* Re: [PATCH net-next-2.6] inetpeer: remove unused list
From: Eric Dumazet @ 2011-06-09 3:47 UTC (permalink / raw)
To: David Miller; +Cc: tim.c.chen, andi, netdev
In-Reply-To: <20110608.170557.2273990802892743973.davem@davemloft.net>
Le mercredi 08 juin 2011 à 17:05 -0700, David Miller a écrit :
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Thu, 09 Jun 2011 01:35:34 +0200
>
> > Andi Kleen and Tim Chen reported huge contention on inetpeer
> > unused_peers.lock, on memcached workload on a 40 core machine, with
> > disabled route cache.
> >
> > It appears we constantly flip peers refcnt between 0 and 1 values, and
> > we must insert/remove peers from unused_peers.list, holding a contended
> > spinlock.
> >
> > Remove this list completely and perform a garbage collection on-the-fly,
> > at lookup time, using the expired nodes we met during the tree
> > traversal.
> >
> > This removes a lot of code, makes locking more standard, and obsoletes
> > two sysctls (inet_peer_gc_mintime and inet_peer_gc_maxtime). This also
> > removes two pointers in inet_peer structure.
> >
> > There is still a false sharing effect because refcnt is in first cache
> > line of object [were the links and keys used by lookups are located], we
> > might move it at the end of inet_peer structure to let this first cache
> > line mostly read by cpus.
> >
> > Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
>
> Didn't expect you to implement this so fast :-)
>
> Applied, thanks!
Thanks David
I would be glad to hear about numbers from Andi and Tim, I am very
jealous of their 80 Threads machine ;)
^ permalink raw reply
* Re: ipv6 hitting route max_size
From: Simon Kirby @ 2011-06-09 4:40 UTC (permalink / raw)
To: David Miller; +Cc: netdev, yoshfuji
In-Reply-To: <20110607.005645.1883989402770213985.davem@davemloft.net>
On Tue, Jun 07, 2011 at 12:56:45AM -0700, David Miller wrote:
> From: Simon Kirby <sim@hostway.ca>
> Date: Mon, 6 Jun 2011 16:15:21 -0700
>
> > Ok, makes sense, but the result is now that ipv4 loads a full Internet
> > table with no adjustments, while ipv6 does not. Would it make sense to
> > change 4096 to 1048576, or would it be better to count only clones of
> > the actual route or something along those lines?
>
> Simon can you give this patch a try?
Didn't apply to 2.6.39, so I tried 3.0-rc2, but I get an Oops when
running the example reproduction case I gave before (
for ((i = 0;i < 4200;i++)); do ip route add unreachable 2000::$i; done
) both with and without your patch applied:
BUG: unable to handle kernel NULL pointer dereference at 00000000000000a0
IP: [<ffffffff8143e2b7>] ip6_route_add+0xe7/0x6b0
PGD 3ed7c8067 PUD 3ed5a1067 PMD 0
Oops: 0002 [#1] SMP
CPU 0
Modules linked in: nf_conntrack_ipv6 nf_defrag_ipv6 xt_state nf_conntrack_ipv4 nf_defrag_ipv4 nf_conntrack tg3 e100 libphy
Pid: 8932, comm: ip Not tainted 3.0.0-rc2-amd64-net #1 To Be Filled By O.E.M. To Be Filled By O.E.M./TYAN High-End Dual AMD Opteron, S2882
RIP: 0010:[<ffffffff8143e2b7>] [<ffffffff8143e2b7>] ip6_route_add+0xe7/0x6b0
RSP: 0018:ffff8803e59939f8 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff8803e5993a58 RCX: 0000000000000038
RDX: 00000000000000a0 RSI: 0000000000000008 RDI: 00000000000000a0
RBP: ffffffff817b3300 R08: ffffffff816c8980 R09: 0000000000000000
R10: 0000000000000001 R11: dead000000200200 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 00000000fffffff4
FS: 00007f5f11908700(0000) GS:ffff8803ffc00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000000000a0 CR3: 00000003edfdd000 CR4: 00000000000006f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process ip (pid: 8932, threadinfo ffff8803e5992000, task ffff8803ed70dfa0)
Stack:
0000000000000000 0000000000000000 ffff8803eecd2a80 0000000000000000
0000000000000000 ffff8803eedabe00 ffff8803ee671600 ffffffff813b8a30
ffff8803fe00ac00 ffff8803e5993b50 0000000000000000 ffffffff8143e89c
Call Trace:
[<ffffffff813b8a30>] ? rtnetlink_rcv+0x30/0x30
[<ffffffff8143e89c>] ? inet6_rtm_newroute+0x1c/0x30
[<ffffffff813cb3b9>] ? netlink_rcv_skb+0x89/0xb0
[<ffffffff813b8a1f>] ? rtnetlink_rcv+0x1f/0x30
[<ffffffff813cb013>] ? netlink_unicast+0x283/0x2d0
[<ffffffff813cb930>] ? netlink_sendmsg+0x230/0x390
[<ffffffff8139639b>] ? sock_sendmsg+0xab/0xe0
[<ffffffff810925eb>] ? __alloc_pages_nodemask+0x10b/0x700
[<ffffffff810a3fc2>] ? __do_fault+0x3e2/0x4c0
[<ffffffff81395b9e>] ? move_addr_to_kernel+0x2e/0x40
[<ffffffff813a1fd9>] ? verify_iovec+0x69/0xd0
[<ffffffff813972e2>] ? __sys_sendmsg+0x172/0x300
[<ffffffff81027465>] ? do_page_fault+0x1a5/0x430
[<ffffffff813cb6be>] ? netlink_autobind+0x8e/0xd0
[<ffffffff81395bfc>] ? move_addr_to_user+0x4c/0x60
[<ffffffff81396f55>] ? sys_getsockname+0xd5/0xe0
[<ffffffff81397634>] ? sys_sendmsg+0x44/0x80
[<ffffffff814a35bb>] ? system_call_fastpath+0x16/0x1b
Code: 31 c9 31 d2 45 31 c0 31 f6 41 bf f4 ff ff ff e8 b0 2d f7 ff 48 8d 90 a0 00 00 00 49 89 c4 b9 38 00 00 00 31 c0 4d 85 e4 48 89 d7 <f3> ab 0f 84 06 03 00 00 66 41 c7 44 24 6a ff ff 31 c0 f6 43 16
RIP [<ffffffff8143e2b7>] ip6_route_add+0xe7/0x6b0
RSP <ffff8803e59939f8>
CR2: 00000000000000a0
---[ end trace 370907621d87fefc ]---
I don't see many changes to ip6_route_add other than c3968a857a6b6c3.
Checking shortly once I get a git tree on this box, but no ipmi and I'm
remote at the moment.
Btw, maybe rt6_alloc_clone or rt6_alloc_cow needs to clear the DST_NOCOUNT
flag from rt->dst.flags for it to count any of them? Didn't verify.
Simon-
^ permalink raw reply
* Re: INFO: suspicious rcu_dereference_check() usage.
From: Justin Mattock @ 2011-06-09 4:49 UTC (permalink / raw)
To: Dave Jones, linux-kernel@vger.kernel.org, netdev
In-Reply-To: <4DDBDB4D.4080606@gmail.com>
this fired off again.. fpasted dmesg.
http://fpaste.org/9bs4/
just pulled the laste Mainline will run and see, will send anything
forward if I see anything.
Note: I am off the lists at the moment due to a transistion, Cc's me
if anything.
--
Justin P. Mattock
^ permalink raw reply
* linux-next: build failure after merge of the net tree
From: Stephen Rothwell @ 2011-06-09 4:56 UTC (permalink / raw)
To: David Miller, netdev; +Cc: linux-next, linux-kernel, Greg KH, Alexey Dobriyan
Hi all,
After merging the net tree, today's linux-next build (x86_64 allmodconfig)
failed like this:
[See summary in patch below.]
[Hmmm, suddenly allmodconfig is building stuff in staging? Ah ha!
Commit fe35a59e16fb ("Staging: remove STAGING_EXCLUDE_BUILD
option") Greg, a bit of warning might have been nice :-( ]
I assume that this is fallout from commit a6b7a407865a ("net: remove
interrupt.h inclusion from netdevice.h").
I added the following patch for today:
>From 520e46c255afa851b33e5c4a2bcf2112ecccbb6f Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Thu, 9 Jun 2011 14:19:09 +1000
Subject: [PATCH] net/staging: add needed interrupt.h and hardirq.h includes
Made necessary by commit a6b7a407865a ("net: remove
interrupt.h inclusion from netdevice.h").
Fixes these build errors:
drivers/staging/brcm80211/brcmfmac/dhd_linux.c:1212:3: error: implicit declaration of function 'in_interrupt'
drivers/staging/brcm80211/brcmfmac/dhd_linux.c:1390:4: error: implicit declaration of function 'tasklet_schedule'
drivers/staging/brcm80211/brcmfmac/dhd_linux.c:1988:3: error: implicit declaration of function 'tasklet_init'
drivers/staging/brcm80211/brcmfmac/dhd_linux.c:2288:5: error: implicit declaration of function 'tasklet_kill'
drivers/staging/brcm80211/brcmfmac/dhd_linux.c:232:24: error: field 'tasklet' has incomplete type
drivers/staging/brcm80211/brcmfmac/wl_iw.c:3670:2: error: implicit declaration of function 'in_interrupt'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:1349:3: error: implicit declaration of function 'free_irq'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:1352:2: error: implicit declaration of function 'tasklet_kill'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:1538:4: error: implicit declaration of function 'tasklet_schedule'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:765:2: error: implicit declaration of function 'tasklet_init'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:817:2: error: implicit declaration of function 'request_irq'
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c:817:31: error: 'IRQF_SHARED' undeclared (first use in this function)
drivers/staging/brcm80211/brcmsmac/wl_mac80211.h:65:24: error: field 'tasklet' has incomplete type
drivers/staging/rtl8187se/ieee80211/ieee80211.h:994:24: error: field 'ps_task' has incomplete type
drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c:1790:3: error: implicit declaration of function 'tasklet_schedule'
drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c:2590:2: error: implicit declaration of function 'tasklet_init'
drivers/staging/rtl8187se/r8180.h:461:24: error: field 'irq_rx_tasklet' has incomplete type
drivers/staging/rtl8187se/r8180_core.c:2764:2: error: implicit declaration of function 'tasklet_init'
drivers/staging/rtl8187se/r8180_core.c:2954:2: error: implicit declaration of function 'request_irq'
drivers/staging/rtl8187se/r8180_core.c:2954:55: error: 'IRQF_SHARED' undeclared (first use in this function)
drivers/staging/rtl8187se/r8180_core.c:3641:4: error: implicit declaration of function 'free_irq'
drivers/staging/rtl8187se/r8180_core.c:4026:3: error: implicit declaration of function 'tasklet_schedule'
drivers/staging/rtl8192e/ieee80211/ieee80211.h:2000:24: error: field 'ps_task' has incomplete type
drivers/staging/rtl8192e/r8192E.h:909:24: error: field 'irq_rx_tasklet' has incomplete type
drivers/staging/rtl8192e/r8192E.h:910:24: error: field 'irq_tx_tasklet' has incomplete type
drivers/staging/rtl8192e/r8192E.h:911:31: error: field 'irq_prepare_beacon_tasklet' has incomplete type
drivers/staging/rtl8192e/r8192E_core.c:2008:2: error: implicit declaration of function 'tasklet_init'
drivers/staging/rtl8192e/r8192E_core.c:2412:54: error: 'IRQF_SHARED' undeclared (first use in this function)
drivers/staging/rtl8192e/r8192E_core.c:2412:9: error: implicit declaration of function 'request_irq'
drivers/staging/rtl8192e/r8192E_core.c:4640:4: error: implicit declaration of function 'free_irq'
drivers/staging/rtl8192e/r8192E_core.c:620:2: error: implicit declaration of function 'synchronize_irq'
drivers/staging/rtl8192e/r8192E_core.c:913:3: error: implicit declaration of function 'tasklet_schedule'
drivers/staging/rtl8192u/ieee80211/ieee80211.h:1993:24: error: field 'ps_task' has incomplete type
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
drivers/staging/brcm80211/brcmfmac/dhd_linux.c | 2 ++
drivers/staging/brcm80211/brcmfmac/wl_iw.c | 1 +
drivers/staging/brcm80211/brcmsmac/wl_mac80211.c | 1 +
drivers/staging/brcm80211/brcmsmac/wl_mac80211.h | 2 ++
drivers/staging/rtl8187se/ieee80211/ieee80211.h | 1 +
.../rtl8187se/ieee80211/ieee80211_softmac.c | 1 +
drivers/staging/rtl8187se/r8180.h | 1 +
drivers/staging/rtl8187se/r8180_core.c | 1 +
drivers/staging/rtl8192e/ieee80211/ieee80211.h | 1 +
drivers/staging/rtl8192e/r8192E.h | 1 +
drivers/staging/rtl8192e/r8192E_core.c | 2 ++
drivers/staging/rtl8192u/ieee80211/ieee80211.h | 1 +
12 files changed, 15 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/brcm80211/brcmfmac/dhd_linux.c b/drivers/staging/brcm80211/brcmfmac/dhd_linux.c
index f356c56..09957bd 100644
--- a/drivers/staging/brcm80211/brcmfmac/dhd_linux.c
+++ b/drivers/staging/brcm80211/brcmfmac/dhd_linux.c
@@ -31,6 +31,8 @@
#include <linux/fcntl.h>
#include <linux/fs.h>
#include <linux/uaccess.h>
+#include <linux/interrupt.h>
+#include <linux/hardirq.h>
#include <bcmdefs.h>
#include <bcmutils.h>
diff --git a/drivers/staging/brcm80211/brcmfmac/wl_iw.c b/drivers/staging/brcm80211/brcmfmac/wl_iw.c
index 15e1b05..35eec91 100644
--- a/drivers/staging/brcm80211/brcmfmac/wl_iw.c
+++ b/drivers/staging/brcm80211/brcmfmac/wl_iw.c
@@ -18,6 +18,7 @@
#include <linux/semaphore.h>
#include <bcmdefs.h>
#include <linux/netdevice.h>
+#include <linux/hardirq.h>
#include <wlioctl.h>
#include <bcmutils.h>
diff --git a/drivers/staging/brcm80211/brcmsmac/wl_mac80211.c b/drivers/staging/brcm80211/brcmsmac/wl_mac80211.c
index 6c6236c..8261229 100644
--- a/drivers/staging/brcm80211/brcmsmac/wl_mac80211.c
+++ b/drivers/staging/brcm80211/brcmsmac/wl_mac80211.c
@@ -24,6 +24,7 @@
#include <linux/pci.h>
#include <linux/sched.h>
#include <linux/firmware.h>
+#include <linux/interrupt.h>
#include <net/mac80211.h>
#include <proto/802.11.h>
diff --git a/drivers/staging/brcm80211/brcmsmac/wl_mac80211.h b/drivers/staging/brcm80211/brcmsmac/wl_mac80211.h
index e703d8b..f7a58b7 100644
--- a/drivers/staging/brcm80211/brcmsmac/wl_mac80211.h
+++ b/drivers/staging/brcm80211/brcmsmac/wl_mac80211.h
@@ -17,6 +17,8 @@
#ifndef _wl_mac80211_h_
#define _wl_mac80211_h_
+#include <linux/interrupt.h>
+
/* BMAC Note: High-only driver is no longer working in softirq context as it needs to block and
* sleep so perimeter lock has to be a semaphore instead of spinlock. This requires timers to be
* submitted to workqueue instead of being on kernel timer
diff --git a/drivers/staging/rtl8187se/ieee80211/ieee80211.h b/drivers/staging/rtl8187se/ieee80211/ieee80211.h
index 16aa6a8..4384d93 100644
--- a/drivers/staging/rtl8187se/ieee80211/ieee80211.h
+++ b/drivers/staging/rtl8187se/ieee80211/ieee80211.h
@@ -32,6 +32,7 @@
#include <linux/semaphore.h>
#include <linux/wireless.h>
#include <linux/ieee80211.h>
+#include <linux/interrupt.h>
#define KEY_TYPE_NA 0x0
#define KEY_TYPE_WEP40 0x1
diff --git a/drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c b/drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c
index 736a140..00ee02f 100644
--- a/drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c
+++ b/drivers/staging/rtl8187se/ieee80211/ieee80211_softmac.c
@@ -20,6 +20,7 @@
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/version.h>
+#include <linux/interrupt.h>
#include <asm/uaccess.h>
#include "dot11d.h"
diff --git a/drivers/staging/rtl8187se/r8180.h b/drivers/staging/rtl8187se/r8180.h
index d15bdf6..a2c46ae 100644
--- a/drivers/staging/rtl8187se/r8180.h
+++ b/drivers/staging/rtl8187se/r8180.h
@@ -18,6 +18,7 @@
#ifndef R8180H
#define R8180H
+#include <linux/interrupt.h>
#define RTL8180_MODULE_NAME "r8180"
#define DMESG(x,a...) printk(KERN_INFO RTL8180_MODULE_NAME ": " x "\n", ## a)
diff --git a/drivers/staging/rtl8187se/r8180_core.c b/drivers/staging/rtl8187se/r8180_core.c
index 2155a77..bae7d85 100644
--- a/drivers/staging/rtl8187se/r8180_core.c
+++ b/drivers/staging/rtl8187se/r8180_core.c
@@ -33,6 +33,7 @@
#include <linux/slab.h>
#include <linux/syscalls.h>
#include <linux/eeprom_93cx6.h>
+#include <linux/interrupt.h>
#include "r8180_hw.h"
#include "r8180.h"
diff --git a/drivers/staging/rtl8192e/ieee80211/ieee80211.h b/drivers/staging/rtl8192e/ieee80211/ieee80211.h
index dbe21ab..82bc59a 100644
--- a/drivers/staging/rtl8192e/ieee80211/ieee80211.h
+++ b/drivers/staging/rtl8192e/ieee80211/ieee80211.h
@@ -31,6 +31,7 @@
#include <linux/timer.h>
#include <linux/sched.h>
#include <linux/semaphore.h>
+#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/wireless.h>
diff --git a/drivers/staging/rtl8192e/r8192E.h b/drivers/staging/rtl8192e/r8192E.h
index 0229031..89fe8fc 100644
--- a/drivers/staging/rtl8192e/r8192E.h
+++ b/drivers/staging/rtl8192e/r8192E.h
@@ -36,6 +36,7 @@
#include <linux/if_arp.h>
#include <linux/random.h>
#include <linux/version.h>
+#include <linux/interrupt.h>
#include <asm/io.h>
#include "ieee80211/rtl819x_HT.h"
#include "ieee80211/ieee80211.h"
diff --git a/drivers/staging/rtl8192e/r8192E_core.c b/drivers/staging/rtl8192e/r8192E_core.c
index 58d800f..19a9a07 100644
--- a/drivers/staging/rtl8192e/r8192E_core.c
+++ b/drivers/staging/rtl8192e/r8192E_core.c
@@ -27,6 +27,8 @@
#include <linux/vmalloc.h>
#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/hardirq.h>
#include <asm/uaccess.h>
#include "r8192E_hw.h"
#include "r8192E.h"
diff --git a/drivers/staging/rtl8192u/ieee80211/ieee80211.h b/drivers/staging/rtl8192u/ieee80211/ieee80211.h
index e716f7b..2333257 100644
--- a/drivers/staging/rtl8192u/ieee80211/ieee80211.h
+++ b/drivers/staging/rtl8192u/ieee80211/ieee80211.h
@@ -31,6 +31,7 @@
#include <linux/timer.h>
#include <linux/sched.h>
#include <linux/semaphore.h>
+#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/wireless.h>
--
1.7.5.3
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
^ permalink raw reply related
* Re: linux-next: build failure after merge of the net tree
From: Greg K-H @ 2011-06-09 5:17 UTC (permalink / raw)
To: Stephen Rothwell
Cc: David Miller, <netdev@vger.kernel.org>,
linux-next@vger.kernel.org, linux-kernel@vger.kernel.org,
Alexey Dobriyan
In-Reply-To: <20110609145657.96fae0d0.sfr@canb.auug.org.au>
On Jun 8, 2011, at 9:56 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> Hi all,
>
> After merging the net tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
>
> [See summary in patch below.]
>
> [Hmmm, suddenly allmodconfig is building stuff in staging? Ah ha!
> Commit fe35a59e16fb ("Staging: remove STAGING_EXCLUDE_BUILD
> option") Greg, a bit of warning might have been nice :-( ]
>
Sorry, this is something that Linus asked for and will be going to him
tomorrow with the rest of the staging-linus branch.
greg k-h
^ permalink raw reply
* RE: [Pv-drivers] [PATCH 1/9] vmxnet3: convert to 64 bit stats interface
From: Scott Goldman @ 2011-06-09 5:17 UTC (permalink / raw)
To: Stephen Hemminger, David S. Miller, Shreyas Bhatewara
Cc: pv-drivers@vmware.com, netdev@vger.kernel.org
In-Reply-To: <20110609005417.324633042@vyatta.com>
> Convert vmxnet3 driver to 64 bit statistics interface.
> This driver was already counting packet per queue in a 64 bit value so not
> a huge change. Eliminate unused old net_device_stats structure.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Thanks for taking care of this.
Signed-off-by: Scott J. Goldman <scottjg@vmware.com>
^ permalink raw reply
* Re: [PATCH net-next-2.6] inetpeer: remove unused list
From: David Miller @ 2011-06-09 5:22 UTC (permalink / raw)
To: eric.dumazet; +Cc: tim.c.chen, andi, netdev
In-Reply-To: <1307591254.3980.32.camel@edumazet-laptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Thu, 09 Jun 2011 05:47:34 +0200
> I would be glad to hear about numbers from Andi and Tim, I am very
> jealous of their 80 Threads machine ;)
Don't be, you can play on my 128 thread machine any time you want :-)
^ permalink raw reply
* Re: [PATCH net-next-2.6] inetpeer: remove unused list
From: Eric Dumazet @ 2011-06-09 5:30 UTC (permalink / raw)
To: David Miller; +Cc: tim.c.chen, andi, netdev
In-Reply-To: <20110608.222251.1893803820744993013.davem@davemloft.net>
Le mercredi 08 juin 2011 à 22:22 -0700, David Miller a écrit :
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Thu, 09 Jun 2011 05:47:34 +0200
>
> > I would be glad to hear about numbers from Andi and Tim, I am very
> > jealous of their 80 Threads machine ;)
>
> Don't be, you can play on my 128 thread machine any time you want :-)
I believe they also have 40Gb NICs to play with huge memcached load, is
your machine still a candidate, with others in the lab to be used as
clients ? :)
^ permalink raw reply
* Re: linux-next: build failure after merge of the net tree
From: Stephen Rothwell @ 2011-06-09 5:39 UTC (permalink / raw)
To: Greg K-H
Cc: David Miller, <netdev@vger.kernel.org>,
linux-next@vger.kernel.org, linux-kernel@vger.kernel.org,
Alexey Dobriyan
In-Reply-To: <7CA9E653-DABC-470C-BCA6-905D6EBF9283@kroah.com>
[-- Attachment #1: Type: text/plain, Size: 694 bytes --]
Hi Greg,
On Wed, 8 Jun 2011 22:17:55 -0700 Greg K-H <greg@kroah.com> wrote:
>
> Sorry, this is something that Linus asked for and will be going to him
> tomorrow with the rest of the staging-linus branch.
Yeah, I remember the discussion, just not a conclusion. At least the
x86_64 allmodconfig builds worked before I merged the net tree, so that is
something.
The powerpc allyesconfig build is going to be interesting :-(
You might like to check the build results at
http://kisskb.ellerman.id.au/kisskb/branch/9/ sometime tomorrow. Or
point your janitors at it :-)
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/
[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply
* Re: [PATCH] vlan: Fix the ingress VLAN_FLAG_REORDER_HDR check v2
From: Jiri Pirko @ 2011-06-09 6:01 UTC (permalink / raw)
To: Changli Gao
Cc: Eric W. Biederman, David Miller, shemminger, greearb,
nicolas.2p.debian, netdev, kaber, fubar, eric.dumazet, andy,
jesse
In-Reply-To: <BANLkTiksNQ-f=qN4PhPP4qX0Saiwb-hY1A@mail.gmail.com>
Thu, Jun 09, 2011 at 01:08:13AM CEST, xiaosuo@gmail.com wrote:
>On Thu, Jun 9, 2011 at 12:28 AM, Jiri Pirko <jpirko@redhat.com> wrote:
>>
>> Why we can't remove this right away?
>>
>
>I think the reason is this patch is a bug fix for the stable kernel,
>so we should not introduce new features or remove old features.
But for net-next if should be ok, right?
What's the reason for possibility of removing the reorder flag? Why
would user want do it?
Thanks
Jirka
>
>--
>Regards,
>Changli Gao(xiaosuo@gmail.com)
^ permalink raw reply
* RE: [PATCH 5/9] netxen: convert to 64 bit statistics
From: Amit Salecha @ 2011-06-09 6:13 UTC (permalink / raw)
To: Stephen Hemminger, David Miller; +Cc: netdev
In-Reply-To: <20110609005417.572148247@vyatta.com>
> From: Stephen Hemminger [mailto:shemminger@vyatta.com]
>
> Change to 64 bit statistics interface, driver was already maintaining
> 64 bit
> value.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
>
Acked-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Thanks Stephen.
^ permalink raw reply
* [PATCH 4/5 net-next] cnic: Improve NETDEV_UP event handling
From: Michael Chan @ 2011-06-09 5:29 UTC (permalink / raw)
To: davem; +Cc: netdev
In-Reply-To: <1307597376-5379-3-git-send-email-mchan@broadcom.com>
During NETDEV_UP, we use symbol_get() to get the net driver's cnic
probe function. This sometimes doesn't work if NETDEV_UP happens
right after NETDEV_REGISTER and the net driver is still running module
init code. As a result, the cnic device may not be discovered. We
fix this by probing on all NETDEV events if the device's netif_running
state is up.
Signed-off-by: Michael Chan <mchan@broadcom.com>
---
drivers/net/cnic.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/cnic.c b/drivers/net/cnic.c
index a529bde..6c544b3 100644
--- a/drivers/net/cnic.c
+++ b/drivers/net/cnic.c
@@ -5342,7 +5342,7 @@ static int cnic_netdev_event(struct notifier_block *this, unsigned long event,
dev = cnic_from_netdev(netdev);
- if (!dev && (event == NETDEV_REGISTER || event == NETDEV_UP)) {
+ if (!dev && (event == NETDEV_REGISTER || netif_running(netdev))) {
/* Check for the hot-plug device */
dev = is_cnic_dev(netdev);
if (dev) {
@@ -5358,7 +5358,7 @@ static int cnic_netdev_event(struct notifier_block *this, unsigned long event,
else if (event == NETDEV_UNREGISTER)
cnic_ulp_exit(dev);
- if (event == NETDEV_UP) {
+ if (event == NETDEV_UP || (new_dev && netif_running(netdev))) {
if (cnic_register_netdev(dev) != 0) {
cnic_put(dev);
goto done;
--
1.6.4.GIT
^ permalink raw reply related
* [PATCH 5/5 net-next] cnic, bnx2: Check iSCSI support early in bnx2_init_one()
From: Michael Chan @ 2011-06-09 5:29 UTC (permalink / raw)
To: davem; +Cc: netdev, Neil Horman
In-Reply-To: <1307597376-5379-4-git-send-email-mchan@broadcom.com>
Based on earlier patch from Neil Horman <nhorman@tuxdriver.com>
If iSCSI is not supported on a bnx2 device, bnx2_cnic_probe() will
return NULL and the cnic device will not be visible to bnx2i. This
will prevent bnx2i from registering and then unregistering during
cnic_start() and cause the warning message:
bnx2 0003:01:00.1: eth1: Failed waiting for ULP up call to complete
Signed-off-by: Michael Chan <mchan@broadcom.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
---
drivers/net/bnx2.c | 7 +++++++
drivers/net/cnic.c | 12 ++----------
2 files changed, 9 insertions(+), 10 deletions(-)
diff --git a/drivers/net/bnx2.c b/drivers/net/bnx2.c
index 57d3293..74580bb 100644
--- a/drivers/net/bnx2.c
+++ b/drivers/net/bnx2.c
@@ -416,6 +416,9 @@ struct cnic_eth_dev *bnx2_cnic_probe(struct net_device *dev)
struct bnx2 *bp = netdev_priv(dev);
struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
+ if (!cp->max_iscsi_conn)
+ return NULL;
+
cp->drv_owner = THIS_MODULE;
cp->chip_id = bp->chip_id;
cp->pdev = bp->pdev;
@@ -8177,6 +8180,10 @@ bnx2_init_board(struct pci_dev *pdev, struct net_device *dev)
bp->timer.data = (unsigned long) bp;
bp->timer.function = bnx2_timer;
+#ifdef BCM_CNIC
+ bp->cnic_eth_dev.max_iscsi_conn =
+ bnx2_reg_rd_ind(bp, BNX2_FW_MAX_ISCSI_CONN);
+#endif
pci_save_state(pdev);
return 0;
diff --git a/drivers/net/cnic.c b/drivers/net/cnic.c
index 6c544b3..363c7f3 100644
--- a/drivers/net/cnic.c
+++ b/drivers/net/cnic.c
@@ -4225,14 +4225,6 @@ static void cnic_enable_bnx2_int(struct cnic_dev *dev)
BNX2_PCICFG_INT_ACK_CMD_INDEX_VALID | cp->last_status_idx);
}
-static void cnic_get_bnx2_iscsi_info(struct cnic_dev *dev)
-{
- u32 max_conn;
-
- max_conn = cnic_reg_rd_ind(dev, BNX2_FW_MAX_ISCSI_CONN);
- dev->max_iscsi_conn = max_conn;
-}
-
static void cnic_disable_bnx2_int_sync(struct cnic_dev *dev)
{
struct cnic_local *cp = dev->cnic_priv;
@@ -4557,8 +4549,6 @@ static int cnic_start_bnx2_hw(struct cnic_dev *dev)
return err;
}
- cnic_get_bnx2_iscsi_info(dev);
-
return 0;
}
@@ -5224,6 +5214,8 @@ static struct cnic_dev *init_bnx2_cnic(struct net_device *dev)
cdev->pcidev = pdev;
cp->chip_id = ethdev->chip_id;
+ cdev->max_iscsi_conn = ethdev->max_iscsi_conn;
+
cp->cnic_ops = &cnic_bnx2_ops;
cp->start_hw = cnic_start_bnx2_hw;
cp->stop_hw = cnic_stop_bnx2_hw;
--
1.6.4.GIT
^ permalink raw reply related
* [PATCH 1/5 net-next] cnic: Fix interrupt logic
From: Michael Chan @ 2011-06-09 5:29 UTC (permalink / raw)
To: davem; +Cc: netdev
We need to keep looping until cnic_get_kcqes() returns 0. cnic_get_kcqes()
returns a maximum of 64 entries. If there are more entries in the queue
and we don't loop back, the remaining entries may not be serviced for a
long time.
Signed-off-by: Michael Chan <mchan@broadcom.com>
---
drivers/net/cnic.c | 13 ++++---------
1 files changed, 4 insertions(+), 9 deletions(-)
diff --git a/drivers/net/cnic.c b/drivers/net/cnic.c
index 11a92af..5c0a669 100644
--- a/drivers/net/cnic.c
+++ b/drivers/net/cnic.c
@@ -2778,13 +2778,10 @@ static u32 cnic_service_bnx2_queues(struct cnic_dev *dev)
/* Tell compiler that status_blk fields can change. */
barrier();
- if (status_idx != *cp->kcq1.status_idx_ptr) {
- status_idx = (u16) *cp->kcq1.status_idx_ptr;
- /* status block index must be read first */
- rmb();
- cp->kwq_con_idx = *cp->kwq_con_idx_ptr;
- } else
- break;
+ status_idx = (u16) *cp->kcq1.status_idx_ptr;
+ /* status block index must be read first */
+ rmb();
+ cp->kwq_con_idx = *cp->kwq_con_idx_ptr;
}
CNIC_WR16(dev, cp->kcq1.io_addr, cp->kcq1.sw_prod_idx);
@@ -2908,8 +2905,6 @@ static u32 cnic_service_bnx2x_kcq(struct cnic_dev *dev, struct kcq_info *info)
/* Tell compiler that sblk fields can change. */
barrier();
- if (last_status == *info->status_idx_ptr)
- break;
last_status = *info->status_idx_ptr;
/* status block index must be read before reading the KCQ */
--
1.6.4.GIT
^ permalink raw reply related
* [PATCH 2/5 net-next] cnic: Fix race conditions with firmware
From: Michael Chan @ 2011-06-09 5:29 UTC (permalink / raw)
To: davem; +Cc: netdev
In-Reply-To: <1307597376-5379-1-git-send-email-mchan@broadcom.com>
During iSCSI connection terminations, if the target is also terminating
at about the same time, the firmware may not complete the driver's
request to close or reset the connection. This is fixed by handling
other events (instead of the expected completion event) as an indication
that the driver's request has been rejected.
Signed-off-by: Michael Chan <mchan@broadcom.com>
---
drivers/net/cnic.c | 14 +++++++++++---
1 files changed, 11 insertions(+), 3 deletions(-)
diff --git a/drivers/net/cnic.c b/drivers/net/cnic.c
index 5c0a669..ea7245c 100644
--- a/drivers/net/cnic.c
+++ b/drivers/net/cnic.c
@@ -3767,7 +3767,13 @@ static void cnic_cm_process_kcqe(struct cnic_dev *dev, struct kcqe *kcqe)
break;
case L4_KCQE_OPCODE_VALUE_CLOSE_RECEIVED:
- cnic_cm_upcall(cp, csk, opcode);
+ /* after we already sent CLOSE_REQ */
+ if (test_bit(CNIC_F_BNX2X_CLASS, &dev->flags) &&
+ !test_bit(SK_F_OFFLD_COMPLETE, &csk->flags) &&
+ csk->state == L4_KCQE_OPCODE_VALUE_CLOSE_COMP)
+ cp->close_conn(csk, L4_KCQE_OPCODE_VALUE_RESET_COMP);
+ else
+ cnic_cm_upcall(cp, csk, opcode);
break;
}
csk_put(csk);
@@ -3821,12 +3827,14 @@ static int cnic_ready_to_close(struct cnic_sock *csk, u32 opcode)
}
/* 1. If event opcode matches the expected event in csk->state
- * 2. If the expected event is CLOSE_COMP, we accept any event
+ * 2. If the expected event is CLOSE_COMP or RESET_COMP, we accept any
+ * event
* 3. If the expected event is 0, meaning the connection was never
* never established, we accept the opcode from cm_abort.
*/
if (opcode == csk->state || csk->state == 0 ||
- csk->state == L4_KCQE_OPCODE_VALUE_CLOSE_COMP) {
+ csk->state == L4_KCQE_OPCODE_VALUE_CLOSE_COMP ||
+ csk->state == L4_KCQE_OPCODE_VALUE_RESET_COMP) {
if (!test_and_set_bit(SK_F_CLOSING, &csk->flags)) {
if (csk->state == 0)
csk->state = opcode;
--
1.6.4.GIT
^ permalink raw reply related
* [PATCH 3/5 net-next] cnic: Randomize initial TCP port for iSCSI connections
From: Michael Chan @ 2011-06-09 5:29 UTC (permalink / raw)
To: davem; +Cc: netdev, Eddie Wai
In-Reply-To: <1307597376-5379-2-git-send-email-mchan@broadcom.com>
From: Eddie Wai <eddie.wai@broadcom.com>
This reduces the likelihood of port re-use when re-loading the driver.
Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
---
drivers/net/cnic.c | 14 +++++++++-----
1 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/drivers/net/cnic.c b/drivers/net/cnic.c
index ea7245c..a529bde 100644
--- a/drivers/net/cnic.c
+++ b/drivers/net/cnic.c
@@ -605,11 +605,12 @@ static int cnic_unregister_device(struct cnic_dev *dev, int ulp_type)
}
EXPORT_SYMBOL(cnic_unregister_driver);
-static int cnic_init_id_tbl(struct cnic_id_tbl *id_tbl, u32 size, u32 start_id)
+static int cnic_init_id_tbl(struct cnic_id_tbl *id_tbl, u32 size, u32 start_id,
+ u32 next)
{
id_tbl->start = start_id;
id_tbl->max = size;
- id_tbl->next = 0;
+ id_tbl->next = next;
spin_lock_init(&id_tbl->lock);
id_tbl->table = kzalloc(DIV_ROUND_UP(size, 32) * 4, GFP_KERNEL);
if (!id_tbl->table)
@@ -3804,14 +3805,17 @@ static void cnic_cm_free_mem(struct cnic_dev *dev)
static int cnic_cm_alloc_mem(struct cnic_dev *dev)
{
struct cnic_local *cp = dev->cnic_priv;
+ u32 port_id;
cp->csk_tbl = kzalloc(sizeof(struct cnic_sock) * MAX_CM_SK_TBL_SZ,
GFP_KERNEL);
if (!cp->csk_tbl)
return -ENOMEM;
+ get_random_bytes(&port_id, sizeof(port_id));
+ port_id %= CNIC_LOCAL_PORT_RANGE;
if (cnic_init_id_tbl(&cp->csk_port_tbl, CNIC_LOCAL_PORT_RANGE,
- CNIC_LOCAL_PORT_MIN)) {
+ CNIC_LOCAL_PORT_MIN, port_id)) {
cnic_cm_free_mem(dev);
return -ENOMEM;
}
@@ -4829,7 +4833,7 @@ static int cnic_start_bnx2x_hw(struct cnic_dev *dev)
pfid = cp->pfid;
ret = cnic_init_id_tbl(&cp->cid_tbl, MAX_ISCSI_TBL_SZ,
- cp->iscsi_start_cid);
+ cp->iscsi_start_cid, 0);
if (ret)
return -ENOMEM;
@@ -4837,7 +4841,7 @@ static int cnic_start_bnx2x_hw(struct cnic_dev *dev)
if (BNX2X_CHIP_IS_E2(cp->chip_id)) {
ret = cnic_init_id_tbl(&cp->fcoe_cid_tbl,
BNX2X_FCOE_NUM_CONNECTIONS,
- cp->fcoe_start_cid);
+ cp->fcoe_start_cid, 0);
if (ret)
return -ENOMEM;
--
1.6.4.GIT
^ permalink raw reply related
* [PATCH net-next-2.6] inetpeer: lower false sharing effect
From: Eric Dumazet @ 2011-06-09 6:26 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Andi Kleen, Tim Chen
Profiles show false sharing in addr_compare() because refcnt/dtime
changes dirty the first inet_peer cache line, where are lying the keys
used at lookup time. If many cpus are calling inet_getpeer() and
inet_putpeer(), or need frag ids, addr_compare() is in 2nd position in
"perf top".
Before patch, my udpflood bench (16 threads) on my 2x4x2 machine :
5784.00 9.7% csum_partial_copy_generic [kernel]
3356.00 5.6% addr_compare [kernel]
2638.00 4.4% fib_table_lookup [kernel]
2625.00 4.4% ip_fragment [kernel]
1934.00 3.2% neigh_lookup [kernel]
1617.00 2.7% udp_sendmsg [kernel]
1608.00 2.7% __ip_route_output_key [kernel]
1480.00 2.5% __ip_append_data [kernel]
1396.00 2.3% kfree [kernel]
1195.00 2.0% kmem_cache_free [kernel]
1157.00 1.9% inet_getpeer [kernel]
1121.00 1.9% neigh_resolve_output [kernel]
1012.00 1.7% dev_queue_xmit [kernel]
# time ./udpflood.sh
real 0m44.511s
user 0m20.020s
sys 11m22.780s
# time ./udpflood.sh
real 0m44.099s
user 0m20.140s
sys 11m15.870s
After patch, no more addr_compare() in profiles :
4171.00 10.7% csum_partial_copy_generic [kernel]
1787.00 4.6% fib_table_lookup [kernel]
1756.00 4.5% ip_fragment [kernel]
1234.00 3.2% udp_sendmsg [kernel]
1191.00 3.0% neigh_lookup [kernel]
1118.00 2.9% __ip_append_data [kernel]
1022.00 2.6% kfree [kernel]
993.00 2.5% __ip_route_output_key [kernel]
841.00 2.2% neigh_resolve_output [kernel]
816.00 2.1% kmem_cache_free [kernel]
658.00 1.7% ia32_sysenter_target [kernel]
632.00 1.6% kmem_cache_alloc_node [kernel]
# time ./udpflood.sh
real 0m41.587s
user 0m19.190s
sys 10m36.370s
# time ./udpflood.sh
real 0m41.486s
user 0m19.290s
sys 10m33.650s
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Andi Kleen <andi@firstfloor.org>
CC: Tim Chen <tim.c.chen@linux.intel.com>
---
include/net/inetpeer.h | 24 +++++++++++++-----------
1 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/include/net/inetpeer.h b/include/net/inetpeer.h
index 1f0966f..39d1230 100644
--- a/include/net/inetpeer.h
+++ b/include/net/inetpeer.h
@@ -32,12 +32,17 @@ struct inet_peer {
struct inet_peer __rcu *avl_left, *avl_right;
struct inetpeer_addr daddr;
__u32 avl_height;
- __u32 dtime; /* the time of last use of not
- * referenced entries */
- atomic_t refcnt;
+
+ u32 metrics[RTAX_MAX];
+ u32 rate_tokens; /* rate limiting for ICMP */
+ unsigned long rate_last;
+ unsigned long pmtu_expires;
+ u32 pmtu_orig;
+ u32 pmtu_learned;
+ struct inetpeer_addr_base redirect_learned;
/*
* Once inet_peer is queued for deletion (refcnt == -1), following fields
- * are not available: rid, ip_id_count, tcp_ts, tcp_ts_stamp, metrics
+ * are not available: rid, ip_id_count, tcp_ts, tcp_ts_stamp
* We can share memory with rcu_head to help keep inet_peer small.
*/
union {
@@ -46,17 +51,14 @@ struct inet_peer {
atomic_t ip_id_count; /* IP ID for the next packet */
__u32 tcp_ts;
__u32 tcp_ts_stamp;
- u32 metrics[RTAX_MAX];
- u32 rate_tokens; /* rate limiting for ICMP */
- unsigned long rate_last;
- unsigned long pmtu_expires;
- u32 pmtu_orig;
- u32 pmtu_learned;
- struct inetpeer_addr_base redirect_learned;
};
struct rcu_head rcu;
struct inet_peer *gc_next;
};
+
+ /* following fields might be frequently dirtied */
+ __u32 dtime; /* the time of last use of not referenced entries */
+ atomic_t refcnt;
};
void inet_initpeers(void) __init;
^ permalink raw reply related
* Re: [PATCH 3/9] veth: convert to 64 bit statistics
From: David Miller @ 2011-06-09 6:26 UTC (permalink / raw)
To: shemminger; +Cc: netdev
In-Reply-To: <20110609005417.449670103@vyatta.com>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 08 Jun 2011 17:53:59 -0700
> Not much change, device was already keeping per cpu statistics.
> Use recent 64 statistics interface.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Applied.
^ permalink raw reply
* Re: [PATCH 8/9] niu: support 64 bit stats interface
From: David Miller @ 2011-06-09 6:27 UTC (permalink / raw)
To: shemminger; +Cc: netdev
In-Reply-To: <20110609005417.753634050@vyatta.com>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 08 Jun 2011 17:54:04 -0700
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Applied.
^ permalink raw reply
* Re: [PATCH 6/9] enic: update to support 64 bit stats
From: David Miller @ 2011-06-09 6:27 UTC (permalink / raw)
To: shemminger; +Cc: benve, vkolluri, roprabhu, dwang2, netdev
In-Reply-To: <20110609005417.630728228@vyatta.com>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 08 Jun 2011 17:54:02 -0700
> The device driver already uses 64 bit statistics, it just
> doesn't use the 64 bit interface.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Applied.
^ permalink raw reply
* Re: [PATCH 5/9] netxen: convert to 64 bit statistics
From: David Miller @ 2011-06-09 6:27 UTC (permalink / raw)
To: shemminger; +Cc: amit.salecha, netdev
In-Reply-To: <20110609005417.572148247@vyatta.com>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 08 Jun 2011 17:54:01 -0700
> Change to 64 bit statistics interface, driver was already maintaining 64 bit
> value.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Applied.
^ permalink raw reply
* Re: [PATCH 7/9] myricom: update to 64 bit stats
From: David Miller @ 2011-06-09 6:27 UTC (permalink / raw)
To: shemminger; +Cc: gallatin, brice, netdev
In-Reply-To: <20110609005417.688902862@vyatta.com>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Wed, 08 Jun 2011 17:54:03 -0700
> Driver was already keeping 64 bit counters, just not using the new interface.
>
> Ps: IMHO drivers should not be duplicating network device
> stats into ethtool stats. It is useless duplication.
>
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next-2.6] inetpeer: lower false sharing effect
From: David Miller @ 2011-06-09 6:31 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev, andi, tim.c.chen
In-Reply-To: <1307600810.3980.85.camel@edumazet-laptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Thu, 09 Jun 2011 08:26:50 +0200
> Profiles show false sharing in addr_compare() because refcnt/dtime
> changes dirty the first inet_peer cache line, where are lying the keys
> used at lookup time. If many cpus are calling inet_getpeer() and
> inet_putpeer(), or need frag ids, addr_compare() is in 2nd position in
> "perf top".
>
> Before patch, my udpflood bench (16 threads) on my 2x4x2 machine :
..
> After patch, no more addr_compare() in profiles :
..
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> CC: Andi Kleen <andi@firstfloor.org>
> CC: Tim Chen <tim.c.chen@linux.intel.com>
Applied, thanks!
^ permalink raw reply
* Re: [PATCH 1/5 net-next] cnic: Fix interrupt logic
From: David Miller @ 2011-06-09 6:52 UTC (permalink / raw)
To: mchan; +Cc: netdev
In-Reply-To: <1307597376-5379-1-git-send-email-mchan@broadcom.com>
From: "Michael Chan" <mchan@broadcom.com>
Date: Wed, 8 Jun 2011 22:29:32 -0700
> We need to keep looping until cnic_get_kcqes() returns 0. cnic_get_kcqes()
> returns a maximum of 64 entries. If there are more entries in the queue
> and we don't loop back, the remaining entries may not be serviced for a
> long time.
>
> Signed-off-by: Michael Chan <mchan@broadcom.com>
Applied.
^ permalink raw reply
* Re: [PATCH 2/5 net-next] cnic: Fix race conditions with firmware
From: David Miller @ 2011-06-09 6:53 UTC (permalink / raw)
To: mchan; +Cc: netdev
In-Reply-To: <1307597376-5379-2-git-send-email-mchan@broadcom.com>
From: "Michael Chan" <mchan@broadcom.com>
Date: Wed, 8 Jun 2011 22:29:33 -0700
> During iSCSI connection terminations, if the target is also terminating
> at about the same time, the firmware may not complete the driver's
> request to close or reset the connection. This is fixed by handling
> other events (instead of the expected completion event) as an indication
> that the driver's request has been rejected.
>
> Signed-off-by: Michael Chan <mchan@broadcom.com>
Applied.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox