From: Cheng Xu <chengyou@linux.alibaba.com>
To: jgg@ziepe.ca, leon@kernel.org
Cc: linux-rdma@vger.kernel.org, KaiShen@linux.alibaba.com
Subject: [PATCH for-next v4 0/5] RDMA/erdma: Support non-contiguous kernel queue buffers
Date: Thu, 17 Sep 2026 20:04:54 +0800 [thread overview]
Message-ID: <20260917120459.13286-1-chengyou@linux.alibaba.com> (raw)
Hi,
This small patchset introduces support for non-contiguous kernel queue
buffers. Kernel QP and CQ buffers currently use a single coherent DMA
allocation, which can fail for large queues when memory is fragmented.
To address this, we allocate page-sized coherent buffers and describe
them with the existing MTT.
- #1 unwinds kernel QP initialization failures explicitly.
- #2 adds support for non-contiguous kernel QP buffers.
- #3 adds support for non-contiguous kernel CQ buffers.
- #4 unifies userspace and kernel space queue buffer management.
- #5 moves kernel QP helpers after the shared memory helpers.
Changes from v3:
- Address Sashiko automated review findings:
- Keep the CPU and DMA addresses returned by dma_alloc_coherent() instead
of deriving a virtual address through scatterlist helpers.
- Sync continuous MTTs after populating their entries.
- Keep DMA MRs and MTT-only FRMRs outside queue buffer backing cleanup.
Changes from v2:
- Add explicit resource unwinding for kernel QP initialization.
- Use struct scatterlist for kernel queue buffer pages.
- Make memory cleanup operate only on initialized backing memory.
- Return -ENOMEM when a kernel QP doorbell record allocation fails.
Changes from v1:
- Fix formatting issues.
Thanks,
Cheng Xu
Cheng Xu (5):
RDMA/erdma: Unwind kernel QP initialization failures
RDMA/erdma: Support non-contiguous kernel QP buffers
RDMA/erdma: Support non-contiguous kernel CQ buffers
RDMA/erdma: Unify userspace and kernel queue buffer management
RDMA/erdma: Move kernel QP helpers after memory helpers
drivers/infiniband/hw/erdma/erdma_cq.c | 18 +-
drivers/infiniband/hw/erdma/erdma_qp.c | 38 +-
drivers/infiniband/hw/erdma/erdma_verbs.c | 569 ++++++++++++----------
drivers/infiniband/hw/erdma/erdma_verbs.h | 65 ++-
4 files changed, 404 insertions(+), 286 deletions(-)
--
2.31.1
next reply other threads:[~2026-09-17 12:05 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-17 12:04 Cheng Xu [this message]
2026-09-17 12:04 ` [PATCH for-next v4 1/5] RDMA/erdma: Unwind kernel QP initialization failures Cheng Xu
2026-09-17 12:11 ` sashiko-bot
2026-09-17 12:04 ` [PATCH for-next v4 2/5] RDMA/erdma: Support non-contiguous kernel QP buffers Cheng Xu
2026-09-17 12:16 ` sashiko-bot
2026-09-17 12:04 ` [PATCH for-next v4 3/5] RDMA/erdma: Support non-contiguous kernel CQ buffers Cheng Xu
2026-09-17 12:13 ` sashiko-bot
2026-09-17 12:04 ` [PATCH for-next v4 4/5] RDMA/erdma: Unify userspace and kernel queue buffer management Cheng Xu
2026-09-17 12:13 ` sashiko-bot
2026-09-17 12:04 ` [PATCH for-next v4 5/5] RDMA/erdma: Move kernel QP helpers after memory helpers Cheng Xu
2026-09-17 12:10 ` sashiko-bot
2026-09-23 10:53 ` [PATCH for-next v4 0/5] RDMA/erdma: Support non-contiguous kernel queue buffers Leon Romanovsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260917120459.13286-1-chengyou@linux.alibaba.com \
--to=chengyou@linux.alibaba.com \
--cc=KaiShen@linux.alibaba.com \
--cc=jgg@ziepe.ca \
--cc=leon@kernel.org \
--cc=linux-rdma@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox