On Fri, May 06, 2022 at 10:38:18AM +0100, Giovanni Cabiddu wrote: > On Fri, May 06, 2022 at 11:24:14AM +0200, Greg KH wrote: > > On Fri, May 06, 2022 at 09:23:17AM +0100, Giovanni Cabiddu wrote: > > > Move the submission loop to a new function, qat_alg_send_message(), and > > > share it between the symmetric and the asymmetric algorithms. > > > > > > If the HW queues are full return -ENOSPC instead of -EBUSY. > > > > > > For both symmetric and asymmetric services set the number of retries > > > before returning an error to a value that works for both, 20 (was 10 for > > > symmetric and 100 for asymmetric). > > > > > > Cc: stable@xxxxxxxxxxxxxxx > > > Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@xxxxxxxxx> > > > Reviewed-by: Marco Chiappero <marco.chiappero@xxxxxxxxx> > > > > What does this "fix" to require it to be backported to stable trees? > For two reasons, (1) the error code returned if the HW queues are full is > incorrect and (2) to facilitate the backport of the next fix "crypto: > qat - add backlog mechanism". > I can be more explicit in the commit message if required. Please always be explicit in the commit message, it's the only way to communicate this type of thing. thanks, greg k-h
