> > > The diff in this patch seems larger then it should be. It seems to
> > > reorder handle_tx and handle_rx and the diff results in ripping out the
> > > current implementations and putting in something new, hiding the actual
> > > refactoring.
> > >
> > > Sören
> >
> > Yes, I agree the reordering is confusing, but I was trying to keep the
> commits I had cherry-picked from the linux-xlnx tree as intact as possible, so
> as to help traceability. I can try refactoring Anirudha's changes to be less
> obfuscatory if you think that's necessary?
>
> I think so. Why isn't this change in mainline? Is it needed in mainline
> at all? The diff makes it pretty much impossible to judge that and the
> commit message suggests that parts of this change are no longer
> applicable upstream. I'd prefer a clean diff that makes it easy to
> understand the changes/refactoring.
Looking at the original commits again, it appears that both yourself and anirudh@xxxxxxxxxx have factored cdns_uart_isr() out into cdns_uart_handle_rx() and cdns_uart_handle_tx(), but in different commits in different repos, with the new functions in different orders, and Anirudha also added some error handling fixes in the same commit. Hence my attempt to reconcile the two forks of the driver have resulted in this messy patch. I'll revise the patch to just include the error handling fixes.
Regards,
Scott.
��.n��������+%������w��{.n�����{��ǫ����{ay� ʇڙ���f���h������_�(�階�ݢj"��������G����?���&��
