Re: [PATCH 5.4 00/90] 5.4.263-rc3 review

[Guenter Roeck]

On 12/5/23 11:22, Greg Kroah-Hartman wrote:
> This is the start of the stable review cycle for the 5.4.263 release.
> There are 90 patches in this series, all will be posted as a response
> to this one.  If anyone has any issues with these being applied, please
> let me know.
>
> Responses should be made by Thu, 07 Dec 2023 18:32:16 +0000.
> Anything received after that time might be too late.
[ ... ]
> Qu Wenruo <wqu@xxxxxxxx>
>      btrfs: add dmesg output for first mount and last unmount of a filesystem

This patch results in the following code in fs/btrfs/disk-io.c:open_ctree():

	struct btrfs_super_block *disk_super;
	... (no access to disk_super)
	btrfs_info(fs_info, "first mount of filesystem %pU", disk_super->fsid);

which I would assume _should_ result in btrfs crashes. No idea why that isn't
actually happening or why gcc doesn't complain. Building allmodconfig with
clang does complain, but doesn't bail out.

s/btrfs/disk-io.c:2832:55: warning: variable 'disk_super' is uninitialized when used here [-Wuninitialized]
        btrfs_info(fs_info, "first mount of filesystem %pU", disk_super->fsid);
                                                             ^~~~~~~~~~

The actual log output is:

[    7.302427] BTRFS info (device nvme0n1): first mount of filesystem (efault)

It might be a good idea to either revert this patch or fix it up
(though I don't know how to fix it up).

Guenter