Bug 21912 - System starts but no IO device works, also no dmesg and no journal logs get written (kernel-4.9.56-1.mga6)
Summary: System starts but no IO device works, also no dmesg and no journal logs get w...
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 6
Hardware: All Linux
Priority: Normal critical
Target Milestone: ---
Assignee: Kernel and Drivers maintainers
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-10-22 09:28 CEST by Bernard MAUDRY
Modified: 2017-10-23 09:21 CEST (History)
2 users (show)

See Also:
Source RPM: kernel-4.9.56-1.mga6.src.rpm
CVE:
Status comment:


Attachments

Description Bernard MAUDRY 2017-10-22 09:28:53 CEST
Description of problem:
After updating to kernel 4.9.56, the system seemed to startnormally, even autologin worked. But when trying to do something, I saw that I could no do anything: keyboard, mouse, network were all dead;
Rebooting with kernel 4.9.80 and the issue disapears.
Looking in the journal: nothing about the previous boot
Looking at dmesg.old: was about the boot the day before

AMD A10-7870K Radeon R7, 12 Compute Cores 4C+8G
Motherboard Asus A88XM+

Version-Release number of selected component (if applicable):
4.9.56

How reproducible:
Every boot

Steps to Reproduce:
1. boot using kernel 4.9.56
2. no input device is usable
Comment 1 Bernard MAUDRY 2017-10-22 09:32:26 CEST
Sorry for the typing mistakes; I cannot find how to edit them.
I rebooted using the kernel 4.9.50
Comment 2 Marja Van Waes 2017-10-22 09:57:16 CEST
(In reply to Bernard MAUDRY from comment #1)
> Sorry for the typing mistakes; I cannot find how to edit them.
> I rebooted using the kernel 4.9.50

Unfortunately, editing a comment is impossible. You did well by adding a new comment with the correction :-)

Assignee: bugsquad => kernel
CC: (none) => marja11
Summary: System starts but no IO device works => System starts but no IO device works, also no dmesg and no journal logs get written (kernel-4.9.56-1.mga6)

Comment 3 Frank Griffin 2017-10-22 14:48:45 CEST
This is a wild guess, but if the kernel detects disk I/O errors on the root partition during boot, it marks it read-only, which might account for some of what you see (no logs, etc.).

Can you log in from a tty and try to "touch /test.file" and see what you get ?

CC: (none) => ftg

Comment 4 Bernard MAUDRY 2017-10-22 22:48:53 CEST
Unfortunately, without keyboard input, I cannot do what you ask.

By the way, is this disk I/O error detection behaviour specific to kernel 4.9.56?
If not, I should not be able to boot kernel 4.9.50 normally.
Comment 5 Frank Griffin 2017-10-22 22:58:24 CEST
You mentioned that autologin worked, so I assumed you were getting a GUI desktop.  Also, you stated that journalctl and dmesg gave no output, so I assumed you had *some* keyboard access that worked.  Are you sure that you can't get a tty session or boot in failsafe mode ?
Comment 6 Frank Griffin 2017-10-22 23:03:37 CEST
(In reply to Bernard MAUDRY from comment #4)
> By the way, is this disk I/O error detection behaviour specific to kernel
> 4.9.56?
> If not, I should not be able to boot kernel 4.9.50 normally.

I suppose the particular sectors accessed on a disk can vary from kernel to kernel.  If a kernel doesn't get an error reading or writing the areas of the disk it tries to read or write, then it won't trigger the read-only behavior.
Comment 7 Bernard MAUDRY 2017-10-23 08:20:47 CEST
> You mentioned that autologin worked, so I assumed you were getting a GUI
> desktop.
I do, but I had no input, either with the mouse or the keyboard

> Also, you stated that journalctl and dmesg gave no output, so I assumed
> you had *some* keyboard access that worked.
Yes, but only after rebooting with kernel 4.9.50

> Are you sure that you can't get a tty session or boot in failsafe mode ?
I am sure that I was unable to switch to a virtual tty as I tried this
I didn't try the failsafe boot. I will do it now.
Comment 8 Bernard MAUDRY 2017-10-23 09:00:31 CEST
I had to re-install kernel 4.9.56 as I removed it to avoid any mistake.

I booted it in safe mode and got access to the maintenance mode.
I quit the console to let the boot process continue.
Everything worked correctly.

I then rebooted kernel 4.9.56 in normal mode, and I am typing this with it.

It seems that there were some corruption during the first installation of this kernel that the second installation fixed.

I will do a check on the boot disk to see if fsck detects some errors.
Comment 9 Bernard MAUDRY 2017-10-23 09:21:05 CEST
fsck -f didn't detect any error on the boot partition

I wonder what happened.

Sorry for the false alert.

Status: UNCONFIRMED => RESOLVED
Resolution: (none) => FIXED


Note You need to log in before you can comment on or make changes to this bug.