Linux RDMA and InfiniBand development
 help / color / mirror / Atom feed
From: Michael Gur <michaelgur@nvidia.com>
To: jgg@ziepe.ca, leon@kernel.org, linux-rdma@vger.kernel.org
Cc: Edward Srouji <edwards@nvidia.com>,
	Yishai Hadas <yishaih@nvidia.com>,
	Patrisious Haddad <phaddad@nvidia.com>,
	Michael Guralnik <michaelgur@nvidia.com>
Subject: [PATCH rdma-next 8/9] RDMA/core: Add ib_frmr_pool_drop for unrecoverable handles
Date: Wed, 10 Jun 2026 03:01:44 +0300	[thread overview]
Message-ID: <20260610000145.820592-9-michaelgur@nvidia.com> (raw)
In-Reply-To: <20260610000145.820592-1-michaelgur@nvidia.com>

From: Michael Guralnik <michaelgur@nvidia.com>

A driver that has popped a handle from an FRMR pool can hit failures
that leave the handle in a state where it can't safely be returned
for reuse. The driver destroys the handle itself, but the pool has
no way to learn about it, so the in_use counter drifts upward.

Add ib_frmr_pool_drop to balance the pool's accounting in this case.
Every pop is now balanced by exactly one push or drop.

Fixes: 36680ef7bceb ("RDMA/mlx5: Switch from MR cache to FRMR pools")
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
---
 drivers/infiniband/core/frmr_pools.c | 15 +++++++++++++++
 include/rdma/frmr_pools.h            |  1 +
 2 files changed, 16 insertions(+)

diff --git a/drivers/infiniband/core/frmr_pools.c b/drivers/infiniband/core/frmr_pools.c
index e214a8273df8..ce8ae4305b9c 100644
--- a/drivers/infiniband/core/frmr_pools.c
+++ b/drivers/infiniband/core/frmr_pools.c
@@ -578,3 +578,18 @@ void ib_frmr_pool_push(struct ib_device *device, struct ib_mr *mr)
 
 }
 EXPORT_SYMBOL(ib_frmr_pool_push);
+
+/*
+ * Drop a handle previously popped from the pool without returning it for
+ * reuse. The caller is responsible for destroying the underlying hardware
+ * resource.
+ */
+void ib_frmr_pool_drop(struct ib_mr *mr)
+{
+	struct ib_frmr_pool *pool = mr->frmr.pool;
+
+	spin_lock(&pool->lock);
+	pool->in_use--;
+	spin_unlock(&pool->lock);
+}
+EXPORT_SYMBOL(ib_frmr_pool_drop);
diff --git a/include/rdma/frmr_pools.h b/include/rdma/frmr_pools.h
index 5b57bafa3636..aed4d69d3841 100644
--- a/include/rdma/frmr_pools.h
+++ b/include/rdma/frmr_pools.h
@@ -35,5 +35,6 @@ int ib_frmr_pools_init(struct ib_device *device,
 void ib_frmr_pools_cleanup(struct ib_device *device);
 int ib_frmr_pool_pop(struct ib_device *device, struct ib_mr *mr);
 void ib_frmr_pool_push(struct ib_device *device, struct ib_mr *mr);
+void ib_frmr_pool_drop(struct ib_mr *mr);
 
 #endif /* FRMR_POOLS_H */
-- 
2.52.0


  parent reply	other threads:[~2026-06-10  0:03 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-10  0:01 [PATCH rdma-next 0/9] FRMR pools fixes Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 1/9] RDMA/mlx5: Fix mkey creation error flow rollback Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 2/9] RDMA/mlx5: Fix TPH extraction in FRMR pool key Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 3/9] RDMA/core: Fix skipped usage for driver built FRMR key Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 4/9] RDMA/core: Fix FRMR aging push to queue error flow Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 5/9] RDMA/core: Fix FRMR set pinned push error path Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 6/9] RDMA/core: Avoid NULL dereference on FRMR bad usage Michael Gur
2026-06-10  0:01 ` [PATCH rdma-next 7/9] RDMA/core: Fix FRMR handle leak on push failure Michael Gur
2026-06-10  0:01 ` Michael Gur [this message]
2026-06-10  0:01 ` [PATCH rdma-next 9/9] RDMA/mlx5: Drop FRMR pool handle on UMR revoke failure Michael Gur

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=20260610000145.820592-9-michaelgur@nvidia.com \
    --to=michaelgur@nvidia.com \
    --cc=edwards@nvidia.com \
    --cc=jgg@ziepe.ca \
    --cc=leon@kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    --cc=phaddad@nvidia.com \
    --cc=yishaih@nvidia.com \
    /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