Alibaba Cloud Linux:What do I do if the Alibaba Cloud Linux 3 operating system is down due to exceptions that occur when I mount an EROFS file system?

Affected environment

Symptom

The system crashes with a kernel NULL pointer dereference when you mount an EROFS file system on a block device. The following commands reproduce the issue:

sudo yum install -y erofs-utils
mkdir -p test mnt
mkfs.erofs foo.erofs test
sudo mount -t erofs -o loop foo.erofs mnt

If the system is affected, it crashes and the following call trace appears:

[  225.747952] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000370
..
[  225.752658] CPU: 3 PID: 5829 Comm: mount Kdump: loaded Not tainted 5.10.134-15.al8.aarch64 #1
[  225.753089] Hardware name: Alibaba Cloud Alibaba Cloud ECS, BIOS 1.0.0 01/01/2017
[  225.754016] pc : __erofs_bread+0x64/0x1d0 [erofs]
[  225.754271] lr : erofs_read_metabuf+0x44/0x80 [erofs]
..
[  225.758508] Call trace:
[  225.758636]  __erofs_bread+0x64/0x1d0 [erofs]
[  225.758859]  erofs_read_metabuf+0x44/0x80 [erofs]
[  225.759112]  erofs_read_superblock+0x60/0x264 [erofs]
[  225.759370]  erofs_fc_fill_super+0xf0/0x310 [erofs]
[  225.759621]  get_tree_bdev+0x15c/0x250
[  225.760109]  erofs_fc_get_tree+0x38/0x54 [erofs]
[  225.760662]  vfs_get_tree+0x2c/0xf0
[  225.761157]  do_new_mount+0x164/0x1d0
[  225.761652]  path_mount+0x1bc/0x570
[  225.762133]  __arm64_sys_mount+0x114/0x140

Root cause

The EROFS feature added in kernel 5.10.134-15.al8 modifies the __erofs_bread() function. The modified function does not correctly handle EROFS mounts on block devices, which causes a NULL pointer dereference and crashes the system.

Solutions

sudo yum install -y kernel-hotfix-18359162-5.10.134-15