Bug 266315
| Summary: | linuxkpi panic after recent updates (13.1-STABLE #0 stable/13-9cbba5950: Wed Sep 7 23:42:41 CEST 2022) | ||
|---|---|---|---|
| Product: | Base System | Reporter: | jakub_lach |
| Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> |
| Status: | Open --- | ||
| Severity: | Affects Only Me | CC: | dufresnep, grahamperrin, lhersch, manu, moonlapse81, x11 |
| Priority: | --- | Keywords: | crash, needs-qa |
| Version: | 13.1-STABLE | ||
| Hardware: | amd64 | ||
| OS: | Any | ||
| See Also: | https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=259670 | ||
|
Description
jakub_lach
2022-09-09 09:09:46 UTC
agp0: <Intel GM45 SVGA controller> on vgapci0
<...>
vgapci0@pci0:0:2:0: class=0x030000 rev=0x07 hdr=0x00 vendor=0x8086 device=0x2a42 subvendor=0x17aa subdevice=0x20e4
vendor = 'Intel Corporation'
device = 'Mobile 4 Series Chipset Integrated Graphics Controller'
class = display
subclass = VGA
vgapci1@pci0:0:2:1: class=0x038000 rev=0x07 hdr=0x00 vendor=0x8086 device=0x2a43 subvendor=0x17aa subdevice=0x20e4
vendor = 'Intel Corporation'
device = 'Mobile 4 Series Chipset Integrated Graphics Controller'
class = display
<...>
Vendor: Intel Open Source Technology Center (0x8086)
Device: Mesa DRI Mobile Intel(R) GM45 Express Chipset (CTG) (0x2a42)
Version: 21.3.8
Accelerated: yes
Video memory: 1536MB
Unified memory: yes
Preferred profile: compat (0x2)
Max core profile version: 0.0
Max compat profile version: 2.1
Max GLES1 profile version: 1.1
Max GLES[23] profile version: 2.0
<...>
mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa
mesa-dri-21.3.8 OpenGL hardware acceleration drivers for DRI2+
mesa-libs-21.3.8 OpenGL libraries that support GLX and EGL clients
<...>
libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5)
xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets
(In reply to jakub_lach from comment #1) $ kldstat Id Refs Address Size Name 1 49 0xffffffff80200000 d56500 kernel 2 1 0xffffffff80f57000 489e8 snd_hda.ko 3 1 0xffffffff80fa0000 74650 if_em.ko 4 1 0xffffffff81920000 3530 fdescfs.ko 5 1 0xffffffff81924000 17f8b8 i915kms.ko 6 1 0xffffffff81aa4000 72bd8 drm.ko 7 4 0xffffffff81b17000 1a170 linuxkpi.ko 8 2 0xffffffff81b32000 2220 backlight.ko 9 2 0xffffffff81b35000 30fc linuxkpi_gplv2.ko 10 3 0xffffffff81b39000 62d8 dmabuf.ko Currently at 13.1-STABLE #0 stable/13-dc96fb072: Fri Sep 9 09:17:51 CEST Something I never saw before - Sep 10 16:15:36 Thinkpad kernel: MCA: Bank 3, Status 0x9020004b0001010a Sep 10 16:15:36 Thinkpad kernel: MCA: Global Cap 0x0000000000000806, Status 0x0000000000000000 Sep 10 16:15:36 Thinkpad kernel: MCA: Vendor "GenuineIntel", ID 0x1067a, APIC ID 0 Sep 10 16:15:36 Thinkpad kernel: MCA: CPU 0 COR (Green) (Yellow) EN GCACHE L2 ERR error (In reply to jakub_lach from comment #0) Another one - FreeBSD 13.1-STABLE #0 stable/13-3f4e44f38 amd 64 <...> mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-21.3.8 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-21.3.8 OpenGL libraries that support GLX and EGL clients <...> drm-510-kmod-5.10.113_6 DRM drivers modules <...> drm-510-kmod-5.10.113_6 DRM drivers modules libdrm-2.4.113,1 Userspace interface to kernel Direct Rendering Module services <...> libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets drmn0: [drm] GPU HANG: ecode 4:1:ca8ffffb, in MainThread [100642] drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 drmn0: [drm] Xorg[100642] context reset due to GPU hang drmn0: [drm] GPU HANG: ecode 4:1:cecffffb, in MainThread [100642] Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 fault virtual address = 0x61 fault code = supervisor read data, page not present instruction pointer = 0x20:0xffffffff8072e388 stack pointer = 0x28:0xfffffe00baa97b60 frame pointer = 0x28:0xfffffe00baa97ba0 code segment = base rx0, limit 0xfffff, type 0x1b = DPL 0, pres 1, long 1, def32 0, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 0 (linuxkpi_short_wq_1) trap number = 12 panic: page fault cpuid = 0 time = 1663143125 Uptime: 22h24m1s The hangs happened before drm-510-kmod _but_ they did not lead to panics. (In reply to jakub_lach from comment #4) FWIW, saving pdfs as files in in Firefox (which triggers dialog box) causes hangs and panics described above in about half cases. Previously, it only lead to hangs (as mentioned). Currently on 13.1-STABLE #0 stable/13-b63021e00, after unsuccessfully trying to replicate with GENERIC/debug on, it looks to me that the problem doesn't exist if there is options COMPAT_LINUXKPI in kernel (linuxkpi module is not loaded). (In reply to jakub_lach from comment #6) Currently FreeBSD 13.1-STABLE #0 stable/13-1815de4fe, $ pkg info | grep 'kmod' drm-510-kmod-5.10.113_8 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.1_1 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.1_1 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.113_8 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services After I compiled in-kernel linuxkpi, I get hard restart/reset when opening dialog boxes in Firefox (sporadically), no trace of restart. (In reply to jakub_lach from comment #7) Currently testing if - #options COMPAT_LINUXKPI and adding linuxkpi and linuxkpi_wlan as modules (the latter just in case - I shouldn't need it) will change the output. I've got another hang with (modules out of kernel) - FreeBSD 13.1-STABLE #0 stable/13-03c82ccba amd64 $ kldstat Id Refs Address Size Name 1 49 0xffffffff80200000 d56f90 kernel 2 1 0xffffffff81596000 74560 if_em.ko 3 1 0xffffffff8160c000 48b20 snd_hda.ko 4 1 0xffffffff81920000 3530 fdescfs.ko 5 1 0xffffffff81924000 17f8b8 i915kms.ko 6 1 0xffffffff81aa4000 72bd8 drm.ko 7 4 0xffffffff81b17000 1d230 linuxkpi.ko 8 2 0xffffffff81b35000 2220 backlight.ko 9 2 0xffffffff81b38000 30fc linuxkpi_gplv2.ko 10 3 0xffffffff81b3c000 62d8 dmabuf.ko $ pkg info | grep 'kmod' drm-510-kmod-5.10.113_8 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.2_1 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.2_1 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.113_8 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services Jan 3 10:38:05 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Jan 3 10:38:05 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:ce8ffffb, in MainThread [100091] Jan 3 10:38:05 Thinkpad kernel: drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 Jan 3 10:38:05 Thinkpad kernel: drmn0: [drm] Xorg[100091] context reset due to GPU hang Jan 3 10:38:05 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:ce8ffffb, in MainThread [100091] Jan 3 10:38:05 Thinkpad kernel: Jan 3 10:38:05 Thinkpad syslogd: last message repeated 1 times Jan 3 10:38:05 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Jan 3 10:38:05 Thinkpad kernel: cpuid = 0; apic id = 00 Jan 3 10:38:05 Thinkpad kernel: fault virtual address = 0x61 Jan 3 10:38:05 Thinkpad kernel: fault code = supervisor read data, page not present Jan 3 10:38:05 Thinkpad kernel: instruction pointer = 0x20:0xffffffff807331b8 Jan 3 10:38:05 Thinkpad kernel: stack pointer = 0x28:0xfffffe00c7046b60 Jan 3 10:38:05 Thinkpad kernel: frame pointer = 0x28:0xfffffe00c7046ba0 Jan 3 10:38:05 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Jan 3 10:38:05 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Jan 3 10:38:05 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Jan 3 10:38:05 Thinkpad kernel: current process = 0 (linuxkpi_short_wq_3) Jan 3 10:38:05 Thinkpad kernel: trap number = 12 Jan 3 10:38:05 Thinkpad kernel: panic: page fault Jan 3 10:38:05 Thinkpad kernel: cpuid = 0 Jan 3 10:38:05 Thinkpad kernel: time = 1672738222 Jan 3 10:38:05 Thinkpad kernel: Uptime: 1d16h37m4s Jan 3 10:38:05 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Jan 3 10:38:05 Thinkpad kernel: Rebooting... Jan 3 10:38:05 Thinkpad kernel: ---<<BOOT>>--- (In reply to jakub_lach from comment #9) FreeBSD 13.1-STABLE #0 stable/13-c84ec3076 amd64 $ pkg info | grep 'kmod' drm-510-kmod-5.10.163 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.3_2 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.3_1 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.163 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services Jan 24 11:49:15 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Jan 24 11:49:15 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:cecffffb, in MainThread [100639] Jan 24 11:49:15 Thinkpad kernel: drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 Jan 24 11:49:15 Thinkpad kernel: drmn0: [drm] Xorg[100639] context reset due to GPU hang Jan 24 11:49:15 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:cecffffb, in MainThread [100639] Jan 24 11:49:15 Thinkpad kernel: Jan 24 11:49:15 Thinkpad syslogd: last message repeated 1 times Jan 24 11:49:15 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Jan 24 11:49:15 Thinkpad kernel: cpuid = 1; apic id = 01 Jan 24 11:49:15 Thinkpad kernel: fault virtual address = 0x61 Jan 24 11:49:15 Thinkpad kernel: fault code = supervisor read data, page not present Jan 24 11:49:15 Thinkpad kernel: instruction pointer = 0x20:0xffffffff80736298 Jan 24 11:49:15 Thinkpad kernel: stack pointer = 0x28:0xfffffe00baa97b60 Jan 24 11:49:15 Thinkpad kernel: frame pointer = 0x28:0xfffffe00baa97ba0 Jan 24 11:49:15 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Jan 24 11:49:15 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Jan 24 11:49:15 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Jan 24 11:49:15 Thinkpad kernel: current process = 0 (linuxkpi_short_wq_3) Jan 24 11:49:15 Thinkpad kernel: trap number = 12 Jan 24 11:49:15 Thinkpad kernel: panic: page fault Jan 24 11:49:15 Thinkpad kernel: cpuid = 1 Jan 24 11:49:15 Thinkpad kernel: time = 1674557051 Jan 24 11:49:15 Thinkpad kernel: Uptime: 3d3h49m42s Jan 24 11:49:15 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Jan 24 11:49:15 Thinkpad kernel: Rebooting... Jan 24 11:49:15 Thinkpad kernel: cpu_reset: Restarting BSP Jan 24 11:49:15 Thinkpad kernel: cpu_reset_proxy: Stopped CPU 1 Jan 24 11:49:15 Thinkpad kernel: ---<<BOOT>>--- (In reply to jakub_lach from comment #10) FreeBSD 13.1-STABLE #0 stable/13-df8c42f5a amd64 $ pkg info | grep 'kmod' drm-510-kmod-5.10.163 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.4 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.4 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_2,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.163 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services Jan 28 11:25:15 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:fdfffffb, in MainThread [100643] Jan 28 11:25:15 Thinkpad kernel: drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 Jan 28 11:25:15 Thinkpad kernel: drmn0: [drm] Xorg[100643] context reset due to GPU hang Jan 28 11:28:55 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Jan 28 11:28:55 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:ca8ffffb, in MainThread [100643] Jan 28 11:28:55 Thinkpad kernel: Jan 28 11:28:55 Thinkpad syslogd: last message repeated 1 times Jan 28 11:28:55 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Jan 28 11:28:55 Thinkpad kernel: cpuid = 0; apic id = 00 Jan 28 11:28:55 Thinkpad kernel: fault virtual address = 0x61 Jan 28 11:28:55 Thinkpad kernel: fault code = supervisor read data, page not present Jan 28 11:28:55 Thinkpad kernel: instruction pointer = 0x20:0xffffffff80736d48 Jan 28 11:28:55 Thinkpad kernel: stack pointer = 0x28:0xfffffe00c5fbdb60 Jan 28 11:28:55 Thinkpad kernel: frame pointer = 0x28:0xfffffe00c5fbdba0 Jan 28 11:28:55 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Jan 28 11:28:55 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Jan 28 11:28:55 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Jan 28 11:28:55 Thinkpad kernel: current process = 0 (linuxkpi_short_wq_1) Jan 28 11:28:55 Thinkpad kernel: trap number = 12 Jan 28 11:28:55 Thinkpad kernel: panic: page fault Jan 28 11:28:55 Thinkpad kernel: cpuid = 0 Jan 28 11:28:55 Thinkpad kernel: time = 1674901517 Jan 28 11:28:55 Thinkpad kernel: Uptime: 1d0h35m14s Jan 28 11:28:55 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Jan 28 11:28:55 Thinkpad kernel: Rebooting... Jan 28 11:28:55 Thinkpad kernel: ---<<BOOT>>--- (In reply to jakub_lach from comment #11) FreeBSD 13.2-PRERELEASE #0 stable/13-3d7a88248 amd64 $ pkg info | grep 'kmod' drm-510-kmod-5.10.163 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.4 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.4 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_3,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.163 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services Feb 7 13:16:42 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Feb 7 13:16:42 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:cecffffb, in MainThread [100645] Feb 7 13:16:42 Thinkpad kernel: drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 Feb 7 13:16:42 Thinkpad kernel: drmn0: [drm] Xorg[100645] context reset due to GPU hang Feb 7 13:16:42 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:cecffffb, in MainThread [100645] Feb 7 13:16:42 Thinkpad kernel: Feb 7 13:16:42 Thinkpad syslogd: last message repeated 1 times Feb 7 13:16:42 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Feb 7 13:16:42 Thinkpad kernel: cpuid = 1; apic id = 01 Feb 7 13:16:42 Thinkpad kernel: fault virtual address = 0x61 Feb 7 13:16:42 Thinkpad kernel: fault code = supervisor read data, page not present Feb 7 13:16:42 Thinkpad kernel: instruction pointer = 0x20:0xffffffff80736578 Feb 7 13:16:42 Thinkpad kernel: stack pointer = 0x28:0xfffffe00c7046b60 Feb 7 13:16:42 Thinkpad kernel: frame pointer = 0x28:0xfffffe00c7046ba0 Feb 7 13:16:42 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Feb 7 13:16:42 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Feb 7 13:16:42 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Feb 7 13:16:42 Thinkpad kernel: current process = 0 (linuxkpi_short_wq_3) Feb 7 13:16:42 Thinkpad kernel: trap number = 12 Feb 7 13:16:42 Thinkpad kernel: panic: page fault Feb 7 13:16:42 Thinkpad kernel: cpuid = 1 Feb 7 13:16:42 Thinkpad kernel: time = 1675772031 Feb 7 13:16:42 Thinkpad kernel: Uptime: 1d23h28m43s Feb 7 13:16:42 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Feb 7 13:16:42 Thinkpad kernel: --> Press a key on the console to reboot, Feb 7 13:16:42 Thinkpad kernel: --> or switch off the system now. Feb 7 13:16:42 Thinkpad kernel: Rebooting... Feb 7 13:16:42 Thinkpad kernel: cpu_reset: Restarting BSP Feb 7 13:16:42 Thinkpad kernel: cpu_reset_proxy: Stopped CPU 1 Feb 7 13:16:42 Thinkpad kernel: ---<<BOOT>>--- FreeBSD 13.2-STABLE #0 stable/13-77733aaa5 amd64 $ pkg info | grep 'kmod' drm-510-kmod-5.10.163 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.5 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.5 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20221108 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_3,1 X.Org legacy driver for Intel integrated graphics chipsets $ pkg info | grep 'drm' drm-510-kmod-5.10.163 DRM drivers modules libdrm-2.4.114,1 Userspace interface to kernel Direct Rendering Module services Feb 13 19:58:44 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Feb 13 19:58:44 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:ca8ffffb, in MainThread [100633] Feb 13 19:58:44 Thinkpad kernel: drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 Feb 13 19:58:44 Thinkpad kernel: drmn0: [drm] Xorg[100633] context reset due to GPU hang Feb 13 19:58:44 Thinkpad kernel: drmn0: [drm] GPU HANG: ecode 4:1:ca8ffffb, in MainThread [100633] Feb 13 19:58:44 Thinkpad kernel: Feb 13 19:58:44 Thinkpad syslogd: last message repeated 1 times Feb 13 19:58:44 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Feb 13 19:58:44 Thinkpad kernel: cpuid = 0; apic id = 00 Feb 13 19:58:44 Thinkpad kernel: fault virtual address = 0x61 Feb 13 19:58:44 Thinkpad kernel: fault code = supervisor read data, page not present Feb 13 19:58:44 Thinkpad kernel: instruction pointer = 0x20:0xffffffff80736c98 Feb 13 19:58:44 Thinkpad kernel: stack pointer = 0x28:0xfffffe00c705ab60 Feb 13 19:58:44 Thinkpad kernel: frame pointer = 0x28:0xfffffe00c705aba0 Feb 13 19:58:44 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Feb 13 19:58:44 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Feb 13 19:58:44 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Feb 13 19:58:44 Thinkpad kernel: current process = 0 (linuxkpi_short_wq_1) Feb 13 19:58:44 Thinkpad kernel: trap number = 12 Feb 13 19:58:44 Thinkpad kernel: panic: page fault Feb 13 19:58:44 Thinkpad kernel: cpuid = 0 Feb 13 19:58:44 Thinkpad kernel: time = 1676312082 Feb 13 19:58:44 Thinkpad kernel: Uptime: 10h35m44s Feb 13 19:58:44 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Feb 13 19:58:44 Thinkpad kernel: Rebooting... Feb 13 19:58:44 Thinkpad kernel: ---<<BOOT>>--- (In reply to jakub_lach from comment #13) Another reboot - first time after updating src, no trace in messages. FreeBSD 13.2-STABLE #0 stable/13-f3f350c5c amd64 $ pkg info | grep 'kmod' drm-510-kmod-5.10.163_2 DRM drivers modules $ pkg info | grep 'mesa' mesa-demos-8.4.0_3 OpenGL demos distributed with Mesa mesa-dri-22.3.5 OpenGL hardware acceleration drivers for DRI2+ mesa-libs-22.3.5 OpenGL libraries that support GLX and EGL clients $ pkg info | grep 'intel' devcpu-data-intel-20230214 Intel CPU microcode updates libva-intel-driver-2.4.1_1 VAAPI legacy driver for Intel GMA 4500 (Gen4) to UHD 630 (Gen9.5) xf86-video-intel-2.99.917.916_3,1 X.Org legacy driver for Intel integrated graphics chipsets now, that i updated to FreeBSD-CURRENT for testing purposes, i had less frequents hangs with master branch of drm-kmod ( which was linux-5.12.x ). Now, after, drm-kmod update to linux-5.15.x hangs and GPU resets returned. System freezes and only hard reset required to bring system back. Only thing i seeing is: drmn0: [drm] GPU HANG: ecode 6:1:0ffe0000, in Renderer [101196] drmn0: [drm] Resetting chip for stopped heartbeat on rcs0 drmn0: [drm] Renderer[101196] context reset due to GPU hang Notice, that i already reported this in: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=259670 To me this looks like a grave issue, which is not fully addressed and simple updates to newer linux graphic stack would not help. Since, it is really hard to get meaningful information and ability to debug by regular users, i propose this issues, which are quite common, to be reported to FreeBSD foundation/core developers as it looks like subtle bug within FreeBSD linuxkpi layer. (In reply to Oleh Vinichenko from comment #15) Do any of you have a way to reproduce this reliably ? I've never encounter this problems on any of my i915 machine (3-4 machines here). problem is that it happens out of nowhere, in random fashion. but always, when running firefox or any application that is using gpu. this happens on this particular box, so it might be specific to it's hardware. in #259670 i provided some information but i do not know how helpful it is, neither i have knowledge how to to debug such events. if there are steps or ideas how to debug, please, share. it was very very reliable and stable prior to drm-kmod-5.4 (In reply to Emmanuel Vadot from comment #16) Usually triggered in Firefox upon opening a dialog window (save file or similar). (In reply to jakub_lach from comment #18) Not always - usually after a longer uptime. (In reply to jakub_lach from comment #19) The hangs happened before drm-510-kmod _but_ they did not lead to panics. i ran 20 random save page as, which saved pdf and it does not trigger anything You both have old i915 chipset so I'm not surprised that only you two have reported this problem. Fun. Seriously. (In reply to Oleh Vinichenko from comment #21) Here it would hang on first instance of dialog box spawn - or not at all (as in your case). To me it looks state-depended, that is, conditional on longer uptime, heavier memory usage (maybe). (In reply to Oleh Vinichenko from comment #15) Do you run GENERIC? I wonder if there is some relation to running custom kernel. I use GENERIC. On Linux a /sys/class/drm/card0/error is created on a GPU Hang... Do we get that file on FreeBSD too? Also found:
torvalds/linux:3d7cb6b0:drivers/gpu/drm/i915/i915_gpu_error.c
}
len = scnprintf(error->error_msg, sizeof(error->error_msg),
"GPU HANG: ecode %d:%x:%08x",
GRAPHICS_VER(error->i915), hung_classes,
generate_ecode(first));
if (first && first->context.pid) {
So first number after ecode would be the version of GPU (I think)... second an indication of the cause , and third I would guess the address.
(In reply to Paul Dufresne from comment #27) looks like we have no sys/class... and I doubt that creating the directory would help the driver write the information. Have you tried the "modesetting" driver:
Like:
paul@dufresnep:~ $ cat /usr/local/etc/X11/xorg.conf.d/20-intel.conf
Section "Device"
Identifier "Card"
#Driver "vesa"
#Driver "intel"
#Driver "scfb"
# does not exist: Driver "glamor"
Driver "modesetting"
EndSection
paul@dufresnep:~ $
(In reply to Paul Dufresne from comment #30) Thanks, but this is hardly a solution. Note, that I had working/stable accelerated intel driver for years prior to drm-510-kmod. I've already experienced old drm, newcons and drm2/kms2 introduction. At that point, intel was more or less stable for me. Last crashes (and corruption) I had was related to forcing 'sna' acceleration method. At some point, I've used drm-legacy-kmod, as I had believed at that time it was last one supporting GM45. It was also stable iirc. I could live with some hang ups as long as they did not crash the system (prior to drm-510-kmod). (In reply to jakub_lach from comment #31) >Note, that I had working/stable accelerated intel driver for years prior to rm-510-kmod. You seems to consider modesetting as no hardware acceleration. Most likely, modesetting would use Glamor as hardware acceleration. A kind of hardware acceleration that use the card 3D to achieve 2D acceleration. Old 2D acceleration, is I believed, a bit forgotten by developers... because they prefer Glamor. Also, and I am less sure... I think Glamor also used the drm stack... so testing with Intel would allows like help to determine if the problem is within the Xorg Intel driver or in the drm stack. But yeah... frankly I think most people should just use modesetting with Glamor acceleration and avoid the Xorg intel driver. (In reply to Paul Dufresne from comment #32) I've accidentally used modesetting a while ago, it was noticeably slower than intel (that's how I've noticed something was not right). (In reply to jakub_lach from comment #33) FWIW, modesetting is now still somewhat slower in real applications (ioquake3), though faster in glxgears (using glamor) and YouTube performance is approximately the same. I think the last time I've checked it could be preglamor. I will probably run modesetting now to check if the linuxkpi problem reoccurs. (In reply to jakub_lach from comment #31) FWIW, I've noticed that intel uses 'UXA' by default here. I've tried forcing SNA (as I had in the past) - there was no corruption and glxgears were faster than with modesetting, however I've got instacrash when trying to run Firefox Feb 22 13:19:51 Thinkpad syslogd: last message repeated 1 times Feb 22 13:25:46 Thinkpad syslogd: kernel boot file is /boot/kernel/kernel Feb 22 13:25:46 Thinkpad kernel: Feb 22 13:25:46 Thinkpad syslogd: last message repeated 1 times Feb 22 13:25:46 Thinkpad kernel: Fatal trap 12: page fault while in kernel mode Feb 22 13:25:46 Thinkpad kernel: cpuid = 1; apic id = 01 Feb 22 13:25:46 Thinkpad kernel: fault virtual address = 0x0 Feb 22 13:25:46 Thinkpad kernel: fault code = supervisor read data, page not present Feb 22 13:25:46 Thinkpad kernel: instruction pointer = 0x20:0xffffffff82055cc7 Feb 22 13:25:46 Thinkpad kernel: stack pointer = 0x28:0xfffffe00c859ad00 Feb 22 13:25:46 Thinkpad kernel: frame pointer = 0x28:0xfffffe00c859ad70 Feb 22 13:25:46 Thinkpad kernel: code segment = base rx0, limit 0xfffff, type 0x1b Feb 22 13:25:46 Thinkpad kernel: = DPL 0, pres 1, long 1, def32 0, gran 1 Feb 22 13:25:46 Thinkpad kernel: processor eflags = interrupt enabled, resume, IOPL = 0 Feb 22 13:25:46 Thinkpad kernel: current process = 0 (i915-userptr-acquir) Feb 22 13:25:46 Thinkpad kernel: trap number = 12 Feb 22 13:25:46 Thinkpad kernel: panic: page fault Feb 22 13:25:46 Thinkpad kernel: cpuid = 1 Feb 22 13:25:46 Thinkpad kernel: time = 1677068559 Feb 22 13:25:46 Thinkpad kernel: Uptime: 6m55s Feb 22 13:25:46 Thinkpad kernel: Automatic reboot in 15 seconds - press a key on the console to abort Feb 22 13:25:46 Thinkpad kernel: --> Press a key on the console to reboot, Feb 22 13:25:46 Thinkpad kernel: --> or switch off the system now. Feb 22 13:25:46 Thinkpad kernel: Rebooting... Feb 22 13:25:46 Thinkpad kernel: cpu_reset: Restarting BSP Feb 22 13:25:46 Thinkpad kernel: cpu_reset_proxy: Stopped CPU 1 Sorry... I am still a bit newbie. Is the information you give are in /var/messages files, or is it you that get it from procedure described in: https://docs.freebsd.org/en/books/developers-handbook/kerneldebug/ I note that: dumpon -l shows the device where crash info is saved in swap area before rebooting after a kernel panic savecore -C -v (to be done after the reboot I think) would show if a crash report exist if it exist: savecore -v should save core crash in /var/crash while savecore -v . would save core crash in current directory then: kgdb -n last would get you in kernel debugger where info you given would be shown. but the interesting part is after, read lines 41 and 53 in the: https://docs.freebsd.org/en/books/developers-handbook/kerneldebug/ where you get shown source code, and the backtrace. (In reply to Paul Dufresne from comment #36) Those are from /var/log/messages, as a side note that SNA is not stable for me with intel on gen4 (GM45). I do not have debug on/crash dumps enabled as of now (iirc compiling Firefox with debug was almost impossible for me). I think files won't happens in /var/crash until you do (after reboot): savecore -v I think you would need to rebuild the kernel with debug symbols: so with a file like: root@dufresnep:/var/crash # cat /etc/make.conf WITH_DEBUG=YES DEBUG_FLAGS= -g -O0 root@dufresnep:/var/crash # Read https://docs.freebsd.org/en/books/developers-handbook/kernelbuild/ Don't know hard it looks to you. (In reply to Paul Dufresne from comment #39) Yes, I know that I would need to (at least) enable debug and configure dump device to debug linuxkpi crash further. See my previous message (comment #6), last time I turned on debug with GENERIC, I couldn't replicate the crash. Thinking aloud... root@dufresnep:/var/crash # pkg which /boot/modules/i915kms.ko /boot/modules/i915kms.ko was installed by package drm-510-kmod-5.10.163_2 So, maybe it is not so much recompiling the kernel with debug symbols, but more: make sure: root@dufresnep:/var/crash # cat /etc/make.conf WITH_DEBUG=YES DEBUG_FLAGS= -g -O0 root@dufresnep:/var/crash # root@dufresnep:/var/crash # cd /usr/ports/graphics/drm-kmod # make # make install This would help to figure out problem with sna acceleration but not so much for GPU hang, because GPU hang have no clear point of failure... What would help maybe with GPU hang is: sysctl compat.linuxkpi.drm_debug:1 Maybe lower: sys.class.drm.card0.engine.rcs0.heartbeat_interval_ms like: sys.class.drm.card0.engine.rcs0.heartbeat_interval_ms:1000 each second, rather than each 2.5 second (2500 ms) I believe bug #268138 is the same. (In reply to jakub_lach from comment #35) I searched: i915-userptr-acquir and come to this patch: https://lore.kernel.org/lkml/20191010083524.098009692@linuxfoundation.org/ With that patch: Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=203317 Fixes: 5cc9ed4b9a7a ("drm/i915: Introduce mapping of user pages into video memory (userptr) ioctl") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: stable@vger.kernel.org Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20190708140327.26825-1-chris@chris-wilson.co.uk Signed-off-by: Sasha Levin <sashal@kernel.org> --- drivers/gpu/drm/i915/i915_gem_userptr.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c index 05ae8c4a8a1b6..9760b67dab28b 100644 --- a/drivers/gpu/drm/i915/i915_gem_userptr.c +++ b/drivers/gpu/drm/i915/i915_gem_userptr.c @@ -691,7 +691,15 @@ i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj, for_each_sgt_page(page, sgt_iter, pages) { if (obj->mm.dirty) - set_page_dirty(page); + /* + * As this may not be anonymous memory (e.g. shmem) + * but exist on a real mapping, we have to lock + * the page in order to dirty it -- holding + * the page reference is not sufficient to + * prevent the inode from being truncated. + * Play safe and take the lock. + */ + set_page_dirty_lock(page); mark_page_accessed(page); put_page(page); -- 2.20.1 Need to find where the equivalent for us is and see if it is applied. (In reply to Paul Dufresne from comment #44) - userptr isn't supported on FreeBSD except the unstable version per https://github.com/FreeBSDDesktop/kms-drm/issues/197 - xf86-video-intel prefers the unstable version since https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/commit/dd66ba8e5666 - drm-515-kmod no longer supports the unstable version since https://github.com/torvalds/linux/commit/c6bcc0c2fdfd Given the above one can try disabling userptr in xf86-video-intel e.g., --- src/sna/kgem.c.orig 2021-01-15 20:59:05 UTC +++ src/sna/kgem.c @@ -68,7 +68,7 @@ search_snoop_cache(struct kgem *kgem, unsigned int num #define DBG_NO_CACHE_LEVEL 0 #define DBG_NO_CPU 0 #define DBG_NO_CREATE2 0 -#define DBG_NO_USERPTR 0 +#define DBG_NO_USERPTR 1 #define DBG_NO_UNSYNCHRONIZED_USERPTR 0 #define DBG_NO_COHERENT_MMAP_GTT 0 #define DBG_NO_LLC 0 For me sna works flawlessly (with uxa I get: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=269606 ). On i3-8100... root@dufresnep:/usr/home/paul # uname -v FreeBSD 13.2-BETA2 releng/13.2-fbb102b2c GENERIC I did not apply a patch like suggested in comment #45. (In reply to Jan Beich from comment #45) To test this you would do (as a newbie helping others): as root (su -): cd /usr/ports/x11-drivers/xf86-video-intel/ make clean make fetch extract patch cp work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c.orig nano work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c [change line 71 to change 0 to 1] Ctrl-O make makepatch make build make deinstall make install exit (to exit root) Hope I am really helping. Computer still working with previous patch... But as it does not need it in my case, erasing it: su cd /usr/ports/x11-drivers/xf86-video-intel/ rm files/patch-src_sna_kgem.c make clean deinstall install libtool --finish /usr/local/lib/xorg/modules/drivers last step suggested by previous command output. (In reply to Paul Dufresne from comment #47) I did mislead you! Don't do: cp work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c.orig nano work/xf86-video-intel-31486f40f8e8f8923ca0799aea84b58799754564/src/sna/kgem.c because the .orig file was existing... and so the file was patched by doing this step I am overwriting the patch... erasing it, at the next step: make makepatch (which will compare .c.orig with .c file). Do that only when the .orig file did not exist. (In reply to Paul Dufresne from comment #48) Don't do that! The file was patched... and by erasing the file you also erase the previous changes for FreeBSD! Please edit the .c file to change back your change. Then: make makepatch make build make deinstall make install Tried NomadBSD with 13.1 version. (with intel sna accel method). Had to place my file in /usr/local/share/X11/xorg.conf.d/ Works well with my i3-8100... But I believe this is not the same driver: nomad@NomadBSD ~> pkg info xf86-video-intel xf86-video-intel-2.99.917.916_2,1 Name : xf86-video-intel Version : 2.99.917.916_2,1 Installed on : Wed Nov 30 13:03:23 2022 EST Origin : x11-drivers/xf86-video-intel Architecture : FreeBSD:13:amd64 Prefix : /usr/local Categories : x11-drivers Licenses : MIT Maintainer : x11@FreeBSD.org WWW : https://01.org/linuxgraphics/ Comment : X.Org legacy driver for Intel integrated graphics chipsets I think you said the legacy version was removed in FreeBSD. So unsure if it is the driver version... or because my GPU is more recent that it works for me. (In reply to Paul Dufresne from comment #51) It would be the same driver, if you would update your system in line with ports, however as yours is a newer card, it might just use sna by default. (In reply to Jan Beich from comment #45) Currently trying to replicate with modesetting, if I try intel again (with sna), will try that, thanks. Trying sna on intel on an other computer:
vgapci1@pci0:0:2:0: class=0x030000 rev=0x06 hdr=0x00 vendor=0x8086 device=0x0412 subvendor=0x103c subdevice=0x18e7
vendor = 'Intel Corporation'
device = 'Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller'
class = display
subclass = VGA
nomad@NomadBSD /v/log [1]> uname -v
FreeBSD 13.1-RELEASE-p5 NOMADBSD
Seems to works fine too.
(In reply to Paul Dufresne from comment #53) with uxa: not testing much... but working... with: Feb 24 12:10:33 NomadBSD kernel: drmn1: <drmn> on vgapci1 Feb 24 12:10:33 NomadBSD kernel: vgapci1: child drmn1 requested pci_enable_io Feb 24 12:10:33 NomadBSD syslogd: last message repeated 1 times Feb 24 12:10:33 NomadBSD kernel: [drm] Unable to create a private tmpfs mount, hugepage support will be disabled(-19). Feb 24 12:10:33 NomadBSD kernel: [drm] Got stolen memory base rxcc200000, size 0x2000000 Feb 24 12:10:33 NomadBSD kernel: sysctl_warn_reuse: can't re-use a leaf (hw.dri.debug)! Feb 24 12:10:33 NomadBSD kernel: [drm] Initialized i915 1.6.0 20200917 for drmn1 on minor 0 Feb 24 12:10:34 NomadBSD kernel: VT: Replacing driver "efifb" with new "fb". Feb 24 12:10:34 NomadBSD kernel: start FB_INFO: Feb 24 12:10:34 NomadBSD kernel: type=11 height=900 width=1440 depth=32 Feb 24 12:10:34 NomadBSD kernel: pbase=0xd0000000 vbase=0xfffff800d0000000 Feb 24 12:10:34 NomadBSD kernel: name=drmn1 flags=0x0 stride=5760 bpp=32 Feb 24 12:10:34 NomadBSD kernel: end FB_INFO Feb 24 12:10:34 NomadBSD kernel: sysctl: unknown oid 'dev.hdac.3.polling' at line 45 Feb 24 12:10:34 NomadBSD kernel: Feb 24 12:10:34 NomadBSD kernel: acpi_hp0: <HP ACPI-WMI Mapping> on acpi_wmi0 Feb 24 12:10:34 NomadBSD kernel: acpi_hp0: HP event GUID detected, installing event handler Feb 24 12:10:34 NomadBSD kernel: acpi_video1: <ACPI video extension> on vgapci1 Feb 24 12:10:34 NomadBSD kernel: Feb 24 12:10:34 NomadBSD kernel: Feb 24 12:10:46 NomadBSD pulseaudio[1649]: [(null)] oss-util.c: '/dev/dsp0' doesn't support full duplex Feb 24 12:10:46 NomadBSD pulseaudio[1649]: [(null)] oss-util.c: '/dev/dsp1' doesn't support full duplex Feb 24 12:10:46 NomadBSD pulseaudio[1649]: [(null)] oss-util.c: '/dev/dsp2' doesn't support full duplex Feb 24 12:10:46 NomadBSD pulseaudio[1649]: [(null)] oss-util.c: '/dev/dsp4' doesn't support full duplex Feb 24 12:10:51 NomadBSD devd[811]: notify_clients: send() failed; dropping unresponsive client (In reply to Paul Dufresne from comment #54) You should verify /var/log/Xorg.0.log to check which accel method is used, Sandybridge and newer will probabaly default to sna. (In reply to Jan Beich from comment #45) FWIW, after applying your patch (added under xf86-video-intel/files to keep it simple), Firefox runs with intel/sna (GM45). No idea about stability. Previously I've dropped sna due to corruption (I believe this is solved) and occasional hangups. As far as glxgears goes - sna > modesetting > uxa. YouTube performance is the same, 1080p being maximum usable resolution. (In reply to jakub_lach from comment #56) FWIW, I'v triggered similar corruption with SNA just now (as previously reported) https://bz-attachments.freebsd.org/attachment.cgi?id=231765 though this time in Firefox save dialog - that's still better than hangup/crash with uxa in similar scenario. (In reply to jakub_lach from comment #55) man intel (on FreeBSD) shows that UXA is default, and I can confirm that from Xorg.0.log: [ 52.583] (II) Module "dri2" already built-in [ 52.583] (II) intel(0): Allocated new frame buffer 1280x1024 stride 5120, tiled [ 52.606] (II) UXA(0): Driver registered support for the following operations: [ 52.606] (II) solid [ 52.606] (II) copy [ 52.606] (II) put_image [ 52.606] (II) get_image [ 52.606] (II) intel(0): [DRI2] Setup complete [ 52.606] (II) intel(0): [DRI2] DRI driver: iris [ 52.606] (II) intel(0): [DRI2] VDPAU driver: va_gl But on Ubuntu: https://manpages.ubuntu.com/manpages/xenial/man4/intel.4.html default is sna, apparently There is also "AccelMethod" "blt": (limited UXA) or "blt" to disable render acceleration and only use the BLT engine I found info on BLT engine in section 3.2.2.2 of: https://www.intel.ca/content/dam/doc/datasheet/atom-d2000-n2000-vol-1-datasheet.pdf You might want to check: Option "RelaxedFencing" "boolean" This option controls whether we attempt to allocate the minimal amount of memory required for the buffers. The reduction in working set has a substantial improvement on system performance. However, this has been demonstrate to be buggy on older hardware (845-865 and 915-945, but ok on PineView and later) so on those chipsets defaults to off. Default: Enabled for G33 (includes PineView), and later, class machines. Should be decided by your GPU, but might make sure. We use UXA by default because: # XXX bug 214593: SNA crashes on pre-SandyBridge hardware CONFIGURE_ARGS+=--with-default-accel=uxa See bug #214593 Running with Firefox for a few hours on a fresh install of:
root@Simbad:/var/log # freebsd-version -ru
13.2-BETA3
13.2-BETA3
not yet officially released, and with a similar GPU than jakub:
root@Simbad:/var/log # pciconf -lev
...
vgapci0@pci0:0:2:0: class=0x030000 rev=0x03 hdr=0x00 vendor=0x8086 device=0x2e12 subvendor=0x103c subdevice=0x3048
vendor = 'Intel Corporation'
device = '4 Series Chipset Integrated Graphics Controller'
class = display
subclass = VGA
Now using UXA, been able to use sna without recompiling anything.
I am thinking about installing 13.1 to see if I can replicate this bug.
(In reply to jakub_lach from comment #34) There have been no problems with modesetting here so far, apart from one message (no other effect) - Feb 25 09:00:33 Thinkpad kernel: drmn0: [drm] *ERROR* CPU pipe A FIFO underrun (In reply to Paul Dufresne from comment #60) The thing I was reporting with intel/uxa with GM45 is not something I could reliably replicate in few hours. The crashes I've reported are few days apart (see also included uptime). (In reply to jakub_lach from comment #62) As far as I can tell, the conclusion is: 1. I cannot replicate linuxkpi crash with modesetting. 2. linuxkpi crash only occurs here with UXA. 3. SNA needs Jan Beich's patch from comment #45, to not crash when starting Firefox. With patch, there were no crashes/hangs, however he corruption appears in similar scenarios to crashes with UXA (dialog spawn in Firefox). |
