From: Stephen Hemminger <shemminger@osdl.org>
To: Jeff Garzik <jgarzik@pobox.com>
Cc: netdev@vger.kernel.org
Subject: [PATCH 10/10] chelesio: transmit locking (plus bug fix).
Date: Fri, 01 Dec 2006 16:36:22 -0800 [thread overview]
Message-ID: <20061202003707.873200000@osdl.org> (raw)
In-Reply-To: 20061202003612.573260000@osdl.org
[-- Attachment #1: chelsio-tx-lock.patch --]
[-- Type: text/plain, Size: 1946 bytes --]
If transmit lock is contended on, then push return code back
and retry at higher level.
Bugfix: If buffer is reallocated because of lack of headroom
and the send is blocked, then drop packet. This is necessary
because caller would end up requeuing a freed skb.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
---
drivers/net/chelsio/sge.c | 19 +++++++++++++++----
1 file changed, 15 insertions(+), 4 deletions(-)
--- chelsio.orig/drivers/net/chelsio/sge.c
+++ chelsio/drivers/net/chelsio/sge.c
@@ -1778,7 +1778,9 @@ static int t1_sge_tx(struct sk_buff *skb
struct cmdQ *q = &sge->cmdQ[qid];
unsigned int credits, pidx, genbit, count, use_sched_skb = 0;
- spin_lock(&q->lock);
+ if (!spin_trylock(&q->lock))
+ return NETDEV_TX_LOCKED;
+
reclaim_completed_tx(sge, q);
pidx = q->pidx;
@@ -1887,6 +1889,8 @@ int t1_start_xmit(struct sk_buff *skb, s
struct sge *sge = adapter->sge;
struct sge_port_stats *st = per_cpu_ptr(sge->port_stats[dev->if_port], smp_processor_id());
struct cpl_tx_pkt *cpl;
+ struct sk_buff *orig_skb = skb;
+ int ret;
if (skb->protocol == htons(ETH_P_CPL5))
goto send;
@@ -1930,8 +1934,6 @@ int t1_start_xmit(struct sk_buff *skb, s
* Complain when this happens but try to fix things up.
*/
if (unlikely(skb_headroom(skb) < dev->hard_header_len - ETH_HLEN)) {
- struct sk_buff *orig_skb = skb;
-
pr_debug("%s: headroom %d header_len %d\n", dev->name,
skb_headroom(skb), dev->hard_header_len);
@@ -1991,7 +1993,16 @@ int t1_start_xmit(struct sk_buff *skb, s
send:
st->tx_packets++;
dev->trans_start = jiffies;
- return t1_sge_tx(skb, adapter, 0, dev);
+ ret = t1_sge_tx(skb, adapter, 0, dev);
+
+ /* If transmit busy, and we reallocated skb's due to headroom limit,
+ * then silently discard to avoid leak.
+ */
+ if (unlikely(ret != NETDEV_TX_OK && skb != orig_skb)) {
+ dev_kfree_skb_any(skb);
+ ret = NETDEV_TX_OK;
+ }
+ return ret;
}
/*
--
next prev parent reply other threads:[~2006-12-02 0:45 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-12-02 0:36 [PATCH 00/10] chelsio: 10G driver Stephen Hemminger
2006-12-02 0:36 ` [PATCH 01/10] chelsio: whitespace fixes Stephen Hemminger
2006-12-02 0:36 ` [PATCH 02/10] chelsio: use kzalloc Stephen Hemminger
2006-12-02 0:36 ` [PATCH 03/10] chelsio: remove unused mutex Stephen Hemminger
2006-12-02 0:36 ` [PATCH 05/10] chelsio: add 1G swcixw aupport Stephen Hemminger
2006-12-02 0:36 ` [PATCH 06/10] chelsio: cleanup pm3393 code Stephen Hemminger
2006-12-02 0:36 ` [PATCH 07/10] chelsio: use standard CRC routines Stephen Hemminger
2006-12-02 0:36 ` [PATCH 08/10] chelsio: add MSI support Stephen Hemminger
2006-12-02 0:36 ` [PATCH 09/10] chelsio: statistics improvement Stephen Hemminger
2006-12-02 0:36 ` Stephen Hemminger [this message]
2006-12-03 10:45 ` [PATCH 10/10] chelesio: transmit locking (plus bug fix) Eric Lemoine
2006-12-04 17:58 ` Stephen Hemminger
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=20061202003707.873200000@osdl.org \
--to=shemminger@osdl.org \
--cc=jgarzik@pobox.com \
--cc=netdev@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;
as well as URLs for NNTP newsgroup(s).