Bug 1811353
Summary: | F-31 regression: memtest86+ reboots after few secs | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Jan Kratochvil <jan.kratochvil> |
Component: | memtest86+ | Assignee: | Jaroslav Škarvada <jskarvad> |
Status: | CLOSED EOL | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | urgent | Docs Contact: | |
Priority: | unspecified | ||
Version: | 31 | CC: | edgar.hoch, jan.kratochvil, jhlavac, jskarvad, kparal, oli, Pascal, promac, robatino, wgianopoulos, yulinux |
Target Milestone: | --- | Keywords: | CommonBugs, Regression |
Target Release: | --- | ||
Hardware: | x86_64 | ||
OS: | Linux | ||
Whiteboard: | https://fedoraproject.org/wiki/Common_F32_bugs#memtest-reboots-freezes | ||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2020-11-24 16:52:01 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | 1598922 | ||
Bug Blocks: | |||
Attachments: |
Description
Jan Kratochvil
2020-03-07 22:27:08 UTC
Reproducibility is easy even in KVM (on a different host so that will be host unrelated I think): FAIL: qemu-kvm -cdrom Fedora-MATE_Compiz-Live-x86_64-31-1.9.iso -m 5G -net none PASS: qemu-kvm -cdrom Fedora-MATE_Compiz-Live-x86_64-31-1.9.iso -m 4G -net none I can confirm this attempting to run memtest86+ from the grub menu on a fully updated F31 box. The same box was affected by the reboot bug before it was fixed. This also affects F32 Beta, and is discussed here: https://lists.fedoraproject.org/archives/list/test@lists.fedoraproject.org/thread/6TXB3XGGHFSCYVHU54HJWMDZ2NN3UAAV/ Jarda, can you please look into fixing this again somehow? Thanks a lot. Please try memtest86+-5.31-0.1.beta.fc33 Created attachment 1688454 [details]
crash screenshot
I have tried it only in qemu-kvm so far but it does crash for me. Does it work for you? Former memtest86+-5.01-27.fc31.x86_64 works for me fine in qemu-kvm this way:
sync;hdparm -f /dev/nvme0n1*;qemu-kvm -hda /dev/nvme0n1 -snapshot -m 2G -net none
I downloaded Fedora-Everything-netinst-x86_64-Rawhide-20200514.n.0.iso which should contain memtest86+-5.31-0.1.beta.fc33 and tested on a desktop Haswell PC and Thinkpads T450s and T480s. On all these PCs I tested all startup options - default, failsafe and SMP. In all these cases, the new memtest build worked fine (I only waited a few minutes, not for the whole test to complete). While previously it froze or rebooted almost instantly (see the link in comment 3). So this seems to fix all issues I saw on bare metal. VMs are still affected unless you force failsafe mode. In other modes, they reboot instantly. I don't think that is a huge problem, though. (In reply to Kamil Páral from comment #6) > VMs are still affected unless you force failsafe mode. In other modes, they > reboot instantly. I don't think that is a huge problem, though. That's a bug somewhere, either memtest86+ or qemu. There is gdbserver stub inside qemu for debugging. Also the bug (Unexpected interrupt) looks as a different one than the previous one (sudden reboot) so maybe it is even bisectable. Fedora-Everything-netinst-x86_64-Rawhide-20200515.n.0.iso works fine for me in qemu-kvm. Still the GRUB-loaded memtest86+-5.01-27.fc31.x86_64 from my HDD according to Comment 5 does crash as Comment 5 shows. memtest86+-5.01-27.fc31.x86_64 loads fine with GRUB (although sure it reboots soon with mem >=5GB in qemu-kvm). Memtest86+ v5.01 included with Fedora 32 installation media hangs in less than a minute. Exact time depends upon amount of installed RAM and number of CPUs enabled, but appears consistent. Same version from https://www.memtest.org and https://www.ultimatebootcd.com works perfectly. v5.31b from https://www.memtest.org also works. I have this issue also. It fails on all of my laptops. ASUS with an AMD-A4 processor. HP and Samsung with AMD-A6 processors Lenovo with intel (core) i3 processor. The failure always occurs somewhere between 40 and 55 % of the very first test on the first pass. Sometimes it spontaneously reboots other times it just hangs/ This is also with SMP disabled. If I upgrade to memtest86+-5.31-0.1.beta.fc33.x86_64.rpm from the rawhide repository, that resolves the issue. I should have mentioned the fedora 32 version of memtest does not resolve the issue. Created attachment 1689600 [details]
This is the rpm file I installed that corrects the issue
Created attachment 1689837 [details]
crash photo - booted by GRUB from HDD on Lenovo X1 6th
Created attachment 1689838 [details]
BIOS/firmware versions photo for the crash photo
BIOS Version: N23ET63W (1.38)
Although sure it could depend also on some BIOS settings.
From Fedora-Everything-netinst-x86_64-Rawhide-20200517.n.1.iso flashdisk on bare iron it really works OK. (Just one needs to boot it in legacy mode, in UEFI mode there is no memtest86+.) Well, that is not a regression. There has NEVER been a memtest for UEFI mode. I still say the version on rawhide works correctly via grub on a non-UEFI installation. The regression is that memtest86+-5.01-27.fc31.x86_64 starts running OK for me from my NVMe GRUB menu while memtest86+-5.31-0.1.beta.fc33.x86_64 immediately crashes with "Unexpected interrupt" from my NVMe GRUB menu. Both work fine from a flashdisk on the same computer (which also boots the memtest by GRUB). I do not understand what is the difference between a boot from my NVMe vs. from flashdisk. Nobody else can reproduce my case? All the cases are non-UEFI (=legacy boot). (In reply to Jan Kratochvil from comment #17) > The regression is that memtest86+-5.01-27.fc31.x86_64 starts running OK for > me from my NVMe GRUB menu while memtest86+-5.31-0.1.beta.fc33.x86_64 > immediately crashes with "Unexpected interrupt" from my NVMe GRUB menu. > Both work fine from a flashdisk on the same computer (which also boots the > memtest by GRUB). > I do not understand what is the difference between a boot from my NVMe vs. > from flashdisk. Nobody else can reproduce my case? > All the cases are non-UEFI (=legacy boot). It's still beta and upstream is gathering feedback on it. Could you please send the report upstream to memtest with the description of the HW where it is failing. Thanks. (In reply to Jan Kratochvil from comment #17) Could you post the menuentry for memtest from you grub2.cfg file? Created attachment 1690658 [details]
FIle for /etc/grub.d to automagially add memetest to the gurb boot menu
Because of issues I have seen with people trying to do this by hand I am once again asking to have this added to /etc/grub.d so the /usr/sbin/grub2-mkconfig creates a grub2 config file that includes memtest in the boot menu.
The problem gets fixed by using INSTALL_ELF=0. The default INSTALL_ELF=1 has the regression as I have described. works (Unexpected interrupt): linux16 /memtest86+-5.31 does not work: knetbsd /elf-memtest86+-5.31 Regarding the Memtest86+ v5.01 included with Fedora 32 installation media: Always hangs when it starts testing RAM above 4 GB if I have 6, 8, or 16 GB RAM installed Runs perfectly if I only have 2 or 4 GB RAM installed (same computer, flash drive, etc.) Appears to be having a problem with the 32bit limit? Hardware was identical for all tests and all results were repeatable: Hangs when accessing over 4GB* RAM: Memtest86+ v5.01 in Fedora-Workstation-Live-x86_64-32-1.6.iso Reboots when accessing over 4GB* RAM: Memtest86+ v5.01 in CentOS-8.1.1911-x86_64-boot.iso Works perfectly with 16GB RAM: Memtest86+ v5.31b in Fedora-Workstation-Live-x86_64-Rawhide-20200601.n.1.iso Works perfectly with 16GB RAM: Memtest86+ v5.01 & v5.31b from https://www.memtest.org *Based upon where the program said it was currently testing. As above: Reboots when accessing over 4GB* RAM: Memtest86+ v5.01 in CentOS-8.2.2004-x86_64-boot.iso Fails for me also on CentOS8.2 - with Intel(R) Core(TM) i5-2500 CPU on DQ67SW motherboard. This message is a reminder that Fedora 31 is nearing its end of life. Fedora will stop maintaining and issuing updates for Fedora 31 on 2020-11-24. It is Fedora's policy to close all bug reports from releases that are no longer maintained. At that time this bug will be closed as EOL if it remains open with a Fedora 'version' of '31'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, simply change the 'version' to a later Fedora version. Thank you for reporting this issue and we are sorry that we were not able to fix it before Fedora 31 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Fedora, you are encouraged change the 'version' to a later Fedora version prior this bug is closed as described in the policy above. Although we aim to fix as many bugs as possible during every release's lifetime, sometimes those efforts are overtaken by events. Often a more recent Fedora release includes newer upstream software that fixes bugs or makes them obsolete. The latest problem is not that memtest86+ reboots, but that it fails immediately. See https://bugzilla.redhat.com/show_bug.cgi?id=1869211 . Fedora 31 changed to end-of-life (EOL) status on 2020-11-24. Fedora 31 is no longer maintained, which means that it will not receive any further security or bug fix updates. As a result we are closing this bug. If you can reproduce this bug against a currently maintained version of Fedora please feel free to reopen this bug against that version. If you are unable to reopen this bug, please file a new report against the current release. If you experience problems, please add a comment to this bug. Thank you for reporting this bug and we are sorry it could not be fixed. |