Hi [apologies if I missed some followup but was not able to find the answer] On Wed, Sep 20, 2023 at 08:32:17AM +0000, Ricky WU wrote: > > On Wed, Sep 20, 2023 at 07:30:00AM +0000, Ricky WU wrote: > > > Hi Greg k-h, > > > > > > This patch is our solution for this issue... > > > And now how can I push this? > > > > Submit it properly like any other patch, what is preventing that from > > happening? > > > > (commit 8ee39ec) some reader no longer force #CLKREQ to low when system need to enter ASPM. > But some platform maybe not implement complete ASPM? I don't know..... it causes problems... > > Like in the past Only the platform support L1ss we release the #CLKREQ. > But new patch we move the judgment (L1ss) to probe, because we met some host will clean the config space from S3 or some power saving mode > And also we think just to read config space one time when the driver start is enough Is there a potential fix which is queued for this or would be the safest option to unbreak the regression to revert the commit in the stable trees temporarily? I'm asking because in Debian we got the report at https://bugs.debian.org/1052063 (and ideally to unbreak the situation for the user I would like to include a fix in the next upload we do, but following what you as upstream will do ideally). Regards, Salvatore
