linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/2] io_uring: unaccount scq mem properly
@ 2019-04-27 11:38 Shenghui Wang
  2019-04-27 11:38 ` [PATCH 2/2] io_uring: free ctx->sq_ring if array size overflow detected in io_allocate_scq_urings Shenghui Wang
  0 siblings, 1 reply; 2+ messages in thread
From: Shenghui Wang @ 2019-04-27 11:38 UTC (permalink / raw)
  To: axboe, viro, linux-block, linux-fsdevel

io_allocate_scq_urings() may fail to allocate scq rings, and fail to
set ctx->sq_entries or ctx->cq_entries. In io_ring_ctx_free(), the code
'
    if (ctx->account_mem)
        io_unaccount_mem(ctx->user,
             ring_pages(ctx->sq_entries, ctx->cq_entries));
'
may not unaccount properly.

E.g, in io_uring_create(), we have ctx allocated with zero filled.
Later, io_allocate_scq_urings() fails, and we may still have
0-valued ctx->sq_entries & ctx->cq_entries. Then the unaccount code
cannot unaccout what io_account_mem() has changed.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
---
 fs/io_uring.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index f65f85d89217..9d382ac27e63 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2741,7 +2741,6 @@ static int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	sq_ring->ring_mask = p->sq_entries - 1;
 	sq_ring->ring_entries = p->sq_entries;
 	ctx->sq_mask = sq_ring->ring_mask;
-	ctx->sq_entries = sq_ring->ring_entries;
 
 	size = array_size(sizeof(struct io_uring_sqe), p->sq_entries);
 	if (size == SIZE_MAX)
@@ -2764,7 +2763,6 @@ static int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	cq_ring->ring_mask = p->cq_entries - 1;
 	cq_ring->ring_entries = p->cq_entries;
 	ctx->cq_mask = cq_ring->ring_mask;
-	ctx->cq_entries = cq_ring->ring_entries;
 	return 0;
 }
 
@@ -2854,6 +2852,8 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p)
 	ctx->compat = in_compat_syscall();
 	ctx->account_mem = account_mem;
 	ctx->user = user;
+	ctx->sq_entries = p->sq_entries;
+	ctx->cq_entries = p->cq_entries;
 
 	ret = io_allocate_scq_urings(ctx, p);
 	if (ret)
-- 
2.20.1





^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [PATCH 2/2] io_uring: free ctx->sq_ring if array size overflow detected in io_allocate_scq_urings
  2019-04-27 11:38 [PATCH 1/2] io_uring: unaccount scq mem properly Shenghui Wang
@ 2019-04-27 11:38 ` Shenghui Wang
  0 siblings, 0 replies; 2+ messages in thread
From: Shenghui Wang @ 2019-04-27 11:38 UTC (permalink / raw)
  To: axboe, viro, linux-block, linux-fsdevel

Free ctx->sq_ring before return -EOVERFLOW.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
---
 fs/io_uring.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 9d382ac27e63..6e1d22cbb029 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2743,8 +2743,10 @@ static int io_allocate_scq_urings(struct io_ring_ctx *ctx,
 	ctx->sq_mask = sq_ring->ring_mask;
 
 	size = array_size(sizeof(struct io_uring_sqe), p->sq_entries);
-	if (size == SIZE_MAX)
+	if (size == SIZE_MAX) {
+		io_mem_free(ctx->sq_ring);
 		return -EOVERFLOW;
+	}
 
 	ctx->sq_sqes = io_mem_alloc(size);
 	if (!ctx->sq_sqes) {
-- 
2.20.1




^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2019-04-27 11:38 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-04-27 11:38 [PATCH 1/2] io_uring: unaccount scq mem properly Shenghui Wang
2019-04-27 11:38 ` [PATCH 2/2] io_uring: free ctx->sq_ring if array size overflow detected in io_allocate_scq_urings Shenghui Wang

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).