From: "Michael Chan" <mchan@broadcom.com>
To: "David Miller" <davem@davemloft.net>
Cc: jeff@garzik.org, "netdev" <netdev@vger.kernel.org>
Subject: [PATCH revised 10/11][TG3]: Eliminate spurious interrupts.
Date: Sat, 05 May 2007 17:19:31 -0700 [thread overview]
Message-ID: <1178410771.4859.50.camel@dell> (raw)
In-Reply-To: <20070504.204700.24899658.davem@davemloft.net>
[TG3]: Eliminate spurious interrupts.
Spurious interrupts are often encountered especially on systems
using the 8259 PIC mode. This is because the I/O write to deassert
the interrupt is posted and won't get to the chip immediately. As
a result, the IRQ may remain asserted after the IRQ handler exits,
causing spurious interrupts.
Flush the interrupt mailbox in non-MSI handlers to de-assert the
IRQ immediately. This seems to be the most straight forward approach
after discussion with Jeff Garzik and David Miller.
Signed-off-by: Michael Chan <mchan@broadcom.com>
diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
index da1ab01..70b8c5e 100644
--- a/drivers/net/tg3.c
+++ b/drivers/net/tg3.c
@@ -3590,8 +3590,12 @@ static irqreturn_t tg3_interrupt(int irq, void *dev_id)
* Writing non-zero to intr-mbox-0 additional tells the
* NIC to stop sending us irqs, engaging "in-intr-handler"
* event coalescing.
+ *
+ * Flush the mailbox to de-assert the IRQ immediately to prevent
+ * spurious interrupts. The flush impacts performance but
+ * excessive spurious interrupts can be worse in some cases.
*/
- tw32_mailbox(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001);
+ tw32_mailbox_f(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001);
if (tg3_irq_sync(tp))
goto out;
sblk->status &= ~SD_STATUS_UPDATED;
@@ -3635,8 +3639,12 @@ static irqreturn_t tg3_interrupt_tagged(int irq, void *dev_id)
* writing non-zero to intr-mbox-0 additional tells the
* NIC to stop sending us irqs, engaging "in-intr-handler"
* event coalescing.
+ *
+ * Flush the mailbox to de-assert the IRQ immediately to prevent
+ * spurious interrupts. The flush impacts performance but
+ * excessive spurious interrupts can be worse in some cases.
*/
- tw32_mailbox(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001);
+ tw32_mailbox_f(MAILBOX_INTERRUPT_0 + TG3_64BIT_REG_LOW, 0x00000001);
if (tg3_irq_sync(tp))
goto out;
if (netif_rx_schedule_prep(dev)) {
next prev parent reply other threads:[~2007-05-05 23:32 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-05-05 0:55 [PATCH 10/11][TG3]: Reduce spurious interrupts Michael Chan
2007-05-05 0:42 ` Jeff Garzik
2007-05-05 1:45 ` Michael Chan
2007-05-05 2:42 ` David Miller
2007-05-05 2:50 ` Michael Chan
2007-05-05 3:47 ` David Miller
2007-05-06 0:15 ` Michael Chan
2007-05-06 0:19 ` Michael Chan [this message]
2007-05-07 7:26 ` [PATCH revised 10/11][TG3]: Eliminate " David Miller
2007-05-06 0:21 ` [PATCH revised 11/11][TG3]: Update version and reldate Michael Chan
2007-05-07 7:26 ` David Miller
2007-05-07 16:39 ` [PATCH 10/11][TG3]: Reduce spurious interrupts Rick Jones
[not found] ` <1178562469.4859.70.camel@dell>
2007-05-07 17:44 ` Rick Jones
2007-05-07 18:41 ` Michael Chan
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=1178410771.4859.50.camel@dell \
--to=mchan@broadcom.com \
--cc=davem@davemloft.net \
--cc=jeff@garzik.org \
--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).