On 2/13/25 22:46, Mina Almasry wrote:
On Thu, Feb 13, 2025 at 2:36 PM Pavel Begunkov <asml.silence@xxxxxxxxx> wrote:
...
+ if (pp->dma_map)
+ return -EOPNOTSUPP;
This condition should be flipped actually. pp->dma_map should be true,
otherwise the provider isn't supported.
It's not implemented in this patch, which is why rejected.
You can think of it as an unconditional failure, even though
io_pp_zc_init is not reachable just yet.
Ah, I see in the follow up patch you flip the condition, that's fine then.
I usually see defensive checks get rejected but I don't see that
Yeah, sounds like that's the rule in net/.
blocking this series, so FWIW:
Reviewed-by: Mina Almasry <almasrymina@xxxxxxxxxx>
Thanks for the review!
--
Pavel Begunkov
