netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers
@ 2024-10-22  6:56 Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 1/4] igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog() Yue Haibing
                   ` (4 more replies)
  0 siblings, 5 replies; 11+ messages in thread
From: Yue Haibing @ 2024-10-22  6:56 UTC (permalink / raw)
  To: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, horms, jacob.e.keller,
	sven.auhagen, alexander.h.duyck
  Cc: intel-wired-lan, netdev, linux-kernel, bpf, yuehaibing

Fixing sparse error in xdp run code by introducing new variable xdp_res
instead of overloading this into the skb pointer as i40e drivers done
in commit 12738ac4754e ("i40e: Fix sparse errors in i40e_txrx.c") and
commit ae4393dfd472 ("i40e: fix broken XDP support").

v3: Fix uninitialized 'xdp_res' in patch 3 and 4 which Reported-by
    kernel test robot
v2: Fix this as i40e drivers done instead of return NULL in xdp run code

Yue Haibing (4):
  igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog()
  igb: Fix passing 0 to ERR_PTR in igb_run_xdp()
  ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp()
  ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp()

 drivers/net/ethernet/intel/igb/igb_main.c     | 22 +++++++-----------
 drivers/net/ethernet/intel/igc/igc_main.c     | 20 ++++++----------
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 23 ++++++++-----------
 .../net/ethernet/intel/ixgbevf/ixgbevf_main.c | 23 ++++++++-----------
 4 files changed, 34 insertions(+), 54 deletions(-)

-- 
2.34.1


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH v3 net 1/4] igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog()
  2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
@ 2024-10-22  6:56 ` Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 2/4] igb: Fix passing 0 to ERR_PTR in igb_run_xdp() Yue Haibing
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 11+ messages in thread
From: Yue Haibing @ 2024-10-22  6:56 UTC (permalink / raw)
  To: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, horms, jacob.e.keller,
	sven.auhagen, alexander.h.duyck
  Cc: intel-wired-lan, netdev, linux-kernel, bpf, yuehaibing

igc_xdp_run_prog() converts customed xdp action to a negative error code
with the sk_buff pointer type which be checked with IS_ERR in
igc_clean_rx_irq(). Remove this error pointer handing instead use plain
int return value to fix this smatch warnings:

drivers/net/ethernet/intel/igc/igc_main.c:2533
 igc_xdp_run_prog() warn: passing zero to 'ERR_PTR'

Fixes: 26575105d6ed ("igc: Add initial XDP support")
Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
---
 drivers/net/ethernet/intel/igc/igc_main.c | 20 +++++++-------------
 1 file changed, 7 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
index 6e70bca15db1..5e44c2546a12 100644
--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -2123,10 +2123,6 @@ static bool igc_cleanup_headers(struct igc_ring *rx_ring,
 				union igc_adv_rx_desc *rx_desc,
 				struct sk_buff *skb)
 {
-	/* XDP packets use error pointer so abort at this point */
-	if (IS_ERR(skb))
-		return true;
-
 	if (unlikely(igc_test_staterr(rx_desc, IGC_RXDEXT_STATERR_RXE))) {
 		struct net_device *netdev = rx_ring->netdev;
 
@@ -2515,8 +2511,7 @@ static int __igc_xdp_run_prog(struct igc_adapter *adapter,
 	}
 }
 
-static struct sk_buff *igc_xdp_run_prog(struct igc_adapter *adapter,
-					struct xdp_buff *xdp)
+static int igc_xdp_run_prog(struct igc_adapter *adapter, struct xdp_buff *xdp)
 {
 	struct bpf_prog *prog;
 	int res;
@@ -2530,7 +2525,7 @@ static struct sk_buff *igc_xdp_run_prog(struct igc_adapter *adapter,
 	res = __igc_xdp_run_prog(adapter, prog, xdp);
 
 out:
-	return ERR_PTR(-res);
+	return res;
 }
 
 /* This function assumes __netif_tx_lock is held by the caller. */
@@ -2585,6 +2580,7 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
 	struct sk_buff *skb = rx_ring->skb;
 	u16 cleaned_count = igc_desc_unused(rx_ring);
 	int xdp_status = 0, rx_buffer_pgcnt;
+	int xdp_res = 0;
 
 	while (likely(total_packets < budget)) {
 		struct igc_xdp_buff ctx = { .rx_ts = NULL };
@@ -2630,12 +2626,10 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
 			xdp_buff_clear_frags_flag(&ctx.xdp);
 			ctx.rx_desc = rx_desc;
 
-			skb = igc_xdp_run_prog(adapter, &ctx.xdp);
+			xdp_res = igc_xdp_run_prog(adapter, &ctx.xdp);
 		}
 
-		if (IS_ERR(skb)) {
-			unsigned int xdp_res = -PTR_ERR(skb);
-
+		if (xdp_res) {
 			switch (xdp_res) {
 			case IGC_XDP_CONSUMED:
 				rx_buffer->pagecnt_bias++;
@@ -2657,7 +2651,7 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
 			skb = igc_construct_skb(rx_ring, rx_buffer, &ctx);
 
 		/* exit if we failed to retrieve a buffer */
-		if (!skb) {
+		if (!xdp_res && !skb) {
 			rx_ring->rx_stats.alloc_failed++;
 			rx_buffer->pagecnt_bias++;
 			set_bit(IGC_RING_FLAG_RX_ALLOC_FAILED, &rx_ring->flags);
@@ -2672,7 +2666,7 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
 			continue;
 
 		/* verify the packet layout is correct */
-		if (igc_cleanup_headers(rx_ring, rx_desc, skb)) {
+		if (xdp_res || igc_cleanup_headers(rx_ring, rx_desc, skb)) {
 			skb = NULL;
 			continue;
 		}
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH v3 net 2/4] igb: Fix passing 0 to ERR_PTR in igb_run_xdp()
  2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 1/4] igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog() Yue Haibing
@ 2024-10-22  6:56 ` Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp() Yue Haibing
                   ` (2 subsequent siblings)
  4 siblings, 0 replies; 11+ messages in thread
From: Yue Haibing @ 2024-10-22  6:56 UTC (permalink / raw)
  To: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, horms, jacob.e.keller,
	sven.auhagen, alexander.h.duyck
  Cc: intel-wired-lan, netdev, linux-kernel, bpf, yuehaibing

igb_run_xdp() converts customed xdp action to a negative error code
with the sk_buff pointer type which be checked with IS_ERR in
igb_clean_rx_irq(). Remove this error pointer handing instead use plain
int return value.

Fixes: 9cbc948b5a20 ("igb: add XDP support")
Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
---
 drivers/net/ethernet/intel/igb/igb_main.c | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index f1d088168723..50c23dfcf304 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -8584,9 +8584,8 @@ static struct sk_buff *igb_build_skb(struct igb_ring *rx_ring,
 	return skb;
 }
 
-static struct sk_buff *igb_run_xdp(struct igb_adapter *adapter,
-				   struct igb_ring *rx_ring,
-				   struct xdp_buff *xdp)
+static int igb_run_xdp(struct igb_adapter *adapter, struct igb_ring *rx_ring,
+		       struct xdp_buff *xdp)
 {
 	int err, result = IGB_XDP_PASS;
 	struct bpf_prog *xdp_prog;
@@ -8626,7 +8625,7 @@ static struct sk_buff *igb_run_xdp(struct igb_adapter *adapter,
 		break;
 	}
 xdp_out:
-	return ERR_PTR(-result);
+	return result;
 }
 
 static unsigned int igb_rx_frame_truesize(struct igb_ring *rx_ring,
@@ -8752,10 +8751,6 @@ static bool igb_cleanup_headers(struct igb_ring *rx_ring,
 				union e1000_adv_rx_desc *rx_desc,
 				struct sk_buff *skb)
 {
-	/* XDP packets use error pointer so abort at this point */
-	if (IS_ERR(skb))
-		return true;
-
 	if (unlikely((igb_test_staterr(rx_desc,
 				       E1000_RXDEXT_ERR_FRAME_ERR_MASK)))) {
 		struct net_device *netdev = rx_ring->netdev;
@@ -8879,6 +8874,7 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget)
 	struct xdp_buff xdp;
 	u32 frame_sz = 0;
 	int rx_buf_pgcnt;
+	int xdp_res = 0;
 
 	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
 #if (PAGE_SIZE < 8192)
@@ -8936,12 +8932,10 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget)
 			/* At larger PAGE_SIZE, frame_sz depend on len size */
 			xdp.frame_sz = igb_rx_frame_truesize(rx_ring, size);
 #endif
-			skb = igb_run_xdp(adapter, rx_ring, &xdp);
+			xdp_res = igb_run_xdp(adapter, rx_ring, &xdp);
 		}
 
-		if (IS_ERR(skb)) {
-			unsigned int xdp_res = -PTR_ERR(skb);
-
+		if (xdp_res) {
 			if (xdp_res & (IGB_XDP_TX | IGB_XDP_REDIR)) {
 				xdp_xmit |= xdp_res;
 				igb_rx_buffer_flip(rx_ring, rx_buffer, size);
@@ -8960,7 +8954,7 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget)
 						&xdp, timestamp);
 
 		/* exit if we failed to retrieve a buffer */
-		if (!skb) {
+		if (!xdp_res && !skb) {
 			rx_ring->rx_stats.alloc_failed++;
 			rx_buffer->pagecnt_bias++;
 			break;
@@ -8974,7 +8968,7 @@ static int igb_clean_rx_irq(struct igb_q_vector *q_vector, const int budget)
 			continue;
 
 		/* verify the packet layout is correct */
-		if (igb_cleanup_headers(rx_ring, rx_desc, skb)) {
+		if (xdp_res || igb_cleanup_headers(rx_ring, rx_desc, skb)) {
 			skb = NULL;
 			continue;
 		}
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp()
  2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 1/4] igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog() Yue Haibing
  2024-10-22  6:56 ` [PATCH v3 net 2/4] igb: Fix passing 0 to ERR_PTR in igb_run_xdp() Yue Haibing
@ 2024-10-22  6:56 ` Yue Haibing
  2024-10-22 11:25   ` Maciej Fijalkowski
  2024-10-22  6:56 ` [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp() Yue Haibing
  2024-10-22  7:32 ` [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Simon Horman
  4 siblings, 1 reply; 11+ messages in thread
From: Yue Haibing @ 2024-10-22  6:56 UTC (permalink / raw)
  To: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, horms, jacob.e.keller,
	sven.auhagen, alexander.h.duyck
  Cc: intel-wired-lan, netdev, linux-kernel, bpf, yuehaibing

ixgbe_run_xdp() converts customed xdp action to a negative error code
with the sk_buff pointer type which be checked with IS_ERR in
ixgbe_clean_rx_irq(). Remove this error pointer handing instead use
plain int return value.

Fixes: 924708081629 ("ixgbe: add XDP support for pass and drop actions")
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 23 ++++++++-----------
 1 file changed, 9 insertions(+), 14 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 8b8404d8c946..78bf97ab0524 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1908,10 +1908,6 @@ bool ixgbe_cleanup_headers(struct ixgbe_ring *rx_ring,
 {
 	struct net_device *netdev = rx_ring->netdev;
 
-	/* XDP packets use error pointer so abort at this point */
-	if (IS_ERR(skb))
-		return true;
-
 	/* Verify netdev is present, and that packet does not have any
 	 * errors that would be unacceptable to the netdev.
 	 */
@@ -2219,9 +2215,9 @@ static struct sk_buff *ixgbe_build_skb(struct ixgbe_ring *rx_ring,
 	return skb;
 }
 
-static struct sk_buff *ixgbe_run_xdp(struct ixgbe_adapter *adapter,
-				     struct ixgbe_ring *rx_ring,
-				     struct xdp_buff *xdp)
+static int ixgbe_run_xdp(struct ixgbe_adapter *adapter,
+			 struct ixgbe_ring *rx_ring,
+			 struct xdp_buff *xdp)
 {
 	int err, result = IXGBE_XDP_PASS;
 	struct bpf_prog *xdp_prog;
@@ -2271,7 +2267,7 @@ static struct sk_buff *ixgbe_run_xdp(struct ixgbe_adapter *adapter,
 		break;
 	}
 xdp_out:
-	return ERR_PTR(-result);
+	return result;
 }
 
 static unsigned int ixgbe_rx_frame_truesize(struct ixgbe_ring *rx_ring,
@@ -2329,6 +2325,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 	unsigned int offset = rx_ring->rx_offset;
 	unsigned int xdp_xmit = 0;
 	struct xdp_buff xdp;
+	int xdp_res = 0;
 
 	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
 #if (PAGE_SIZE < 8192)
@@ -2374,12 +2371,10 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 			/* At larger PAGE_SIZE, frame_sz depend on len size */
 			xdp.frame_sz = ixgbe_rx_frame_truesize(rx_ring, size);
 #endif
-			skb = ixgbe_run_xdp(adapter, rx_ring, &xdp);
+			xdp_res = ixgbe_run_xdp(adapter, rx_ring, &xdp);
 		}
 
-		if (IS_ERR(skb)) {
-			unsigned int xdp_res = -PTR_ERR(skb);
-
+		if (xdp_res) {
 			if (xdp_res & (IXGBE_XDP_TX | IXGBE_XDP_REDIR)) {
 				xdp_xmit |= xdp_res;
 				ixgbe_rx_buffer_flip(rx_ring, rx_buffer, size);
@@ -2399,7 +2394,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 		}
 
 		/* exit if we failed to retrieve a buffer */
-		if (!skb) {
+		if (!xdp_res && !skb) {
 			rx_ring->rx_stats.alloc_rx_buff_failed++;
 			rx_buffer->pagecnt_bias++;
 			break;
@@ -2413,7 +2408,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 			continue;
 
 		/* verify the packet layout is correct */
-		if (ixgbe_cleanup_headers(rx_ring, rx_desc, skb))
+		if (xdp_res || ixgbe_cleanup_headers(rx_ring, rx_desc, skb))
 			continue;
 
 		/* probably a little skewed due to removing CRC */
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp()
  2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
                   ` (2 preceding siblings ...)
  2024-10-22  6:56 ` [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp() Yue Haibing
@ 2024-10-22  6:56 ` Yue Haibing
  2024-10-22 11:27   ` Maciej Fijalkowski
  2024-10-22  7:32 ` [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Simon Horman
  4 siblings, 1 reply; 11+ messages in thread
From: Yue Haibing @ 2024-10-22  6:56 UTC (permalink / raw)
  To: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, horms, jacob.e.keller,
	sven.auhagen, alexander.h.duyck
  Cc: intel-wired-lan, netdev, linux-kernel, bpf, yuehaibing

ixgbevf_run_xdp() converts customed xdp action to a negative error code
with the sk_buff pointer type which be checked with IS_ERR in
ixgbevf_clean_rx_irq(). Remove this error pointer handing instead use
plain int return value.

Fixes: c7aec59657b6 ("ixgbevf: Add XDP support for pass and drop actions")
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
---
 .../net/ethernet/intel/ixgbevf/ixgbevf_main.c | 23 ++++++++-----------
 1 file changed, 10 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 149911e3002a..183d2305d058 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -732,10 +732,6 @@ static bool ixgbevf_cleanup_headers(struct ixgbevf_ring *rx_ring,
 				    union ixgbe_adv_rx_desc *rx_desc,
 				    struct sk_buff *skb)
 {
-	/* XDP packets use error pointer so abort at this point */
-	if (IS_ERR(skb))
-		return true;
-
 	/* verify that the packet does not have any known errors */
 	if (unlikely(ixgbevf_test_staterr(rx_desc,
 					  IXGBE_RXDADV_ERR_FRAME_ERR_MASK))) {
@@ -1044,9 +1040,9 @@ static int ixgbevf_xmit_xdp_ring(struct ixgbevf_ring *ring,
 	return IXGBEVF_XDP_TX;
 }
 
-static struct sk_buff *ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
-				       struct ixgbevf_ring  *rx_ring,
-				       struct xdp_buff *xdp)
+static int ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
+			   struct ixgbevf_ring *rx_ring,
+			   struct xdp_buff *xdp)
 {
 	int result = IXGBEVF_XDP_PASS;
 	struct ixgbevf_ring *xdp_ring;
@@ -1080,7 +1076,7 @@ static struct sk_buff *ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
 		break;
 	}
 xdp_out:
-	return ERR_PTR(-result);
+	return result;
 }
 
 static unsigned int ixgbevf_rx_frame_truesize(struct ixgbevf_ring *rx_ring,
@@ -1122,6 +1118,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 	struct sk_buff *skb = rx_ring->skb;
 	bool xdp_xmit = false;
 	struct xdp_buff xdp;
+	int xdp_res = 0;
 
 	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
 #if (PAGE_SIZE < 8192)
@@ -1165,11 +1162,11 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 			/* At larger PAGE_SIZE, frame_sz depend on len size */
 			xdp.frame_sz = ixgbevf_rx_frame_truesize(rx_ring, size);
 #endif
-			skb = ixgbevf_run_xdp(adapter, rx_ring, &xdp);
+			xdp_res = ixgbevf_run_xdp(adapter, rx_ring, &xdp);
 		}
 
-		if (IS_ERR(skb)) {
-			if (PTR_ERR(skb) == -IXGBEVF_XDP_TX) {
+		if (xdp_res) {
+			if (xdp_res == IXGBEVF_XDP_TX) {
 				xdp_xmit = true;
 				ixgbevf_rx_buffer_flip(rx_ring, rx_buffer,
 						       size);
@@ -1189,7 +1186,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 		}
 
 		/* exit if we failed to retrieve a buffer */
-		if (!skb) {
+		if (!xdp_res && !skb) {
 			rx_ring->rx_stats.alloc_rx_buff_failed++;
 			rx_buffer->pagecnt_bias++;
 			break;
@@ -1203,7 +1200,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 			continue;
 
 		/* verify the packet layout is correct */
-		if (ixgbevf_cleanup_headers(rx_ring, rx_desc, skb)) {
+		if (xdp_res || ixgbevf_cleanup_headers(rx_ring, rx_desc, skb)) {
 			skb = NULL;
 			continue;
 		}
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers
  2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
                   ` (3 preceding siblings ...)
  2024-10-22  6:56 ` [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp() Yue Haibing
@ 2024-10-22  7:32 ` Simon Horman
  2024-10-22 19:17   ` Jacob Keller
  4 siblings, 1 reply; 11+ messages in thread
From: Simon Horman @ 2024-10-22  7:32 UTC (permalink / raw)
  To: Yue Haibing
  Cc: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, jacob.e.keller,
	sven.auhagen, alexander.h.duyck, intel-wired-lan, netdev,
	linux-kernel, bpf

On Tue, Oct 22, 2024 at 02:56:19PM +0800, Yue Haibing wrote:
> Fixing sparse error in xdp run code by introducing new variable xdp_res
> instead of overloading this into the skb pointer as i40e drivers done
> in commit 12738ac4754e ("i40e: Fix sparse errors in i40e_txrx.c") and
> commit ae4393dfd472 ("i40e: fix broken XDP support").
> 
> v3: Fix uninitialized 'xdp_res' in patch 3 and 4 which Reported-by
>     kernel test robot
> v2: Fix this as i40e drivers done instead of return NULL in xdp run code

Hi Yue Haibing, all,

I like these changes a lot. But I do wonder if it would
be more appropriate to target them at net-next (or iwl-next)
rather than net, without Fixes tags. This is because they
don't seem to be fixing (user-visible) bugs. Am I missing something?

...

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp()
  2024-10-22  6:56 ` [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp() Yue Haibing
@ 2024-10-22 11:25   ` Maciej Fijalkowski
  0 siblings, 0 replies; 11+ messages in thread
From: Maciej Fijalkowski @ 2024-10-22 11:25 UTC (permalink / raw)
  To: Yue Haibing
  Cc: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, vedang.patel,
	jithu.joseph, andre.guedes, horms, jacob.e.keller, sven.auhagen,
	alexander.h.duyck, intel-wired-lan, netdev, linux-kernel, bpf

On Tue, Oct 22, 2024 at 02:56:22PM +0800, Yue Haibing wrote:
> ixgbe_run_xdp() converts customed xdp action to a negative error code
> with the sk_buff pointer type which be checked with IS_ERR in
> ixgbe_clean_rx_irq(). Remove this error pointer handing instead use
> plain int return value.
> 
> Fixes: 924708081629 ("ixgbe: add XDP support for pass and drop actions")
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Signed-off-by: Yue Haibing <yuehaibing@huawei.com>

Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>

> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 23 ++++++++-----------
>  1 file changed, 9 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> index 8b8404d8c946..78bf97ab0524 100644
> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> @@ -1908,10 +1908,6 @@ bool ixgbe_cleanup_headers(struct ixgbe_ring *rx_ring,
>  {
>  	struct net_device *netdev = rx_ring->netdev;
>  
> -	/* XDP packets use error pointer so abort at this point */
> -	if (IS_ERR(skb))
> -		return true;
> -
>  	/* Verify netdev is present, and that packet does not have any
>  	 * errors that would be unacceptable to the netdev.
>  	 */
> @@ -2219,9 +2215,9 @@ static struct sk_buff *ixgbe_build_skb(struct ixgbe_ring *rx_ring,
>  	return skb;
>  }
>  
> -static struct sk_buff *ixgbe_run_xdp(struct ixgbe_adapter *adapter,
> -				     struct ixgbe_ring *rx_ring,
> -				     struct xdp_buff *xdp)
> +static int ixgbe_run_xdp(struct ixgbe_adapter *adapter,
> +			 struct ixgbe_ring *rx_ring,
> +			 struct xdp_buff *xdp)
>  {
>  	int err, result = IXGBE_XDP_PASS;
>  	struct bpf_prog *xdp_prog;
> @@ -2271,7 +2267,7 @@ static struct sk_buff *ixgbe_run_xdp(struct ixgbe_adapter *adapter,
>  		break;
>  	}
>  xdp_out:
> -	return ERR_PTR(-result);
> +	return result;
>  }
>  
>  static unsigned int ixgbe_rx_frame_truesize(struct ixgbe_ring *rx_ring,
> @@ -2329,6 +2325,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
>  	unsigned int offset = rx_ring->rx_offset;
>  	unsigned int xdp_xmit = 0;
>  	struct xdp_buff xdp;
> +	int xdp_res = 0;
>  
>  	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
>  #if (PAGE_SIZE < 8192)
> @@ -2374,12 +2371,10 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
>  			/* At larger PAGE_SIZE, frame_sz depend on len size */
>  			xdp.frame_sz = ixgbe_rx_frame_truesize(rx_ring, size);
>  #endif
> -			skb = ixgbe_run_xdp(adapter, rx_ring, &xdp);
> +			xdp_res = ixgbe_run_xdp(adapter, rx_ring, &xdp);
>  		}
>  
> -		if (IS_ERR(skb)) {
> -			unsigned int xdp_res = -PTR_ERR(skb);
> -
> +		if (xdp_res) {
>  			if (xdp_res & (IXGBE_XDP_TX | IXGBE_XDP_REDIR)) {
>  				xdp_xmit |= xdp_res;
>  				ixgbe_rx_buffer_flip(rx_ring, rx_buffer, size);
> @@ -2399,7 +2394,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
>  		}
>  
>  		/* exit if we failed to retrieve a buffer */
> -		if (!skb) {
> +		if (!xdp_res && !skb) {
>  			rx_ring->rx_stats.alloc_rx_buff_failed++;
>  			rx_buffer->pagecnt_bias++;
>  			break;
> @@ -2413,7 +2408,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
>  			continue;
>  
>  		/* verify the packet layout is correct */
> -		if (ixgbe_cleanup_headers(rx_ring, rx_desc, skb))
> +		if (xdp_res || ixgbe_cleanup_headers(rx_ring, rx_desc, skb))
>  			continue;
>  
>  		/* probably a little skewed due to removing CRC */
> -- 
> 2.34.1
> 

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp()
  2024-10-22  6:56 ` [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp() Yue Haibing
@ 2024-10-22 11:27   ` Maciej Fijalkowski
  0 siblings, 0 replies; 11+ messages in thread
From: Maciej Fijalkowski @ 2024-10-22 11:27 UTC (permalink / raw)
  To: Yue Haibing
  Cc: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, vedang.patel,
	jithu.joseph, andre.guedes, horms, jacob.e.keller, sven.auhagen,
	alexander.h.duyck, intel-wired-lan, netdev, linux-kernel, bpf

On Tue, Oct 22, 2024 at 02:56:23PM +0800, Yue Haibing wrote:
> ixgbevf_run_xdp() converts customed xdp action to a negative error code
> with the sk_buff pointer type which be checked with IS_ERR in
> ixgbevf_clean_rx_irq(). Remove this error pointer handing instead use
> plain int return value.
> 
> Fixes: c7aec59657b6 ("ixgbevf: Add XDP support for pass and drop actions")
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Signed-off-by: Yue Haibing <yuehaibing@huawei.com>

Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>

> ---
>  .../net/ethernet/intel/ixgbevf/ixgbevf_main.c | 23 ++++++++-----------
>  1 file changed, 10 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
> index 149911e3002a..183d2305d058 100644
> --- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
> +++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
> @@ -732,10 +732,6 @@ static bool ixgbevf_cleanup_headers(struct ixgbevf_ring *rx_ring,
>  				    union ixgbe_adv_rx_desc *rx_desc,
>  				    struct sk_buff *skb)
>  {
> -	/* XDP packets use error pointer so abort at this point */
> -	if (IS_ERR(skb))
> -		return true;
> -
>  	/* verify that the packet does not have any known errors */
>  	if (unlikely(ixgbevf_test_staterr(rx_desc,
>  					  IXGBE_RXDADV_ERR_FRAME_ERR_MASK))) {
> @@ -1044,9 +1040,9 @@ static int ixgbevf_xmit_xdp_ring(struct ixgbevf_ring *ring,
>  	return IXGBEVF_XDP_TX;
>  }
>  
> -static struct sk_buff *ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
> -				       struct ixgbevf_ring  *rx_ring,
> -				       struct xdp_buff *xdp)
> +static int ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
> +			   struct ixgbevf_ring *rx_ring,
> +			   struct xdp_buff *xdp)
>  {
>  	int result = IXGBEVF_XDP_PASS;
>  	struct ixgbevf_ring *xdp_ring;
> @@ -1080,7 +1076,7 @@ static struct sk_buff *ixgbevf_run_xdp(struct ixgbevf_adapter *adapter,
>  		break;
>  	}
>  xdp_out:
> -	return ERR_PTR(-result);
> +	return result;
>  }
>  
>  static unsigned int ixgbevf_rx_frame_truesize(struct ixgbevf_ring *rx_ring,
> @@ -1122,6 +1118,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
>  	struct sk_buff *skb = rx_ring->skb;
>  	bool xdp_xmit = false;
>  	struct xdp_buff xdp;
> +	int xdp_res = 0;
>  
>  	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
>  #if (PAGE_SIZE < 8192)
> @@ -1165,11 +1162,11 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
>  			/* At larger PAGE_SIZE, frame_sz depend on len size */
>  			xdp.frame_sz = ixgbevf_rx_frame_truesize(rx_ring, size);
>  #endif
> -			skb = ixgbevf_run_xdp(adapter, rx_ring, &xdp);
> +			xdp_res = ixgbevf_run_xdp(adapter, rx_ring, &xdp);
>  		}
>  
> -		if (IS_ERR(skb)) {
> -			if (PTR_ERR(skb) == -IXGBEVF_XDP_TX) {
> +		if (xdp_res) {
> +			if (xdp_res == IXGBEVF_XDP_TX) {
>  				xdp_xmit = true;
>  				ixgbevf_rx_buffer_flip(rx_ring, rx_buffer,
>  						       size);
> @@ -1189,7 +1186,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
>  		}
>  
>  		/* exit if we failed to retrieve a buffer */
> -		if (!skb) {
> +		if (!xdp_res && !skb) {
>  			rx_ring->rx_stats.alloc_rx_buff_failed++;
>  			rx_buffer->pagecnt_bias++;
>  			break;
> @@ -1203,7 +1200,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
>  			continue;
>  
>  		/* verify the packet layout is correct */
> -		if (ixgbevf_cleanup_headers(rx_ring, rx_desc, skb)) {
> +		if (xdp_res || ixgbevf_cleanup_headers(rx_ring, rx_desc, skb)) {
>  			skb = NULL;
>  			continue;
>  		}
> -- 
> 2.34.1
> 

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers
  2024-10-22  7:32 ` [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Simon Horman
@ 2024-10-22 19:17   ` Jacob Keller
  2024-10-25  3:30     ` Yue Haibing
  0 siblings, 1 reply; 11+ messages in thread
From: Jacob Keller @ 2024-10-22 19:17 UTC (permalink / raw)
  To: Simon Horman, Yue Haibing
  Cc: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, sven.auhagen,
	alexander.h.duyck, intel-wired-lan, netdev, linux-kernel, bpf



On 10/22/2024 12:32 AM, Simon Horman wrote:
> On Tue, Oct 22, 2024 at 02:56:19PM +0800, Yue Haibing wrote:
>> Fixing sparse error in xdp run code by introducing new variable xdp_res
>> instead of overloading this into the skb pointer as i40e drivers done
>> in commit 12738ac4754e ("i40e: Fix sparse errors in i40e_txrx.c") and
>> commit ae4393dfd472 ("i40e: fix broken XDP support").
>>
>> v3: Fix uninitialized 'xdp_res' in patch 3 and 4 which Reported-by
>>     kernel test robot
>> v2: Fix this as i40e drivers done instead of return NULL in xdp run code
> 
> Hi Yue Haibing, all,
> 
> I like these changes a lot. But I do wonder if it would
> be more appropriate to target them at net-next (or iwl-next)
> rather than net, without Fixes tags. This is because they
> don't seem to be fixing (user-visible) bugs. Am I missing something?
> 
> ...

Yea, these do seem like next candidates.

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers
  2024-10-22 19:17   ` Jacob Keller
@ 2024-10-25  3:30     ` Yue Haibing
  2024-10-25  4:11       ` Przemek Kitszel
  0 siblings, 1 reply; 11+ messages in thread
From: Yue Haibing @ 2024-10-25  3:30 UTC (permalink / raw)
  To: Jacob Keller, Simon Horman
  Cc: anthony.l.nguyen, przemyslaw.kitszel, davem, edumazet, kuba,
	pabeni, ast, daniel, hawk, john.fastabend, maciej.fijalkowski,
	vedang.patel, jithu.joseph, andre.guedes, sven.auhagen,
	alexander.h.duyck, intel-wired-lan, netdev, linux-kernel, bpf

On 2024/10/23 3:17, Jacob Keller wrote:
> 
> 
> On 10/22/2024 12:32 AM, Simon Horman wrote:
>> On Tue, Oct 22, 2024 at 02:56:19PM +0800, Yue Haibing wrote:
>>> Fixing sparse error in xdp run code by introducing new variable xdp_res
>>> instead of overloading this into the skb pointer as i40e drivers done
>>> in commit 12738ac4754e ("i40e: Fix sparse errors in i40e_txrx.c") and
>>> commit ae4393dfd472 ("i40e: fix broken XDP support").
>>>
>>> v3: Fix uninitialized 'xdp_res' in patch 3 and 4 which Reported-by
>>>     kernel test robot
>>> v2: Fix this as i40e drivers done instead of return NULL in xdp run code
>>
>> Hi Yue Haibing, all,
>>
>> I like these changes a lot. But I do wonder if it would
>> be more appropriate to target them at net-next (or iwl-next)
>> rather than net, without Fixes tags. This is because they
>> don't seem to be fixing (user-visible) bugs. Am I missing something?
>>
>> ...
> 
> Yea, these do seem like next candidates.

Should I resend this serial target to iwl-next?
> 
> .

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers
  2024-10-25  3:30     ` Yue Haibing
@ 2024-10-25  4:11       ` Przemek Kitszel
  0 siblings, 0 replies; 11+ messages in thread
From: Przemek Kitszel @ 2024-10-25  4:11 UTC (permalink / raw)
  To: Yue Haibing, Jacob Keller, Simon Horman
  Cc: anthony.l.nguyen, davem, edumazet, kuba, pabeni, ast, daniel,
	hawk, john.fastabend, maciej.fijalkowski, vedang.patel,
	jithu.joseph, andre.guedes, sven.auhagen, alexander.h.duyck,
	intel-wired-lan, netdev, linux-kernel, bpf

On 10/25/24 05:30, Yue Haibing wrote:
> On 2024/10/23 3:17, Jacob Keller wrote:
>>
>>
>> On 10/22/2024 12:32 AM, Simon Horman wrote:
>>> On Tue, Oct 22, 2024 at 02:56:19PM +0800, Yue Haibing wrote:
>>>> Fixing sparse error in xdp run code by introducing new variable xdp_res
>>>> instead of overloading this into the skb pointer as i40e drivers done
>>>> in commit 12738ac4754e ("i40e: Fix sparse errors in i40e_txrx.c") and
>>>> commit ae4393dfd472 ("i40e: fix broken XDP support").
>>>>
>>>> v3: Fix uninitialized 'xdp_res' in patch 3 and 4 which Reported-by
>>>>      kernel test robot
>>>> v2: Fix this as i40e drivers done instead of return NULL in xdp run code
>>>
>>> Hi Yue Haibing, all,
>>>
>>> I like these changes a lot. But I do wonder if it would
>>> be more appropriate to target them at net-next (or iwl-next)
>>> rather than net, without Fixes tags. This is because they
>>> don't seem to be fixing (user-visible) bugs. Am I missing something?
>>>
>>> ...
>>
>> Yea, these do seem like next candidates.
> 
> Should I resend this serial target to iwl-next?
yes, please
(please also link to v3 from v4)

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2024-10-25  4:11 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-10-22  6:56 [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Yue Haibing
2024-10-22  6:56 ` [PATCH v3 net 1/4] igc: Fix passing 0 to ERR_PTR in igc_xdp_run_prog() Yue Haibing
2024-10-22  6:56 ` [PATCH v3 net 2/4] igb: Fix passing 0 to ERR_PTR in igb_run_xdp() Yue Haibing
2024-10-22  6:56 ` [PATCH v3 net 3/4] ixgbe: Fix passing 0 to ERR_PTR in ixgbe_run_xdp() Yue Haibing
2024-10-22 11:25   ` Maciej Fijalkowski
2024-10-22  6:56 ` [PATCH v3 net 4/4] ixgbevf: Fix passing 0 to ERR_PTR in ixgbevf_run_xdp() Yue Haibing
2024-10-22 11:27   ` Maciej Fijalkowski
2024-10-22  7:32 ` [PATCH v3 net 0/4] Fix passing 0 to ERR_PTR in intel ether drivers Simon Horman
2024-10-22 19:17   ` Jacob Keller
2024-10-25  3:30     ` Yue Haibing
2024-10-25  4:11       ` Przemek Kitszel

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).