public inbox for linux-kselftest@vger.kernel.org
 help / color / mirror / Atom feed
From: fklassen at appneta.com (Fred Klassen)
Subject: [PATCH net 1/4] net/udp_gso: Allow TX timestamp with UDP GSO
Date: Thu, 23 May 2019 14:06:48 -0700	[thread overview]
Message-ID: <20190523210651.80902-2-fklassen@appneta.com> (raw)
In-Reply-To: <20190523210651.80902-1-fklassen@appneta.com>

Fixes an issue where TX Timestamps are not arriving on the error queue
when UDP_SEGMENT CMSG type is combined with CMSG type SO_TIMESTAMPING.
This can be illustrated with an updated updgso_bench_tx program which
includes the '-T' option to test for this condition.

    ./udpgso_bench_tx -4ucTPv -S 1472 -l2 -D 172.16.120.18
    poll timeout
    udp tx:      0 MB/s        1 calls/s      1 msg/s

The "poll timeout" message above indicates that TX timestamp never
arrived.

It also appears that other TX CMSG types cause similar issues, for
example trying to set SOL_IP/IP_TOS.

    ./udpgso_bench_tx -4ucPv -S 1472 -q 182 -l2 -D 172.16.120.18
    poll timeout
    udp tx:      0 MB/s        1 calls/s      1 msg/s

This patch preserves tx_flags for the first UDP GSO segment. This
mirrors the stack's behaviour for IPv4 fragments.

Fixes: ee80d1ebe5ba ("udp: add udp gso")
Signed-off-by: Fred Klassen <fklassen at appneta.com>
---
 net/ipv4/udp_offload.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index 065334b41d57..33de347695ae 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -228,6 +228,10 @@ struct sk_buff *__udp_gso_segment(struct sk_buff *gso_skb,
 	seg = segs;
 	uh = udp_hdr(seg);
 
+	/* preserve TX timestamp and zero-copy info for first segment */
+	skb_shinfo(seg)->tskey = skb_shinfo(gso_skb)->tskey;
+	skb_shinfo(seg)->tx_flags = skb_shinfo(gso_skb)->tx_flags;
+
 	/* compute checksum adjustment based on old length versus new */
 	newlen = htons(sizeof(*uh) + mss);
 	check = csum16_add(csum16_sub(uh->check, uh->len), newlen);
-- 
2.11.0

WARNING: multiple messages have this Message-ID (diff)
From: fklassen@appneta.com (Fred Klassen)
Subject: [PATCH net 1/4] net/udp_gso: Allow TX timestamp with UDP GSO
Date: Thu, 23 May 2019 14:06:48 -0700	[thread overview]
Message-ID: <20190523210651.80902-2-fklassen@appneta.com> (raw)
Message-ID: <20190523210648.SAOsA1xo9RUOlfHl5cWtNCk9EbI2RlEqviDV0sTPt60@z> (raw)
In-Reply-To: <20190523210651.80902-1-fklassen@appneta.com>

Fixes an issue where TX Timestamps are not arriving on the error queue
when UDP_SEGMENT CMSG type is combined with CMSG type SO_TIMESTAMPING.
This can be illustrated with an updated updgso_bench_tx program which
includes the '-T' option to test for this condition.

    ./udpgso_bench_tx -4ucTPv -S 1472 -l2 -D 172.16.120.18
    poll timeout
    udp tx:      0 MB/s        1 calls/s      1 msg/s

The "poll timeout" message above indicates that TX timestamp never
arrived.

It also appears that other TX CMSG types cause similar issues, for
example trying to set SOL_IP/IP_TOS.

    ./udpgso_bench_tx -4ucPv -S 1472 -q 182 -l2 -D 172.16.120.18
    poll timeout
    udp tx:      0 MB/s        1 calls/s      1 msg/s

This patch preserves tx_flags for the first UDP GSO segment. This
mirrors the stack's behaviour for IPv4 fragments.

Fixes: ee80d1ebe5ba ("udp: add udp gso")
Signed-off-by: Fred Klassen <fklassen at appneta.com>
---
 net/ipv4/udp_offload.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index 065334b41d57..33de347695ae 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -228,6 +228,10 @@ struct sk_buff *__udp_gso_segment(struct sk_buff *gso_skb,
 	seg = segs;
 	uh = udp_hdr(seg);
 
+	/* preserve TX timestamp and zero-copy info for first segment */
+	skb_shinfo(seg)->tskey = skb_shinfo(gso_skb)->tskey;
+	skb_shinfo(seg)->tx_flags = skb_shinfo(gso_skb)->tx_flags;
+
 	/* compute checksum adjustment based on old length versus new */
 	newlen = htons(sizeof(*uh) + mss);
 	check = csum16_add(csum16_sub(uh->check, uh->len), newlen);
-- 
2.11.0

  parent reply	other threads:[~2019-05-23 21:06 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-23 21:06 [PATCH net 0/4] Allow TX timestamp with UDP GSO fklassen
2019-05-23 21:06 ` Fred Klassen
2019-05-23 21:06 ` fklassen [this message]
2019-05-23 21:06   ` [PATCH net 1/4] net/udp_gso: " Fred Klassen
2019-05-23 21:39   ` willemdebruijn.kernel
2019-05-23 21:39     ` Willem de Bruijn
2019-05-24  1:38     ` fklassen
2019-05-24  1:38       ` Fred Klassen
2019-05-24  4:53       ` willemdebruijn.kernel
2019-05-24  4:53         ` Willem de Bruijn
2019-05-24 16:34         ` fklassen
2019-05-24 16:34           ` Fred Klassen
2019-05-24 19:29           ` willemdebruijn.kernel
2019-05-24 19:29             ` Willem de Bruijn
2019-05-24 22:01             ` fklassen
2019-05-24 22:01               ` Fred Klassen
2019-05-25 15:20               ` willemdebruijn.kernel
2019-05-25 15:20                 ` Willem de Bruijn
2019-05-25 18:47                 ` fklassen
2019-05-25 18:47                   ` Fred Klassen
2019-05-27  1:30                   ` willemdebruijn.kernel
2019-05-27  1:30                     ` Willem de Bruijn
2019-05-27  2:09                     ` willemdebruijn.kernel
2019-05-27  2:09                       ` Willem de Bruijn
2019-05-25 20:46     ` fklassen
2019-05-25 20:46       ` Fred Klassen
2019-05-23 21:59   ` willemdebruijn.kernel
2019-05-23 21:59     ` Willem de Bruijn
2019-05-25 20:09     ` fklassen
2019-05-25 20:09       ` Fred Klassen
2019-05-25 20:47     ` fklassen
2019-05-25 20:47       ` Fred Klassen
2019-05-23 21:06 ` [PATCH net 2/4] net/udpgso_bench_tx: options to exercise TX CMSG fklassen
2019-05-23 21:06   ` Fred Klassen
2019-05-23 21:45   ` willemdebruijn.kernel
2019-05-23 21:45     ` Willem de Bruijn
2019-05-23 21:52   ` willemdebruijn.kernel
2019-05-23 21:52     ` Willem de Bruijn
2019-05-24  2:10     ` fklassen
2019-05-24  2:10       ` Fred Klassen
2019-05-23 21:06 ` [PATCH net 3/4] net/udpgso_bench_tx: fix sendmmsg on unconnected socket fklassen
2019-05-23 21:06   ` Fred Klassen
2019-05-23 21:06 ` [PATCH net 4/4] net/udpgso_bench_tx: audit error queue fklassen
2019-05-23 21:06   ` Fred Klassen
2019-05-23 21:56   ` willemdebruijn.kernel
2019-05-23 21:56     ` Willem de Bruijn
2019-05-24  1:27     ` fklassen
2019-05-24  1:27       ` Fred Klassen
2019-05-24  5:02       ` willemdebruijn.kernel
2019-05-24  5:02         ` Willem de Bruijn
2019-05-27 21:30     ` fklassen
2019-05-27 21:30       ` Fred Klassen
2019-05-27 21:46       ` willemdebruijn.kernel
2019-05-27 21:46         ` Willem de Bruijn
2019-05-27 22:56         ` fklassen
2019-05-27 22:56           ` Fred Klassen
2019-05-28  1:15           ` willemdebruijn.kernel
2019-05-28  1:15             ` Willem de Bruijn
2019-05-28  5:19             ` fklassen
2019-05-28  5:19               ` Fred Klassen
2019-05-28 15:08               ` willemdebruijn.kernel
2019-05-28 15:08                 ` Willem de Bruijn
2019-05-28 16:57                 ` fklassen
2019-05-28 16:57                   ` Fred Klassen
2019-05-28 17:07                   ` willemdebruijn.kernel
2019-05-28 17:07                     ` Willem de Bruijn
2019-05-28 17:11                     ` willemdebruijn.kernel
2019-05-28 17:11                       ` Willem de Bruijn

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190523210651.80902-2-fklassen@appneta.com \
    --to=linux-kselftest@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox