On Tue, Aug 27, 2024 at 4:07 AM Yu Kuai <yukuai1@xxxxxxxxxxxxxxx> wrote:
>
[...]
>
> Test result: about 10 times faster:
> Before this patch: 50943374 microseconds
> After this patch: 5096347 microseconds
>
> BTW, commit 611d5cbc0b35 ("md: fix deadlock between mddev_suspend and flush
> bio") claims to fix the problem introduced by commit fa2bbff7b0b4 ("md:
> synchronize flush io with array reconfiguration"), which is wrong, the
> problem is actually indroduced by commit 409c57f38017 ("md: enable
> suspend/resume of md devices."), hence older kernels will be affected by
> CVE-2024-43855.
>
> What's worse, the CVE patch can't be backported to older kernels due to
> a lot of relied patches, and this patch can be backported to olders
> kernels to fix the CVE instead.
>
> Signed-off-by: Yu Kuai <yukuai3@xxxxxxxxxx>
Applied to md-next.
Thanks,
Song
