On 2021/7/8 上午2:14, David Sterba wrote:

On Wed, Jul 07, 2021 at 01:41:46PM -0400, Neal Gompa wrote:

quoted

quoted

But on the other hand, I'm not sure what's the proper way to introduce a
fix for v5.15 window.

Should I just disable readahead for compression read (which just needs
two lines to return 0 for subpage case in add_ra_bio_pages())?

Or should I add the proper fix into the patchset?

If this is going into 5.15 instead of 5.14, just add the proper fix
into this patch set. But if we want to land this in 5.14, I would
suggest disabling it for now and then having a separate patch set for
that.

5.14 is not possible, there are other subpage changes already merged so
fixes to existing level of support (still limited) can go to 5.14-rcs
but this whole patchses it is targeting 5.15.

quoted

You're already targeting 5.15 (though I kind of want this in 5.14...),
so I suggest going with adding the fix to the patch set.

For 5.15 it's free to do any change, eg. fold the fix or do a separate
patch explaining all the details.

Another reason why I'm considering to disable that readahead ability is,
it really makes little sense for 64K page size.

Our maximum compressed extent size is just 128K, at most 3 pages for 64K
page size (that's when the extent is not page aligned).

Just saving two reads (which may only be one or 2 4K sectors) doesn't
really make it worthy, especially considering the change to ra itself is
going to affecting common code path.

Thus I prefer to disable for the initial support, and only add the
proper fix for the compression write support.

In general the development changes must be ready at the rc5 time at the
latest (with a week or two for some testing and catching last bugs).
Once it's due there's like two months until the next code freeze.

Great, I'll just add a simple disabling patch with comments on why the
current code doesn't work and why it's fine to disable it for 64K page size.

Thanks,
Qu