Re: Sequential read(8K) from compressed files are very slow

[Christoph Hellwig <hch@infradead.org>]

On Wed, Jun 04, 2025 at 02:49:19PM -0700, Boris Burkov wrote:
> unlocked the perf on the compressed 8k (and 4k) case without regressing
> 128k or bs=1M for a file with enormous extents. I get 1.4GB/s on both.
> Still less on the incompressible file (though higher than without the
> patch. 900MB/s vs 600MB/s)
> 
> I don't think this really perfectly makes sense for an actually proper
> version, as readahead_folio() will advance the index and we will keep
> "expanding" it to be the size of the compressed extent_map. On the other
> hand, if it is stupid and it works, is it stupid?

The way readahead_expand works is that it should basically be no-op
in that case as the two while loops doing the actual work are skipped.
But it might be worth double checking that this is really the case.

You'll probably also want to remove add_ra_bio_pages entirely.

> Very promising... I'm going to study this API further and try to make
> a real version. Very open to ideas/advice on how the API is best meant
> to be used and if this result might be "cheating" in some way.

As far as I can tell this is how it is intended to be used.