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

Groups > linux.debian.bugs.dist > #1248796 > unrolled thread

Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm

Back to article view | Back to linux.debian.bugs.dist

Contents

  Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Paul Gevers <elbrus@debian.org> - 2025-06-02 22:50 +0200
    Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Ben Hutchings <ben@decadent.org.uk> - 2025-06-03 01:10 +0200
      Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Paul Gevers <elbrus@debian.org> - 2025-06-03 08:00 +0200
        Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Paul Gevers <elbrus@debian.org> - 2025-06-03 08:10 +0200
          Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Paul Gevers <elbrus@debian.org> - 2025-06-03 18:50 +0200
            Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm Ben Hutchings <ben@decadent.org.uk> - 2025-06-09 17:30 +0200

#1248796 — Bug#1107194: [s390x] trixie's lsinitramfs fails on initramfs build on bookworm

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

Package: initramfs-tools-core
Version: 0.147
Severity: normal
File: /usr/bin/lsinitramfs
X-Debbugs-Cc: debian-s390@lists.debian.org
User: debian-s390@lists.debian.org
Usertags: s390x

Hi,

Two days ago, I upgraded the s390x ci.debian.net host from bookworm to
trixie. The system was already running the kernel from backports, so
the upgrade was mostly of the userland. That upgrade hasn't been a
nice experience, because the system failed to boot afterwards. I was
able to boot the previous kernel eventually, but while trying to
figure out if the initramfs was somehow missing pieces I tried to run
lsinitramfs on both the old (current) and new (failing) initramfs, but
that failed on the old one:
$ lsinitramfs /boot/initrd.img.old
cpio: premature end of archive

Running the same command on the new (but non booting initramfs) worked
as expected.

Paul

-- System Information:
Debian Release: 13.0
   APT prefers testing
   APT policy: (990, 'testing'), (500, 'unstable'), (1, 'experimental')
Architecture: s390x

Kernel: Linux 6.12.22+bpo-s390x (SMP w/10 CPU threads)
Kernel taint flags: TAINT_WARN
Locale: LANG=C, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages initramfs-tools-core depends on:
ii  coreutils       9.7-2
ii  cpio            2.15+dfsg-2
ii  dracut-install  106-6
ii  e2fsprogs       1.47.2-1+b1
ii  klibc-utils     2.0.14-1
ii  kmod            34.2-2
ii  logsave         1.47.2-1+b1
ii  udev            257.6-1

Versions of packages initramfs-tools-core recommends:
ii  busybox-static [busybox]  1:1.37.0-4
ii  zstd                      1.5.7+dfsg-1

Versions of packages initramfs-tools-core suggests:
ii  bash-completion  1:2.16.0-7

-- no debconf information

[toc] | [next] | [standalone]

#1248815

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

Control: tag -1 moreinfo

On Mon, 2025-06-02 at 22:23 +0200, Paul Gevers wrote:
> Package: initramfs-tools-core
> Version: 0.147
> Severity: normal
> File: /usr/bin/lsinitramfs
> X-Debbugs-Cc: debian-s390@lists.debian.org
> User: debian-s390@lists.debian.org
> Usertags: s390x
> 
> Hi,
> 
> Two days ago, I upgraded the s390x ci.debian.net host from bookworm to
> trixie. The system was already running the kernel from backports, so
> the upgrade was mostly of the userland. That upgrade hasn't been a
> nice experience, because the system failed to boot afterwards. I was
> able to boot the previous kernel eventually, but while trying to
> figure out if the initramfs was somehow missing pieces I tried to run
> lsinitramfs on both the old (current) and new (failing) initramfs, but
> that failed on the old one:
> $ lsinitramfs /boot/initrd.img.old
> cpio: premature end of archive
[...]

Are you quite sure this is actually the old, working, initramfs?  (What
is that a xymlink to?)

The current version of initramfs-tools (0.148.1) has a rewritten
unmkinitramfs and an autopkgtest script for it.  But the implementation
in 0.147 also passes all test cases except those where all component
archives are uncompressed, which is not something initramfs-tools ever
generated.

Does the new implementation of unmkinitramfs work any better on the
working initramfs image?

If not, can you upload the image somewhere for me to inspect?

Ben.

-- 
Ben Hutchings
The two most common things in the universe are hydrogen and stupidity.

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

#1248827

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

Control: tag -1 - moreinfo

Hi,

On 03-06-2025 00:33, Ben Hutchings wrote:
> Are you quite sure this is actually the old, working, initramfs?  (What
> is that a xymlink to?)


root@ci-worker-s390x-01:~# ls -al /boot/
total 106288
drwxr-xr-x  4 root root     4096 Jun  1 19:58 .
drwxr-xr-x 19 root root     4096 Jun  1 18:41 ..
-rw-r--r--  1 root root       83 Apr 25 01:42 System.map-6.12.22+bpo-s390x
-rw-r--r--  1 root root       83 May  6 19:33 System.map-6.12.27-s390x
-rw-r--r--  1 root root       83 Apr 25 01:42 System.map.bookworm
-rw-------  1 root root   119808 Jun  1 19:58 bootmap
-rw-r--r--  1 root root   161926 Apr 25 01:42 config-6.12.22+bpo-s390x
-rw-r--r--  1 root root   164356 May  6 19:33 config-6.12.27-s390x
-rw-r--r--  1 root root   161926 Apr 25 01:42 config.bookworm
lrwxrwxrwx  1 root root       24 May 31 18:50 initrd.img -> 
initrd.img-6.12.27-s390x
-rw-r--r--  1 root root 26903848 May 31 06:25 initrd.img-6.12.22+bpo-s390x
-rw-r--r--  1 root root 30448110 Jun  1 19:38 initrd.img-6.12.27-s390x
-rw-r--r--  1 root root 26903848 May 31 06:25 initrd.img.bookworm
lrwxrwxrwx  1 root root       28 May 31 18:50 initrd.img.old -> 
initrd.img-6.12.22+bpo-s390x
drwx------  2 root root    16384 Apr 24  2022 lost+found
lrwxrwxrwx  1 root root       21 May 31 18:50 vmlinuz -> 
vmlinuz-6.12.27-s390x
-rw-r--r--  1 root root  7688192 Apr 25 01:42 vmlinuz-6.12.22+bpo-s390x
-rw-r--r--  1 root root  8540160 May  6 19:33 vmlinuz-6.12.27-s390x
-rw-r--r--  1 root root  7688192 Apr 25 01:42 vmlinuz.bookworm
lrwxrwxrwx  1 root root       25 May 31 18:50 vmlinuz.old -> 
vmlinuz-6.12.22+bpo-s390x
drwxr-xr-x  2 root root     4096 Jun  1 19:43 working

(Those bookworm lines are copies of the bpo lines. I added an extra line 
to the boot menu in the hope that things can remain working).

> The current version of initramfs-tools (0.148.1) has a rewritten
> unmkinitramfs and an autopkgtest script for it.  But the implementation
> in 0.147 also passes all test cases except those where all component
> archives are uncompressed, which is not something initramfs-tools ever
> generated.
> 
> Does the new implementation of unmkinitramfs work any better on the
> working initramfs image?


That at least doesn't fail. Yes, it seems to extract the content.

> If not, can you upload the image somewhere for me to inspect?

Given the answer above, do you still want me to do this?

Paul

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

#1248828

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

Hi,

On 03-06-2025 07:54, Paul Gevers wrote:
> On 03-06-2025 00:33, Ben Hutchings wrote:
>> Are you quite sure this is actually the old, working, initramfs?  (What
>> is that a xymlink to?)


I could have sworn that I checked this, but is seems that lsinitramfs 
doesn't work on the symlink, but does work on the actual file. 
Interesting for initrd.img (the new one) it doesn't matter if I point at 
the symlink or the file.

Paul

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

#1248901

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

Control: merge -1 1015954

On 03-06-2025 08:03, Paul Gevers wrote:
> I could have sworn that I checked this, but is seems that lsinitramfs 
> doesn't work on the symlink, but does work on the actual file. 
> Interesting for initrd.img (the new one) it doesn't matter if I point at 
> the symlink or the file.


Seems like this is a duplicate of bug 1015954, which even has a patch 
and maybe is moot with the new version in stable.

Paul

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

#1249500

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

On Tue, 3 Jun 2025 18:44:21 +0200 Paul Gevers <elbrus@debian.org> wrote:
> Control: merge -1 1015954
> 
> On 03-06-2025 08:03, Paul Gevers wrote:
> > I could have sworn that I checked this, but is seems that
lsinitramfs 
> > doesn't work on the symlink, but does work on the actual file. 
> > Interesting for initrd.img (the new one) it doesn't matter if I
point at 
> > the symlink or the file.
> 
> 
> Seems like this is a duplicate of bug 1015954, which even has a patch 
> and maybe is moot with the new version in stable.

Right.  Thanks for working that out.

Ben.

-- 
Ben Hutchings
It is a miracle that curiosity survives formal education.
                                                      - Albert Einstein

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.bugs.dist

csiph-web