SQUASHFS error: squashfs_read_data failed to read block

[Chris Read]

We've had system relying on reading data from squashfs that's been stable for a number of years. We've recently noticed that when running on machines that include the patch set from https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2015511 that we start to see error messages of the following shape:

SQUASHFS error: squashfs_read_data failed to read block

Machines showing the problem are all running kernel 5.4.0-150-generic

Machines NOT having any problems with the SAME workloads are running a mix of the following kernels:

5.15.0-52-generic
5.15.0-53-generic
5.4.0-100-generic
5.4.0-104-generic
5.4.0-105-generic
5.4.0-125-generic
5.4.0-137-generic
5.4.0-147-generic

I'm not able to reproduce the failure on demand.

What is suspicious to me from that change set is that the following 2 patches are applied:

squashfs: harden sanity check in squashfs_read_xattr_id_table
Squashfs: fix handling and sanity checking of xattr_ids count

and then only 1 of them is reverted, although the patches in question appear to be related:

revert "squashfs: harden sanity check in squashfs_read_xattr_id_table"

In addition to 5.4.0-149.166, it appears the same surviving patch was also pulled into the following Ubuntu kernel builds:

5.15.0-72.79
5.19.0-42.43

Additional data from one of the problem machines:

# lsb_release -rd
Description: Ubuntu 20.04.6 LTS
Release: 20.04

# cat /proc/version_signature
Ubuntu 5.4.0-150.167-generic 5.4.233

Relevant lines from dmesg:

[964961.979932] SQUASHFS error: squashfs_read_data failed to read block 0x52ca3e42
[964961.987247] SQUASHFS error: Unable to read data cache entry [52ca3e42]
[964961.993872] SQUASHFS error: Unable to read page, block 52ca3e42, size d674
[1051130.734432] SQUASHFS error: squashfs_read_data failed to read block 0xdd6d9c3
[1100614.200893] SQUASHFS error: squashfs_read_data failed to read block 0x239673b7
[1100614.208305] SQUASHFS error: Unable to read data cache entry [239673b7]
[1100614.215017] SQUASHFS error: Unable to read page, block 239673b7, size 12fba
[1100614.222170] SQUASHFS error: Unable to read data cache entry [239673b7]
[1100614.228870] SQUASHFS error: Unable to read page, block 239673b7, size 12fba
[1100614.236007] SQUASHFS error: Unable to read data cache entry [239673b7]
[1100614.242711] SQUASHFS error: Unable to read page, block 239673b7, size 12fba
[1100634.392881] SQUASHFS error: squashfs_read_data failed to read block 0x1746ee61
[1100634.400287] SQUASHFS error: Unable to read data cache entry [1746ee61]
[1100634.406994] SQUASHFS error: Unable to read page, block 1746ee61, size 111f3
[1100634.414163] SQUASHFS error: Unable to read data cache entry [1746ee61]
[1100634.420867] SQUASHFS error: Unable to read page, block 1746ee61, size 111f3
[1100634.428044] SQUASHFS error: Unable to read data cache entry [1746ee61]
[1100634.434757] SQUASHFS error: Unable to read page, block 1746ee61, size 111f3
[1100634.647016] SQUASHFS error: squashfs_read_data failed to read block 0x3afd7129

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 2023925

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Chris Read (chris-read)]

Not able to add a full apport-collect output to this bug.