From: Andi Kleen [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 07:11:06 +0200
The hash sizing code needs far more tweaks. iirc it can still
allocate several GB hash tables on large memory systems (i've seen
that once in the boot log of a 2TB system). Even on smaller systems
it is usually too much.
Whereas it should probably go:
if (max == 0) {
max = (flags HASH_HIGHMEM) ? nr_all_pages : nr_kernel_pages;
max = (max PAGE_SHIFT) 4;
do_div(max, bucketsize);
}
or something like that.
That's still too big. Consider a 2TB
From: Andi Kleen [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 08:53:03 +0200
That's still too big. Consider a 2TB machine, with all memory in LOWMEM.
Andi I agree with you, route.c should pass in a suitable limit.
I'm just suggesting a fix for a seperate problem.
-
To unsubscribe from this list:
== Changelog ==
The value is_setbyuser from struct ip_options is never used and set
only one time (http://linux-net.osdl.org/index.php/TODO#IPV4).
This little patch removes it from the kernel source.
Signed-off-by: Louis Nyffenegger [EMAIL PROTECTED]
== Patch ==
---
Theodore Tso wrote:
Thanks, that description was very helpful. Would you accept a patch
with adding a comment describing this?
I will put it on my queue to add some comments for ASF.
It appears that there is no way of disabling ASF; is that a true
statement?
Turning off ASF is just a
From: Louis Nyffenegger [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 09:02:44 +0200
== Changelog ==
The value is_setbyuser from struct ip_options is never used and set
only one time (http://linux-net.osdl.org/index.php/TODO#IPV4).
This little patch removes it from the kernel source.
Generic event handling mechanism.
I send this patchset for comments and review, it still contains AIO and
aio_sendfile() implementation on top of get_block() abstraction, which was
decided to postpone for a while (it is simpler right now to generate patchset
as a whole,
when kevent will be
Core files.
This patch includes core kevent files:
- userspace controlling
- kernelspace interfaces
- initialization
- notification state machines
It might also inlclude parts from other subsystem (like network related
syscalls, so it is possible that it will not compile without other
poll/select() notifications. Timer notifications.
This patch includes generic poll/select and timer notifications.
kevent_poll works simialr to epoll and has the same issues (callback
is invoked not from internal state machine of the caller, but through
process awake).
Timer notifications can
Network AIO, socket notifications.
This patchset includes socket notifications and network asynchronous IO.
Network AIO is based on kevent and works as usual kevent storage on top
of inode.
Signed-off-by: Evgeniy Polyakov [EMAIL PROTECTED]
diff --git a/include/asm-i386/socket.h
The same without gmail problems
== Changelog ==
The value is_setbyuser from struct ip_options is never used and set
only one time (http://linux-net.osdl.org/index.php/TODO#IPV4).
This little patch removes it from the kernel source.
Signed-off-by: Louis Nyffenegger [EMAIL PROTECTED]
== Patch ==
AIO, aio_sendfile() implementation.
This patch includes asynchronous propagation of file's data into VFS
cache and aio_sendfile() implementation.
Network aio_sendfile() works lazily - it asynchronously populates pages
into the VFS cache (which can be used for various tricks with adaptive
On Mon, 07 Aug 2006 11:45:28 -0700
Jeremy Fitzhardinge [EMAIL PROTECTED] wrote:
With 2.6.18-rc3-mm2, I get a bogus device name for my e1000 device,
which I would expect to be eth0:
: ezr:pts/0; ifconfig -a
�6f� Link encap:Ethernet HWaddr 00:16:D3:20:D2:0B
UP BROADCAST
Andrew Morton wrote:
e1000 seems OK here. Don't know, sorry.
It's happening to all my ethernet-like devices: the Atheros wireless
comes up as a mess too. It's different each time, so it looks like
random uninitialized crud.
I did a clean rebuild, but it still happens. I guess I'll
David Miller wrote:
we quickly discover this GIT commit:
424c4b70cc4ff3930ee36a2ef7b204e4d704fd26
[IPV4]: Use the fancy alloc_large_system_hash() function for route hash table
- rt hash table allocated using alloc_large_system_hash() function.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
From: Kirill Korotaev [EMAIL PROTECTED]
Date: Tue, 08 Aug 2006 12:17:57 +0400
at least for i686 num_physpages includes highmem, so IMHO this bug
was there for years:
Correct, I misread the x86 code.
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to
Hi:
[IPV4]: Use network-order dport for all visible inet_lookup_*
Right now most inet_lookup_* functions take a host-order hnum instead
of a network-order dport because that's how it is represented internally.
This means that users of these functions have to be careful about using
the right
Hi:
[IPV4]: Uninline inet_lookup_listener
By modern standards this function is way too big to be inlined. It's
even bigger than __inet_lookup_listener :)
Signed-off-by: Herbert Xu [EMAIL PROTECTED]
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmVHI~} [EMAIL
On Tuesday 08 August 2006 05:42, David Miller wrote:
From: Alexey Kuznetsov [EMAIL PROTECTED]
Date: Mon, 7 Aug 2006 20:48:42 +0400
The patch looks OK. But I am not sure too.
To be honest, I do not understand the sense of HASH_HIGHMEM flag.
At the first sight, hash table eats low memory,
From: Eric Dumazet [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 10:57:31 +0200
I think we had discussion about being able to dynamically resize
route hash table (or tcp hash table), using RCU. Did someone worked
on this ? For most current machines (ram size = 1GB) , the default
hash table sizes
From: Herbert Xu [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 18:46:14 +1000
[IPV4]: Uninline inet_lookup_listener
By modern standards this function is way too big to be inlined. It's
even bigger than __inet_lookup_listener :)
Signed-off-by: Herbert Xu [EMAIL PROTECTED]
Applied, thanks.
-
To
From: Herbert Xu [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 18:47:29 +1000
So this patch changes all visible inet_lookup functions to take a
dport and move all dport-hnum conversion inside them.
This isn't so nice because we will now byte-swap the port twice when
we try looking up a listening
Only aironet lockdep related report I could find was
http://marc.theaimsgroup.com/?l=linux-netdevm=115406279721287w=2
this looks a bit different:
Linux version 2.6.17-1.2528.fc6 ([EMAIL PROTECTED]) (gcc version 4.1.1 20060802
(Red Hat 4.1.1-14)) #1 SMP Sun Aug 6 01:43:42 EDT 2006
On Tuesday 08 August 2006 09:44, Evgeniy Polyakov wrote:
Network AIO, socket notifications.
This patchset includes socket notifications and network asynchronous IO.
Network AIO is based on kevent and works as usual kevent storage on top
of inode.
diff --git a/include/net/sock.h
On Tue, Aug 08, 2006 at 11:52:53AM +0200, Eric Dumazet ([EMAIL PROTECTED])
wrote:
On Tuesday 08 August 2006 09:44, Evgeniy Polyakov wrote:
Network AIO, socket notifications.
This patchset includes socket notifications and network asynchronous IO.
Network AIO is based on kevent and works
On Tue, Aug 08, 2006 at 02:24:43AM -0700, David Miller wrote:
This isn't so nice because we will now byte-swap the port twice when
we try looking up a listening socket. Once for the established lookup
and once for the listening hash lookup.
It should be easy to do the byte-swap once at a
Jouni Malinen wrote:
Depends on what exactly you mean with on the fly. We have indeed
changed between doing software and hardware crypto for some cases, e.g.,
when enabling another BSS while one BSS is using static WEP (which would
need default WEP keys in hwaccel) in one BSS, we may disable
On Tue, 2006-08-08 at 12:44 +0300, Pekka Pietikainen wrote:
Only aironet lockdep related report I could find was
http://marc.theaimsgroup.com/?l=linux-netdevm=115406279721287w=2
Shouldn't this be fixed by the wireless events patch that was a result
of the orinoco lockdep issues that davej
On Tue, Aug 08, 2006 at 09:16:13PM +1000, Herbert Xu wrote:
Pekka Pietikainen [EMAIL PROTECTED] wrote:
Only aironet lockdep related report I could find was
http://marc.theaimsgroup.com/?l=linux-netdevm=115406279721287w=2
this looks a bit different:
Linux version 2.6.17-1.2528.fc6
On Wed, Aug 02, 2006 at 10:50:08AM -0700, Ben Greear wrote:
Currently, the bridge hook logic is something like:
if (bridge-consumed-pkt) {
return
}
// drop through to other layers
There are several other hooks I'd like to see added (pktgen receive
processing,
mac-vlans,
Is the patch below making it in? As i pointed out in another email
tests reveal it is good.
cheers,
jamal
On Sun, 2006-06-08 at 12:51 +1000, Herbert Xu wrote:
On Sat, Aug 05, 2006 at 07:36:50PM -0400, jamal wrote:
I know the qlen is = 0 otherwise i will hit the BUG().
A qlen of 0 will
On Sun, 6 Aug 2006, David Miller wrote:
From: Steven Rostedt [EMAIL PROTECTED]
Date: Mon, 7 Aug 2006 01:34:56 -0400 (EDT)
My suggestion would be to separate that tg3_timer into 4 different
timers, which is what it actually looks like.
Timers have non-trivial cost. It's cheaper to have
On Tue, Aug 08, 2006 at 08:21:58AM -0400, jamal wrote:
Is the patch below making it in? As i pointed out in another email
tests reveal it is good.
I'll write up a proper description and submit it.
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmVHI~} [EMAIL
David Miller wrote:
From: Andi Kleen [EMAIL PROTECTED]
Date: Tue, 8 Aug 2006 08:53:03 +0200
That's still too big. Consider a 2TB machine, with all memory in LOWMEM.
Andi I agree with you, route.c should pass in a suitable limit.
I'm just suggesting a fix for a seperate problem.
So
On Tue, Aug 08, 2006 at 02:38:51PM +0300, Pekka Pietikainen wrote:
Hmm... I retried with a 2.6.18rc4-based rawhide kernel and the warning
is still there, previous one was rc3-git7.
Are you sure it's the same warning? The one you quoted earlier is
caused by the lock validator mixing up spin
3) should we limit TCP hashe and hashb size the same way?
Yes - or best in fact all hashes handled by alloc_large_system_hash()
-Andi
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at
On Tue, Aug 08, 2006 at 10:52:22PM +1000, Herbert Xu wrote:
On Tue, Aug 08, 2006 at 02:38:51PM +0300, Pekka Pietikainen wrote:
Could be http://marc.theaimsgroup.com/?l=linux-netdevm=115461336523555w=2
which isn't upstream yet, right?
That's a separate bug fix which should produce an
On Tue, Aug 08, 2006 at 09:11:40PM +1000, Herbert Xu wrote:
On Tue, Aug 08, 2006 at 02:24:43AM -0700, David Miller wrote:
This isn't so nice because we will now byte-swap the port twice when
we try looking up a listening socket. Once for the established lookup
and once for the listening
Christoph Hellwig [EMAIL PROTECTED] wrote:
[IPV4]: Use network-order dport for all visible inet_lookup_*
Shouldn't it use __be16 types then?
Yes it should.
However, we can't easily annotate the stack piecemeal. So this should
be annotated along with the rest of the stack (e.g., tcphdr,
According to Documentation/pci.txt in 2.6.18 pci_request_regions()
should be called _before_ pci_enable_device().
And pci_release_regions() should be called after pci_disable_device().
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Hi,
Here are some minor patches for rt2x00 for wireless-dev,
nothing fancy in there, just some optimizations and fixes that
are now possible due to some changes in the dscape stack. :)
Ivo
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL
Move scan structure initialization and handling into
some generic handlers in rt2x00.h.
This also fixed some obscure coding when waiting for the
empty txrings before starting a scan.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Also check for equality with txpower limits.
It doesn't do any harm, but it does prevent compiler
warnings in certain conditions.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
wireless-dev-rt2x00-perm-addr/drivers/net/wireless/d80211/rt2x00/rt2400pci.h
Small misc fixes,
- remove unwanted whitespaces
- limit lines to 80 characters
- byteordering fix
- comment fix
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
wireless-dev-rt2x00-scan/drivers/net/wireless/d80211/rt2x00/rt2500pci.c
When reading the MAC addr from MAC register or EEPROM
it is always read as little endian. Since we want to store it in a u8 array
we don't require byte ordering as long as we correctly read it into
an u8 array directly.
Also copy the address to perm_addr and enable ethtool to read it.
d80211 no longer calls ieee80211_hw-config() fom interrupt context.
Make gratefully use of this and remove the workqueue scheduling
for the config changes.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
wireless-dev-rt2x00-pci/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
Another question that bothers me is the bonding code multicast related
behavior when it does fail-over.
From what I see in bond_mc_swap(), set_multicast_list() is well
handled: dev_mc_delete() is called for the old slave (so if in the
future the old slave has a link, it will leave the
On Tue, 8 Aug 2006, Steven Rostedt wrote:
On Sun, 6 Aug 2006, David Miller wrote:
From: Steven Rostedt [EMAIL PROTECTED]
Date: Mon, 7 Aug 2006 01:34:56 -0400 (EDT)
My suggestion would be to separate that tg3_timer into 4 different
timers, which is what it actually looks like.
Jeff,
This is the same as previous, but with a number of indentation issues addressed
in arch/ppc/platforms/mpc*. I think it makes sense to do it right now, since
all the files are touched
anyway. Haven't noticed before, sorry about that. Please find the patch inlined
below.
This contains
Hi,
These patches are for wireless-dev
and will add a crc-itu-t implementation to the lib/ directory.
Some drivers are currently using there own version
of this crc routine so there could be usefull to add
this to lib/
The second patch is to make rt2x00 use this crc-itu-t library
and removes its
Now crc-itu-t is available rt2x00 should use that implementation
instead of its own private version.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rNU3 wireless-dev-rt2x00/drivers/net/wireless/d80211/rt2x00/Kconfig
wireless-dev-rt2x00-crc/drivers/net/wireless/d80211/rt2x00/Kconfig
Add the CRC-ITU-T implementation to lib/
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff --git a/include/linux/crc-itu-t.h b/include/linux/crc-itu-t.h
new file mode 100644
index 000..3e6d80c
--- /dev/null
+++ b/include/linux/crc-itu-t.h
@@ -0,0 +1,27 @@
+/*
+ * crc-itu-t.h - CRC
Hello,
I'm sorry to tell you that my previous message was bogus. The problem
still persists.
Sorry to have bothered you.
Regards,
Lutz Urban
Lutz Urban schrieb:
Hello,
I got your email-addresses from my /var/log/boot.msg.
The kernel switch 'pci=routeirq' solved the following problem:
Now rfkill is available rt2x00 should make use of it let
the hardware button status be periodically probed by rfkill.
This makes the ACPI implementation as done in rt2x00 no longer required.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rNU3
This will add the rfkill driver to the input/misc section of the kernel.
rfkill is usefull for newtwork devices that contain a hardware button
to enable or disable the radio.
With rfkill a generic interface is created for the network drivers,
as well as providing a uniform way for userspace to
Hi,
With the previous Request for comments for rfkill it seemed
most people did not have any objectiosn against rfkill.
So now I wish to submit rfkill to wireless-dev. :)
First patch will add rfkill to input/misc/ while the second
patch will make rt2x00 use rfkill.
Ivo
-
To unsubscribe from
On Tue, Aug 08, 2006 at 01:35:22PM +0200, Johannes Berg wrote:
Jouni Malinen wrote:
Depends on what exactly you mean with on the fly. We have indeed
changed between doing software and hardware crypto for some cases, e.g.,
when enabling another BSS while one BSS is using static WEP (which would
On Tuesday 08 August 2006 17:57, Jouni Malinen wrote:
On Tue, Aug 08, 2006 at 01:35:22PM +0200, Johannes Berg wrote:
Jouni Malinen wrote:
Depends on what exactly you mean with on the fly. We have indeed
changed between doing software and hardware crypto for some cases, e.g.,
when enabling
Arjan van de Ven wrote:
and you're also sure this is not your userspace using interface
renaming...
(could be an initscripts bug for name-by-MAC ethernet device naming)
It's definitely in-kernel, since its specific to this version. And it
seems to have gone away since I turned on slab
Jeremy Fitzhardinge wrote:
Andrew Morton wrote:
e1000 seems OK here. Don't know, sorry.
It's happening to all my ethernet-like devices: the Atheros wireless
comes up as a mess too. It's different each time, so it looks like
random uninitialized crud.
is this the binary atheros
Jeremy Fitzhardinge wrote:
Arjan van de Ven wrote:
Jeremy Fitzhardinge wrote:
Andrew Morton wrote:
e1000 seems OK here. Don't know, sorry.
It's happening to all my ethernet-like devices: the Atheros wireless
comes up as a mess too. It's different each time, so it looks like
random
Arjan van de Ven wrote:
Jeremy Fitzhardinge wrote:
Andrew Morton wrote:
e1000 seems OK here. Don't know, sorry.
It's happening to all my ethernet-like devices: the Atheros wireless
comes up as a mess too. It's different each time, so it looks like
random uninitialized crud.
is
Hello!
This fix goes against the old historical comments about UNIX98 semantics
but without this fix SOCK_DGRAM is broken and useless. So either ANK's
interpretation was incorect or UNIX98 standard was wrong.
Just found this reference to me. :-)
The comment migrated from tcp.c. It is only
Christophe Devriese wrote:
On Wed, Aug 02, 2006 at 10:50:08AM -0700, Ben Greear wrote:
Currently, the bridge hook logic is something like:
if (bridge-consumed-pkt) {
return
}
// drop through to other layers
There are several other hooks I'd like to see added (pktgen receive
On Tue, 8 Aug 2006 20:36:18 +0400
Alexey Kuznetsov [EMAIL PROTECTED] wrote:
Hello!
This fix goes against the old historical comments about UNIX98 semantics
but without this fix SOCK_DGRAM is broken and useless. So either ANK's
interpretation was incorect or UNIX98 standard was wrong.
On Fri, Aug 04, 2006 at 04:31:11PM -0700, David Miller wrote:
Yes, it's meant to catch unintented races.
The queueing layer that calls -hard_start_xmit() technically has no
need to support NETDEV_TX_BUSY as a return value, since the device
is able to prevent this.
If we could avoid
Or Gerlitz [EMAIL PROTECTED] wrote:
[...]
As i don't see any buffering of the IGMP packets, i understand there's no
reply of sending them during fail-over and this means that only when the
router would do IGMP query on this node it will learn on the fail-over. Is
it indeed what's going on? if i
On Mon, 2006-08-07 at 20:59 -0700, David Miller wrote:
From: Sridhar Samudrala [EMAIL PROTECTED]
Date: Mon, 07 Aug 2006 16:00:32 -0700
Update sunrpc to use in-kernel sockets API.
Signed-off-by: Sridhar Samudrala [EMAIL PROTECTED]
Acked-by: James Morris [EMAIL PROTECTED]
Applied,
I figured that the word ad ver tise must throw off the spam filter,
so I hope this gets through now that I've removed it from the subject.
Note this is an ethtool patch, not a kernel one :)
Cheers,
Auke
-- Forwarded message --
From: Jeff Kirsher [EMAIL PROTECTED]
Date: Aug 7,
Em Tue, Aug 08, 2006 at 08:36:18PM +0400, Alexey Kuznetsov escreveu:
Hello!
This fix goes against the old historical comments about UNIX98 semantics
but without this fix SOCK_DGRAM is broken and useless. So either ANK's
interpretation was incorect or UNIX98 standard was wrong.
Just
Ther is no point in using a more expensive reader/writer lock
for a low contention lock like the fib_info_lock. The only
reader case is in handling route redirects.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- tcp-rcu.orig/net/ipv4/fib_semantics.c
+++ tcp-rcu/net/ipv4/fib_semantics.c
From a recent rc3-git kernel.
Dave
--
http://www.codemonkey.org.uk
---BeginMessage---
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug report.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=201560
It's been a week since -rc3, so now we have a -rc4.
2.6.18 comes with ipt_statistic, but there is no way from userspace
(iptables) to use it (libipt_statistic.so simply does not come with the
latest iptables from svn). Does someone know what's going on?
Jan Engelhardt
--
-
To unsubscribe
Jan Engelhardt wrote:
It's been a week since -rc3, so now we have a -rc4.
2.6.18 comes with ipt_statistic, but there is no way from userspace
(iptables) to use it (libipt_statistic.so simply does not come with the
latest iptables from svn). Does someone know what's going on?
I still have
On Monday 07 August 2006 23:04, Adrian Bunk wrote:
This patch removes three no longer used functions (that are even
generating gcc warnings).
This patch doesn't look right, but it is the result of
58e5528ee464d38040b9489e10033c9387a10d56 in git-netdev...
Hm, can't find that commit in a
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/r8169.c | 16 ++--
1 file changed, 10 insertions(+), 6 deletions(-)
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
arch/um/drivers/net_kern.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
Index:
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/3c59x.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
Index:
Update the driver to make use of the NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/tg3.c |2 ++
1 file changed, 2 insertions(+)
Index: linux-2.6/drivers/net/tg3.c
Use sk_set_memalloc() on the nbd socket.
Limit each request to 1 page, so that the request throttling also limits the
number of in-flight pages and force the IO scheduler to NOOP as anything else
doesn't make sense anyway.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel
From: Daniel Phillips [EMAIL PROTECTED]
Recently, Peter Zijlstra and I have been busily collaborating on a
solution to the memory deadlock problem described here:
http://lwn.net/Articles/144273/
Kernel Summit 2005: Convergence of network and storage paths
We believe that an approach very
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/e100.c |5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
Index:
Update the driver to make use of the NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/e1000/e1000_main.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
Index:
The core of the VM deadlock avoidance framework.
From the 'user' side of things it provides a function to mark a 'struct sock'
as SOCK_MEMALLOC, meaning this socket may dip into the memalloc reserves on
the receive side.
From the net_device side of things, the extra 'struct net_device *'
Update UML with a proper 'pfn_to_kaddr()' definition, the VM deadlock
avoidance framework uses it.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
include/asm-um/page.h |2 ++
1 file changed, 2 insertions(+)
Index:
On Tue, Aug 08, 2006 at 08:32:37PM +0200, Michael Buesch wrote:
On Monday 07 August 2006 23:04, Adrian Bunk wrote:
This patch removes three no longer used functions (that are even
generating gcc warnings).
This patch doesn't look right, but it is the result of
Signed-off-by: Thomas Graf [EMAIL PROTECTED]
Index: net-2.6.19.git/net/ipv4/fib_semantics.c
===
--- net-2.6.19.git.orig/net/ipv4/fib_semantics.c
+++ net-2.6.19.git/net/ipv4/fib_semantics.c
@@ -286,7 +286,7 @@ void rtmsg_fib(int
Fixes various unvalidated netlink attributes causing memory
corruptions when left empty by userspace applications.
Signed-off-by: Thomas Graf [EMAIL PROTECTED]
Index: net-2.6.19.git/include/net/ip_fib.h
===
---
Replaces the rather ugly struct kern_rta + rtmsg with a new
struct fib_config while converting everything to the new netlink
api. Simplifies the FIB module interface quite a bit and allows
changing internals while keeping a stable netlink interface.
Gets rid of things like passing on
Introduces struct fib_config replacing the ugly struct kern_rta
prone to ordering issues. Avoids creating faked netlink messages
for auto generated routes or requests via ioctl.
A new interface net/nexthop.h is added to help navigate through
nexthop configuration arrays.
Since the netlink source
On Tue, 8 Aug 2006, Andi Kleen wrote:
The hash sizing code needs far more tweaks. iirc it can still allocate
several GB hash tables on large memory systems (i've seen that once in
the boot log of a 2TB system). Even on smaller systems it is usually
too much.
Yes. Linear growth with
Hi
I am currently working on porting 3945 driver to use d80211. The porting
is going well, thanks for the great stack. I ran into some problems and
issues that I had to use some work around to have it functioning. I hope
I can find the help to point me to right way of using the stack.
1- I
Evgeniy Polyakov wrote:
Generic event handling mechanism.
I send this patchset for comments and review, it still contains AIO and
aio_sendfile() implementation on top of get_block() abstraction, which was
decided to postpone for a while (it is simpler right now to generate patchset
as a
On Tuesday 08 August 2006 14:27, Mohamed Abbas wrote:
2- Scanning; in 3945 driver, we can not tune to other channels while we
are connected, this will cause a firmware error. The firmware provides a
scanning command we call so the firmware will take care of switching of
the available channels
Peter Zijlstra wrote:
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
this should be done in two separate patches. I should take care of the
netdev_alloc_skb()
part too for e100 (which I've already queued internally), also since ixgb still
needs
On Tue, 2006-08-08 at 13:13 -0700, Auke Kok wrote:
Peter Zijlstra wrote:
Update the driver to make use of the netdev_alloc_skb() API and the
NETIF_F_MEMALLOC feature.
this should be done in two separate patches. I should take care of the
netdev_alloc_skb()
part too for e100 (which I've
Peter Zijlstra wrote:
Update the driver to make use of the NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/e1000/e1000_main.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
Index:
On Tue, 2006-08-08 at 13:57 -0700, Stephen Hemminger wrote:
On Tue, 08 Aug 2006 21:33:45 +0200
Peter Zijlstra [EMAIL PROTECTED] wrote:
The core of the VM deadlock avoidance framework.
From the 'user' side of things it provides a function to mark a 'struct
sock'
as SOCK_MEMALLOC,
* Peter Zijlstra [EMAIL PROTECTED] 2006-08-08 21:33
+struct sk_buff *__netdev_alloc_skb(struct net_device *dev,
+ unsigned length, gfp_t gfp_mask)
+{
+ struct sk_buff *skb;
+
+ if (dev (dev-flags IFF_MEMALLOC)) {
+ WARN_ON(gfp_mask (__GFP_NOMEMALLOC |
On Tue, 2006-08-08 at 13:50 -0700, Auke Kok wrote:
Peter Zijlstra wrote:
Update the driver to make use of the NETIF_F_MEMALLOC feature.
Signed-off-by: Peter Zijlstra [EMAIL PROTECTED]
Signed-off-by: Daniel Phillips [EMAIL PROTECTED]
---
drivers/net/e1000/e1000_main.c | 11
1 - 100 of 179 matches
Mail list logo