Re: [PATCH for-rc] RDMA/efa: Fix PBL chunk length computation
Leon Romanovsky <[email protected]> Mon, 27 Jul 2026 07:49:43 -0400
| Newsgroups | org.kernel.vger.linux-rdma |
|---|---|
| Message-ID | <[email protected]> |
On Mon, 27 Jul 2026 09:02:55 +0000, Yonatan Nachum wrote:
> On register MR, when creating the PBL, if it's an indirect PBL we create
> a chunk list to hold the PBL pages pointers. Each chunk is 4KB in size
> and can hold 510 addresses (EFA_PTRS_PER_CHUNK) and has a 12-byte
> control buffer at the end of it holding the next chunk's pointer and its
> length.
>
> If the PBL number of pages is a multiple of EFA_PTRS_PER_CHUNK, the
> calculated last chunk length is wrongly computed as 0, even though that
> chunk is fully populated with 510 real page pointers. This wrong length
> is used both to DMA map the chunk and is propagated to the device,
> causing the device to see the chunk as empty and reject the memory
> registration.
>
> [...]
Applied, thanks!
[1/1] RDMA/efa: Fix PBL chunk length computation
https://git.kernel.org/rdma/rdma/c/983c450ae26317
Best regards,
--
Leon Romanovsky <[email protected]>