Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #85027 > unrolled thread

Bug#1092103: Linux 6.12 freeze on small machine

Back to article view | Back to linux.debian.kernel

Contents

  Bug#1092103: Linux 6.12 freeze on small machine GMX <ebhb010@gmx.com> - 2025-01-04 16:10 +0100
    Bug#1092103: switch to important GMX <ebhb010@gmx.com> - 2025-01-04 21:20 +0100
    Bug#1092103: switch to important GMX <ebhb010@gmx.com> - 2025-01-04 21:20 +0100
    Bug#1092103: switch to important GMX <ebhb010@gmx.com> - 2025-01-04 21:40 +0100
    Processed: switch to important "Debian Bug Tracking System" <owner@bugs.debian.org> - 2025-01-04 21:40 +0100
    Processed: Re: Bug#1092103: Linux 6.12 freeze on small machine "Debian Bug Tracking System" <owner@bugs.debian.org> - 2025-05-03 07:50 +0200
    Bug#1092103: Linux 6.12 freeze on small machine Salvatore Bonaccorso <carnil@debian.org> - 2025-05-03 07:50 +0200
    Processed: Re: Bug#1092103: issue seems gone. "Debian Bug Tracking System" <owner@bugs.debian.org> - 2025-05-03 13:50 +0200
    Bug#1092103: issue seems gone. Salvatore Bonaccorso <carnil@debian.org> - 2025-05-03 13:50 +0200

#85027 — Bug#1092103: Linux 6.12 freeze on small machine

[Multipart message — attachments visible in raw view] — view raw

Package: linux-image-6.12.6-amd64
Version: 6.12.6-1

Found a case where cpu freeze(100% usage) will occur.

(Tested)Environment:
CPU: 1 or 2 cores.
Memory: <= 1GB

Only tested in Virtual Machine(KVM).

Procedure:
1. Start iperf3 as server on the test machine.
2. On any client machine, start a reversed(download) test to the test machine.
To trigger the case, use 4 threads and 256k block size. Higer value slightly increase the chance.
3.Wait for iperf3 to run.
(4). After a few iteration or at test end, the server will stop responding and completly locked.(no kernel panic log, console just freeze)

 From monitoring platform , the process contained the already freezed vm uses 100% cpu(50% on 2 cores). Charts show the vm linearly increase memory usage to watermark point and never freed them(Because cpu is freezed ?).

Tested on vanilla debian 6.12 and xanmod(no debian patches) 6.12, both freezed.
Tested on xanmod 6.6, no such problem.
Tested on vanilla debian 6.1, no such problem.

[toc] | [next] | [standalone]

#85030 — Bug#1092103: switch to important

[Multipart message — attachments visible in raw view] — view raw

Control: severity -1|important |

[toc] | [prev] | [next] | [standalone]

#85031 — Bug#1092103: switch to important

[Multipart message — attachments visible in raw view] — view raw

Control: severity -1|important I don't know if it's the correct level but this bug cause all
small machines to lockup eventually. |

[toc] | [prev] | [next] | [standalone]

#85032 — Bug#1092103: switch to important

Control: severity -1 important

[toc] | [prev] | [next] | [standalone]

#85033 — Processed: switch to important

Processing control commands:

> severity -1 important
Bug #1092103 [linux-image-6.12.6-amd64] Linux 6.12 freeze on small machine
Severity set to 'important' from 'normal'

-- 
1092103: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1092103
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#87321 — Processed: Re: Bug#1092103: Linux 6.12 freeze on small machine

Processing control commands:

> tags -1 + moreinfo
Bug #1092103 [src:linux] Linux 6.12 freeze on small machine
Added tag(s) moreinfo.

-- 
1092103: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1092103
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#87322

Control: tags -1 + moreinfo

Hi,

On Sat, Jan 04, 2025 at 11:08:41PM +0800, GMX wrote:
> Package: linux-image-6.12.6-amd64
> Version: 6.12.6-1
> 
> Found a case where cpu freeze(100% usage) will occur.
> 
> (Tested)Environment:
> CPU: 1 or 2 cores.
> Memory: <= 1GB
> 
> Only tested in Virtual Machine(KVM).
> 
> Procedure:
> 1. Start iperf3 as server on the test machine.
> 2. On any client machine, start a reversed(download) test to the test machine.
> To trigger the case, use 4 threads and 256k block size. Higer value slightly increase the chance.
> 3.Wait for iperf3 to run.
> (4). After a few iteration or at test end, the server will stop responding and completly locked.(no kernel panic log, console just freeze)
> 
> From monitoring platform , the process contained the already freezed vm uses 100% cpu(50% on 2 cores). Charts show the vm linearly increase memory usage to watermark point and never freed them(Because cpu is freezed ?).
> 
> Tested on vanilla debian 6.12 and xanmod(no debian patches) 6.12, both freezed.
> Tested on xanmod 6.6, no such problem.
> Tested on vanilla debian 6.1, no such problem.

Was unable here to reproduce the problem with a small VM.

Can you please describe in more details both the sizing of the choosen
VM ideally the full qemu commandline which is invoked as well when
starting and the vm description.

Additionally please the exact test setup which triggers the problem
for you with the iperf3 invocations.

Thanks already,

Regards,
Salvatore

[toc] | [prev] | [next] | [standalone]

#87329 — Processed: Re: Bug#1092103: issue seems gone.

Processing control commands:

> tags -1 - moreinfo
Bug #1092103 [src:linux] Linux 6.12 freeze on small machine
Removed tag(s) moreinfo.
> fixed -1 6.12.25-1
Bug #1092103 [src:linux] Linux 6.12 freeze on small machine
Marked as fixed in versions linux/6.12.25-1.

-- 
1092103: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1092103
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#87330 — Bug#1092103: issue seems gone.

Control: tags -1 - moreinfo
Control: fixed -1 6.12.25-1

Hi,

On Sat, May 03, 2025 at 01:57:48PM +0800, GMX wrote:
> Hi. I'm currently using 6.12.25 from xanmod.
> 
> The issue seems no longer there. Either tested locally or on cloud vps, no
> such issue now. So I think we can assume that this is fixed now.

Thanks for reporting back, accordingly I will close the bug.

One note about the problem: I did as well test kernel back to the
6.12.6-1 as mentioned initially in the report, and was not able to
trigger there either. So right now I'm not conviced a change between
6.12.6 and 6.12.25 is "fixing" the problem, but I will still close it
now with 6.12.25-1.

Regards,
Salvatore

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web