Re: [PATCH v3 1/3] mm/swapfile: unuse_pte can map random data if swap read fails

[Andrew Morton]

On Tue, 10 May 2022 14:58:05 +0800 Miaohe Lin <linmiaohe@xxxxxxxxxx> wrote:

> >> Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
> >> Acked-by: David Hildenbrand <david@xxxxxxxxxx>
> > 
> > When I reproduced the issue (generated read error with dm-dust), I saw
> > infinite loop in the while loop in shmem_unuse_inode() (and this happens
> > even with this patch). I confirmed that shmem_swapin_page() returns -EIO,
> > but shmem_unuse_swap_entries() does not return the error to the callers,
> > so the while loop in shmem_unuse_inode() seems not break.
> > 
> 
> Many thanks for your report! I didn't test the shmem case because I saw -EIO
> is returned. So I just focus on the normal page case. Sorry about it. :(
> 
> > So maybe you need more code around shmem_unuse_inode() to handle the error?
> 
> I will try to reproduce it and come up a fixup patch asap! And if you like, you
> can kindly solve this issue too. ;)

Seems that this patch didn't cause the infinite loop, so as far as I
can tell it is good to be merged up.  But the problem it solves isn't
urgent and fixing that infinite loop might impact this change so I
think I'll drop this version.