netdev - [PATCH bpf-next v2] xsk: fix xsk_build_skb() error: 'skb' dereferencing possible ERR

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date: Tue, 15 Aug 2023 20:33:25 +0530
From: Tirthendu Sarkar <tirthendu.sarkar@...el.com>
To: bpf@...r.kernel.org
Cc: netdev@...r.kernel.org,
	bjorn@...nel.org,
	magnus.karlsson@...el.com,
	maciej.fijalkowski@...el.com,
	jonathan.lemon@...il.com,
	davem@...emloft.net,
	kuba@...nel.org,
	pabeni@...hat.com,
	ast@...nel.org,
	daniel@...earbox.net,
	martin.lau@...ux.dev,
	dan.carpenter@...aro.org,
	sdf@...gle.com
Subject: [PATCH bpf-next v2] xsk: fix xsk_build_skb() error: 'skb' dereferencing possible ERR_PTR()

xsk_build_skb_zerocopy() may return an error other than -EAGAIN and this
is received as skb and used later in xsk_set_destructor_arg() and
xsk_drop_skb() which must operate on a valid skb.

Set -EOVERFLOW as error when MAX_SKB_FRAGS are exceeded and packet needs
to be dropped and use this to distinguish against all other error cases
where allocation needs to be retried.

Signed-off-by: Tirthendu Sarkar <tirthendu.sarkar@...el.com>
Reported-by: kernel test robot <lkp@...el.com>
Reported-by: Dan Carpenter <dan.carpenter@...aro.org>
Closes: https://lore.kernel.org/r/202307210434.OjgqFcbB-lkp@intel.com/
Fixes: cf24f5a5feea ("xsk: add support for AF_XDP multi-buffer on Tx path")

Changelog:
	v1 -> v2:
	- Removed err as a parameter to xsk_build_skb_zerocopy()
	[Stanislav Fomichev]
	- use explicit error to distinguish packet drop vs retry
---
 net/xdp/xsk.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
index fcfc8472f73d..55f8b9b0e06d 100644
--- a/net/xdp/xsk.c
+++ b/net/xdp/xsk.c
@@ -602,7 +602,7 @@ static struct sk_buff *xsk_build_skb_zerocopy(struct xdp_sock *xs,
 
 	for (copied = 0, i = skb_shinfo(skb)->nr_frags; copied < len; i++) {
 		if (unlikely(i >= MAX_SKB_FRAGS))
-			return ERR_PTR(-EFAULT);
+			return ERR_PTR(-EOVERFLOW);
 
 		page = pool->umem->pgs[addr >> PAGE_SHIFT];
 		get_page(page);
@@ -655,15 +655,17 @@ static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
 			skb_put(skb, len);
 
 			err = skb_store_bits(skb, 0, buffer, len);
-			if (unlikely(err))
+			if (unlikely(err)) {
+				kfree_skb(skb);
 				goto free_err;
+			}
 		} else {
 			int nr_frags = skb_shinfo(skb)->nr_frags;
 			struct page *page;
 			u8 *vaddr;
 
 			if (unlikely(nr_frags == (MAX_SKB_FRAGS - 1) && xp_mb_desc(desc))) {
-				err = -EFAULT;
+				err = -EOVERFLOW;
 				goto free_err;
 			}
 
@@ -690,12 +692,14 @@ static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
 	return skb;
 
 free_err:
-	if (err == -EAGAIN) {
-		xsk_cq_cancel_locked(xs, 1);
-	} else {
-		xsk_set_destructor_arg(skb);
-		xsk_drop_skb(skb);
+	if (err == -EOVERFLOW) {
+		/* Drop the packet */
+		xsk_set_destructor_arg(xs->skb);
+		xsk_drop_skb(xs->skb);
 		xskq_cons_release(xs->tx);
+	} else {
+		/* Let application retry */
+		xsk_cq_cancel_locked(xs, 1);
 	}
 
 	return ERR_PTR(err);
@@ -738,7 +742,7 @@ static int __xsk_generic_xmit(struct sock *sk)
 		skb = xsk_build_skb(xs, &desc);
 		if (IS_ERR(skb)) {
 			err = PTR_ERR(skb);
-			if (err == -EAGAIN)
+			if (err != -EOVERFLOW)
 				goto out;
 			err = 0;
 			continue;
-- 
2.34.1