Reviewed-by: Pavel Shilovsky <pshilov@xxxxxxxxxxxxx> -- Best regards, Pavel Shilovsky вс, 27 сент. 2020 г. в 20:41, Steve French <smfrench@xxxxxxxxx>: > > merged into cifs-2.6.git for-next > > On Fri, Sep 25, 2020 at 2:33 AM Rohith Surabattula > <rohiths.msft@xxxxxxxxx> wrote: > > > > As this status code is returned when there is an internal > > unavailability. So for any transaction, this status code can be > > returned and EJUKEBOX check needs to be added at many places to > > support this. > > > > Respined the patch with signoff flag and attached. > > > > On Fri, Sep 25, 2020 at 11:42 AM Steve French <smfrench@xxxxxxxxx> wrote: > > > > > > Ronnie also mentioned EJUKEBOX as a possibly better error mapping to > > > return (and then check for). EJUKEBOX implies waiting and then > > > backoff. > > > > > > On Fri, Sep 18, 2020 at 1:20 AM ronnie sahlberg > > > <ronniesahlberg@xxxxxxxxx> wrote: > > > > > > > > On Fri, Sep 18, 2020 at 4:08 PM rohiths msft <rohiths.msft@xxxxxxxxx> wrote: > > > > > > > > > > Hi All, > > > > > > > > > > This fix is to handle STATUS_IO_TIMEOUT status code. This status code > > > > > is returned by the server in case of unavailability(internal > > > > > disconnects,etc) and is not treated by linux clients as retriable. So, > > > > > this fix maps the status code as retriable error and also has a check > > > > > to drop the connection to not overload the server. > > > > > > > > > > > > > Do we need a new method for this? Wouldn't it be enough to just do the > > > > remap-to-EAGAIN and have it handled as all other retryable errors? > > > > > > > > > > > > > Regards, > > > > > Rohith > > > > > > > > > > > > -- > > > Thanks, > > > > > > Steve > > > > -- > Thanks, > > Steve
