On Jan 8, 2008 5:23 AM, David Miller [EMAIL PROTECTED] wrote:
I am going to push this patch upstream, it is correct and we
have a positive test case that failed before, so overall it's
a net improvement even if we still don't exactly know why
Adolfo's case still fails.
Sorry it took so long,
Andrew Morton a écrit :
- First up, why was this added at all? We have percpu_counter.h which
has several years development invested in it. afaict it would suit the
present applications of pcounters.
If some deficiency in percpu_counters has been identified, is it
possible to correct
Thanks, i will try it.
You think lockdep can be buggy?
On Sat, 16 Feb 2008 09:00:36 +0100, Jarek Poplawski wrote
Denys Fedoryshchenko wrote, On 02/13/2008 09:13 AM:
It is very difficult to reproduce, happened after running about 1month.
No
changes done in classes at time of crash.
On Sat, 16 Feb 2008 11:07:29 +0100 Eric Dumazet [EMAIL PROTECTED] wrote:
Andrew, pcounter is a temporary abstraction.
It's buggy! Main problems are a) possible return of negative numbers b)
some of the API can't be from preemptible code c) excessive interrupt-off
time on some machines if
#part type=text/plain nofile=yes
Any comments on this? Apparently introduced in 2.1.68, so there's not
much hurry. But I'd still like to hear whether that analysis is correct
or not...
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary===-=-=
The bug did bite at least one user, who did
Andrew Morton a écrit :
On Sat, 16 Feb 2008 11:07:29 +0100 Eric Dumazet [EMAIL PROTECTED] wrote:
Andrew, pcounter is a temporary abstraction.
It's buggy! Main problems are a) possible return of negative numbers b)
some of the API can't be from preemptible code c) excessive interrupt-off
Also suggested by Herbert Xu, using workqueue is better than timer
for net/xfrm/xfrm_state.c, so replace them with delayed_work.
Note that, this patch is not fully tested, just compile and
run as a whole on an Intel Core Duo matchine. So should be
in -mm first.
Signed-off-by: WANG Cong [EMAIL
As suggested by Herbert, using workqueue is better than timer
for net/xfrm/xfrm_policy.c, so replace them with delayed_work.
Note that, this patch is not fully tested, just compile and
run as a whole on an Intel Core Duo matchine. So should be
in -mm first.
Signed-off-by: WANG Cong [EMAIL
The mv643xx_eth driver can be loaded as a platform device, as is done by
various Orion (ARM) based devices. The driver needs to define a module
alias for the platform driver so udev will load it automatically.
Tested with Debian on a QNAP TS-209.
Signed-off-by: Martin Michlmayr [EMAIL PROTECTED]
On Sat, 16 Feb 2008 11:26:18 -0800
Andrew Morton [EMAIL PROTECTED] wrote:
indirect functions calls are everywhere in kernel, network, fs,
everywhere.
That doesn't make them fast.
just to emphasize this: an indirect function call is at least as expensive as
an atomic operation on
x86
Here's a patch with cleanups and without various encoding bugs.
Can you verify it still works?
Also, some of the Prolific chips use some bizarre control requests,
which by all rights should not be needed. They seem to exist only
to cope with things the device firmware should have handled ...
On Sat, Feb 16, 2008 at 12:25:31PM +0200, Denys Fedoryshchenko wrote:
Thanks, i will try it.
You think lockdep can be buggy?
Just like every code... But the main reason is it has quite meaningful
overhead, so could be right in production only after lockups happen.
But if it doesn't report
[ 58.097923] =
[ 58.098883] [ INFO: possible recursive locking detected ]
[ 58.099603] 2.6.25-rc2-1avb #1
[ 58.100013] -
[ 58.100672] wpa_supplicant/2682 is trying to acquire lock:
[ 58.100672]
beclib header files.
---
diff -uprN orig/linux-2.6.24.2/drivers/message/beclib/pch.h
benet/linux-2.6.24.2/drivers/message/beclib/pch.h
--- orig/linux-2.6.24.2/drivers/message/beclib/pch.h1970-01-01
05:30:00.0 +0530
+++
Ether ring creation function in beclib.
---
diff -uprN orig/linux-2.6.24.2/drivers/message/beclib/ethtx_ll.c
benet/linux-2.6.24.2/drivers/message/beclib/ethtx_ll.c
--- orig/linux-2.6.24.2/drivers/message/beclib/ethtx_ll.c 1970-01-01
05:30:00.0 +0530
System abstraction function for Linux. beclib uses these helper routines /
macros to do OS dependent functions.
diff -uprN orig/linux-2.6.24.2/drivers/message/beclib/sa/setypes.h
benet/linux-2.6.24.2/drivers/message/beclib/sa/setypes.h
---
F/W header files.
--
diff -uprN
orig/linux-2.6.24.2/drivers/message/beclib/fw/bmap/iscsi_initiator_host_struct_bmap.h
benet/linux-2.6.24.2/drivers/message/beclib/fw/bmap/iscsi_initiator_host_struct_bmap.h
---
F/W header files.
---
diff -uprN
orig/linux-2.6.24.2/drivers/message/beclib/fw/bmap/ioctl_types_bmap.h
benet/linux-2.6.24.2/drivers/message/beclib/fw/bmap/ioctl_types_bmap.h
--- orig/linux-2.6.24.2/drivers/message/beclib/fw/bmap/ioctl_types_bmap.h
1970-01-01
F/W header files.
---
diff -uprN
orig/linux-2.6.24.2/drivers/message/beclib/fw/bmap/iscsi_common_host_struct_bmap.h
benet/linux-2.6.24.2/drivers/message/beclib/fw/bmap/iscsi_common_host_struct_bmap.h
---
Patch 0/16 and 15/16 of this series are getting dropped by the
spam filter. I am trying to get them across with changes
that will please the spam filter. Sorry about the inconvenience.
Below is the 0/16 of the series.
--
Hi,
I am sending a patch for network driver for
From: Andrew Morton [EMAIL PROTECTED]
Date: Sat, 16 Feb 2008 11:26:18 -0800
On Sat, 16 Feb 2008 13:03:54 +0100 Eric Dumazet [EMAIL PROTECTED] wrote:
Yes, per connection basis. Some workloads want to open/close more than 1000
sockets per second.
ie: slowpath
Definitely not slow path in
21 matches
Mail list logo