Bug 16120 - AB-INT: do_bootimg: cannot lseek hddimg/rootfs.img: Invalid argument
Summary: AB-INT: do_bootimg: cannot lseek hddimg/rootfs.img: Invalid argument
Status: RESOLVED FIXED
Alias: None
Product: Functional (self) Testing
Classification: QA/Testing
Component: oe-selftest (show other bugs)
Version: unspecified
Hardware: x86 Multiple
: Medium normal
Target Milestone: 6.0
Assignee: Paul Barker
QA Contact:
URL:
Whiteboard: AB-INT
Depends on:
Blocks:
 
Reported: 2026-01-05 08:56 UTC by Mathieu Dubois-Briand
Modified: 2026-02-26 16:11 UTC (History)
2 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: No (bug/feature does not impact docs)


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Mathieu Dubois-Briand 2026-01-05 08:56:50 UTC
ERROR: core-image-minimal-1.0-r0 do_bootimg: Execution of '/srv/pokybuild/yocto-worker/oe-selftest-debian/build/build-st-1120930/tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0/temp/run.build_hddimg.4042060' failed with exit code 1
...
| NOTE: Trying to install /srv/pokybuild/yocto-worker/oe-selftest-debian/build/build-st-1120930/tmp/deploy/images/qemux86-64/bzImage as /srv/pokybuild/yocto-worker/oe-selftest-debian/build/build-st-1120930/tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0/sources/core-image-minimal-1.0/hddimg/bzImage
| install: cannot lseek '/srv/pokybuild/yocto-worker/oe-selftest-debian/build/build-st-1120930/tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0/sources/core-image-minimal-1.0/hddimg/rootfs.img': Invalid argument
| WARNING: exit code 1 from a shell command.
Comment 1 Mathieu Dubois-Briand 2026-01-05 09:48:20 UTC
oe-selftest-debian ubuntu2510-vk-1 mathieu/master-next completed at 2026-01-01 22:35:18+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/2937/steps/15/logs/stdio
Comment 2 Randy MacLeod 2026-01-08 16:02:15 UTC
Let's see if it happens again, on master even...
Comment 3 Antonin Godard 2026-01-15 13:21:57 UTC
oe-selftest-fedora alma8-vk-2 master completed at 2026-01-15 03:37:30+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/2912/steps/16/logs/stdio

oe-selftest-fedora alma8-vk-2 master&master-next completed at 2026-01-14 20:54:08+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/2909/steps/16/logs/stdio
Comment 4 Antonin Godard 2026-01-20 14:42:12 UTC
oe-selftest-fedora opensuse155-vk-2 agodard/master-next completed at 2026-01-15 21:19:02+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/2918/steps/16/logs/stdio

oe-selftest-debian ubuntu2204-vk-2 agodard/master-next-bis&agodard/master-next completed at 2026-01-16 21:12:33+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/3036/steps/15/logs/stdio
Comment 5 Paul Barker 2026-01-20 18:24:07 UTC
I've looked into this in some detail but can't reproduce the failure locally.

There is a known issue in coreutils 9.9 (https://github.com/coreutils/coreutils/issues/159), with two upstream commits we should take to fix this:

- https://github.com/coreutils/coreutils/commit/bd528f923482223649aa84be7d131e69356149da

- https://github.com/coreutils/coreutils/commit/33bc44e1ba3aa4c70f3cd16aa9c41331543986dd

The issue reported upstream is an infinite loop, we're seeing an invalid argument passed to lseek. But I suspect these issues are at least related, if not the same.

There are a couple of places that the "cannot lseek" error message gets printed in src/copy-file-data.c from coreutils 9.9. We specifically see a failure to seek on the destination file, which means the failure is in `create_hole()` (https://github.com/coreutils/coreutils/blob/v9.9/src/copy-file-data.c#L65), and I can figure out that it's being given bad arguments by `lseek_copy()` (https://github.com/coreutils/coreutils/blob/v9.9/src/copy-file-data.c#L317).

Trying to fully understand the loop in `lseek_copy()` and prove that this is our bug is a lot of work. We need to prioritise, so let's take the upstream fixes for now and see if the intermittent build failure re-occurs. If it does occur again, we've narrowed down the potential causes by applying these fixes.

I'll send patches...
Comment 6 Paul Barker 2026-01-21 10:46:35 UTC
Patch sent to the list: https://lore.kernel.org/openembedded-core/20260121-fix-coreutils-v1-1-30a1769b5733@pbarker.dev/T/#u

I've ran the test that failed (runqemu.RunqemuTests.test_boot_compressed_wic_by_path) 20 times on the autobuilder - 10x oe-selftest-debian + 10x oe-selftest-fedora. None of those runs hit the issue so hopefully it's fixed. Once this is merged to master we can mark as resolved and re-open if we see the issue again.
Comment 7 Antonin Godard 2026-01-21 15:34:32 UTC
oe-selftest-fedora alma8-vk-1 agodard/master-next completed at 2026-01-20 15:24:00+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/2951/steps/16/logs/stdio

oe-selftest-debian ubuntu2204-vk-4 agodard/master-next completed at 2026-01-21 00:37:17+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/3061/steps/15/logs/stdio

oe-selftest-fedora opensuse155-vk-2 master-next&master completed at 2026-01-21 03:43:57+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/2969/steps/16/logs/stdio
Comment 8 Paul Barker 2026-02-26 16:11:37 UTC
Merged to master and things have been fine since so let's close this.

https://git.openembedded.org/openembedded-core/commit/?id=20f2ea3756aa5f0460cb114d0473bbc656e8b97d