[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1099461: marked as done (initramfs-tools: _call_dracut_install breaks set -e hooks without /sys)

Your message dated Thu, 24 Apr 2025 17:42:09 +0200
with message-id <22f1f1e0705f268778e5abb090cc453ee57ede09.camel@decadent.org.uk>
and subject line Re: initramfs-tools: _call_dracut_install breaks set -e hooks without /sys
has caused the Debian Bug report #1099461,
regarding initramfs-tools: _call_dracut_install breaks set -e hooks without /sys
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)


-- 
1099461: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1099461
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message --- 
Source: initramfs-tools
Version: 0.146
X-Debbugs-CC: team@grml.org
Control: affects -1 live-boot

Dear initramfs-tools maintainers,
daily builds of Grml started failing today. I've debugged it a bit, and it seems to boil down to this combination:
1) our builds run in a containerized environment, and thus have no /sys or /proc. This is /good/ for building Live images, as leaking anything from the build host would be wrong anyway. 

2) live-boot hook.live does "set -e" right at the start:
https://sources.debian.org/src/live-boot/1%3A20250225/backend/initramfs-tools/live.hook/#L3

3) hook-functions _call_dracut_install reads from /sys/module/firmware_class/parameters/path
without checking if the file exists (or otherwise allowing for set -e), and this causes the entire process to abort. 

Happens here:
https://sources.debian.org/src/initramfs-tools/0.146/hook-functions/?hl=106#L122
In add_firmware there's a similar code path, I assume that will also be a problem: 

https://sources.debian.org/src/initramfs-tools/0.146/hook-functions/?hl=106#L65
I understand that this combination might be somewhat surprising, but it is very valuable for building live "CD" images. And so far it worked really well. 

Please consider short-circuiting the read from /sys/module/firmware_class/parameters/path
somehow, and please for trixie.

Best,
Chris

--- End Message ---
--- Begin Message --- 
Version: 0.147

The fix for this was included in the above version.

Ben.

-- 
Ben Hutchings
compatible: Gracefully accepts erroneous data from any source

Attachment: signature.asc
Description: This is a digitally signed message part


--- End Message ---
Reply to: