On 7/8/2022, Michael Grzeschik wrote: > Cc: Thinh.Nguyen@xxxxxxxxxxxx > > On Fri, Jul 08, 2022 at 02:45:30PM +0200, Greg KH wrote: >> On Mon, Jul 04, 2022 at 04:18:12PM +0200, Michael Grzeschik wrote: >>> For High-Speed Transfers the prepare_one_trb function is calculating >>> the >>> multiplier setting for the trb based on the length parameter of the trb >>> currently prepared. This assumption is wrong. For trbs with a sg list, >>> the length of the actual request has to be taken instead. >>> >>> Signed-off-by: Michael Grzeschik <m.grzeschik@xxxxxxxxxxxxxx> >>> --- >>> v1 -> v2: - added refactor patch before this patch >>> - using req->request.length as condition value >> >> Does this need to be backported to older kernels or is it ok for >> 5.20-rc1? > > We can add this line to the patch. > > Fixes: 40d829fb2ec6 ("usb: dwc3: gadget: Correct ISOC DATA PIDs for > short packets") > > I would say this could be backported. > > Thinh? Hi Michael, Since you have the refactoring patch prior to this, it's a bit difficult to backport cleanly without rewriting the patch. It would be nice to be able to backport fixes. But for this particular fix, I don't think will affect too many users. Not many application would use SG for simply 3KB or less requests. (Note that periodic ep can only transfer up to 3KB/interval in HS). Typically a page size is 4KB, and I don't think SG is needed. Anyway, I think Greg already picked up your patches. Thanks for the fix! BR, Thinh