On Mon, Feb 27, 2023 at 11:47:24AM -0800, Darrick J. Wong wrote:
> OTOH, it also means that we've learned the hard way that pagecache
> operations need a means to revalidate mappings to avoid write races.
> This applies both to the initial pagecache write and to scheduling
> writeback, but the mechanisms for each were developed separately and
> years apart. See iomap::validity_cookie and
> xfs_writepage_ctx::{data,cow}_seq for what I'm talking about.
> We (xfs developers) ought to figure out if these two mechanisms should
> be merged before more filesystems start using iomap for buffered io.
That puts a good yield notice to some conversion efforts, thanks, this
already alone is very useulf.
> I'd like to have a discussion about how to clean up and clarify the
> iomap interfaces, and a separate one about how to port the remaining 35+
> filesystems. I don't know how exactly to split this into LSF sessions,
> other than to suggest at least two.
>From a conversion perspective, ideally if it was obvious I think we
should be able to do some of it ala coccinelle, but I have yet to see
any remotely obvious pattern. And it makes me wonder, should we strive
to make the conversion as painless / obvious ? Is that a good litmus
for when we should be ready to start converting other filesystems?
> If hch or dchinner show up, I also want to drag them into this. :)
And here I've been thinking I had to go to Australia to see you all together.
Luis
