[RFC 1/7] io_uring: add io_mapped_ubuf caches

[Pavel Begunkov <asml.silence@xxxxxxxxx>]

We'll be allocating lots of io_mapped_ubuf shortly, add caches

Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
---
 include/linux/io_uring_types.h |  2 ++
 io_uring/io_uring.c            |  9 ++++++++
 io_uring/rsrc.c                | 39 ++++++++++++++++++++++++++++++++--
 io_uring/rsrc.h                |  6 +++++-
 4 files changed, 53 insertions(+), 3 deletions(-)

diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index 1b2a20a42413..3d103a00264c 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -336,6 +336,8 @@ struct io_ring_ctx {
 	struct wait_queue_head		rsrc_quiesce_wq;
 	unsigned			rsrc_quiesce;
 
+	struct io_alloc_cache		reg_buf_cache;
+
 	struct list_head		io_buffers_pages;
 
 	#if defined(CONFIG_UNIX)
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 3d43df8f1e4e..fdd62dbfd0ba 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -312,6 +312,8 @@ static __cold struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
 	INIT_LIST_HEAD(&ctx->io_buffers_cache);
 	io_alloc_cache_init(&ctx->rsrc_node_cache, IO_NODE_ALLOC_CACHE_MAX,
 			    sizeof(struct io_rsrc_node));
+	io_alloc_cache_init(&ctx->reg_buf_cache, IO_NODE_ALLOC_CACHE_MAX,
+			    sizeof(struct io_async_msghdr));
 	io_alloc_cache_init(&ctx->apoll_cache, IO_ALLOC_CACHE_MAX,
 			    sizeof(struct async_poll));
 	io_alloc_cache_init(&ctx->netmsg_cache, IO_ALLOC_CACHE_MAX,
@@ -2827,6 +2829,11 @@ static void io_rsrc_node_cache_free(struct io_cache_entry *entry)
 	kfree(container_of(entry, struct io_rsrc_node, cache));
 }
 
+static void io_reg_buf_cache_free(struct io_cache_entry *entry)
+{
+	kvfree(container_of(entry, struct io_mapped_ubuf, cache));
+}
+
 static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx)
 {
 	io_sq_thread_finish(ctx);
@@ -2865,6 +2872,8 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx)
 	WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list));
 
 	io_alloc_cache_free(&ctx->rsrc_node_cache, io_rsrc_node_cache_free);
+	io_alloc_cache_free(&ctx->reg_buf_cache, io_reg_buf_cache_free);
+
 	if (ctx->mm_account) {
 		mmdrop(ctx->mm_account);
 		ctx->mm_account = NULL;
diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c
index ddee7adb4006..fef94f8d788d 100644
--- a/io_uring/rsrc.c
+++ b/io_uring/rsrc.c
@@ -33,6 +33,8 @@ static int io_sqe_buffer_register(struct io_ring_ctx *ctx, struct iovec *iov,
 #define IORING_MAX_FIXED_FILES	(1U << 20)
 #define IORING_MAX_REG_BUFFERS	(1U << 14)
 
+#define IO_BUF_CACHE_MAX_BVECS	64
+
 int __io_account_mem(struct user_struct *user, unsigned long nr_pages)
 {
 	unsigned long page_limit, cur_pages, new_pages;
@@ -78,6 +80,39 @@ static int io_account_mem(struct io_ring_ctx *ctx, unsigned long nr_pages)
 	return 0;
 }
 
+static void io_put_reg_buf(struct io_ring_ctx *ctx, struct io_mapped_ubuf *imu)
+{
+	lockdep_assert_held(&ctx->uring_lock);
+
+	if ((imu->max_bvecs != IO_BUF_CACHE_MAX_BVECS) ||
+	    !io_alloc_cache_put(&ctx->reg_buf_cache, &imu->cache))
+		kvfree(imu);
+}
+
+static struct io_mapped_ubuf *io_alloc_reg_buf(struct io_ring_ctx *ctx,
+					       int nr_bvecs)
+{
+	struct io_cache_entry *entry;
+	struct io_mapped_ubuf *imu;
+
+	lockdep_assert_held(&ctx->uring_lock);
+
+	if (nr_bvecs > IO_BUF_CACHE_MAX_BVECS) {
+do_alloc:
+		imu = kvmalloc(struct_size(imu, bvec, nr_bvecs), GFP_KERNEL);
+		if (!imu)
+			return NULL;
+	} else {
+		nr_bvecs = IO_BUF_CACHE_MAX_BVECS;
+		entry = io_alloc_cache_get(&ctx->reg_buf_cache);
+		if (!entry)
+			goto do_alloc;
+		imu = container_of(entry, struct io_mapped_ubuf, cache);
+	}
+	imu->max_bvecs = nr_bvecs;
+	return imu;
+}
+
 static int io_copy_iov(struct io_ring_ctx *ctx, struct iovec *dst,
 		       void __user *arg, unsigned index)
 {
@@ -137,7 +172,7 @@ static void io_buffer_unmap(struct io_ring_ctx *ctx, struct io_mapped_ubuf **slo
 			unpin_user_page(imu->bvec[i].bv_page);
 		if (imu->acct_pages)
 			io_unaccount_mem(ctx, imu->acct_pages);
-		kvfree(imu);
+		io_put_reg_buf(ctx, imu);
 	}
 	*slot = NULL;
 }
@@ -1134,7 +1169,7 @@ static int io_sqe_buffer_register(struct io_ring_ctx *ctx, struct iovec *iov,
 		}
 	}
 
-	imu = kvmalloc(struct_size(imu, bvec, nr_pages), GFP_KERNEL);
+	imu = io_alloc_reg_buf(ctx, nr_pages);
 	if (!imu)
 		goto done;
 
diff --git a/io_uring/rsrc.h b/io_uring/rsrc.h
index 0a8a95e9b99e..f34de451a79a 100644
--- a/io_uring/rsrc.h
+++ b/io_uring/rsrc.h
@@ -50,9 +50,13 @@ struct io_rsrc_node {
 };
 
 struct io_mapped_ubuf {
-	u64		ubuf;
+	union {
+		struct io_cache_entry		cache;
+		u64				ubuf;
+	};
 	u64		ubuf_end;
 	unsigned int	nr_bvecs;
+	unsigned int	max_bvecs;
 	unsigned long	acct_pages;
 	struct bio_vec	bvec[];
 };
-- 
2.40.0