1

Until recently I was using Fedora 26 with kernel 4.13.16-202.fc26.x86_64 without problems. But the last two kernel upgrades (4.14.4-200.fc26.x86_64 and 4.14.5-200.fc26.x86_64) are giving me kernel panic. This is the only thing I was able to get before the screen froze:

 [0.061000] ? acpi_sleep_proc_init+0x2a/0x2a
 [0.061000] ? set_debug_rodata+0x1?/0x1?
 [0.061000] acpi_bus_scan+0x?f/0x90
 [0.061000] acpi_scan_init+0xea/0XZZC
 [0.061000] ? acpi_sleep_proc_init+0x2a/0x2a
 [0.061000] acpi_init+0x2f8/0x356
 [0.061000] d0_one_initcall+0x50/0x190
 [0.061000] kernel_init_freeable+0x1a8/0x245
 [0.061000] ? rest_init+0xb0/0xb0
 [0.061000] kcrne1_init+0xe/0x101
 [0.061000] ret_from_fork+0x25/0x30
 [0.061000] Code: c? 01 00 00 00 00 ?4 le 7? Oh 41 83 £0 08 75 58 0f b6 03 ch
   18 41 83 fc 20 ?4 10 41 83 fc 40 75 4? 48 8b 03 ch 0? 0f b7 03 ch 02 <8b> 03 48
   89 01 ch 36 41 83 fc 10 ?4 23 7? 0d 41 83 fc 08 75 28
 [0.061000] RIP: acpi_ex_sgstem_memorg_space_handler+0x16f/0x1c1 RSP: ffffb?0 140006640
 [0.061000] CR2: ffffb701406a6000
 [0.061000] ---[ end trace be3326226d4700d? ]â~@~Tâ~@~Tâ~@~T
 [0.061010] Kernel panic â~@~T not syncing: attempted to kill init! exitcode=0x00 000009
 [0.061010]
 [0.061169] ---[ end Kernel panic - not syncing: Attempted to kill init! exit code=0x00000009
 [0.061169]

Suggestions?

Luis A. Florit
  • 111
  • 3
  • 1
    You should prob try on https://superuser.com - This looks off topic for here since Fedora is a user os. – user3788685 Dec 19 '17 at 18:59
  • 2
    @user3788685 Fedora is not a "user OS". It's a fully functional metadistribution used in numerous production environments. The use of something other than your preferred distribution does not make a question off-topic. – Spooler Dec 19 '17 at 19:10
  • Regarding this panic, anyone involved would need to know a lot more about your workload and the conditions it's in. Have you communicated with any Fedora specific resources? The community is pretty active. Here's a link that should point you towards the right people to ask: https://fedoraproject.org/wiki/Communicating_and_getting_help – Spooler Dec 19 '17 at 19:14
  • @SmallLoanOf1M: Although the community is active, my experience with (system) bug reports in Fedora is horrendous. The vast majority of reports are simply ignored, virtually all of them expire when the release enters EOL without answers. I began using RedHat 2, and it was fine at the time, but as the system evolved this became worse and worse. I completely gave up reporting any system bugs, it's a waste of time. – Luis A. Florit Dec 20 '17 at 12:44
  • I just asked in askfedora, but there are very few answers there for each topic. – Luis A. Florit Dec 20 '17 at 12:57

1 Answers1

1

The kernel crashed because the init process is terminating which isn't supposed to happen and would leave kernel data structures in a corrupt state.

That means the problem isn't necessarily caused by the kernel itself. Try booting with an older kernel version that has been working in the past.

What to do next depends on whether the problem turns out to be caused by the kernel version or something else.

If it isn't a kernel problem

Boot into a shell by passing init=/bin/bash or init=/bin/sh as argument for the kernel. Then start looking for clues in logfiles.

If it is a kernel problem

Boot the latest working kernel version. Download the sources of the latest working kernel version and the first broken kernel version. Verify that you get the same results when using a kernel that you build directly from those sources. Next start looking at the differences between the two versions to identify which of the differences actually apply to your system. (Changes in drivers for hardware that you don't have would be an obvious example of something unlikely to be relevant).

If you feel up to the task you can try creating a separate source tree with some files from the working version and some from the broken version to identify which file was causing the breakage.

kasperd
  • 29,894
  • 16
  • 72
  • 122
  • I don't have the computer with me (only in January), but what I can tell you is that I can still boot with the old kernel `4.13.16` without any problems, so it seems to be a kernel problem. – Luis A. Florit Dec 23 '17 at 00:50
  • @LuisA.Florit Does the problem happen during boot or later? – kasperd Dec 23 '17 at 00:53
  • As soon as it passes the bootloader, it panics and everything halts. That's why the only thing I can see is what I wrote in the original post. It is not even in the logs. – Luis A. Florit Dec 23 '17 at 01:38
  • @LuisA.Florit Logging to the serial port may help capture more of the output. – kasperd Dec 23 '17 at 08:58