[PATCH net] igc: only strip RX timestamp header from first buffer

[PATCH net] igc: only strip RX timestamp header from first buffer

[Paul Moses]

igc_clean_rx_irq() strips IGC_TS_HDR_LEN whenever a descriptor reports
IGC_RXDADV_STAT_TSIP. For multi-buffer packets, continuation descriptors
retain TSIP even though the inline timestamp is present only in the first
RX buffer.

Subtracting the header length from each continuation buffer truncates
jumbo packets by 16 bytes per continuation and leaves the packet length
larger than the received data.

Only consume the timestamp header when skb is NULL, which identifies the
first buffer of a new packet. An skb carried in rx_ring->skb remains
non-NULL when packet assembly resumes in a later NAPI poll.

Link: https://lore.kernel.org/all/20260625-igb-rx-ts-fix-v3-1-99b3efa08dca@aweta.nl/
Fixes: e1ed4f92a625 ("igc: Refactor Rx timestamp handling")
Cc: stable@vger.kernel.org
Signed-off-by: Paul Moses <p@1g4.org>
---
#
# REPRO
# 
# Receiver: igc / enp1s0 / 10.0.5.175
# Sender:   RTL8125B / enP4p65s0 / 10.0.5.165

# Both hosts
sudo ip link set dev enp1s0 mtu 9000
sudo ip link set dev enP4p65s0 mtu 9000

# Baseline: RX timestamping off
sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
ping -n -M do -c 5 -s 8972 10.0.5.175
# 5/5 received

# Enable RX timestamping only
sudo hwstamp_ctl -i enp1s0 -t 0 -r 1

# Receiver
sudo timeout 20 tcpdump -ni enp1s0 -s 0 -w /tmp/igc-jumbo.pcap \
  'icmp and host 10.0.5.165'

# Sender
ping -n -M do -c 5 -W 2 -s 8972 10.0.5.175
# 0/5 received

# Receiver
sudo tcpdump -nn -e -vvv -r /tmp/igc-jumbo.pcap
# length 8950
# [total length 9000 > length 8936] (invalid)
# wrong icmp cksum

# Expected Ethernet length: 9014
# Captured length:          8950
# Loss:                       64 = 4 continuation buffers * 16 bytes

# Disable RX timestamping again
sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
ping -n -M do -c 5 -s 8972 10.0.5.175
# 5/5 received

---

 drivers/net/ethernet/intel/igc/igc_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
index 1fb5f3cbe93c..41b293dc10c3 100644
--- a/drivers/net/ethernet/intel/igc/igc_main.c
+++ b/drivers/net/ethernet/intel/igc/igc_main.c
@@ -2639,7 +2639,8 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
 
 		pktbuf = page_address(rx_buffer->page) + rx_buffer->page_offset;
 
-		if (igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
+		if (!skb &&
+		    igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
 			ctx.rx_ts = pktbuf;
 			pkt_offset = IGC_TS_HDR_LEN;
 			size -= IGC_TS_HDR_LEN;
-- 
2.55.GIT

RE: [PATCH net] igc: only strip RX timestamp header from first buffer

[Loktionov, Aleksandr]

> -----Original Message-----
> From: Paul Moses <p@1g4.org>
> Sent: Wednesday, August 26, 2026 1:56 AM
> To: intel-wired-lan@lists.osuosl.org
> Cc: netdev@vger.kernel.org; linux-kernel@vger.kernel.org; Nguyen,
> Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> <przemyslaw.kitszel@intel.com>; andrew+netdev@lunn.ch;
> davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com; Fijalkowski, Maciej <maciej.fijalkowski@intel.com>;
> andre.guedes@intel.com; vedang.patel@intel.com;
> jithu.joseph@intel.com; richardcochran@gmail.com; tkusters@aweta.nl;
> stable@vger.kernel.org
> Subject: [PATCH net] igc: only strip RX timestamp header from first
> buffer
> 
> igc_clean_rx_irq() strips IGC_TS_HDR_LEN whenever a descriptor reports
> IGC_RXDADV_STAT_TSIP. For multi-buffer packets, continuation
> descriptors
> retain TSIP even though the inline timestamp is present only in the
> first
> RX buffer.
> 
> Subtracting the header length from each continuation buffer truncates
> jumbo packets by 16 bytes per continuation and leaves the packet
> length
> larger than the received data.
> 
> Only consume the timestamp header when skb is NULL, which identifies
> the
> first buffer of a new packet. An skb carried in rx_ring->skb remains
> non-NULL when packet assembly resumes in a later NAPI poll.
> 
> Link: https://lore.kernel.org/all/20260625-igb-rx-ts-fix-v3-1-
> 99b3efa08dca@aweta.nl/
> Fixes: e1ed4f92a625 ("igc: Refactor Rx timestamp handling")
> Cc: stable@vger.kernel.org
> Signed-off-by: Paul Moses <p@1g4.org>
> ---
> #
> # REPRO
> #
> # Receiver: igc / enp1s0 / 10.0.5.175
> # Sender:   RTL8125B / enP4p65s0 / 10.0.5.165
> 
> # Both hosts
> sudo ip link set dev enp1s0 mtu 9000
> sudo ip link set dev enP4p65s0 mtu 9000
> 
> # Baseline: RX timestamping off
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
> ping -n -M do -c 5 -s 8972 10.0.5.175
> # 5/5 received
> 
> # Enable RX timestamping only
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 1
> 
> # Receiver
> sudo timeout 20 tcpdump -ni enp1s0 -s 0 -w /tmp/igc-jumbo.pcap \
>   'icmp and host 10.0.5.165'
> 
> # Sender
> ping -n -M do -c 5 -W 2 -s 8972 10.0.5.175
> # 0/5 received
> 
> # Receiver
> sudo tcpdump -nn -e -vvv -r /tmp/igc-jumbo.pcap
> # length 8950
> # [total length 9000 > length 8936] (invalid)
> # wrong icmp cksum
> 
> # Expected Ethernet length: 9014
> # Captured length:          8950
> # Loss:                       64 = 4 continuation buffers * 16 bytes
> 
> # Disable RX timestamping again
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
> ping -n -M do -c 5 -s 8972 10.0.5.175
> # 5/5 received
> 
> ---
> 
>  drivers/net/ethernet/intel/igc/igc_main.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c
> b/drivers/net/ethernet/intel/igc/igc_main.c
> index 1fb5f3cbe93c..41b293dc10c3 100644
> --- a/drivers/net/ethernet/intel/igc/igc_main.c
> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> @@ -2639,7 +2639,8 @@ static int igc_clean_rx_irq(struct igc_q_vector
> *q_vector, const int budget)
> 
>  		pktbuf = page_address(rx_buffer->page) + rx_buffer-
> >page_offset;
> 
> -		if (igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
> +		if (!skb &&
> +		    igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
>  			ctx.rx_ts = pktbuf;
>  			pkt_offset = IGC_TS_HDR_LEN;
>  			size -= IGC_TS_HDR_LEN;
> --
> 2.55.GIT
> 

Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>

Re: [PATCH net] igc: only strip RX timestamp header from first buffer

[Maciej Fijalkowski]

On Tue, Aug 25, 2026 at 11:55:40PM +0000, Paul Moses wrote:
> igc_clean_rx_irq() strips IGC_TS_HDR_LEN whenever a descriptor reports
> IGC_RXDADV_STAT_TSIP. For multi-buffer packets, continuation descriptors
> retain TSIP even though the inline timestamp is present only in the first
> RX buffer.
> 
> Subtracting the header length from each continuation buffer truncates
> jumbo packets by 16 bytes per continuation and leaves the packet length
> larger than the received data.
> 
> Only consume the timestamp header when skb is NULL, which identifies the
> first buffer of a new packet. An skb carried in rx_ring->skb remains
> non-NULL when packet assembly resumes in a later NAPI poll.
> 
> Link: https://lore.kernel.org/all/20260625-igb-rx-ts-fix-v3-1-99b3efa08dca@aweta.nl/
> Fixes: e1ed4f92a625 ("igc: Refactor Rx timestamp handling")
> Cc: stable@vger.kernel.org
> Signed-off-by: Paul Moses <p@1g4.org>

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

> ---
> #
> # REPRO
> # 
> # Receiver: igc / enp1s0 / 10.0.5.175
> # Sender:   RTL8125B / enP4p65s0 / 10.0.5.165
> 
> # Both hosts
> sudo ip link set dev enp1s0 mtu 9000
> sudo ip link set dev enP4p65s0 mtu 9000
> 
> # Baseline: RX timestamping off
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
> ping -n -M do -c 5 -s 8972 10.0.5.175
> # 5/5 received
> 
> # Enable RX timestamping only
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 1
> 
> # Receiver
> sudo timeout 20 tcpdump -ni enp1s0 -s 0 -w /tmp/igc-jumbo.pcap \
>   'icmp and host 10.0.5.165'
> 
> # Sender
> ping -n -M do -c 5 -W 2 -s 8972 10.0.5.175
> # 0/5 received
> 
> # Receiver
> sudo tcpdump -nn -e -vvv -r /tmp/igc-jumbo.pcap
> # length 8950
> # [total length 9000 > length 8936] (invalid)
> # wrong icmp cksum
> 
> # Expected Ethernet length: 9014
> # Captured length:          8950
> # Loss:                       64 = 4 continuation buffers * 16 bytes
> 
> # Disable RX timestamping again
> sudo hwstamp_ctl -i enp1s0 -t 0 -r 0
> ping -n -M do -c 5 -s 8972 10.0.5.175
> # 5/5 received
> 
> ---
> 
>  drivers/net/ethernet/intel/igc/igc_main.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
> index 1fb5f3cbe93c..41b293dc10c3 100644
> --- a/drivers/net/ethernet/intel/igc/igc_main.c
> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> @@ -2639,7 +2639,8 @@ static int igc_clean_rx_irq(struct igc_q_vector *q_vector, const int budget)
>  
>  		pktbuf = page_address(rx_buffer->page) + rx_buffer->page_offset;
>  
> -		if (igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
> +		if (!skb &&
> +		    igc_test_staterr(rx_desc, IGC_RXDADV_STAT_TSIP)) {
>  			ctx.rx_ts = pktbuf;
>  			pkt_offset = IGC_TS_HDR_LEN;
>  			size -= IGC_TS_HDR_LEN;
> -- 
> 2.55.GIT
> 
>