From: Leon Romanovsky <leon@kernel.org>
To: Doug Ledford <dledford@redhat.com>, Jason Gunthorpe <jgg@mellanox.com>
Cc: Leon Romanovsky <leonro@mellanox.com>,
RDMA mailing list <linux-rdma@vger.kernel.org>,
Guy Levi <guyle@mellanox.com>, Moni Shoua <monis@mellanox.com>
Subject: [PATCH rdma-next 02/12] RDMA/odp: Iterate over the whole rbtree directly
Date: Mon, 19 Aug 2019 14:17:00 +0300 [thread overview]
Message-ID: <20190819111710.18440-3-leon@kernel.org> (raw)
In-Reply-To: <20190819111710.18440-1-leon@kernel.org>
From: Jason Gunthorpe <jgg@mellanox.com>
Instead of intersecting a full interval, just iterate over every element
directly. This is faster and clearer.
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
drivers/infiniband/core/umem_odp.c | 51 ++++++++++++++++--------------
drivers/infiniband/hw/mlx5/odp.c | 41 +++++++++++-------------
2 files changed, 47 insertions(+), 45 deletions(-)
diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/umem_odp.c
index 8358eb8e3a26..b9bebef00a33 100644
--- a/drivers/infiniband/core/umem_odp.c
+++ b/drivers/infiniband/core/umem_odp.c
@@ -72,35 +72,41 @@ static void ib_umem_notifier_end_account(struct ib_umem_odp *umem_odp)
mutex_unlock(&umem_odp->umem_mutex);
}
-static int ib_umem_notifier_release_trampoline(struct ib_umem_odp *umem_odp,
- u64 start, u64 end, void *cookie)
-{
- /*
- * Increase the number of notifiers running, to
- * prevent any further fault handling on this MR.
- */
- ib_umem_notifier_start_account(umem_odp);
- umem_odp->dying = 1;
- /* Make sure that the fact the umem is dying is out before we release
- * all pending page faults. */
- smp_wmb();
- complete_all(&umem_odp->notifier_completion);
- umem_odp->umem.context->invalidate_range(
- umem_odp, ib_umem_start(umem_odp), ib_umem_end(umem_odp));
- return 0;
-}
-
static void ib_umem_notifier_release(struct mmu_notifier *mn,
struct mm_struct *mm)
{
struct ib_ucontext_per_mm *per_mm =
container_of(mn, struct ib_ucontext_per_mm, mn);
+ struct rb_node *node;
down_read(&per_mm->umem_rwsem);
- if (per_mm->active)
- rbt_ib_umem_for_each_in_range(
- &per_mm->umem_tree, 0, ULLONG_MAX,
- ib_umem_notifier_release_trampoline, true, NULL);
+ if (!per_mm->active)
+ goto out;
+
+ for (node = rb_first_cached(&per_mm->umem_tree); node;
+ node = rb_next(node)) {
+ struct ib_umem_odp *umem_odp =
+ rb_entry(node, struct ib_umem_odp, interval_tree.rb);
+
+ /*
+ * Increase the number of notifiers running, to prevent any
+ * further fault handling on this MR.
+ */
+ ib_umem_notifier_start_account(umem_odp);
+
+ umem_odp->dying = 1;
+ /*
+ * Make sure that the fact the umem is dying is out before we
+ * release all pending page faults.
+ */
+ smp_wmb();
+ complete_all(&umem_odp->notifier_completion);
+ umem_odp->umem.context->invalidate_range(
+ umem_odp, ib_umem_start(umem_odp),
+ ib_umem_end(umem_odp));
+ }
+
+out:
up_read(&per_mm->umem_rwsem);
}
@@ -760,4 +766,3 @@ int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root,
return ret_val;
}
-EXPORT_SYMBOL(rbt_ib_umem_for_each_in_range);
diff --git a/drivers/infiniband/hw/mlx5/odp.c b/drivers/infiniband/hw/mlx5/odp.c
index b0c5de39d186..3922fced41ec 100644
--- a/drivers/infiniband/hw/mlx5/odp.c
+++ b/drivers/infiniband/hw/mlx5/odp.c
@@ -539,34 +539,31 @@ struct mlx5_ib_mr *mlx5_ib_alloc_implicit_mr(struct mlx5_ib_pd *pd,
return imr;
}
-static int mr_leaf_free(struct ib_umem_odp *umem_odp, u64 start, u64 end,
- void *cookie)
+void mlx5_ib_free_implicit_mr(struct mlx5_ib_mr *imr)
{
- struct mlx5_ib_mr *mr = umem_odp->private, *imr = cookie;
-
- if (mr->parent != imr)
- return 0;
-
- ib_umem_odp_unmap_dma_pages(umem_odp, ib_umem_start(umem_odp),
- ib_umem_end(umem_odp));
+ struct ib_ucontext_per_mm *per_mm = mr_to_per_mm(imr);
+ struct rb_node *node;
- if (umem_odp->dying)
- return 0;
+ down_read(&per_mm->umem_rwsem);
+ for (node = rb_first_cached(&per_mm->umem_tree); node;
+ node = rb_next(node)) {
+ struct ib_umem_odp *umem_odp =
+ rb_entry(node, struct ib_umem_odp, interval_tree.rb);
+ struct mlx5_ib_mr *mr = umem_odp->private;
- WRITE_ONCE(umem_odp->dying, 1);
- atomic_inc(&imr->num_leaf_free);
- schedule_work(&umem_odp->work);
+ if (mr->parent != imr)
+ continue;
- return 0;
-}
+ ib_umem_odp_unmap_dma_pages(umem_odp, ib_umem_start(umem_odp),
+ ib_umem_end(umem_odp));
-void mlx5_ib_free_implicit_mr(struct mlx5_ib_mr *imr)
-{
- struct ib_ucontext_per_mm *per_mm = mr_to_per_mm(imr);
+ if (umem_odp->dying)
+ continue;
- down_read(&per_mm->umem_rwsem);
- rbt_ib_umem_for_each_in_range(&per_mm->umem_tree, 0, ULLONG_MAX,
- mr_leaf_free, true, imr);
+ WRITE_ONCE(umem_odp->dying, 1);
+ atomic_inc(&imr->num_leaf_free);
+ schedule_work(&umem_odp->work);
+ }
up_read(&per_mm->umem_rwsem);
wait_event(imr->q_leaf_free, !atomic_read(&imr->num_leaf_free));
--
2.20.1
next prev parent reply other threads:[~2019-08-19 11:17 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-19 11:16 [PATCH rdma-next 00/12] Improvements for ODP Leon Romanovsky
2019-08-19 11:16 ` [PATCH rdma-next 01/12] RDMA/odp: Use the common interval tree library instead of generic Leon Romanovsky
2019-08-19 11:17 ` Leon Romanovsky [this message]
2019-08-21 17:15 ` [PATCH rdma-next 02/12] RDMA/odp: Iterate over the whole rbtree directly Jason Gunthorpe
2019-08-21 17:27 ` Leon Romanovsky
2019-08-21 17:35 ` Jason Gunthorpe
2019-08-21 17:47 ` Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 03/12] RDMA/odp: Make it clearer when a umem is an implicit ODP umem Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 04/12] RMDA/odp: Consolidate umem_odp initialization Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 05/12] RDMA/odp: Make the three ways to create a umem_odp clear Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 06/12] RDMA/odp: Split creating a umem_odp from ib_umem_get Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 07/12] RDMA/odp: Provide ib_umem_odp_release() to undo the allocs Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 08/12] RDMA/odp: Check for overflow when computing the umem_odp end Leon Romanovsky
2019-08-26 16:42 ` Nathan Chancellor
2019-08-26 16:55 ` Jason Gunthorpe
2019-08-27 19:25 ` Nathan Chancellor
2019-08-19 11:17 ` [PATCH rdma-next 09/12] RDMA/odp: Use kvcalloc for the dma_list and page_list Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 10/12] RDMA/core: Make invalidate_range a device operation Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 11/12] RDMA/mlx5: Use ib_umem_start instead of umem.address Leon Romanovsky
2019-08-19 11:17 ` [PATCH rdma-next 12/12] RDMA/mlx5: Use odp instead of mr->umem in pagefault_mr Leon Romanovsky
2019-08-21 16:42 ` [PATCH rdma-next 00/12] Improvements for ODP Jason Gunthorpe
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=20190819111710.18440-3-leon@kernel.org \
--to=leon@kernel.org \
--cc=dledford@redhat.com \
--cc=guyle@mellanox.com \
--cc=jgg@mellanox.com \
--cc=leonro@mellanox.com \
--cc=linux-rdma@vger.kernel.org \
--cc=monis@mellanox.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;
as well as URLs for NNTP newsgroup(s).