[Bug 1177384] New: Black screen after updating to kernel 5.3.18-lp152.44-default
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 Bug ID: 1177384 Summary: Black screen after updating to kernel 5.3.18-lp152.44-default Classification: openSUSE Product: openSUSE Distribution Version: Leap 15.2 Hardware: i686 OS: openSUSE Leap 15.2 Status: NEW Severity: Critical Priority: P5 - None Component: Kernel Assignee: kernel-bugs@opensuse.org Reporter: grewin.mathias@online.de QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- After updating to kernel 5.3.18-lp152.44-default (Security Update) and rebooting, my monitor displays "Out of Range!" No login to text-console or graphics is possible and I can't get into the system. My Card is a Radeon HD 7750 (AMD) and the Radeon open source driver is running. So I turned back to the formerly kernel-version 5.3.18-lp152.41-default and all works fine. In the OpenSuSE-Forum also this problem is reported with other AMD graphic cards. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 Stephan Hemeier <Sauerlandlinux@gmx.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |Sauerlandlinux@gmx.de -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c1 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |grewin.mathias@online.de, | |tiwai@suse.com Flags| |needinfo?(grewin.mathias@on | |line.de) --- Comment #1 from Takashi Iwai <tiwai@suse.com> --- Could you give the dmesg output from both working and non-working boots? Also, please upload the hwinfo output, too. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c2 --- Comment #2 from Mathias Grewin <grewin.mathias@online.de> --- It's only possible to get into the running system with the elder kernel and the information you want is too much expensive for this forum - so I can't send them here (I tried!). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c3 --- Comment #3 from Takashi Iwai <tiwai@suse.com> --- (In reply to Mathias Grewin from comment #2)
It's only possible to get into the running system with the elder kernel and the information you want is too much expensive for this forum - so I can't send them here (I tried!).
Can you login via ssh from another machine? Also, it's not clear how "expensive"; if you think about the size, you can compress the dmesg output beforehand, of course. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c4 --- Comment #4 from Takashi Iwai <tiwai@suse.com> --- Or, you can get the output from journalctl or syslog of the previous (non-working) boot, too. That should suffice. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c5 Felix Miata (offline until ???) <mrmazda@earthlink.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mrmazda@earthlink.net --- Comment #5 from Felix Miata (offline until ???) <mrmazda@earthlink.net> --- Created attachment 842338 --> http://bugzilla.opensuse.org/attachment.cgi?id=842338&action=edit DVI failing journalctl -b and dmesg with chip ID: 1002:6779 Forum thread: https://forums.opensuse.org/showthread.php/545527-Screen-is-black-after-upda... Same thing with my AMD Caicos [Radeon HD 6450/7450/8450 / R5 230 OEM] vendor: Dell driver: radeon v: kernel chip ID: 1002:6779. DisplayPort works as expected, DVI port does not. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c6 Felix Miata (offline until ???) <mrmazda@earthlink.net> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(grewin.mathias@on | |line.de) | --- Comment #6 from Felix Miata (offline until ???) <mrmazda@earthlink.net> --- Created attachment 842339 --> http://bugzilla.opensuse.org/attachment.cgi?id=842339&action=edit DisplayPort journalctl -b, dmesg and hwinfo --gfxcard with chip ID: 1002:6779 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c7 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo? --- Comment #7 from Takashi Iwai <tiwai@suse.com> --- Thanks. Unfortunately there seems no obvious errors relevant with that through a quick glance. To check whether it's a regression in radeon kernel module or not, could you try the following as root? % mkdir /lib/modules/5.3.18-lp152.44-default/extra % cp /lib/modules/5.3.18-lp152.36-default/kernel/drivers/gpu/drm/radeon/radeon.ko /lib/modules/5.3.18-lp152.44-default/extra/ % depmod 5.3.18-lp152.44-default % mkinitrd Basically this copies the old radeon module onto the new kernel. Then reboot and retest with *.44 kernel. If this works, it's a regression in radeon driver. If this doesn't fix, it's something else (likely in drm core modules). You can do similar tests by copying drivers/gpu/drm*.ko files, too. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c8 --- Comment #8 from Mathias Grewin <grewin.mathias@online.de> --- Yes - excuse me, but there is no other machine by the time. I know if I need help I have to support it well. Because I'm not the only one with this problem I started this bug report but now time is my greatest problem. So I thank you for your help ... I will go with your proposal. The machine I use works fine with SuSE since eight years all the time - and also it does with the elder kernel. But we need information about the messages what happens with the "up-to-date" System. I'm interesting for it too! Thank you! -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c9 --- Comment #9 from Mathias Grewin <grewin.mathias@online.de> --- Now the new kernel 5.3.18-lp152.44-default is running, but only it does when I set "nomodeset" as a new Kernel Parameter! The resolution is only 1400 * 1050 by 77 Hz and there is no other setting possible. So I think, the kernel delivers wrong settings (without nomodeset). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c10 --- Comment #10 from Takashi Iwai <tiwai@suse.com> --- (In reply to Mathias Grewin from comment #9)
Now the new kernel 5.3.18-lp152.44-default is running, but only it does when I set "nomodeset" as a new Kernel Parameter! The resolution is only 1400 * 1050 by 77 Hz and there is no other setting possible. So I think, the kernel delivers wrong settings (without nomodeset).
The nomodeset option essentially disables the native graphics, hence it's no wonder that works around the problem but results in the fixed mode. Did you try the procedure in comment 7, and it still leads to a blank screen? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c11 Felix Miata (offline until ???) <mrmazda@earthlink.net> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo? | --- Comment #11 from Felix Miata (offline until ???) <mrmazda@earthlink.net> --- (In reply to Takashi Iwai from comment #7)
% mkdir /lib/modules/5.3.18-lp152.44-default/extra % cp /lib/modules/5.3.18-lp152.36-default/kernel/drivers/gpu/drm/radeon/radeon.ko /lib/modules/5.3.18-lp152.44-default/extra/ % depmod 5.3.18-lp152.44-default % mkinitrd
I erred in following the steps. I copied 5.3.18-lp152.41-default/kernel/drivers/gpu/drm/radeon/radeon.ko as .radeon.ko41 in the .44 /kernel/drivers/gpu/drm/radeon/ directory and 5.3.18-lp152.36-default/kernel/drivers/gpu/drm/radeon/radeon.ko and as .radeon.ko36 after renaming .44's radeon.ko .radeon.ko44. I then proceeded without the next steps (the /extra/ directory and running % depmod and % mkinitrd) to shutdown, swap DVI cable from DP cable, booted .44, and all seems to be working completely normally. I did note that radeon.ko for all of .36, .41 and .44 have the same byte count, 2,894,595. I see that .33's is 2894691. Then I copied .33's radeon.ko to 5.3.18-lp152.44-default/extra, but '% depmod 5.3.18-lp152.44-default' reported bash: fg: %: no such job My kernel last installed on this host before .44 is .20, so I suppose the regression may be from .36. So I'm paused to comment here. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c12 Felix Miata (offline until ???) <mrmazda@earthlink.net> changed: What |Removed |Added ---------------------------------------------------------------------------- Hardware|i686 |x86-64 --- Comment #12 from Felix Miata (offline until ???) <mrmazda@earthlink.net> --- I undid what I reported copied in comment 11, then followed the steps in comment 7 more precisely. The deviations: 1- didn't use % anywhere 2- used radeon.ko from kernel .33 rather than from .36 Problem with .44 is gone with the result. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c13 --- Comment #13 from Takashi Iwai <tiwai@suse.com> --- (In reply to Felix Miata (offline until ???) from comment #12)
I undid what I reported copied in comment 11, then followed the steps in comment 7 more precisely. The deviations:
1- didn't use % anywhere
2- used radeon.ko from kernel .33 rather than from .36
Problem with .44 is gone with the result.
OK, thanks, this information helps a lot. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c14 --- Comment #14 from Takashi Iwai <tiwai@suse.com> --- So there are only two patches for radeon between *.33 and *.36 kernels. I'm building a test kernel in OBS home:tiwai:bsc1177384 repo with the revert of those two. A package will appear later at http://download.opensuse.org/repositories/home:/tiwai:/bsc1177384/standard/ Please give it a try later. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c15 --- Comment #15 from Takashi Iwai <tiwai@suse.com> --- Oh, and the test kernel above is based on the latest Leap 15.2 KOTD. Please test the kernel below beforehand, and confirm that the problem persists. http://download.opensuse.org/repositories/Kernel:/openSUSE-15.2/standard/ If KOTD is broken and my test kernel works, it means that the revert is the fix. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c16 --- Comment #16 from Felix Miata (offline until ???) <mrmazda@earthlink.net> --- kernel-default-5.3.18-lp152.123.1.g27f6195.x86_64.rpm 03-Oct-2020 09:57 (standard) and kernel-default-5.3.18-lp152.1.1.g3d4aad4.x86_64.rpm 07-Oct-2020 10:00 (tiwai) both fail. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c17 Peter Kirchgeßner <peter@kirchgessner.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |peter@kirchgessner.net --- Comment #17 from Peter Kirchgeßner <peter@kirchgessner.net> --- Just a question about the kernel versions and radeon versions. Like the OP, I have kernel .44 and .41 on my system. Because of that problem with the kernel .44, I opened the related thread in the openSUSE forum. As fallback I used kernel .41, which works fine. I can see that the radeon.ko have the same size for kernels .44 and .41, but a different content. When I follow the instructions of mkdir/cp/depmod/mkinitrd with radeon.ko from kernel .41, the kernel .44 works too. So I would expect that there are some modifications to radeon from kernel .41 to .44, that are responsible for that problem. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c18 --- Comment #18 from Mathias Grewin <grewin.mathias@online.de> --- (In reply to Felix Miata (offline until ???) from comment #16)
kernel-default-5.3.18-lp152.123.1.g27f6195.x86_64.rpm 03-Oct-2020 09:57 (standard) and kernel-default-5.3.18-lp152.1.1.g3d4aad4.x86_64.rpm 07-Oct-2020 10:00 (tiwai) both fail.
Both kernels fail, I confirm message of comment 16. (By the way ... I lost my kernel 5.3.18-lp152.41-default trying this - it was overwritten and so I couldn't get into my system anymore. But a snapshot and a rollback by using btrfs-Filesystem was no problem. All ok!) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c19 --- Comment #19 from Takashi Iwai <tiwai@suse.com> --- Ah, good to know that the last working version is *.41. Then there is a known patch that can be the culprit, the commit fc8c70526bd30733ea8667adb8b8ffebea30a8ed I'm building yet another kernel in OBS home:tiwai:bsc1177384-radeon repo. A new test package will appear later at http://download.opensuse.org/repositories/home:/tiwai:/bsc1177384-radeon/sta... Please give it a try later. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 Takashi Iwai <tiwai@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c20 --- Comment #20 from Peter Kirchgeßner <peter@kirchgessner.net> --- kernel-default-5.3.18-lp152.1.1.g9749897.x86_64.rpm works. I get 1920x1200 on DVI-port. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c21 --- Comment #21 from Takashi Iwai <tiwai@suse.com> --- OK, thanks for confirmation. The fix was queued now. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c22 --- Comment #22 from Mathias Grewin <grewin.mathias@online.de> --- (In reply to Takashi Iwai from comment #21)
OK, thanks for confirmation. The fix was queued now.
Yes it's fine - thank you so much and to all users which are involved in this problem and also helped with their comments! -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1177384 http://bugzilla.opensuse.org/show_bug.cgi?id=1177384#c23 Mathias Grewin <grewin.mathias@online.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #23 from Mathias Grewin <grewin.mathias@online.de> --- (In reply to Mathias Grewin from comment #22)
(In reply to Takashi Iwai from comment #21)
OK, thanks for confirmation. The fix was queued now.
Yes it's fine - thank you so much and to all users which are involved in this problem and also helped with their comments!
Problem resolved! -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@suse.com