diff mbox series

block: Fix potential page reference leak in __bio_iov_append_get_pages()

Message ID 20200905094006.1925-1-linmiaohe@huawei.com (mailing list archive)
State New, archived
Headers show
Series block: Fix potential page reference leak in __bio_iov_append_get_pages() | expand

Commit Message

Miaohe Lin Sept. 5, 2020, 9:40 a.m. UTC
When bio_add_hw_page() failed, we left page reference still held in pages.
Release these references and advance the iov_iter according to what we
have done successfully yet.

Fixes: 0512a75b98f8 ("block: Introduce REQ_OP_ZONE_APPEND")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 block/bio.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

Comments

Johannes Thumshirn Sept. 9, 2020, 7:09 a.m. UTC | #1
On 05/09/2020 11:41, Miaohe Lin wrote:
> When bio_add_hw_page() failed, we left page reference still held in pages.

I'd add "from iov_iter_get_pages()" to the above sentence.

Otherwise
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
diff mbox series

Patch

diff --git a/block/bio.c b/block/bio.c
index e113073958cb..a323a5446221 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1080,7 +1080,7 @@  static int __bio_iov_append_get_pages(struct bio *bio, struct iov_iter *iter)
 		len = min_t(size_t, PAGE_SIZE - offset, left);
 		if (bio_add_hw_page(q, bio, page, len, offset,
 				max_append_sectors, &same_page) != len)
-			return -EINVAL;
+			goto put_pages;
 		if (same_page)
 			put_page(page);
 		offset = 0;
@@ -1088,6 +1088,16 @@  static int __bio_iov_append_get_pages(struct bio *bio, struct iov_iter *iter)
 
 	iov_iter_advance(iter, size);
 	return 0;
+put_pages:
+	iov_iter_advance(iter, size - left);
+	for (; left > 0; left -= len, i++) {
+		struct page *page = pages[i];
+
+		len = min_t(size_t, PAGE_SIZE - offset, left);
+		put_page(page);
+		offset = 0;
+	}
+	return -EINVAL;
 }
 
 /**