DORMANTno replies

[PATCH] rapidio: potential overflow in riocm_ch_send()

From: Dan Carpenter <hidden>
Date: 2021-06-01 09:15:39
Also in: lkml
Subsystem: rapidio subsystem, the rest · Maintainers: Matt Porter, Alexandre Bounine, Linus Torvalds

The "buf" variable has "len" bytes, and the size is controlled by the
user in cm_chan_msg_send().  If the length is fewer than sizeof(*hdr)
then it could lead to memory corruption.

Signed-off-by: Dan Carpenter <redacted>
---
Strictly speaking the last two bytes of length are reserved and not
written to but it's simpler and better to check "< sizeof(*hdr)" instead
of "< sizeof(*hdr) - 2".  This is better for future proofing.

 drivers/rapidio/rio_cm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/rapidio/rio_cm.c b/drivers/rapidio/rio_cm.c
index db4c265287ae..5c332b9867e1 100644
--- a/drivers/rapidio/rio_cm.c
+++ b/drivers/rapidio/rio_cm.c
@@ -784,7 +784,8 @@ static int riocm_ch_send(u16 ch_id, void *buf, int len)
 	struct rio_ch_chan_hdr *hdr;
 	int ret;
 
-	if (buf == NULL || ch_id == 0 || len == 0 || len > RIO_MAX_MSG_SIZE)
+	if (buf == NULL || ch_id == 0 ||
+	    len < sizeof(*hdr) || len > RIO_MAX_MSG_SIZE)
 		return -EINVAL;
 
 	ch = riocm_get_channel(ch_id);
-- 
2.30.2
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help