 22.545129] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [ 22.545130] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 22.545130] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [ 22.545131] GPU crash dump saved to /sys/class/drm/card0/error [ 22.545169] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 [ 22.648000] i915 0000:00:02.0: plasmashell[702] context reset due to GPU hang [ 29.586964] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xorg [553] [ 29.587007] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 [ 29.688047] i915 0000:00:02.0: Xorg[553] context reset due to GPU hang [ 32.574413] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in kscreenlocker_g [691] [ 32.574453] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 [ 32.677250] i915 0000:00:02.0: kscreenlocker_g[691] context reset due to GPU hang [ 38.547064] i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in QSGRenderThread [840] [ 38.547688] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 [ 38.650605] i915 0000:00:02.0: QSGRenderThread[840] context reset due to GPU hang [ 44.520284] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xorg [553] [ 44.520325] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 [ 44.623911] i915 0000:00:02.0: Xorg[553] context reset due to GPU hangmentioned in issue #2177 (closed)
Acer C720 with Fedora 32 and Kernel 5.7.10 does have the error when playing back video with mpv and hw accel/vaapi, using 5.6.19 atm since Kernel 5.7.x are unusable with that error because graphics corruption+lags on Desktop upon hang
i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in mpv/vo [2831] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. drm/i915 developers can then reassign to the right component if it's not a kernel issue. The GPU crash dump is required to analyze GPU hangs, so please always attach it. GPU crash dump saved to /sys/class/drm/card0/error i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 i915 0000:00:02.0: mpv/vo[2831] context reset due to GPU hang i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in mpv [2818] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 i915 0000:00:02.0: mpv[2818] context reset due to GPU hang i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in mpv/vo [2831] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 i915 0000:00:02.0: mpv/vo[2831] context reset due to GPU hang i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in mpv/vo [2831] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 i915 0000:00:02.0: mpv/vo[2831] context reset due to GPU hang i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in mpv [2818] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 i915 0000:00:02.0: mpv[2818] context reset due to GPU hangHW: Acer C720 CPU: Intel(R) Celeron(R) 2955U @ 1.40GHz
Linux fedora-laptop 5.8.0-0.rc7.1.vanilla.1.fc32.x86_64 #1 (moved) SMP Mon Jul 27 17:48:09 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
Kernel from: https://repos.fedorapeople.org/repos/thl/kernel-vanilla-mainline/fedora-32/x86_64/
different error but still an issue with 5.8.0-rc7
[ 569.195453] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in mpv [2694] [ 569.195455] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 569.195455] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. [ 569.195456] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [ 569.195457] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 569.195457] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [ 569.195458] GPU crash dump saved to /sys/class/drm/card0/error [ 569.198346] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 569.301155] i915 0000:00:02.0: [drm] mpv[2694] context reset due to GPU hangcrash dump: gpu_hang_k580rc7_mpv.txt
Is this relevant with this problem ?
https://bbs.archlinux.org/viewtopic.php?id=256520 [SOLVED] Kernel 5.7 Intel graphics will not boot without nomodeset (This title doesn't describe the problem... but it's the same...)
For everyone that has this problem the proposed solution is to add intel_iommu=on to the kernel options.
or, in french (?!) https://forums.archlinux.fr/viewtopic.php?f=15&t=21988
I tried booting with several combinations of kernel options on the 5.7.12-arch1-1 kernel, such as nomodeset, i915.modeset=0, intel_iommu=on, intel_iommu=igfs_off, intel_iommu=on,igfx_off.
Nothing I tried improved anything, and some options caused my C720 to not boot. Everything still works fine with the 5.4.55-1 lts kernel
I am also having this issue as of the past 1-2 months on a system I'm working on. I will update my post when I can access the system... The system is on Arch linux latest pacman updates. I tried various kernel boot options as well such as intel_iommu=on intel_iommu=off i915.enable_fbc=1
Kernel current version is 5.7.12-arch1-1.
The system becomes unstable/unresponsive and often needs to be manually restarted after this. It seems to happen randomly, but with audio and web browsers. With web browsers it happens constantly with Brave, and frequently (but not as much) with firefox.
GPU crash dump: error
This is the error portion of journal logs:
Jul 26 23:07:38 fred audit: BPF prog-id=16 op=UNLOAD Jul 26 23:07:38 fred audit: BPF prog-id=15 op=UNLOAD Jul 26 23:07:38 fred kernel: audit: type=1334 audit(1595822858.563:73): prog-id=16 op=UNLOAD Jul 26 23:07:38 fred kernel: audit: type=1334 audit(1595822858.563:74): prog-id=15 op=UNLOAD Jul 26 23:07:42 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in brave [1422] Jul 26 23:07:42 fred kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jul 26 23:07:42 fred kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jul 26 23:07:42 fred kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jul 26 23:07:42 fred kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jul 26 23:07:42 fred kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jul 26 23:07:42 fred kernel: GPU crash dump saved to /sys/class/drm/card0/error Jul 26 23:07:42 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:07:42 fred kernel: i915 0000:00:02.0: brave[1422] context reset due to GPU hang Jul 26 23:07:45 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79 Jul 26 23:07:45 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:07:51 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd Jul 26 23:07:51 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:07:53 fred dbus-daemon[315]: [system] Failed to activate service 'org.bluez': timed out (service_start_timeout=25000ms) Jul 26 23:07:54 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in brave [1544] Jul 26 23:07:54 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:00 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in brave [1544] Jul 26 23:08:00 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:00 fred kernel: i915 0000:00:02.0: brave[1544] context reset due to GPU hang Jul 26 23:08:06 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd Jul 26 23:08:06 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd Jul 26 23:08:06 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on bcs0 Jul 26 23:08:09 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc7b, in Xorg [338] Jul 26 23:08:09 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc7b, in Xorg [338] Jul 26 23:08:09 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:09 fred kernel: i915 0000:00:02.0: Xorg[338] context reset due to GPU hang Jul 26 23:08:12 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in brave [1573] Jul 26 23:08:12 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:15 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xorg [338] Jul 26 23:08:15 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on bcs0 Jul 26 23:08:18 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in brave [1573] Jul 26 23:08:18 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:18 fred kernel: i915 0000:00:02.0: brave[1573] context reset due to GPU hang Jul 26 23:08:24 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in brave [1573] Jul 26 23:08:24 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:24 fred kernel: i915 0000:00:02.0: brave[1573] context reset due to GPU hang Jul 26 23:08:30 fred kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc7b, in Xorg [338] Jul 26 23:08:30 fred kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jul 26 23:08:30 fred kernel: i915 0000:00:02.0: Xorg[338] context reset due to GPU hang Jul 26 23:21:42 fred systemd[1]: Starting Cleanup of Temporary Directories... Jul 26 23:21:43 fred systemd[1]: systemd-tmpfiles-clean.service: Succeeded. Jul 26 23:21:43 fred systemd[1]: Finished Cleanup of Temporary Directories. Jul 26 23:21:43 fred audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295```Edited by semcorestill happening on 5.9.0-rc1-15308-gda2968ff879b
[ 19.540009] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79, in kscreenlocker_g [718] [ 19.540011] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 19.540012] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. [ 19.540012] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [ 19.540013] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 19.540013] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [ 19.540014] GPU crash dump saved to /sys/class/drm/card0/error [ 19.540048] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 19.642021] i915 0000:00:02.0: [drm] kscreenlocker_g[718] context reset due to GPU hang [ 25.519871] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc7b, in kscreenlocker_g [718] [ 25.519919] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 25.621270] i915 0000:00:02.0: [drm] kscreenlocker_g[718] context reset due to GPU hangnow on 5.9.0-rc1-15466-gf873db9acd3c this other thing
[ 29.556835] Asynchronous wait on fence 0000:00:02.0:kscreenlocker_g[690]:38 timed out (hint:intel_atomic_commit_ready [i915]) [ 29.556845] ------------[ cut here ]------------ [ 29.556854] WARNING: CPU: 1 PID: 0 at kernel/sched/core.c:4796 default_wake_function+0x16/0x30 [ 29.556855] Modules linked in: nls_iso8859_1 nls_cp437 vfat fat ccm algif_aead des_generic libdes arc4 algif_skcipher cmac md4 algif_hash af_alg raid1 dm_raid raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx md_mod dm_thin_pool dm_persistent_data dm_bufio dm_bio_prison dm_mod btrfs blake2b_generic libcrc32c ath9k_htc mac80211 xor intel_rapl_msr ath9k_common intel_rapl_common ath9k_hw x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel ath ofpart cmdlinepart intel_spi_platform kvm cfg80211 intel_spi spi_nor raid6_pq snd_hda_codec_hdmi at24 mtd iTCO_wdt iTCO_vendor_support mei_hdcp irqbypass ppdev crct10dif_pclmul crc32_pclmul ghash_clmulni_intel cryptd rapl intel_cstate input_leds intel_uncore i915 pcspkr snd_hda_codec_realtek intel_gtt i2c_algo_bit snd_hda_codec_generic ledtrig_audio drm_kms_helper i2c_i801 rfkill r8169 snd_hda_intel i2c_smbus snd_intel_dspcfg libarc4 lpc_ich realtek mdio_devres snd_hda_codec of_mdio fixed_phy snd_hwdep libphy syscopyarea [ 29.556923] snd_hda_core snd_pcm sysfillrect snd_timer sysimgblt fb_sys_fops mei_me snd cec soundcore mei ir_rc6_decoder wmi rc_rc6_mce parport_pc ite_cir parport rc_core evdev mac_hid drm pkcs8_key_parser crypto_user agpgart ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 serio_raw atkbd libps2 crc32c_intel xhci_pci xhci_hcd ehci_pci ehci_hcd i8042 serio [ 29.556958] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 5.9.0-rc1-15466-gf873db9acd3c #34 [ 29.556960] Hardware name: ECS Z87H3-A4/Z87H3-A4, BIOS 4.6.5 05/29/2014 [ 29.556965] RIP: 0010:default_wake_function+0x16/0x30 [ 29.556970] Code: e8 6f ee 3d 00 eb 99 0f 0b 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 f7 c2 fe ff ff ff 75 09 48 8b 7f 08 e9 0a f9 ff ff <0f> 0b 48 8b 7f 08 e9 ff f8 ff ff 66 66 2e 0f 1f 84 00 00 00 00 00 [ 29.556973] RSP: 0018:ffff998a8001ee68 EFLAGS: 00010082 [ 29.556977] RAX: ffffffffaa6e1fa0 RBX: ffff998a8026fd30 RCX: ffff998a8001ee80 [ 29.556979] RDX: 00000000ffffff92 RSI: 0000000000000003 RDI: ffff998a8026fd30 [ 29.556981] RBP: ffff96152a099568 R08: 0000000000000406 R09: 0000000000000004 [ 29.556984] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000046 [ 29.556986] R13: ffff96152a099560 R14: ffff998a8001ee80 R15: ffff9615194bff28 [ 29.556990] FS: 0000000000000000(0000) GS:ffff961557b00000(0000) knlGS:0000000000000000 [ 29.556992] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 29.556995] CR2: 00007f6f742cfce0 CR3: 000000007ca0e006 CR4: 00000000000706e0 [ 29.556997] Call Trace: [ 29.557001] <IRQ> [ 29.557009] autoremove_wake_function+0xe/0x30 [ 29.557087] __i915_sw_fence_complete+0x156/0x1b0 [i915] [ 29.557162] ? i915_sw_fence_complete+0x20/0x20 [i915] [ 29.557231] ? i915_sw_fence_complete+0x20/0x20 [i915] [ 29.557235] call_timer_fn+0x2d/0x160 [ 29.557306] ? i915_sw_fence_complete+0x20/0x20 [i915] [ 29.557311] __run_timers+0x189/0x280 [ 29.557318] run_timer_softirq+0x2b/0x50 [ 29.557328] __do_softirq+0xff/0x344 [ 29.557336] asm_call_on_stack+0x12/0x20 [ 29.557339] </IRQ> [ 29.557347] do_softirq_own_stack+0x5f/0x80 [ 29.557354] irq_exit_rcu+0xd8/0x120 [ 29.557360] sysvec_apic_timer_interrupt+0x47/0xe0 [ 29.557367] asm_sysvec_apic_timer_interrupt+0x12/0x20 [ 29.557375] RIP: 0010:cpuidle_enter_state+0xb6/0x420 [ 29.557379] Code: a0 e1 23 55 e8 eb 74 8f ff 49 89 c7 0f 1f 44 00 00 31 ff e8 2c 83 8f ff 80 7c 24 0f 00 0f 85 06 02 00 00 fb 66 0f 1f 44 00 00 <45> 85 e4 0f 88 e9 01 00 00 49 63 d4 4c 2b 7c 24 10 48 8d 04 52 48 [ 29.557382] RSP: 0018:ffff998a800afe98 EFLAGS: 00000246 [ 29.557385] RAX: ffff961557b00000 RBX: ffff961557b35400 RCX: 000000000000001f [ 29.557387] RDX: 0000000000000000 RSI: ffffffffab71328f RDI: ffffffffab71d859 [ 29.557390] RBP: ffffffffabace0c0 R08: 00000006e1b77c1b R09: 00000006e1b3a619 [ 29.557392] R10: 0000000000000074 R11: 00000000000014f5 R12: 0000000000000004 [ 29.557394] R13: ffff961557b35400 R14: 0000000000000004 R15: 00000006e1b77c1b [ 29.557406] ? cpuidle_enter_state+0xa4/0x420 [ 29.557413] cpuidle_enter+0x29/0x40 [ 29.557420] do_idle+0x1fb/0x2c0 [ 29.557427] cpu_startup_entry+0x19/0x20 [ 29.557433] secondary_startup_64+0xb6/0xc0 [ 29.557441] ---[ end trace 482fae973e5f5285 ]---#2024 (comment 530965) seems to fix the issue with drm-tip
Edited by plntykI did try to bisect this but it does not work.
HW: Chromebook Acer C720 "Peppy" with Intel(R) Celeron(R) 2955U / Haswell
Known good: (Fedora-packaged) 5.6.19 , drm-tip (today) g7296072da95d with patch, v5.6 (linus)
known bad: drm-tip g7296072da95d, 5.7.16 (Fedora), 5.7.14(Fedora), multiple 5.7.x series, v5.7 (linus tree), v5.8 , v5.9-rc1, v5.9-rc2,
some other from bisection:
git bisect bad 71fa42fac5d67773d83d6bc058d992a6226dbb01 git bisect good 0419c450e1bb5eb1c68ba10efeefddaf556cde85 git bisect bad 57a333f8f8896ca66e9d8153930247a3cfd7b592 git bisect good 2c523b344dfa65a3738e7039832044aa133c75fbfinally tested 2 times both non working:
git bisect bad 9f880327160feb695de03caa29604883b0d00087Merge tag 'amd-drm-next-5.6-2020-02-05' of git://people.freedesktop.org/~agd5f/linux into drm-next
This last commit makes git bisect error. (but its unrelated to Intel stuff it seems)
latest drm-tip without patch crashes badly and I cannot change virtual tty like I could with other versions during bisection
Testing was done with a blender movie and mpv cmdline: mpv -vo vaapi bbb_sunflower_1080p_30fps_normal.mp4
Any tips how to bisect further ?
Edit: Some bisection targets are difficult due to GCC 10 changes around that time and compile errors that do occur because of that.
Edited by plntykCheck out "Cherries required" I listed above. Anyway the problematic series is https://patchwork.freedesktop.org/series/74363/. (As of v5.8, at least; maybe more after that)
Coincidentally, btw, https://www.phoronix.com/scan.php?page=news_item&px=GNU-Linux-Libre-5.7-Released
Processor Intel G3260, Integrated video driver only, Fedora 31 (KDE Plasma), Kernel 5.7.15. After i enter my password in login manager (SDDM) i only see a black screen with cursor, then my system goes back to login screen. The same effect on the Kernel 5.7.8-100. Works fine with the Kernel 5.6.19-200.
Aug 24 19:43:55 localhost.localdomain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 24 19:43:55 localhost.localdomain kernel: i915 0000:00:02.0: kwin_x11[1404] context reset due to GPU hang Aug 24 19:43:55 localhost.localdomain kernel: [drm:intel_gt_verify_workarounds [i915]] *ERROR* GT workaround lost on init! (e184=0/0, expected 2000200) Aug 24 19:43:55 localhost.localdomain korgac[1462]: org.kde.pim.akonadicore: Job error: "" for collection: QVector() Aug 24 19:43:56 localhost.localdomain plasmashell[1416]: Connecting to deprecated signal QDBusConnectionInterface::serviceOwnerChanged(QString,QString,QString) Aug 24 19:43:56 localhost.localdomain kdeconnectd[1424]: kdeconnect.core: Could not query capabilities from notifications server Aug 24 19:43:57 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:43:57 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:43:57 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:43:57 localhost.localdomain plasmashell[1416]: trying to show an empty dialog Aug 24 19:43:57 localhost.localdomain plasmashell[1416]: trying to show an empty dialog Aug 24 19:43:58 localhost.localdomain plasmashell[1416]: Component "DBusProperty" is specified with non-relative URL "qrc:///kdeconnect/qml/DBusProperty.qml" in a qmldir file. > Aug 24 19:43:58 localhost.localdomain plasmashell[1416]: Component "PluginChecker" is specified with non-relative URL "qrc:///kdeconnect/qml/PluginChecker.qml" in a qmldir file> Aug 24 19:43:58 localhost.localdomain plasmashell[1416]: Component "RemoteKeyboard" is specified with non-relative URL "qrc:///kdeconnect/qml/RemoteKeyboard.qml" in a qmldir fi> Aug 24 19:43:59 localhost.localdomain systemd[1233]: dbus-:1.2-org.freedesktop.Notifications@0.service: Succeeded. Aug 24 19:43:59 localhost.localdomain cupsd[797]: REQUEST localhost - - "POST / HTTP/1.1" 200 564 Create-Printer-Subscriptions successful-ok Aug 24 19:43:59 localhost.localdomain systemd[1]: Starting PackageKit Daemon... Aug 24 19:43:59 localhost.localdomain plasmashell[1416]: QQuickItem::stackAfter: Cannot stack StatusNotifierItem_QMLTYPE_291(0x556731b506c0, parent=0x5567317f4ec0, geometry=0,0> Aug 24 19:43:59 localhost.localdomain PackageKit[1694]: daemon start Aug 24 19:43:59 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:43:59 localhost.localdomain plasmashell[1416]: file:///usr/share/plasma/plasmoids/org.kde.plasma.battery/contents/ui/CompactRepresentation.qml:53:17: Unable to assign> Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: trying to show an empty dialog Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: file:///usr/share/plasma/shells/org.kde.plasma.desktop/contents/views/Desktop.qml:146:19: QML Loader: Binding loop dete> Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: file:///usr/share/plasma/shells/org.kde.plasma.desktop/contents/views/Desktop.qml:146:19: QML Loader: Binding loop dete> Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: KAStatsFavoritesModel::setFavorites is ignored Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "kontact.desktop" QSharedPointer(0x556732fab4e0) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.ktpcontactlist.desktop" QSharedPointer(0x556732fab4e0) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "ktpcontactlist.desktop" QSharedPointer(0x556732fab4e0) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.kate.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "kate.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.apper.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "apper.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "kontact.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.ktpcontactlist.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "ktpcontactlist.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.kate.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "kate.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "org.kde.apper.desktop" QSharedPointer(0x556732fa1c10) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: Entry is not valid "apper.desktop" QSharedPointer(0x556732f9f230) Aug 24 19:44:00 localhost.localdomain plasmashell[1416]: trying to show an empty dialog Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=packagekit comm="systemd" ex> Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: Both point size and pixel size set. Using pixel size. Aug 24 19:44:01 localhost.localdomain systemd[1]: Started PackageKit Daemon. Aug 24 19:44:01 localhost.localdomain plasmashell[1416]: trying to show an empty dialog Aug 24 19:44:01 localhost.localdomain kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xorg [840] Aug 24 19:44:01 localhost.localdomain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 24 19:44:01 localhost.localdomain pulseaudio[1475]: E: [pulseaudio] bluez5-util.c: GetManagedObjects() failed: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. > Aug 24 19:44:01 localhost.localdomain kernel: i915 0000:00:02.0: Xorg[840] context reset due to GPU hang Aug 24 19:44:01 localhost.localdomain kernel: [drm:intel_gt_verify_workarounds [i915]] *ERROR* GT workaround lost on init! (e184=0/0, expected 2000200) Aug 24 19:44:01 localhost.localdomain PackageKit[1694]: uid 1000 is trying to obtain org.freedesktop.packagekit.system-sources-refresh auth (only_trusted:0) Aug 24 19:44:01 localhost.localdomain PackageKit[1694]: uid 1000 obtained auth for org.freedesktop.packagekit.system-sources-refresh Aug 24 19:44:03 localhost.localdomain PackageKit[1694]: refresh-cache transaction /1503_ccbbbcee from uid 1000 finished with success after 1912ms Aug 24 19:44:05 localhost.localdomain kwin_x11[1404]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1092, resource id: 52428805, major code: 18 (ChangePropert> Aug 24 19:44:07 localhost.localdomain kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xorg [840] Aug 24 19:44:07 localhost.localdomain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 24 19:44:07 localhost.localdomain kernel: i915 0000:00:02.0: Xorg[840] context reset due to GPU hang Aug 24 19:44:07 localhost.localdomain kernel: [drm:intel_gt_verify_workarounds [i915]] *ERROR* GT workaround lost on init! (e184=0/0, expected 2000200) Aug 24 19:44:10 localhost.localdomain kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc7b, in Xorg [840] Aug 24 19:44:10 localhost.localdomain kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 24 19:44:10 localhost.localdomain kernel: i915 0000:00:02.0: Xorg[840] context reset due to GPU hang Aug 24 19:44:10 localhost.localdomain kernel: [drm:intel_gt_verify_workarounds [i915]] *ERROR* GT workaround lost on init! (e184=0/0, expected 2000200) Aug 24 19:44:10 localhost.localdomain korgac[1462]: The X11 connection broke (error 1). Did the X11 server die?Reproducible for me when the attached image is played with:
mpv --cscale=ewa_lanczos sample.jpgAlso attaching the logs and dumps:
Aug 25 08:33:30 archlinux kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in systemd-logind [431] Aug 25 08:33:30 archlinux kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 25 08:33:30 archlinux kernel: i915 0000:00:02.0: systemd-logind[431] context reset due to GPU hang Aug 25 08:33:36 archlinux kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in systemd-logind [431] Aug 25 08:33:36 archlinux kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 25 08:33:36 archlinux kernel: i915 0000:00:02.0: systemd-logind[431] context reset due to GPU hang Aug 25 08:33:39 archlinux kernel: i915 0000:00:02.0: GPU HANG: ecode 7:1:85ddfffd, in systemd-logind [431] Aug 25 08:33:39 archlinux kernel: i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Aug 25 08:33:39 archlinux kernel: i915 0000:00:02.0: systemd-logind[431] context reset due to GPU hangAug 25 08:26:51 archlinux kernel: Asynchronous wait on fence 0000:00:02.0:systemd-logind[358]:117 timed out (hint:intel_atomic_commit_ready [i915]) Aug 25 08:26:51 archlinux kernel: ------------[ cut here ]------------ Aug 25 08:26:51 archlinux kernel: WARNING: CPU: 0 PID: 0 at kernel/sched/core.c:4796 default_wake_function+0x16/0x30 Aug 25 08:26:51 archlinux kernel: Modules linked in: fuse hid_logitech_hidpp intel_rapl_msr intel_rapl_common nls_iso8859_1 nls_cp437 vfat fat x86_pkg_temp_thermal intel_powerclamp i915 coretemp kvm_intel ofpart cmdlinepart snd_virtuoso kvm intel_spi_platform dm_mod intel_spi eeepc_wmi snd_oxygen_lib snd_mpu401_uart asus_wmi spi_nor snd_rawmidi mei_hdcp irqbypass battery iTCO_wdt mtd sparse_keymap iTCO_vendor_support i2c_algo_bit at24 drm_kms_helper snd_seq_device crct10dif_pclmul rfkill bridge wmi_bmof snd_pcm cec r8169 crc32c_intel stp llc ghash_clmulni_intel rc_core realtek aesni_intel snd_timer crypto_simd intel_gtt mdio_devres cryptd glue_helper of_mdio syscopyarea snd mei_me fixed_phy rapl intel_cstate input_leds sysfillrect mousedev sysimgblt i2c_i801 joydev fb_sys_fops libphy soundcore intel_uncore i2c_smbus mei lpc_ich evdev wmi mac_hid hid_logitech_dj drm sg agpgart ip_tables x_tables f2fs hid_generic usbhid hid uas usb_storage sr_mod crc32_pclmul cdrom ehci_pci ehci_hcd xhci_pci xhci_hcd Aug 25 08:26:51 archlinux kernel: CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.9.0-rc2-1-git #1 Aug 25 08:26:51 archlinux kernel: Hardware name: ASUS All Series/H87-PRO, BIOS 2104 10/28/2014 Aug 25 08:26:51 archlinux kernel: RIP: 0010:default_wake_function+0x16/0x30 Aug 25 08:26:51 archlinux kernel: Code: e8 ef 36 3e 00 eb 99 0f 0b 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 f7 c2 fe ff ff ff 75 09 48 8b 7f 08 e9 0a f9 ff ff <0f> 0b 48 8b 7f 08 e9 ff f8 ff ff 66 66 2e 0f 1f 84 00 00 00 00 00 Aug 25 08:26:51 archlinux kernel: RSP: 0018:ffff9ea480003e58 EFLAGS: 00010082 Aug 25 08:26:51 archlinux kernel: RAX: ffffffff9eae4fb0 RBX: ffff9ea4802dbd30 RCX: ffff9ea480003e70 Aug 25 08:26:51 archlinux kernel: RDX: 00000000ffffff92 RSI: 0000000000000003 RDI: ffff9ea4802dbd30 Aug 25 08:26:51 archlinux kernel: RBP: ffff8af4d71ca568 R08: 0000000000000423 R09: 0000000000000004 Aug 25 08:26:51 archlinux kernel: R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000046 Aug 25 08:26:51 archlinux kernel: R13: ffff8af4d71ca560 R14: ffff9ea480003e70 R15: ffff8af4d229d2a8 Aug 25 08:26:51 archlinux kernel: FS: 0000000000000000(0000) GS:ffff8af4dfa00000(0000) knlGS:0000000000000000 Aug 25 08:26:51 archlinux kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Aug 25 08:26:51 archlinux kernel: CR2: 00007f2104119028 CR3: 00000003ffe0e001 CR4: 00000000001706f0 Aug 25 08:26:51 archlinux kernel: Call Trace: Aug 25 08:26:51 archlinux kernel: <IRQ> Aug 25 08:26:51 archlinux kernel: autoremove_wake_function+0xe/0x30 Aug 25 08:26:51 archlinux kernel: __i915_sw_fence_complete+0x156/0x1b0 [i915] Aug 25 08:26:51 archlinux kernel: ? i915_sw_fence_complete+0x20/0x20 [i915] Aug 25 08:26:51 archlinux kernel: ? i915_sw_fence_complete+0x20/0x20 [i915] Aug 25 08:26:51 archlinux kernel: call_timer_fn+0x2d/0x160 Aug 25 08:26:51 archlinux kernel: ? i915_sw_fence_complete+0x20/0x20 [i915] Aug 25 08:26:51 archlinux kernel: __run_timers+0x189/0x280 Aug 25 08:26:51 archlinux kernel: run_timer_softirq+0x2b/0x50 Aug 25 08:26:51 archlinux kernel: __do_softirq+0x10f/0x356 Aug 25 08:26:51 archlinux kernel: asm_call_on_stack+0x12/0x20 Aug 25 08:26:51 archlinux kernel: </IRQ> Aug 25 08:26:51 archlinux kernel: do_softirq_own_stack+0x5f/0x80 Aug 25 08:26:51 archlinux kernel: irq_exit_rcu+0xd8/0x120 Aug 25 08:26:51 archlinux kernel: sysvec_apic_timer_interrupt+0x47/0xe0 Aug 25 08:26:51 archlinux kernel: asm_sysvec_apic_timer_interrupt+0x12/0x20 Aug 25 08:26:51 archlinux kernel: RIP: 0010:cpuidle_enter_state+0xb6/0x420 Aug 25 08:26:51 archlinux kernel: Code: 60 d4 e1 60 e8 bb 97 8d ff 49 89 c7 0f 1f 44 00 00 31 ff e8 fc a5 8d ff 80 7c 24 0f 00 0f 85 06 02 00 00 fb 66 0f 1f 44 00 00 <45> 85 e4 0f 88 e9 01 00 00 49 63 d4 4c 2b 7c 24 10 48 8d 04 52 48 Aug 25 08:26:51 archlinux kernel: RSP: 0018:ffffffffa0003e40 EFLAGS: 00000246 Aug 25 08:26:51 archlinux kernel: RAX: ffff8af4dfa00000 RBX: ffff8af4dfa37600 RCX: 000000000000001f Aug 25 08:26:51 archlinux kernel: RDX: 0000000000000000 RSI: ffffffff9fd4a54f RDI: ffffffff9fd54b59 Aug 25 08:26:51 archlinux kernel: RBP: ffffffffa00cebc0 R08: 0000000847cecb74 R09: 0000000000000008 Aug 25 08:26:51 archlinux kernel: R10: 0000000000002c2f R11: 0000000000020a91 R12: 0000000000000005 Aug 25 08:26:51 archlinux kernel: R13: ffff8af4dfa37600 R14: 0000000000000005 R15: 0000000847cecb74 Aug 25 08:26:51 archlinux kernel: cpuidle_enter+0x29/0x40 Aug 25 08:26:51 archlinux kernel: do_idle+0x1fb/0x2c0 Aug 25 08:26:51 archlinux kernel: cpu_startup_entry+0x19/0x20 Aug 25 08:26:51 archlinux kernel: start_kernel+0x85b/0x880 Aug 25 08:26:51 archlinux kernel: secondary_startup_64+0xb6/0xc0 Aug 25 08:26:51 archlinux kernel: ---[ end trace 3c93c68908efdd8f ]--- Aug 25 08:27:13 archlinux kernel: i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in systemd-logind [358] Aug 25 08:27:13 archlinux kernel: GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Aug 25 08:27:13 archlinux kernel: Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Aug 25 08:27:13 archlinux kernel: Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Aug 25 08:27:13 archlinux kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue. Aug 25 08:27:13 archlinux kernel: The GPU crash dump is required to analyze GPU hangs, so please always attach it. Aug 25 08:27:13 archlinux kernel: GPU crash dump saved to /sys/class/drm/card0/error Aug 25 08:27:13 archlinux kernel: i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 Aug 25 08:27:13 archlinux kernel: i915 0000:00:02.0: [drm] systemd-logind[358] context reset due to GPU hang Aug 25 08:27:19 archlinux kernel: i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in systemd-logind [358] Aug 25 08:27:19 archlinux kernel: i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 Aug 25 08:27:19 archlinux kernel: i915 0000:00:02.0: [drm] systemd-logind[358] context reset due to GPU hangSame ecodes as reported above:
7:1:85ddfffd with kernel 5.7.17-200.fc32.x86_64 x86_64
7:1:8edcfc79 with 5.8.6-201.fc32.x86_64 x86_64
Results in a machine entirely without graphics from boot, the only means of interaction being ttys2-6.
GPU HANG: ecode 7:1:8edcfc79, in ksplashqml [1229] Kernel: 5.8.6-201.fc32.x86_64 x86_64 Driver: 20200515 Time: 1599592553 s 982336 us Boottime: 43 s 835163 us Uptime: 39 s 394040 us Capture: 4294711040 jiffies; 335255 ms ago Active process (on ring rcs0): ksplashqml [1229] Reset count: 0 Suspend count: 0 Platform: HASWELLGPU HANG: ecode 7:1:85ddfffd, in kwin_x11 [1215] Kernel: 5.7.17-200.fc32.x86_64 x86_64 Driver: 20200313 Time: 1598715941 s 22107 us Boottime: 43 s 837975 us Uptime: 39 s 471519 us Capture: 4294711040 jiffies; 934957 ms ago Active process (on ring rcs0): kwin_x11 [1215] Reset count: 0 Suspend count: 0 Platform: HASWELLWhat are these people playing at?
Tested again with drm-tip 5.9.0-rc5-01152-g14dc141acc19, still hangs.
Tested the
&& 0change suggested in #2024 (comment 530965) on top of 5.9.0-rc5, that one fixes the issue (for me, G3220).I'm also experiencing this since 5.7.* (also a problem on 5.8.*). I'm stuck on 5.6.14 which was the last working version for me. CPU is Pentium G3258.
Not sure if this adds anything, but from my log:
Jun 23 19:35:04 gk-pc kernel: [37480.928936] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in sway [4646] Jun 23 19:35:04 gk-pc kernel: [37480.928937] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jun 23 19:35:04 gk-pc kernel: [37480.928937] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. Jun 23 19:35:04 gk-pc kernel: [37480.928937] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. Jun 23 19:35:04 gk-pc kernel: [37480.928938] drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jun 23 19:35:04 gk-pc kernel: [37480.928938] The GPU crash dump is required to analyze GPU hangs, so please always attach it. Jun 23 19:35:04 gk-pc kernel: [37480.928939] GPU crash dump saved to /sys/class/drm/card0/error Jun 23 19:35:04 gk-pc kernel: [37480.928993] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:04 gk-pc kernel: [37481.031547] i915 0000:00:02.0: sway[4646] context reset due to GPU hang Jun 23 19:35:07 gk-pc kernel: [37483.939766] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xwayland [4683] Jun 23 19:35:07 gk-pc kernel: [37483.939808] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:07 gk-pc kernel: [37484.042102] i915 0000:00:02.0: Xwayland[4683] context reset due to GPU hang Jun 23 19:35:13 gk-pc kernel: [37489.953712] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in sway [4646] Jun 23 19:35:13 gk-pc kernel: [37489.953978] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:13 gk-pc kernel: [37490.055109] i915 0000:00:02.0: sway[4646] context reset due to GPU hang Jun 23 19:35:16 gk-pc kernel: [37492.962312] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in Xwayland [4683] Jun 23 19:35:16 gk-pc kernel: [37492.962429] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:16 gk-pc kernel: [37493.065940] i915 0000:00:02.0: Xwayland[4683] context reset due to GPU hang Jun 23 19:35:22 gk-pc kernel: [37498.913709] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc7b, in sway [4646] Jun 23 19:35:22 gk-pc kernel: [37498.913950] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:22 gk-pc kernel: [37499.017178] i915 0000:00:02.0: sway[4646] context reset due to GPU hang Jun 23 19:35:59 gk-pc kernel: [37535.905480] i915 0000:00:02.0: GPU HANG: ecode 7:1:8edcfc79, in alacritty [8198] Jun 23 19:35:59 gk-pc kernel: [37535.905741] i915 0000:00:02.0: Resetting chip for stopped heartbeat on rcs0 Jun 23 19:35:59 gk-pc kernel: [37536.007892] i915 0000:00:02.0: alacritty[8198] context reset due to GPU hangEdited by George KettleboroughI tried running 5.8.9 with the patch in #2024 (comment 530965)
I no longer had a GPU hang, but my CPU was stuck at its base clock speed of 800MHz instead of boosting to 4.2GHz! Very strange!
mentioned in issue mesa/mesa#3532 (closed)
Forget it. Freeloaders in Intel are even too lazy to revert their broken and obsecure "security fix". Basically they just throw out garbage (both chip and code) that doesn't work and say "hey I fixed my broken chip already", and expect you to buy newer (but also likely-broken) chip before you need to.
So it won't help even if we were all willing to give the world to them, if the world has nothing to do with money.
Edited by Tom Yan@tomty89: Once Ubuntu 20.10 releases (which has 5.8), I imagine there will be quite a few people affected. I hope that will encourage the solution for this problem. ):
I too am seeing this on Fedora 32 kernel 5.8.13-200.fc32.x86_64:
[ 3.574142] i915 0000:00:02.0: [drm] PipeC fused off [ 3.574157] i915 0000:00:02.0: vgaarb: deactivate vga console [ 3.577216] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem [ 3.589414] [drm] Initialized i915 1.6.0 20200515 for 0000:00:02.0 on minor 0 [ 3.616442] fbcon: i915drmfb (fb0) is primary device [ 3.708859] i915 0000:00:02.0: fb0: i915drmfb frame buffer device [ 19.677829] snd_hda_intel 0000:00:03.0: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915]) [ 8338.767033] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8338.767035] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [ 8338.767035] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 8338.767092] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8341.772995] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8341.773043] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8378.773008] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8378.774993] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8384.790346] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8384.791417] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8390.813150] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8390.813978] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8396.812580] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8396.816947] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8399.784202] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8399.784671] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8405.800070] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8405.800163] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8411.816034] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8411.818195] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8414.806940] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8414.812192] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8420.826741] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8420.827625] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8426.792204] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8426.798220] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8429.798311] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8429.799985] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8435.814342] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8435.815930] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8441.831924] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8441.836754] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8444.823371] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8444.823487] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8450.787492] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8450.788712] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8456.807788] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8456.814258] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8459.815809] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8459.816391] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8465.831703] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8465.834990] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8490.791873] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8490.793799] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8496.807663] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8496.809380] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8502.798842] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8502.798889] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8508.775461] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8508.775581] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8685.770942] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8685.773476] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8688.778915] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 8688.778966] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8696.780037] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in gnome-shell [3013] [ 8696.784409] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 8696.885587] i915 0000:00:02.0: [drm] gnome-shell[3013] context reset due to GPU hang [ 8702.794850] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:8edcfc79 [ 8702.799103] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 9057.801259] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in Xorg [9023] [ 9057.801310] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 9057.904403] i915 0000:00:02.0: [drm] Xorg[9023] context reset due to GPU hang [ 9063.753362] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd, in Xorg [9023] [ 9063.753413] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [ 9063.855312] i915 0000:00:02.0: [drm] Xorg[9023] context reset due to GPU hang [ 9069.768753] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85ddfffd [ 9069.773305] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0on a:
00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller (rev 06)Same here with OpenSuSE Tumbleweed. Can only run 5.6.x kernels
System: Host: moz Kernel: 5.6.14-1-default x86_64 bits: 64 compiler: gcc v: 10.1.1 parameters: BOOT_IMAGE=/vmlinuz-5.6.14-1-default root=UUID=4e8b26ec-ea5c-4859-8bd1-b662b471a500 splash=silent quiet showopts Desktop: KDE Plasma 5.19.5 tk: Qt 5.15.1 wm: kwin_x11 dm: SDDM Distro: openSUSE Tumbleweed 20200923 CPU: Topology: Dual Core model: Intel Pentium G3258 bits: 64 type: MCP arch: Haswell family: 6 model-id: 3C (60) stepping: 3 microcode: 28 L2 cache: 3072 KiB flags: lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 12770 Speed: 979 MHz min/max: 800/3200 MHz Core speeds (MHz): 1: 877 2: 867 Vulnerabilities: Type: itlb_multihit status: KVM: Split huge pages Type: l1tf mitigation: PTE Inversion; VMX: conditional cache flushes, SMT disabled Type: mds mitigation: Clear CPU buffers; SMT disabled Type: meltdown mitigation: PTI Type: spec_store_bypass mitigation: Speculative Store Bypass disabled via prctl and seccomp Type: spectre_v1 mitigation: usercopy/swapgs barriers and __user pointer sanitization Type: spectre_v2 mitigation: Full generic retpoline, IBPB: conditional, IBRS_FW, STIBP: disabled, RSB filling Type: tsx_async_abort status: Not affected Graphics: Device-1: Intel Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics vendor: Gigabyte driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:0402 Display: x11 server: X.Org 1.20.9 compositor: kwin_x11 driver: intel unloaded: fbdev,modesetting,vesa display ID: :0 screens: 1 Screen-1: 0 s-res: 1680x1050 s-dpi: 96 s-size: 444x277mm (17.5x10.9") s-diag: 523mm (20.6") Monitor-1: HDMI1 res: 1680x1050 hz: 60 dpi: 99 size: 430x270mm (16.9x10.6") diag: 508mm (20") OpenGL: renderer: Mesa DRI Intel HD Graphics (HSW GT1) v: 4.5 Mesa 20.1.8 compat-v: 3.0 direct render: YesEdited by E RECIOmentioned in issue #2578 (closed)
As a Kodi user I stumbled across this bug report because it was mentioned here: https://github.com/LibreELEC/LibreELEC.tv/pull/4583
Since I am using kernel release 5.7 (version 5.7.19 atm) for several weeks I was wondering why my system is not affected by this bug. Apparently all users reporting this bug are using the "HSW GT1" chipset according to the IDs posted here which I compared with https://cgit.freedesktop.org/mesa/mesa/tree/include/pci_ids/i965_pci_ids.h
My system uses the "HSW GT2" chipset as quoted from kodi.log:
2020-10-14 20:38:22.129 T:140702184286048 NOTICE: GL_RENDERER = Mesa DRI Intel(R) HD Graphics 4600 (HSW GT2)
with this CPU: Intel(R) Core(TM) i3-4330T CPU
Does the bug only occur with "HSW GT1" chipsets?
Does the bug only occur with "HSW GT1" chipsets?
There's a gpu hang with the 5.7 kernel on an i5-2520M/Intel HD 3000, reported above. That processor is not a Haswell, nor is it GT1.
Edited by Aleister KonradHere's a quick summary. Seems like most do have Haswell GT1, but not all:
Haswell reports:
- @amfern, @plntyk, @zcat7777 and I have Celeron 2955U and GT1 iGPU.
- @trizen and @glaubersm have the Celeron G1820, also GT1
- @Harpo3, @emrecio and @georgek have Pentium G3258, also GT1
- @gdamjan has a G3220, also GT1
- @leonid-naumenko has a G3260, also GT1
- @lenovouser1 has a Pentium 3550M, also GT1
- @salz in #2177 (closed) has a Pentium G3450, also GT1
- @semcore and @brianjmurrell have Haswell, but I couldn't figure which.
Sandy bridge:
- @arvl and @intel_junk have i5-2520M and HD3000
Braswell:
- @hayasaka has a Celeron N3050
Edited by nariox
I would add myself:
- @salz in #2177 (closed) has a Pentium G3450, also GT1
Thanks @salz and sorry, somehow I ended up missing yours.
In my case the
intel_iommu=onflag does not help. I still have the crash. Also, I read in one of the other threads that this causes power consumption to go up significantly, can you confirm?I am starting to wonder if we have 2 or 3 different issues going on: a Braswell that you and @guajardoscar are having, a Haswell (possibly exclusively GT1) and maybe a Sandy Bridge one (although could be related to either one of the other ones). My clues are that
intel_iommuworks on some cases, but not others and that @guajardoscar stated that his issue happens with 5.6 as well, while in my cases (and I think others too), kernel 5.6 works perfectly fine.Haha かわいい.
What happens if you apply the patch to 5.7? Does it also work? If it doesn't, maybe Cherryview/Braswell has two bugs. I have found the post that reported higher power consumption, it is here: https://bbs.archlinux.org/viewtopic.php?id=256520&p=3
I think
intel_iommuenabled additional virtualization abilities. I remember seeing it for GPU passthrough (connecting a GPU directly to a VM), which would require Vt-d, but I don't know ifintel_iommurequires Vt-d. In that case, it wouldn't make sense for my 2955U (also no Vt-d). But I am not sure exactly what all options even mean (soft, on, pt, igfx_off, etc).EDIT: Found some more info on Gentoo's wiki: https://wiki.gentoo.org/wiki/IOMMU_SWIOTLB
Edited by nariox
As a summary. Seems like there are multiple problems (perhaps leading to the same effect - GPU hang):
- The braswell/cherryview related one, which seems to be fixed in 5.9.1 (according to #2024 (comment 669575)) and can be worked around with intel_iommu=on
- A haswell one (possibly affecting only GT1) one that is not fixed in 5.9.1 and where the intel_iommu workaround does not help, but some have reported that the patch in #2024 (comment 530965) helps, but at least one person reported a clocking issue.
- Sandy Bridge one? Maybe it's a different bug?
Seems like there are multiple problems
Well, er...
Tried the
&& 0patch on kernel 5.9 this time and no crash but got the clocking error again. My G3258 is only running at 800MHz instead of 4.2GHz. 5.6.14 remains the last working version.Edited by George Kettleborough
G3220 on 5.10.0-rc1 still hangs. Part of my config:
CONFIG_INTEL_IOMMU_DEFAULT_ON=y CONFIG_INTEL_IOMMU_FLOPPY_WA=y # CONFIG_INTEL_IOMMU_SCALABLE_MODE_DEFAULT_ON is not set CONFIG_DRM_I915=m CONFIG_DRM_I915_FORCE_PROBE="*" CONFIG_DRM_I915_CAPTURE_ERROR=y CONFIG_DRM_I915_COMPRESS_ERROR=y CONFIG_DRM_I915_USERPTR=y CONFIG_DRM_I915_GVT=y # CONFIG_DRM_I915_WERROR is not set # CONFIG_DRM_I915_DEBUG is not set # CONFIG_DRM_I915_DEBUG_MMIO is not set # CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS is not set # CONFIG_DRM_I915_SW_FENCE_CHECK_DAG is not set # CONFIG_DRM_I915_DEBUG_GUC is not set # CONFIG_DRM_I915_SELFTEST is not set # CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS is not set # CONFIG_DRM_I915_DEBUG_VBLANK_EVADE is not set # CONFIG_DRM_I915_DEBUG_RUNTIME_PM is not set CONFIG_DRM_I915_FENCE_TIMEOUT=10000 CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND=250 CONFIG_DRM_I915_HEARTBEAT_INTERVAL=2500 CONFIG_DRM_I915_PREEMPT_TIMEOUT=640 CONFIG_DRM_I915_MAX_REQUEST_BUSYWAIT=8000 CONFIG_DRM_I915_STOP_TIMEOUT=100 CONFIG_DRM_I915_TIMESLICE_DURATION=1 CONFIG_SND_HDA_I915=yWith the
&& 0patch it works fine. And the CPU is not pegged to a frequency, I can see it going to 3GHz and lower.@ickle posted a fix for 5.7+ here: https://patchwork.freedesktop.org/patch/395580/?series=82783&rev=1
Edited by Francesco BalestrieriThis patch on 5.9.2 works for my G3258 too.
My frequency scaling problem was fixed by setting the default scheduler to "ondemand" rather than "userspace". Turns out this was nothing to do with the patches or anything discussed here, but an unrelated change to the way intel_pstate works on older CPUs.
Edited by George Kettleborough
mentioned in issue #2580 (closed)
I sadly encountered another GPU hang today on 5.9.2 with the patch (https://patchwork.freedesktop.org/patch/395580/?series=82783&rev=1) applied.
[878391.747986] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:85dfbff8, in Compositor [25772] [878391.747987] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [878391.747988] Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new. [878391.747988] Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details. [878391.747989] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [878391.747989] The GPU crash dump is required to analyze GPU hangs, so please always attach it. [878391.747990] GPU crash dump saved to /sys/class/drm/card0/error [878391.748025] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [878391.850535] i915 0000:00:02.0: [drm] Compositor[25772] context reset due to GPU hang [878493.777763] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:dddff7ff, in Xwayland [13123] [878493.777984] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [878493.879694] i915 0000:00:02.0: [drm] Xwayland[13123] context reset due to GPU hang [878501.777727] i915 0000:00:02.0: [drm] GPU HANG: ecode 7:1:dddff7ff, in Xwayland [13123] [878501.777803] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0 [878501.879826] i915 0000:00:02.0: [drm] Xwayland[13123] context reset due to GPU hangIt seemed to be triggered by Firefox somehow, but even after killing Firefox I had to reboot before I could use my window manager again.
Edited by George Kettleborough
Please register or sign in to reply