[Bug 1206316] New: [RKL] add_drivers+=" i915 " included in /etc/dracut.d/*, cause drm/dri issue(s) that delay initial X startup ~90 seconds
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316
Bug ID: 1206316 Summary: [RKL] add_drivers+=" i915 " included in /etc/dracut.d/*, cause drm/dri issue(s) that delay initial X startup ~90 seconds Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: x86-64 OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-bugs@opensuse.org Reporter: mrmazda@earthlink.net QA Contact: qa-bugs@suse.de CC: sndirsch@suse.com, tzimmermann@suse.com Found By: --- Blocker: ---
Created attachment 863458 --> http://bugzilla.opensuse.org/attachment.cgi?id=863458&action=edit .tgz of Xorg.0.log + journalctl -b (with drm.debug=0x1e log_buf_len=1M)
Initial summary: [RKL] add_drivers+=" i915 " included in /etc/dracut.d/*, cause drm/dri issue(s) that delay initial X startup ~90 seconds
To reproduce: 1-include add_drivers+=" i915 " in /etc/dracut.d/local.conf 2-dracut -f 3-reboot
Actual behavior: 1-~90 seconds between shell prompt appearance on tty1 and TDM greeter appearance on tty7 2-Grep of /dev/dri/card from the attachment: [ 93.509] (II) xfree86: Adding drm device (/dev/dri/card1) [ 93.575] (II) modeset(0): using drv /dev/dri/card1 Dec 12 03:57:11 ab560 systemd[1]: Timed out waiting for device /dev/dri/card0. 3-# dmesg | grep -i Faile [ 2.410971] simple-framebuffer simple-framebuffer.0: [drm:drm_atomic_helper_check_planes] [CRTC:34:crtc-0] atomic driver check failed [ 2.410979] simple-framebuffer simple-framebuffer.0: [drm:drm_atomic_check_only] atomic driver check for 000000008eb548ce failed: -22 [ 2.548403] i915 0000:00:02.0: [drm:intel_hdmi_set_edid [i915]] HDMI GMBUS EDID read failed, retry using GPIO bit-banging [ 2.738080] i915 0000:00:02.0: [drm:intel_hdmi_set_edid [i915]] HDMI GMBUS EDID read failed, retry using GPIO bit-banging [ 2.814605] i915 0000:00:02.0: [drm:intel_hdmi_set_edid [i915]] HDMI GMBUS EDID read failed, retry using GPIO bit-banging [ 3.640125] platform regulatory.0: Direct firmware load for regulatory.db failed with error -2 [ 3.640127] cfg80211: failed to load regulatory.db [ 3.870867] asus_wmi: fan_curve_get_factory_default (0x00110024) failed: -61 [ 3.870938] asus_wmi: fan_curve_get_factory_default (0x00110025) failed: -61 [ 93.481800] i915 0000:00:02.0: [drm:i915_perf_remove_config_ioctl [i915]] Failed to remove unknown OA config [ 93.554205] i915 0000:00:02.0: [drm:intel_hdmi_set_edid [i915]] HDMI GMBUS EDID read failed, retry using GPIO bit-banging [ 93.626006] i915 0000:00:02.0: [drm:intel_hdmi_set_edid [i915]] HDMI GMBUS EDID read failed, retry using GPIO bit-banging 4-in the attachment
Expected behavior: 1-# systemd-analyze critical-chain (kernel 6.0.10 without including driver i915 in initrd) .... graphical.target @3.640s ������multi-user.target @3.640s ������kbdsettings.service @1.564s +2.075s ������basic.target @1.560s ������sockets.target @1.560s ������telnet.socket @1.560s ������sysinit.target @1.558s ������systemd-timesyncd.service @1.538s +20ms ������systemd-tmpfiles-setup.service @1.528s +9ms ������local-fs.target @1.527s ������run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount @1.529s ������local-fs-pre.target @514ms ������lvm2-monitor.service @378ms +135ms ������dm-event.socket @315ms ������system.slice ������-.slice 2-Xorg.0.log begins at ~[ 4.7s] 3-Nice, fast boot to greeter
# inxi -CSGz System: Kernel: 6.0.10-1-default arch: x86_64 bits: 64 Desktop: Trinity v: R14.0.13 Distro: openSUSE Tumbleweed 20221210 CPU: Info: 6-core model: 11th Gen Intel Core i5-11400 bits: 64 type: MT MCP cache: L2: 3 MiB Speed (MHz): avg: 2600 min/max: 800/4400 cores: 1: 2600 2: 2600 3: 2600 4: 2600 5: 2600 6: 2600 7: 2600 8: 2600 9: 2600 10: 2600 11: 2600 12: 2600 Graphics: Device-1: Intel RocketLake-S GT1 [UHD Graphics 730] driver: i915 v: kernel Display: x11 server: X.Org v: 21.1.4 driver: X: loaded: modesetting unloaded: fbdev,vesa dri: iris gpu: i915 resolution: 1: 2560x1440~60Hz 2: 1920x1200~60Hz 3: 1680x1050~60Hz API: OpenGL v: 4.6 Mesa 22.2.4 renderer: Mesa Intel Graphics (RKL GT1)
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c1
Felix Miata mrmazda@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|[RKL] add_drivers+=" i915 " |add_drivers+=" i915 " |included in |included in |/etc/dracut.d/*, cause |/etc/dracut.d/*, cause |drm/dri issue(s) that delay |drm/dri issue(s) that delay |initial X startup ~90 |initial X startup ~90 |seconds |seconds
--- Comment #1 from Felix Miata mrmazda@earthlink.net --- Same problem on Kaby Lake: # inxi -CDGSz System: Kernel: 6.0.12-1-default arch: x86_64 bits: 64 Desktop: Trinity v: R14.0.13 Distro: openSUSE Tumbleweed 20221212 CPU: Info: quad core model: Intel Core i5-7500T bits: 64 type: MCP cache: L2: 1024 KiB Speed (MHz): avg: 2700 min/max: 800/3300 cores: 1: 2700 2: 2700 3: 2700 4: 2700 Graphics: Device-1: Intel HD Graphics 630 driver: i915 v: kernel Display: x11 server: X.Org v: 21.1.4 driver: X: loaded: modesetting unloaded: fbdev,vesa dri: iris gpu: i915 resolution: 1: 2560x1440~60Hz 2: 1680x1050~60Hz 3: 1920x1200~60Hz API: OpenGL v: 4.6 Mesa 22.2.4 renderer: Mesa Intel HD Graphics 630 (KBL GT2) Drives: Local Storage: total: 119.24 GiB used: 33.42 GiB (28.0%) ID-1: /dev/nvme0n1 vendor: ZTC model: PCIEG3-128G size: 119.24 GiB # hdparm -t /dev/nvme0n1 /dev/nvme0n1: Timing buffered disk reads: 3406 MB in 3.00 seconds = 1135.21 MB/sec # systemd-analyze critical-chain ... graphical.target @1min 30.743s ������display-manager.service @1min 30.648s +94ms ������systemd-user-sessions.service @1.848s +4ms ������basic.target @1.758s ������sockets.target @1.757s ������telnet.socket @1.756s ������sysinit.target @1.714s ������systemd-timesyncd.service @1.666s +46ms ������systemd-tmpfiles-setup.service @1.640s +21ms ������local-fs.target @1.603s ������run-user-0.mount @17.198s ������local-fs-pre.target @545ms ������lvm2-monitor.service @478ms +66ms ������dm-event.socket @443ms ������system.slice ������-.slice
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c2
--- Comment #2 from Felix Miata mrmazda@earthlink.net --- Created attachment 863514 --> http://bugzilla.opensuse.org/attachment.cgi?id=863514&action=edit Xorg.0.log + tdm.log + dmesg | grep ailed + journalctl -b booted with drm.debug=0x1e log_buf_len=1M
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c3
--- Comment #3 from Felix Miata mrmazda@earthlink.net --- Created attachment 863515 --> http://bugzilla.opensuse.org/attachment.cgi?id=863515&action=edit list of added content of initrd constructed using add_drivers+=" i915 " following "early KMS" suggestion in bug 1188954
I used lsinitrd on 6.0.8 initrds constructed before and after including add_drivers+=" i915 " in /etc/dracut.d/, then used diff -y -W 366 on the output of those two commands, then removed the content lines that did not differ between the two except by their timestamps. 46 lines were added.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c4
--- Comment #4 from Stefan Dirsch sndirsch@suse.com --- May I ask why you add this? AFAIK the required KMS modules are added automatically to initrd.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316
Stefan Dirsch sndirsch@suse.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(mrmazda@earthlink | |.net)
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c5
Felix Miata mrmazda@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(mrmazda@earthlink | |.net) |
--- Comment #5 from Felix Miata mrmazda@earthlink.net --- 1-attempted bug 1188954 workaround 2-can't count on i915 being included in initrd without it 3-display goes black as soon as initrd finishes loading and stays that way until i915 loads, which happens so late that only ~15-30 boot message lines appear before shell prompt appears
Above is not across the board vis-a-vis Intel GPUs. I haven't tried to inventory which GPU's produce which combinations of above listed behaviors, or none of them at all.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206316 http://bugzilla.opensuse.org/show_bug.cgi?id=1206316#c7
--- Comment #7 from Felix Miata mrmazda@earthlink.net --- (In reply to Takashi Iwai from comment #6)
The manual tweak like yours might have some side effect, so likely a WONTFIX issue.
If bug 1188954 would go away I wouldn't need the tweak.
Why when they do start X is it happening on /dev/dri/card1 instead of /dev/dri/card0? I don't have to do this only on RKL, but also both my KBLs at least. I've lost track, but I think this is the general rule here no matter the GPUor chipset. e.g., KBL ab250: # systemd-analyze Startup finished in 14.862s (firmware) + 2.698s (loader) + 2.024s (kernel) + 2.606s (initrd) + 1min 30.834s (userspace) = 1min 53.027s graphical.target reached after 1min 30.802s in userspace. # systemd-analyze critical-chain ... graphical.target @1min 30.802s ������display-manager.service @1min 30.703s +98ms ������systemd-logind.service @1.696s +54ms ������basic.target @1.677s ������sockets.target @1.676s ������telnet.socket @1.675s ������sysinit.target @1.635s ������systemd-timesyncd.service @1.595s +38ms ������systemd-tmpfiles-setup.service @1.573s +18ms ������local-fs.target @1.571s ������run-user-0.mount @43.172s ������local-fs-pre.target @606ms ������systemd-tmpfiles-setup-dev.service @561ms +43ms ������kmod-static-nodes.service @497ms +38ms ������systemd-journald.socket ������system.slice ������-.slice # Little different from comment #1 KBL host gb250.
kernel-bugs@lists.opensuse.org
-
bugzilla_noreply＠suse.com