| Summary: | floating point exception in udevadm breaks booting | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Alan Secker <alan> |
| Component: | RPM Packages | Assignee: | Colin Guthrie <mageia> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | critical | ||
| Priority: | release_blocker | CC: | cooker, eeeemail, filip.komar, fri, guillomovitch, jlnorman, mageia, micheelsen, mitya, thierry.vignaud, tmb, zen25000 |
| Version: | Cauldron | ||
| Target Milestone: | Mageia 3 | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | systemd | CVE: | |
| Status comment: | |||
|
Description
Alan Secker
2013-04-10 16:59:01 CEST
After I executed two updates this morning on my test machine. I rebooted using the new default kernel 3.6.8.2_mga3. It failed during the boot process. There was a "dracut warning: Could not boot" followed by: "/dev/disk/by-uuid/9aaa8201-dcc5-4684-a936-19d24757f1f6" does not exist. There was no fstab in the /etc and that device reference certainly didn't  show up in "my' fstab when I rebooted using an earlier kernel. Priority:
Normal =>
High maybe a duplicate of the bug 'installer adds usk key in the fstab' ? I seriously don't know CC:
(none) =>
mageia, thierry.vignaud, tmb No it isn't. But I confirm the bug. I think it's due to new udev Assignee:
bugsquad =>
mageia
Thierry Vignaud
2013-04-10 21:29:23 CEST
Summary:
MGA 3 beta 4 =>
floating point exception in udevadm breaks booting Confirming. From within the initrd shell I've discovered there were no entries under /dev/disk/*, hence mount failure. However, blkid works and outputs correct UUIDs. From the shell, I was able to mount my disks using /dev/sd* entries. Supplying "root=/dev/sdX resume=/dev/sdX" doesn't work, this seems to be ignored by initrd. CC:
(none) =>
mitya
Johnny A. Solbu
2013-04-10 23:18:22 CEST
CC:
(none) =>
cooker Raised to release_blocker since Anne had raised 9691, which has been closed as a dup of this bug. https://ml.mageia.org/l/arc/bugs/2013-04/msg01012.html Priority:
High =>
release_blocker Hmm, I'm actually thinking this is a kernel problem. I can reproduce the problem here with the latest kernel, but regenerating an initrd for an older kernel and it works fine. Hmm, scratch that. I'm sure it worked but I can no longer reproduce that even on older kernels. Will keep digging. Should be fixed when systemd is pushed. Please double check. Remember you need to regenerate your initrds to test it! Indeed it's fixed once rebuilding the initrds Status:
NEW =>
RESOLVED The kernel now boots, but "Floating point exception" is still printed once, and Plymouth graphical splash is interrupted for some seconds. Though purely cosmetic, the effect is quite unpleasant. Also valid for the new 3.8.7 (both -desktop and -server). Are you sure you have regenerated your initrds properly? I don't get any printouts of the "Floating point exception" here. Yep, my current 3.8.7 kernel has been installed after systemd/udev update (systemd-195-20). The "Floating point exception" is printed out only once, and it disrupts the splash for a moment. After that, the splash is repainted and the boot continues OK. I'm using an i586 distro on Inter Core 2 Duo E4700 CPU. Since updating this morning and booting into 3.8.7-desktop-1.mga3, the hanging issue has gone. Thanks to all concerned.
Hans Micheelsen
2013-04-17 20:46:41 CEST
CC:
(none) =>
micheelsen The "Floating point exception" is still printed, exactly once. Kernel 3.8.8, both -desktop and -server. On my system with fresh install of DVD mga3 RC from USB and after updates "Floating point exception" is printed too. I didn't notice any other strange behaviour. CC:
(none) =>
filip.komar It's still displayed indeed but the booting issue is fixed I'll attempt to reproduce it on my laptop at home so I can fix it. Reopening as per doktor5000's suggestion following forum discussion. The 'floating point exception' still occurs, but seems to be cpu specific. I clean installed 3 rc on two machines. One is a 990fx system with 955BE cpu. The other is a x79 system with i7 3820 cpu. The first system boots up, no error. The second system gives a 'Floating point Exception' message interrupting the plymouth boot splash upon every boot. Installing all updates, which included an updated 3.8.8 kernel, didn't fix; tried regenerating initrds (in spite of it being a new install)-also no effect. Tried the 3.8.7 kernel-linus kernel-same error. Status:
RESOLVED =>
REOPENED Oops, forgot forum thread: https://forums.mageia.org/en/viewtopic.php?f=15&t=4796 Please do not repopen this bug report. The boot issue is _fixed_. Status:
REOPENED =>
RESOLVED Don't worry, this is still on my radar and I'll be taking a further look and asking those affected to retest. (In reply to Colin Guthrie from comment #24) > Don't worry, this is still on my radar and I'll be taking a further look and > asking those affected to retest. Better to continue this in Bug 9850 CC:
(none) =>
zen25000 |