Re: [PATCH mptcp-next 2/7] mptcp: move the stale logic out of retrans scheduler

[Paolo Abeni <pabeni@redhat.com>]

On 8/6/26 11:31 AM, Geliang Tang wrote:
> Thank you for this new version of the series. I have rebased the MPTCP
> KTLS code onto it, and all tests passed.
> 
> On Wed, 2026-08-05 at 18:17 +0200, Paolo Abeni wrote:
>> This allow separating the stale logic invocation and the retrans
>> scheduler, and will simplify the next patch.
>>
>> It's also a cleaner design as the retrans scheduler has currently
>> too many side effects. As a possible downside, the retrans work will
>> now traverse the subflows list additional times; that does not matter
>> much, as this is slowpath.
> 
> However, this patch makes the KTLS selftests significantly slower -
> specifically, this test case now takes several hundred seconds to
> complete, whereas it previously finished in just a few seconds:
> 
> 	chunked_sendfile(_metadata, self, 1, 4096);
> 
> Is there any way we can make it run faster?
I guess you test on top of the whole series?

If you tested on top of the whole series, I think the slowdown is due to
the issue in patch 3/7 noted by sashiko gemini and should be fixed in v2.

/P