Re: [PATCH v4] mm/zswap: store <PAGE_SIZE compression failed page as-is

[Herbert Xu]

On Tue, Aug 19, 2025 at 10:20:45PM -0700, Chris Li wrote:
>
> In that case it is fair game for the caller to check for the error
> other than -ENOSPC. If the return value is -EINPROGRESS during
> synchronous crypto compression, it is a bug. We are not encouraged to
> use WARN_ON, then having an error count to check that something truly
> unlikely but possible to happen is better than folding it into
> incompressible and pretend the error never happened. At least we can
> know such an error case happened.

Sure warning on -EINPROGRESS is fine.  But every other non-zero
return value should be treated as incompressible.

Cheers,
-- 
Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt