* [PATCH] net/pcap: fix handling of transmit errors
@ 2026-11-16 23:28 Stephen Hemminger
2026-11-17 18:32 ` [PATCH v2] " Stephen Hemminger
0 siblings, 1 reply; 2+ messages in thread
From: Stephen Hemminger @ 2026-11-16 23:28 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
The pcap PMD assumed that every error from pcap_sendpacket()
was from back pressure. But in the case of bad mbuf or file system
full or other hard errors, the packet needs to be dropped.
Bugzilla ID: 2132
Fixes: cf91cce204ea ("net/pcap: fix error accounting and backpressure on Tx")
Reported-by: Anthropic OSS scanner (ANT-2026-DLCI8IK6)
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
drivers/net/pcap/pcap_ethdev.c | 20 +++++++++-----------
1 file changed, 9 insertions(+), 11 deletions(-)
diff --git a/drivers/net/pcap/pcap_ethdev.c b/drivers/net/pcap/pcap_ethdev.c
index fb18f0932b..e7a8f007e0 100644
--- a/drivers/net/pcap/pcap_ethdev.c
+++ b/drivers/net/pcap/pcap_ethdev.c
@@ -607,16 +607,6 @@ eth_tx_drop(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
/*
* Send a burst of packets to a pcap device.
- *
- * On Linux, pcap_sendpacket() calls send() on a blocking PF_PACKET
- * socket with default kernel buffer sizes and no TX ring (PACKET_TX_RING).
- * The send() call only blocks when the kernel socket send buffer is full,
- * providing limited backpressure.
- *
- * On error, pcap_sendpacket() returns non-zero and the loop breaks,
- * leaving remaining packets unsent.
- *
- * Bottom line: backpressure is not an error.
*/
static uint16_t
eth_pcap_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
@@ -666,10 +656,18 @@ eth_pcap_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
data = rte_pktmbuf_read(mbuf, 0, len, temp_data);
RTE_ASSERT(data != NULL);
+ errno = 0;
if (unlikely(pcap_sendpacket(pcap, data, len) != 0)) {
+ /* retry on soft errors */
+ if (errno == EINTR || errno == EAGAIN ||
+ errno == EWOULDBLOCK)
+ break;
+
/* Assume failure is backpressure */
PMD_LOG(ERR, "pcap_sendpacket() failed: %s", pcap_geterr(pcap));
- break;
+ rte_pktmbuf_free(mbuf);
+ tx_queue->tx_stat.err_pkts++;
+ continue;
}
num_tx++;
tx_bytes += len;
--
2.53.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* [PATCH v2] net/pcap: fix handling of transmit errors
2026-11-16 23:28 [PATCH] net/pcap: fix handling of transmit errors Stephen Hemminger
@ 2026-11-17 18:32 ` Stephen Hemminger
0 siblings, 0 replies; 2+ messages in thread
From: Stephen Hemminger @ 2026-11-17 18:32 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger, stable
The pcap PMD assumed every pcap_sendpacket() failure was back
pressure and stopped the burst. A hard error such as EMSGSIZE or
ENETDOWN caused the same mbuf to be returned to the caller on every
burst, stalling the queue.
Only treat EAGAIN (socket send buffer full) as back pressure.
Drop and count the packet on any other error, including ENOBUFS
which means the packet was already dropped by the qdisc.
Use PMD_TX_LOG for the error message, as other errors in the Tx burst path do.
Bugzilla ID: 2132
Fixes: cf91cce204ea ("net/pcap: fix error accounting and backpressure on Tx")
Cc: stable@dpdk.org
Reported-by: Anthropic OSS scanner (ANT-2026-DLCI8IK6)
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
v2 - fix AI review feedback, use PMD_TX_LOG
drivers/net/pcap/pcap_ethdev.c | 25 ++++++++++++-------------
1 file changed, 12 insertions(+), 13 deletions(-)
diff --git a/drivers/net/pcap/pcap_ethdev.c b/drivers/net/pcap/pcap_ethdev.c
index fb18f0932b..1b8d6dafa2 100644
--- a/drivers/net/pcap/pcap_ethdev.c
+++ b/drivers/net/pcap/pcap_ethdev.c
@@ -607,16 +607,6 @@ eth_tx_drop(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
/*
* Send a burst of packets to a pcap device.
- *
- * On Linux, pcap_sendpacket() calls send() on a blocking PF_PACKET
- * socket with default kernel buffer sizes and no TX ring (PACKET_TX_RING).
- * The send() call only blocks when the kernel socket send buffer is full,
- * providing limited backpressure.
- *
- * On error, pcap_sendpacket() returns non-zero and the loop breaks,
- * leaving remaining packets unsent.
- *
- * Bottom line: backpressure is not an error.
*/
static uint16_t
eth_pcap_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
@@ -666,10 +656,19 @@ eth_pcap_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
data = rte_pktmbuf_read(mbuf, 0, len, temp_data);
RTE_ASSERT(data != NULL);
+ /* pcap_sendpacket() does not document errno, but it is the only way
+ * to tell back pressure from a hard error.
+ */
+ errno = 0;
if (unlikely(pcap_sendpacket(pcap, data, len) != 0)) {
- /* Assume failure is backpressure */
- PMD_LOG(ERR, "pcap_sendpacket() failed: %s", pcap_geterr(pcap));
- break;
+ /* nothing sent, caller should retry */
+ if (errno == EAGAIN || errno == EWOULDBLOCK)
+ break;
+
+ PMD_TX_LOG(ERR, "pcap_sendpacket() failed: %s", pcap_geterr(pcap));
+ rte_pktmbuf_free(mbuf);
+ tx_queue->tx_stat.err_pkts++;
+ continue;
}
num_tx++;
tx_bytes += len;
--
2.53.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-17 18:32 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16 23:28 [PATCH] net/pcap: fix handling of transmit errors Stephen Hemminger
2026-11-17 18:32 ` [PATCH v2] " Stephen Hemminger
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox