mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Chris Metcalf <cmetcalf@tilera.com>
To: <linux-kernel@vger.kernel.org>
Cc: "Américo Wang" <xiyou.wangcong@gmail.com>,
	"Eric Dumazet" <eric.dumazet@gmail.com>,
	netdev <netdev@vger.kernel.org>, "Cypher Wu" <cypher.w@gmail.com>
Subject: [PATCH] arch/tile: fix rwlock so would-be write lockers don't block new readers
Date: Mon, 15 Nov 2010 09:18:49 -0500	[thread overview]
Message-ID: <201011151425.oAFEPU3W005682@farm-0010.internal.tilera.com> (raw)
In-Reply-To: <AANLkTind92uwcigzmDn8yn9a22exDy7zcreGQ5-6NLV-@mail.gmail.com>

This avoids a deadlock in the IGMP code where one core gets a read
lock, another core starts trying to get a write lock (thus blocking
new readers), and then the first core tries to recursively re-acquire
the read lock.

We still try to preserve some degree of balance by giving priority
to additional write lockers that come along while the lock is held
for write, so they can all complete quickly and return the lock to
the readers.

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
---
This should apply relatively cleanly to 2.6.26.7 source code too.

 arch/tile/lib/spinlock_32.c |   29 ++++++++++++++++++-----------
 1 files changed, 18 insertions(+), 11 deletions(-)

diff --git a/arch/tile/lib/spinlock_32.c b/arch/tile/lib/spinlock_32.c
index 485e24d..5cd1c40 100644
--- a/arch/tile/lib/spinlock_32.c
+++ b/arch/tile/lib/spinlock_32.c
@@ -167,23 +167,30 @@ void arch_write_lock_slow(arch_rwlock_t *rwlock, u32 val)
 	 * when we compare them.
 	 */
 	u32 my_ticket_;
+	u32 iterations = 0;
 
-	/* Take out the next ticket; this will also stop would-be readers. */
-	if (val & 1)
-		val = get_rwlock(rwlock);
-	rwlock->lock = __insn_addb(val, 1 << WR_NEXT_SHIFT);
+	/*
+	 * Wait until there are no readers, then bump up the next
+	 * field and capture the ticket value.
+	 */
+	for (;;) {
+		if (!(val & 1)) {
+			if ((val >> RD_COUNT_SHIFT) == 0)
+				break;
+			rwlock->lock = val;
+		}
+		delay_backoff(iterations++);
+		val = __insn_tns((int *)&rwlock->lock);
+	}
 
-	/* Extract my ticket value from the original word. */
+	/* Take out the next ticket and extract my ticket value. */
+	rwlock->lock = __insn_addb(val, 1 << WR_NEXT_SHIFT);
 	my_ticket_ = val >> WR_NEXT_SHIFT;
 
-	/*
-	 * Wait until the "current" field matches our ticket, and
-	 * there are no remaining readers.
-	 */
+	/* Wait until the "current" field matches our ticket. */
 	for (;;) {
 		u32 curr_ = val >> WR_CURR_SHIFT;
-		u32 readers = val >> RD_COUNT_SHIFT;
-		u32 delta = ((my_ticket_ - curr_) & WR_MASK) + !!readers;
+		u32 delta = ((my_ticket_ - curr_) & WR_MASK);
 		if (likely(delta == 0))
 			break;
 
-- 
1.6.5.2


  parent reply	other threads:[~2010-11-15 14:25 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-11 13:49 Kernel rwlock design, Multicore and IGMP Cypher Wu
2010-11-11 15:23 ` Eric Dumazet
2010-11-11 15:32   ` Eric Dumazet
2010-11-12  3:32   ` Cypher Wu
2010-11-12  6:28     ` Américo Wang
2010-11-12  7:13     ` Américo Wang
2010-11-12  7:27       ` Eric Dumazet
2010-11-12  8:19         ` Américo Wang
2010-11-12  9:09           ` Yong Zhang
2010-11-12  9:18             ` Américo Wang
2010-11-12 11:06               ` Cypher Wu
2010-11-13  6:35                 ` Américo Wang
2010-11-12 13:00               ` Yong Zhang
2010-11-13  6:28                 ` Américo Wang
2010-11-12  9:22           ` Eric Dumazet
2010-11-12  9:33             ` Américo Wang
2010-11-12 13:34             ` [PATCH net-next-2.6] igmp: RCU conversion of in_dev->mc_list Eric Dumazet
2010-11-12 14:26               ` Eric Dumazet
2010-11-12 15:46                 ` [PATCH net-next-2.6 V2] " Eric Dumazet
2010-11-12 21:19                   ` David Miller
2010-11-13  6:44                   ` Américo Wang
2010-11-13 22:54           ` Kernel rwlock design, Multicore and IGMP Peter Zijlstra
2010-11-12 11:10         ` Cypher Wu
2010-11-12 11:25           ` Eric Dumazet
2010-11-13 22:53     ` Peter Zijlstra
     [not found]     ` <ZXmP8hjgLHA.4648@exchange1.tad.internal.tilera.com>
2010-11-13 23:03       ` Chris Metcalf
2010-11-15  7:22         ` Cypher Wu
2010-11-15 11:18           ` Cypher Wu
2010-11-15 11:31             ` Eric Dumazet
2010-11-17  1:30               ` Cypher Wu
2010-11-17  4:43                 ` Eric Dumazet
2010-11-15 14:18           ` Chris Metcalf [this message]
2010-11-15 14:52             ` [PATCH] arch/tile: fix rwlock so would-be write lockers don't block new readers Eric Dumazet
2010-11-15 15:10               ` Chris Metcalf
2010-11-22  5:39             ` Cypher Wu
2010-11-22 13:35               ` Chris Metcalf
2010-11-23  1:36                 ` Cypher Wu
2010-11-23 21:02                   ` Chris Metcalf
2010-11-24  2:53                     ` Cypher Wu
2010-11-24 14:09                       ` Chris Metcalf
2010-11-24 16:37                         ` Cypher Wu
2010-11-13 22:52 ` Kernel rwlock design, Multicore and IGMP Peter Zijlstra

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=201011151425.oAFEPU3W005682@farm-0010.internal.tilera.com \
    --to=cmetcalf@tilera.com \
    --cc=cypher.w@gmail.com \
    --cc=eric.dumazet@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=xiyou.wangcong@gmail.com \
    /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

all inboxes | Powered by JetHome®