| Summary: | needs to mount sysfs before mount -a | ||
|---|---|---|---|
| Product: | buildroot | Reporter: | stsp <stsp2> |
| Component: | Other | Assignee: | unassigned |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | buildroot |
| Priority: | P5 | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Host: | Target: | ||
| Build: | |||
|
Description
stsp
2021-11-16 10:17:34 UTC
This was fixed in util-linux: https://github.com/util-linux/util-linux/issues/1492 Please consider updating, or adding an explicit /sys mount before "mount -a". Can you confirm that the following patch fix your issue: https://patchwork.ozlabs.org/project/buildroot/patch/20211121094254.3425966-1-fontaine.fabrice@gmail.com/ (In reply to Fabrice Fontaine from comment #2) Applying 0002-libblkid-don-t-mark-cache-as-probed-if-sys-not-available.patch using patch: patching file libblkid/src/devname.c Hunk #1 succeeded at 456 (offset 27 lines). Hunk #2 succeeded at 562 with fuzz 1 (offset 27 lines). Hunk #3 succeeded at 583 (offset 29 lines). Hunk #4 succeeded at 608 (offset 29 lines). Hunk #5 succeeded at 626 (offset 29 lines). ... is what I see when rebuilding util-linux. I would rather suggest to update entire util-linux from git, or, better yet, just mount /sys by hands, the same way you already do for /proc. But other than that - yes, the patched buildroot works as expected, thanks! Thanks for confirming that the patch works. We can't bump util-linux as we're too close to the release of 2021.11. Indeed, a bump to the latest git commit could add its own build or runtime issues. Mounting /sys does not seem to be an option for 2021.11 either, see initial comments here: https://patchwork.ozlabs.org/project/buildroot/patch/20211120222857.3110528-1-fontaine.fabrice@gmail.com/ (In reply to stsp from comment #1) Do you know if this is a regression in 2.37.x? We use 2.36.2 on the 2021.02.x branch, and the patch doesn't directly apply there. (In reply to Peter Korsgaard from comment #6) I am quite sure it is a regression. I hit it only after the buildroot update I did in my project: diff --git a/buildroot b/buildroot index b928074..76c7fed 160000 --- a/buildroot +++ b/buildroot @@ -1 +1 @@ -Subproject commit b928074e2d83f5b873bf1ba381b212dfad9bf207 +Subproject commit 76c7fed85f7a822663cf0c8828e7d282198d0803 Unfortunately this is all I know. (In reply to stsp from comment #7) Ok, thanks - then it most likely came from: commit 27a46793b1a0c6c13c00b07d07bd5743e2c9c7b6 Author: Adam Duskett <aduskett@gmail.com> Date: Thu Oct 14 15:58:39 2021 -0700 package/util-linux: bump to version 2.37.2 |