Re: [PATCH 4/5] iomap: support write completions from interrupt context

[Christoph Hellwig <hch@lst.de>]

On Thu, Nov 13, 2025 at 01:06:34PM +0100, Jan Kara wrote:
> On Thu 13-11-25 11:06:30, Christoph Hellwig wrote:
> > On Thu, Nov 13, 2025 at 10:54:46AM +0100, Jan Kara wrote:
> > > > You mean drop the common helper?  How would that be better and less
> > > > fragile?   Note that I care strongly, but I don't really see the point.
> > > 
> > > Sorry I was a bit terse. What I meant is that the two users of
> > > iomap_dio_is_overwrite() actually care about different things and that
> > > results in that function having a bit odd semantics IMHO. The first user
> > > wants to figure out whether calling generic_write_sync() is needed upon io
> > > completion to make data persistent (crash safe).
> > 
> > Yes.
> > 
> > > The second user cares
> > > whether we need to do metadata modifications upon io completion to make data
> > > visible at all.
> > 
> > Not quite.  It cares if either generic_write_sync needs be called,
> > or we need a metadata modification, because both require the workqueue.
> 
> I agree but generic_write_sync() calling is handled by 
> 
> +       else if (dio->flags & IOMAP_DIO_NEED_SYNC)
> +               dio->flags &= ~IOMAP_DIO_INLINE_COMP;
> 
> in your patch. So I assumed (maybe wrongly) that the second call to
> iomap_dio_is_overwrite() in iomap_dio_bio_iter() is only about detecting a
> need of metadata modification. And my argument is that the patch could use
> IOMAP_DIO_UNWRITTEN | IOMAP_DIO_COW the same way as it uses
> IOMAP_DIO_NEED_SYNC instead of calling iomap_dio_is_overwrite().
> 
> But if you don't like that I don't think it makes a huge difference and the
> code is correct as is so feel free to add:

I'll take a look if there is a way to clear thing up a bit.