Re: [RFC PATCH v4 12/12] iomap: Handle deadlock due to repeating folios in RWF_WRITETHROUGH

[Pankaj Raghav (Samsung)]

On Mon, Sep 28, 2026 at 05:33:13PM +0530, Ojaswin Mujoo wrote:
> In iomap_writethrough_iter() we might encounter repeating folios across
> multiple iterations. Repeating folios can occur if, example,
> copy_folio_from_iter_atomic() does a short copy due to userspace pages
> not faulted in. This is an issue because a previous loop might have
> started writeback on them but not yet issued the IO. In the next
> iteration trying to get the same folio with FGP_STABLE will result in a
> deadlock. Since repeating folios will always be encountered back to
> back, we can just use a simple cur != prev check to detect them.
> 
> Use this to avoid waiting for writeback or starting writeback on folios
> we have already processed. Note that in ->endio() we might end up
> calling folio_end_writethrough() twice on the same folio which can cause
> issues with folio_xor_flags_has_waiters(). For simplicity, just change
> the folio_xor_flags_has_waiters() call to an idempotent variant.

Is this an outdated comment?

> 
> Reported-by: Pankaj Raghav <pankaj.raghav@xxxxxxxxx>

Thanks for fixing it. I think I was able to trigger this in a VM with
limited memory where the chances of short copy is probably higher. The
fixes makes sense to me.

I will run some tests this week and report back if there is an issue.

--
Pankaj