[SRU][Focal/Lunar][PATCH 0/1] CVE-2023-3212

[Yuxuan Luo]

On 8/17/23 03:49, Stefan Bader wrote:
> On 16.08.23 20:53, Yuxuan Luo wrote:
>>
>> On 8/16/23 04:28, Stefan Bader wrote:
>>> On 15.08.23 18:24, Yuxuan Luo wrote:
>>>> [Impact]
>>>> Yang Lan discovered that the GFS2 file system implementation in the
>>>> Linux kernel could attempt to dereference a null pointer in some
>>>> situations. An attacker could use this to construct a malicious GFS2
>>>> image that, when mounted and operated on, could cause a denial of
>>>> service (system crash).
>>>
>>> Why is Jammy not affected?
>>
>> Good catch. UCT shows that Jammy is `pending (5.15.0-83.92)`. My local
>>
>> cranky/master-next:linux-meta/debian/changelog only shows up to 83.80,
>
> You should re-fetch. I was temporarily behind for master-next but the 
> branch is updated now.
Sorry I didn't make it clear, I re-fetched from my local repo and it 
says it is up-to-date (5.15.0-83.92), that's why I assumed it was 
patched/fetched somewhere else.
>
>>
>> I assume it has already been fetched in Jammy's tree somewhere else.
>
> Looks like it was included via v5.15.116 upstream stable.

No wonders.

>>
>>>
>>>>
>>>> [Backport]
>>>> For Lunar, it is a clean cherry pick.
>>>> For Focal, there is a conflict revolving the `if
>>>> (test_bit(GIF_ALLOC_FAILED, &ip->i_flags)) {` line which requires
>>>> 9e73330f298a (“gfs2: Try harder to delete inodes locally”) commit.
>>>> However, even though the aforementioned conflicting line was 
>>>> modified to
>>>> what it is now in the fix commit, it is not related to the fix, ignore
>>>> it.
>>>>
>>>> [Test]
>>>> Tested via mount and umount gfs2.
>>>>
>>>> [Potential Regression]
>>>> Expect no regression.
>>>>
>>>>
>>>> Bob Peterson (1):
>>>>    gfs2: Don't deref jdesc in evict
>>>>
>>>>   fs/gfs2/super.c | 8 ++++++++
>>>>   1 file changed, 8 insertions(+)
>>>>
>>>
>
> Thanks,
> - Stefan
>