David S. Miller a écrit :
From: Eric Dumazet [EMAIL PROTECTED]
Date: Wed, 08 Feb 2006 16:06:31 +0100
loopback driver carefully uses per_cpu storage for statistics but updates
loopback_dev.last_rx
This has been discussed before, this is an attribute every
driver must keep uptodate. Things
Eric Dumazet wrote:
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
--- a/include/linux/netdevice.h 2006-02-07 11:55:42.0 +0100
+++ b/include/linux/netdevice.h 2006-02-09 09:23:15.0 +0100
@@ -649,6 +649,19
Kristian Slavov wrote:
Hi,
The following patch seems to get the work done. Patch is against 2.6.15.3.
Here's the signed-off tag I forgot to add, should that matter anymore.
Signed-off-by: Kristian Slavov [EMAIL PROTECTED]
BR,
--
Krisu
-
To unsubscribe from this list: send the line
From: Eric Dumazet [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 09:00:21 +0100
+#ifdef CONFIG_SMP
+ /*
+ * In order to avoid cache line ping pongs on last_rx, we check
+ * if the device is a slave,
+ * and if last_rx really has to be modified
+ */
+ if (!(dev-flags
Herbert Xu a écrit :
David S. Miller [EMAIL PROTECTED] wrote:
From: Nicolas DICHTEL [EMAIL PROTECTED]
Date: Mon, 06 Feb 2006 12:00:30 +0100
in the same way of this patch, why dst_entry are stored for
RAW socket ? In case of specific IPSec rules for ICMPv6,
xfrm state can be different for
From: Nicolas DICHTEL [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 10:39:12 +0100
Yes, that's what I mean ;-) It was the same problem for not-connected
socket in UDP.
Ok, I see. Nicolas, can you resend your patch to me via
private email? Thanks.
-
To unsubscribe from this list: send the line
David S. Miller a écrit :
From: Eric Dumazet [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 09:00:21 +0100
+#ifdef CONFIG_SMP
+ /*
+* In order to avoid cache line ping pongs on last_rx, we check
+* if the device is a slave,
+* and if last_rx really has to be modified
+
Hello!
I am not aware of anyone using it
So, do not worry. The feature is dead. Sending to multiple groups
never was used in real life and now it just does not exist, as it should
happen with useless features.
It is sad, that dead nl_groups stays in struct sockaddr_nl, of course.
Alexey
-
To
From: Eric Dumazet [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 10:51:00 +0100
I'm sorry but last_rx is not available to User Land.
You need a kernel debugger to access it, or mess with /proc/kcore
There are tons of pieces of state tracked for net devices
which we don't export to userland.
This
On Thursday 09 February 2006 02:33, Jesse Brandeburg wrote:
My experience with alloc_page/put_page in the packet split e1000 code is
that it is slower to call alloc_page/put_page (they show up in top
10 oprofile)
Did you resolve it down to specific lines in alloc_page/put_page?
Perhaps they
On Thursday 09 February 2006 12:05, David S. Miller wrote:
From: Eric Dumazet [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 10:51:00 +0100
I'm sorry but last_rx is not available to User Land.
You need a kernel debugger to access it, or mess with /proc/kcore
There are tons of pieces of
On Thu, 2006-09-02 at 13:08 +0300, Alexey Kuznetsov wrote:
Hello!
I am not aware of anyone using it
So, do not worry. The feature is dead.
I misunderstood your question. I thought you were asking about the 32
groups.
Sending to multiple groups
never was used in real life and now it
Paul Mackerras writes:
And the solution is to treat it as a boolean instead?! I'm not sure
which is more ugly.
Why wouldn't explicit comparison against NULL be the preferred fix?
I just think this whole you shouldn't compare a pointer to 0 thing
is completely silly,
Indeed.
jamal wrote:
It is sad, that dead nl_groups stays in struct sockaddr_nl, of
course.
Its value has been highly reduced for sure.
Let me know your thoughts on the xcast idea.
All routing protocols suite implementations use it to select rtnetlink
multicast groups they want to listen in
Nework asynchronous IO.
This project was evolved from reeiving zero-copy support [1].
Network AIO is based on kevent and works as usual
kevent storage on top of inode.
When new socket is created it is associated with that inode
and when some activity is detected appropriate notifications
are
On Thu, 2006-09-02 at 15:18 +0200, Hasso Tepper wrote:
jamal wrote:
It is sad, that dead nl_groups stays in struct sockaddr_nl, of
course.
Its value has been highly reduced for sure.
Let me know your thoughts on the xcast idea.
All routing protocols suite implementations use it to
On Thursday 09 February 2006 14:53, John W. Linville wrote:
On Thu, Feb 09, 2006 at 12:33:11PM +0100, Andi Kleen wrote:
Same would probably apply to schedulers and classifiers. So I cannot
imagine a good use of this field. What does bonding use it for anyways?
ARP-based monitoring of
Andi Kleen a écrit :
On Thursday 09 February 2006 14:53, John W. Linville wrote:
On Thu, Feb 09, 2006 at 12:33:11PM +0100, Andi Kleen wrote:
Same would probably apply to schedulers and classifiers. So I cannot
imagine a good use of this field. What does bonding use it for anyways?
ARP-based
On Thu, 2006-09-02 at 15:21 +0100, Eric Dumazet wrote:
Andi Kleen a écrit :
On Thursday 09 February 2006 14:53, John W. Linville wrote:
On Thu, Feb 09, 2006 at 12:33:11PM +0100, Andi Kleen wrote:
Same would probably apply to schedulers and classifiers. So I cannot
imagine a good use of
David S. Miller a écrit :
There are tons of pieces of state tracked for net devices
which we don't export to userland.
This one is generally useful. Right now, it may be the case
that bonding is the only consumer in the kernel, but that
might not always be true. I can very easily envison
On Thu, Feb 09, 2006 at 04:56:11PM +0300, Evgeniy Polyakov wrote:
Hello.
I'm pleased to announce following projects:
1/2 - Kevent subsystem.
This subsystem incorporates several AIO/kqueue design notes and ideas.
Kevent can be used both for edge and level notifications. It
From: Ingo Oeser [EMAIL PROTECTED]
Here are some possible (and trivial) cleanups.
- use kzalloc() where possible
- remove unused label
- invert allocation failure test like
if (object) {
/* Rest of function here */
}
to
if (object == NULL)
return NULL;
/* Rest of
On Thu, Feb 09, 2006 at 04:16:54PM +0100, Eric Dumazet ([EMAIL PROTECTED])
wrote:
I had this idea after reading recent Van Jacobson paper about net channels.
I'm surprised you see the benefit of net channels (which first assumption
is about the 50-100 ns cost of a memory cache miss) but
Hello.
In article [EMAIL PROTECTED] (at Thu, 9 Feb 2006 17:36:16 +0100), Ingo Oeser
[EMAIL PROTECTED] says:
Here are some possible (and trivial) cleanups.
:
- remove unused label
@@ -3193,7 +3190,6 @@ static int inet6_fill_ifinfo(struct sk_b
kfree(array);
return skb-len;
Hello!
_I_ use it although i am not aware of anyone else (but since you put it
in libnetlink, i suspect there may be other people
No, really. It just was easy, but there were no specific applications.
It is important for listener to be able to subscribe to many groups,
but sending is
On Thu, 9 Feb 2006 19:42:17 +0300
Evgeniy Polyakov [EMAIL PROTECTED] wrote:
On Thu, Feb 09, 2006 at 04:16:54PM +0100, Eric Dumazet ([EMAIL PROTECTED])
wrote:
I had this idea after reading recent Van Jacobson paper about net channels.
I'm surprised you see the benefit of net channels
Begin forwarded message:
Date: Thu, 9 Feb 2006 06:46:33 -0800
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: [Bugme-new] [Bug 6036] New: mmap'ed write to socket hangs when
connection remote end is broken
http://bugzilla.kernel.org/show_bug.cgi?id=6036
Summary: mmap'ed
Hello, Jesse Brandeburg.
On 08.02.2006 23:07 you said the following:
whats the relevance of e1000?
I though Herbert had fixed these
Nope :( I had this messages on 2.6.14.2 and now I have it on 2.6.15.3.
--
Boris B. Zhmurov
mailto: [EMAIL PROTECTED]
wget
On 2/10/06, Boris B. Zhmurov [EMAIL PROTECTED] wrote:
Hello, Jesse Brandeburg.
On 08.02.2006 23:07 you said the following:
whats the relevance of e1000?
I though Herbert had fixed these
Nope :( I had this messages on 2.6.14.2 and now I have it on 2.6.15.3.
For what it's worth I had
Hello, Ian McDonald.
On 09.02.2006 22:25 you said the following:
Is it possible for you to download 2.6.16-rc2 or similar and see if it
goes away?
It'll be better, if I get only patch fixs that problem, not all 2.6.16-rc2.
--
Boris B. Zhmurov
mailto: [EMAIL PROTECTED]
wget
On Thu, Feb 09, 2006 at 10:59:23AM -0500, Benjamin LaHaise ([EMAIL PROTECTED])
wrote:
On Thu, Feb 09, 2006 at 04:56:11PM +0300, Evgeniy Polyakov wrote:
Hello.
I'm pleased to announce following projects:
1/2 - Kevent subsystem.
This subsystem incorporates several AIO/kqueue
The rcvbuf lock should probably be honored here.
Signed-off-by: John Heffner [EMAIL PROTECTED]
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -456,7 +456,8 @@ void tcp_rcv_space_adjust(struct sock *s
tp-rcvq_space.space =
On Thu, Feb 09, 2006 at 08:03:26PM +0300, Evgeniy Polyakov wrote:
It is completely different things.
The only common is that they _require_ some kind of notification
mechanism,
but none provide them.
epoll() can not be used for AIO and timers.
True, that is a disappointment. There really
On Thu, Feb 09, 2006 at 03:55:18PM -0500, Benjamin LaHaise wrote:
Actually, there are patches for generic thread based aio so that
other file descriptors can be used for AIO, perhaps they should be
dusted off?
Threads is exactly what we're trying to get away from, I think.
--L
-
To
This enables PnP support for the nsc-ircc chipset.
Since we can't fetch the chipset cfg_base from the PnP layer, we just use
the PnP information as one more hint when probing the chip.
Signed-off-by: Jean Tourrilhes [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git
This patch brings the nsc-ircc code to a more up to date power management
scheme, following the current device model.
Signed-off-by: Dmitry Torokhov [EMAIL PROTECTED]
Signed-off-by: Rudolf Marek [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git
This patch simply adds support for a variation of the nsc-ircc PC8739x
chipset, found in some IBM Thinkpad laptops.
Signed-off-by: Jean Tourrilhes [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git a/drivers/net/irda/nsc-ircc.c b/drivers/net/irda/nsc-ircc.c
index
This patch set IrDA's addr_len properly, i.e to 4 bytes, the size of the
IrLAP device address.
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git a/include/net/irda/irlap.h b/include/net/irda/irlap.h
index f55e86e..2127cae 100644
--- a/include/net/irda/irlap.h
+++
Hi Dave,
I'm currently working together with Jean for taking over the IrDA
maintainership, since he has no longer that much time to work on it. You
were looking for a cell phone maniac, here I am ;-)
So, next will come a set of 7 patches. Some of them are patches that have
been floating around
This patch replaces the deprecated sti/cli routines with the corresponding
spin_lock ones.
Signed-off-by: David Chosrova [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git a/drivers/net/irda/ep7211_ir.c b/drivers/net/irda/ep7211_ir.c
index 3189626..6edf988 100644
---
This patch fixes an out of range array access in irnet_irda.c.
Author: David Binderman [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
diff --git a/net/irda/irnet/irnet_irda.c b/net/irda/irnet/irnet_irda.c
index 07ec326..e477a8d 100644
--- a/net/irda/irnet/irnet_irda.c
+++
On Thu, Feb 09, 2006 at 10:59:23AM -0500, Benjamin LaHaise wrote:
On Thu, Feb 09, 2006 at 04:56:11PM +0300, Evgeniy Polyakov wrote:
Hello.
I'm pleased to announce following projects:
1/2 - Kevent subsystem.
This subsystem incorporates several AIO/kqueue design notes and ideas.
On Wed, Feb 08, 2006 at 08:06:33PM -0500, John W. Linville wrote:
On Mon, Jan 30, 2006 at 02:28:43PM -0800, Jean Tourrilhes wrote:
On Mon, Jan 30, 2006 at 01:28:44PM -0500, Dan Williams wrote:
I'll post a revert patch for the patch I originally sent so that we go
back to the original
Drivers using mii_check_media (via-rhine in particular) and also
forcing link parameters with ethtool can reach a state where the link
goes down and never comes back up. This is because mii_check_media
short-circuits early if mii-force_media != 0. This was discussed
in a couple of past threads,
On Thu, Feb 09, 2006 at 03:55:18PM -0500, Benjamin LaHaise ([EMAIL PROTECTED])
wrote:
On Thu, Feb 09, 2006 at 08:03:26PM +0300, Evgeniy Polyakov wrote:
It is completely different things.
The only common is that they _require_ some kind of notification
mechanism,
but none provide them.
From: Eric Dumazet [EMAIL PROTECTED]
Date: Thu, 09 Feb 2006 16:16:54 +0100
I had this idea after reading recent Van Jacobson paper about net channels.
I'm surprised you see the benefit of net channels (which first assumption is
about the 50-100 ns cost of a memory cache miss) but keep this
On Thu, Feb 09, 2006 at 05:19:45PM -0500, John W. Linville wrote:
Drivers using mii_check_media (via-rhine in particular) and also
forcing link parameters with ethtool can reach a state where the link
goes down and never comes back up. This is because mii_check_media
short-circuits early if
From: Alexey Kuznetsov [EMAIL PROTECTED]
Date: Wed, 8 Feb 2006 17:51:51 +0300
netlink overrun was broken while improvement of netlink.
Destination socket is used in the place where it was meant to be source
socket,
so that now overrun is never sent to user netlink sockets, when it should be,
In the future, please include a Signed-off-by: line in your patches
or git commits...thanks!
John
On Wed, Feb 01, 2006 at 10:51:00AM +0100, Danny van Dyk wrote:
John, please
git pull rsync://pitr.amd64.dev.gentoo.org/kugelfang/wireless-2.6.git
which will provide this changeset:
Danny
From: Samuel Ortiz [EMAIL PROTECTED]
Date: Thu, 9 Feb 2006 23:32:16 +0200 (EET)
This patch set IrDA's addr_len properly, i.e to 4 bytes, the size of the
IrLAP device address.
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
Bug fix, so applied to net-2.6
-
To unsubscribe from this list: send
From: Samuel Ortiz [EMAIL PROTECTED]
Date: Thu, 9 Feb 2006 23:32:43 +0200 (EET)
This patch fixes an out of range array access in irnet_irda.c.
Author: David Binderman [EMAIL PROTECTED]
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
Bug fix, so applied to net-2.6
thanks
-
To unsubscribe from
From: John Heffner [EMAIL PROTECTED]
Date: Thu, 9 Feb 2006 15:54:12 -0500
The rcvbuf lock should probably be honored here.
Signed-off-by: John Heffner [EMAIL PROTECTED]
Yes, this is consistent with what we do on the send side.
I'll apply this to net-2.6, thanks a lot.
-
To unsubscribe from
From: Samuel Ortiz [EMAIL PROTECTED]
Date: Thu, 9 Feb 2006 23:31:31 +0200 (EET)
So, next will come a set of 7 patches. Some of them are patches that have
been floating around the irda mailing list and others are fixes or
improvements based on Jean's latest TODO list.
I applied patches 4 and 7
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Mon, 06 Feb 2006 14:27:33 -0800
Change Bridge receive path to correctly handle RCU removal of device
from bridge. Also fixes deadlock between carrier_check and del_nbp.
This replaces the previous deleted flag fix.
Signed-off-by: Stephen
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Mon, 06 Feb 2006 14:27:34 -0800
Bridge netfilter code needs to handle the case where device is
removed from bridge while packet in process. In these cases the
bridge_parent can be come null while processing.
This should fix:
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Mon, 06 Feb 2006 14:27:35 -0800
Refactor how the bridge code interacts with kobject system.
It should still use kobjects even if not using sysfs.
Fix the error unwind handling in br_add_if.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
From: Herbert Xu [EMAIL PROTECTED]
Date: Tue, 07 Feb 2006 12:20:39 +1100
David S. Miller [EMAIL PROTECTED] wrote:
.
But, thinking about this some more, it might actually be best to just
add them anyways to TcpExt: and get netstat's limitation fixed.
Any other opinions?
Use
57 matches
Mail list logo