Ingo Oeser wrote:
Hi there,
here are some suggested cleanups, which compiled without problems using
allyesconfig on the latest net-2.6 git tree from David S. Miller.
Please review and apply, if convienient.
I think the patch should wait until 0.57 is merged: There are 4 patches
on their
On Mon, 2006-04-10 at 02:16 +0200, Michael Buesch wrote:
Does kbuild perhaps have some magic to handle that?
This needs to be solved soon, but I have no idea how.
You want to have bcm43xx and bcm43xx-dscape conflict anyway, since
there's no point in building both into the kernel, only one can
On Tue, 11 Apr 2006, Meelis Roos wrote:
Detected unconfigured Toshiba Satellite 1800 SMSC IrDA chip, pre-configuring
device.
SORRY: Toshiba Satellite 1800 has an unsupported bridge controller (ALi): not
pre-configured.
smsc-ircc2, Preconfiguration failed !
Yep we don't have support for ALi
On Mon, 2006-04-10 at 15:04 +0100, Daniel Drake wrote:
I think netif_stop_queue() is only supposed to be used directly from
inside a hard_start_xmit function. I think we should be using
netif_disable_tx() here.
Looks about right to me, since that actually does a spinlock :)
I'll send a new
Detected unconfigured Toshiba Satellite 1800 SMSC IrDA chip,
pre-configuring device.
SORRY: Toshiba Satellite 1800 has an unsupported bridge controller (ALi):
not pre-configured.
smsc-ircc2, Preconfiguration failed !
Yep we don't have support for ALi bridges yet, the Toshiba 1800 still needs
block? For example, typhoon.c:
spin_lock(tp-state_lock);
+#if defined(CONFIG_VLAN_8021Q) || defined (CONFIG_VLAN_8021Q_MODULE)
if(tp-vlgrp != NULL rx-rxStatus TYPHOON_RX_VLAN)
vlan_hwaccel_receive_skb(new_skb,
... but it doesn't, so those addresses are treated as local which is bad.
This problem doesn't exist with ipv6 and the patch is based on the solution
found there.
--- linux-2.4.32.orig/net/ipv4/devinet 2004-08-08 01:26:06.0 +0200
+++ linux-2.4.32/net/ipv4/devinet.c 2006-04-11
From: Matyas Koszik [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 09:44:55 +0200 (CEST)
... but it doesn't, so those addresses are treated as local which is bad.
No, shutting down an interface should not remove ipv4 addresses. The
user must explicitly remove them.
This behavior has been around
On Tuesday 11 April 2006 10:44, Denis Vlasenko wrote:
On Tuesday 11 April 2006 10:43, Denis Vlasenko wrote:
These patches exclude VLAN code from netdevice drivers
and from bonding module, and even remove vlan-related
members of struct netdevice if VLAN is not configured.
Compile tested on
Hi,
After davem and Dave Dillow comments I realized that
a lot of drivers try to do VLAN-related things even on
non-VLAN-enabled kernels.
These patches exclude VLAN code from netdevice drivers
and from bonding module, and even remove vlan-related
members of struct netdevice if VLAN is not
On Tuesday 11 April 2006 10:43, Denis Vlasenko wrote:
These patches exclude VLAN code from netdevice drivers
and from bonding module, and even remove vlan-related
members of struct netdevice if VLAN is not configured.
Compile tested on allyesconfig kernel with CONFIG_8021Q=y,m,n.
This one
On Tue, 11 Apr 2006, David S. Miller wrote:
From: Matyas Koszik [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 09:44:55 +0200 (CEST)
... but it doesn't, so those addresses are treated as local which is bad.
No, shutting down an interface should not remove ipv4 addresses. The
user must
From: Matyas Koszik [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 10:08:01 +0200 (CEST)
Then it maybe shouldn't affect the flow of packets while the
interface is down - or is it also something people depend on?
Yes, people probably do depend upon it.
Addresses are owned by the host not a
On Tuesday 11 April 2006 10:58, David S. Miller wrote:
This is not very nice, there is no way I'm applying these patches.
I think the current situation is far better than the large pile of
ifdefs these patches are adding to the tree.
Let's just leave things the way they are ok?
:(
Ok,
From: Denis Vlasenko [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 11:11:12 +0300
Ok, one last try. Would you like this smallish patch instead?
It takes care of those BIG inlines.
You're putting vlan stuff into a net/core/*.c file, that
is not correct.
If we're not going to do the ifdef mess,
On Sun, Apr 09, 2006 at 08:40:05PM +0200, Patrick McHardy wrote:
Adrian Bunk wrote:
This patch contains the overdue removal of ip{,6}_queue.
I think this is too early, none of the distributions seem to have
picked up the compat library already. I believe (not sure, Harald?)
that it
Hi Manfred,
Manfred Spraul wrote:
I think the patch should wait until 0.57 is merged: There are 4 patches
on their way to Jeff.
These patches contain several bugfixes, they should have the higher
priority.
Fine with me. Will resubmit after next week.
Apart from that: looks good.
Many
It seems the discussion has died. How is this to move forward?
I'm not sure who is responsible for this part of the kernel.
There are a few options:
1. Broadcast ARP by checking the IP range in kernel (current patch)
2. Broadcast ARP if the scope parameter is link
3. Broadcast ARP if a new (to be
Hi Denis,
Denis Vlasenko wrote:
+#if defined(CONFIG_VLAN_8021Q) || defined(CONFIG_VLAN_8021Q_MODULE)
if(vlan_tx_tag_present(skb)) {
first_txd-processFlags |=
TYPHOON_TX_PF_INSERT_VLAN | TYPHOON_TX_PF_VLAN_PRIORITY;
@@ -844,6 +849,7 @@
Maybe it's unrelated to this problem, but it is interesting observation,
at least for me.
All boxes running for two weeks now and spitting these assert messages
have about 1,5GB of slab size allocated, with skbuff_head_cache entry
being the largest entry. After rebooting, it is all nice and
Subject: linux-source-2.6.16: inconsistent device detetion: eth0=eth1
Package: linux-source-2.6.16
Version: 2.6.16-2
Severity: normal
I have two ethernet cards on my box, an onboard SiS 7012 and a pci ne-2k
compatible. When I boot the system, sometimes the SiS is detected as
eth0 and the pci as
Hi Jamal:
On Mon, Apr 10, 2006 at 11:02:06PM -0400, jamal wrote:
Ok, if both you can provide feedback on the attached patch (untested but
compiles) I will make any necessary changes, test and push this +
documentation to Dave.
Looks good to me.
diff --git a/net/xfrm/xfrm_state.c
Is this pending in your tree Jeff?
It fixes compile regressions with HOTPLUG disabled.
Sam
On Mon, Apr 03, 2006 at 04:28:46PM -0700, Randy.Dunlap wrote:
From: Randy Dunlap [EMAIL PROTECTED]
If CONFIG_HOTPLUG=n, gcc doesn't like some __initdata to be
const (rodata) and other
On Tuesday 11 April 2006 11:36, David S. Miller wrote:
From: Denis Vlasenko [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 11:11:12 +0300
Ok, one last try. Would you like this smallish patch instead?
It takes care of those BIG inlines.
You're putting vlan stuff into a net/core/*.c file, that
Below patch allows using iw_mode auto with softmac. bcm43xx forces managed
so this bug wasn't noticed earlier, but this was one of the problems why
zd1211 didn't work earlier.
Signed-off-by: Johannes Berg [EMAIL PROTECTED]
--- wireless-2.6.orig/net/ieee80211/softmac/ieee80211softmac_io.c
Below patch was developed after discussion with Daniel Drake who
mentioned to me that wireless tools expect an EAGAIN return from getscan
so that they can wait for the scan to finish before printing out the
results.
Signed-off-by: Johannes Berg [EMAIL PROTECTED]
---
Here are a few patches to softmac for issues that came up in discussions
with Daniel and others.
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi,
On Tuesday 11 April 2006 05.02, jamal wrote:
Ok, if both you can provide feedback on the attached patch (untested but
compiles) I will make any necessary changes, test and push this +
documentation to Dave.
Looks ok, although I only had a quick look at it.
--
Regards,
Krisztian
josh wrote:
I have two ethernet cards on my box, an onboard SiS 7012 and a pci ne-2k
compatible. When I boot the system, sometimes the SiS is detected as
eth0 and the pci as eth1, and sometimes the SiS is eth1 and the pci
eth0, *without my changing anything*. I just reboot and it's different.
Hi Denis,
Denis Vlasenko wrote:
On Tuesday 11 April 2006 12:49, Ingo Oeser wrote:
#if defined(CONFIG_VLAN_8021Q) || defined(CONFIG_VLAN_8021Q_MODULE)
static inline has_vlan_group(...) {
/* get VLAN group */
}
#else
static inline has_vlan_group(...) {return 0;}
#endif
With
On Tue, 2006-04-11 at 10:28 +0300, Denis Vlasenko wrote:
block? For example, typhoon.c:
spin_lock(tp-state_lock);
+#if defined(CONFIG_VLAN_8021Q) || defined (CONFIG_VLAN_8021Q_MODULE)
if(tp-vlgrp != NULL rx-rxStatus
TYPHOON_RX_VLAN)
Freek Dijkstra wrote:
It seems the discussion has died. How is this to move forward?
I'm not sure who is responsible for this part of the kernel.
There are a few options:
1. Broadcast ARP by checking the IP range in kernel (current patch)
2. Broadcast ARP if the scope parameter is link
3.
Harald Welte wrote:
On Sun, Apr 09, 2006 at 08:40:05PM +0200, Patrick McHardy wrote:
I think this is too early, none of the distributions seem to have
picked up the compat library already. I believe (not sure, Harald?)
that it requires at least recompilation of programs using libipq,
so the 6
Hi all!
Can you give me some guideline about how and where
netrand ... influences an action in the tc sources?
(I've also tried LARTC mailing list, but did not get
any response)
When I last checked, iproute2 sources had a tc gact
module with 2 statistical netrand methods:
- random
- determ
jamal wrote:
On Thu, 2006-06-04 at 18:29 +0200, Patrick McHardy wrote:
[..]
There are basically two possibilities how to implement this. The less
intrusive, but IMO more hackish one is to just handle this inside the
qdiscs that require this operation by not requeueing the packet to
the
On Mon, 2006-04-10 at 16:39 -0700, David S. Miller wrote:
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Mon, 10 Apr 2006 16:24:29 -0700
This change allows link local packets (like 802.3ad and Spanning Tree
Protocol) to be processed even when the bridge is not using the port.
It fixes
This patch fixes the problem of some Dlink cards picking the wrong
driver. It looks like these cards use Yukon 1 chipset, not Yukon 2.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- sky2-2.6.17.orig/drivers/net/skge.c 2006-03-24 09:56:05.0 -0800
+++
Stephen Hemminger wrote:
This patch fixes the problem of some Dlink cards picking the wrong
driver. It looks like these cards use Yukon 1 chipset, not Yukon 2.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- sky2-2.6.17.orig/drivers/net/skge.c 2006-03-24 09:56:05.0 -0800
+++
I have searched the netdev archives and have not found any RFC like
this.
This proposal is to add selective diag/test modes to ethtool with
minimal
overhead/bloat. Please provide feedback on the proposal. This proposed
selective test mode is being requested more and more from our Linux
users,
so
Run the ethernet common code through indent, and fix other whitespace
issues.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- net-2.6.orig/net/ethernet/eth.c 2006-04-11 10:07:37.0 -0700
+++ net-2.6/net/ethernet/eth.c 2006-04-11 10:18:37.0 -0700
@@ -66,55 +66,46 @@
Minor optimizations and cleanups to the common ethernet
header routines.
--
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Don't need to use __constant_htons here.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- net-2.6.orig/net/ethernet/eth.c 2006-04-11 10:18:37.0 -0700
+++ net-2.6/net/ethernet/eth.c 2006-04-11 10:21:40.0 -0700
@@ -208,7 +208,7 @@
eth = (struct ethhdr *)
Change the ethernet support routines to use constant address size.
This generates smaller faster code.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- net-2.6.orig/net/ethernet/eth.c 2006-04-11 09:43:40.0 -0700
+++ net-2.6/net/ethernet/eth.c 2006-04-11 10:07:37.0
On Tue, Apr 11, 2006 at 04:42:50PM +0200, Patrick McHardy wrote:
Maybe we should add a printk ('app foo is using obsolete ip_queue
system').
Good idea, that will probably help speed it up. But I still think
we need to give them at least another six month.
ok. I'll prepare a patch for both
On Tue, 2006-04-11 at 18:35 +0200, Michael Buesch wrote:
Sure. I would probably say that both m should conflict, too.
Nah, you probably want both m for testing.
And one y and the other m, too.
Yeah that should conflict too.
This can be done with ugly depends statements.
I have no idea.
On Mon, 10 Apr 2006 21:41:39 -0400 (EDT)
George P Nychis [EMAIL PROTECTED] wrote:
Hi,
I am trying to install a proprietary qdisc made for research, it is not
publically released yet, however its been used several times so i know it
works.
The files included are:
q_xcp.c:
static
Hi,
Since sometime after 2.6.16, some websites have been very slow to load.
Examples include:
http://zd1211.ath.cx
http://developer.osdl.org/shemminger/blog/
http://www.reactivated.net/weblog
On a good kernel, wget http://zd1211.ath.cx; says:
20:23:38 (90.44 KB/s) - `index.html' saved
On Tue, 11 Apr 2006 20:30:46 +0100
Daniel Drake [EMAIL PROTECTED] wrote:
Hi,
Since sometime after 2.6.16, some websites have been very slow to load.
Examples include:
http://zd1211.ath.cx
http://developer.osdl.org/shemminger/blog/
http://www.reactivated.net/weblog
On a good
Daniel Drake wrote:
Hi,
Since sometime after 2.6.16, some websites have been very slow to load.
Examples include:
http://zd1211.ath.cx
http://developer.osdl.org/shemminger/blog/
http://www.reactivated.net/weblog
On a good kernel, wget http://zd1211.ath.cx; says:
20:23:38 (90.44 KB/s) -
John Heffner wrote:
I'm not seeing this behavior myself. What are the values of
/proc/sys/net/ipv4/tcp_wmem, tcp_rmem, and tcp_mem? How much memory
does this system have? (A binary tcpdump might be good, too.)
tcp_wmem: 409616384 131072
tcp_rmem: 409687380 174760
tcp_mem: 98304
Daniel Drake wrote:
John Heffner wrote:
I'm not seeing this behavior myself. What are the values of
/proc/sys/net/ipv4/tcp_wmem, tcp_rmem, and tcp_mem? How much memory
does this system have? (A binary tcpdump might be good, too.)
tcp_wmem: 409616384 131072
tcp_rmem: 409687380
On Mon, 10 Apr 2006 21:41:39 -0400 (EDT) George P Nychis
[EMAIL PROTECTED] wrote:
Hi,
I am trying to install a proprietary qdisc made for research, it is not
publically released yet, however its been used several times so i know
it works.
The files included are: q_xcp.c: static int
I am getting closer...
I added debugging, and noticed that it looks for:
snprintf(buf, sizeof(buf), %s_qdisc_util, str);
However in q_xcp.c it had:
struct qdisc_util xcp_util = {
so I changed that to xcp_qdisc_util, and now i run tc:
lanthanum-ini tc # tc qdisc add dev ath0 root
John Heffner wrote:
tcp_wmem: 409616384 131072
tcp_rmem: 409687380 174760
tcp_mem: 98304 131072 196608
These are (I assume) with the patch reversed. What are the values with
the patch applied?
Yes- that was on a good kernel, with the patch reversed.
On a bad kernel, with
I think allowing DMA mask range limiting in the IOMMU layer is going
to set a very bad precedence, just don't do it.
It's 2006, we should be way past the era of not putting the full 32
PCI DMA address bits in devices. In this day and age it is simply
inexscusable.
Maybe we could
This is almost certainly due to a buggy firewall that doesn't understand
TCP window scaling. I've usually seen this in the past with OpenBSD
firewalls. Do you have one of these in your path?
Thanks,
-John
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a
Addresses are owned by the host not a particular device, even
though they are assosciated with a particular interface.
Linux defaults to using the host based addressing model instead of the
interface based addressing model.
It is true for IPv4, but IPv6 addresses are removed when interface
Addresses are owned by the host not a particular device, even
though they are assosciated with a particular interface.
Linux defaults to using the host based addressing model instead of the
interface based addressing model.
It is true for IPv4, but IPv6 addresses are removed when
From: Randy Dunlap [EMAIL PROTECTED]
Fix printk format warnings:
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:456: warning: format ‘%u’
expects type ‘unsigned int’, but argument 3 has type ‘size_t’
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:460: warning: format ‘%08x’
expects type
From: Randy Dunlap [EMAIL PROTECTED]
Use depends on to make all bcm43 driver options be listed
at the same level.
Signed-off-by: Randy Dunlap [EMAIL PROTECTED]
---
drivers/net/wireless/bcm43xx/Kconfig |3 +++
1 files changed, 3 insertions(+)
---
From: David Stevens [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 14:33:38 -0700
Addresses are owned by the host not a particular device, even
though they are assosciated with a particular interface.
Linux defaults to using the host based addressing model instead of the
interface based
From: Benjamin Herrenschmidt [EMAIL PROTECTED]
Date: Wed, 12 Apr 2006 06:49:00 +1000
I would tend to agree... except that the broadcom is _the_ wireless
card shipped by Apple with all of their machines for the last few
years, and thus, the problem will be hit by pretty much any G5 user
trying
John Heffner wrote:
This is almost certainly due to a buggy firewall that doesn't understand
TCP window scaling. I've usually seen this in the past with OpenBSD
firewalls. Do you have one of these in your path?
At home I'm behind a Linux gateway box currently running 2.6.15-rc6 - I
am
On Tue, 2006-04-11 at 14:34 -0700, David S. Miller wrote:
I still think we shouldn't reward shit hardware by complicating
up our DMA mappings internals. :-)
Heh, it's a good point but in that specific case, it's a bit difficult
to tell that to users who don't have a choice of what card to put
I still think we shouldn't reward shit hardware by complicating
up our DMA mappings internals. :-)
BTW. In the meantime, can't that driver work in PIO only mode ?
Ben.
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More
On Wed, Apr 12, 2006 at 08:21:17AM +1000, Benjamin Herrenschmidt wrote:
I still think we shouldn't reward shit hardware by complicating
up our DMA mappings internals. :-)
BTW. In the meantime, can't that driver work in PIO only mode ?
yes, I think you just have to have the
On Tue, 11 Apr 2006 23:20:42 +0100
Daniel Drake [EMAIL PROTECTED] wrote:
John Heffner wrote:
This is almost certainly due to a buggy firewall that doesn't understand
TCP window scaling. I've usually seen this in the past with OpenBSD
firewalls. Do you have one of these in your path?
On Wed, 2006-04-12 at 00:30 +0200, Benoit Boissinot wrote:
On Wed, Apr 12, 2006 at 08:21:17AM +1000, Benjamin Herrenschmidt wrote:
I still think we shouldn't reward shit hardware by complicating
up our DMA mappings internals. :-)
BTW. In the meantime, can't that driver work in PIO
Stephen Hemminger wrote:
This is very familiar, and I just found the article I was thinking of:
http://lwn.net/Articles/92727/
I was also hit by that bug, on the same collection of websites, but that
particular problem was fixed for 2.6.8 or so. So I guess it is extremely
likely that my ISP
On Wed, 12 Apr 2006 01:06:09 +0100
Daniel Drake [EMAIL PROTECTED] wrote:
Stephen Hemminger wrote:
This is very familiar, and I just found the article I was thinking of:
http://lwn.net/Articles/92727/
I was also hit by that bug, on the same collection of websites, but that
particular
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Tue, 11 Apr 2006 11:10:54 -0700
Change the ethernet support routines to use constant address size.
This generates smaller faster code.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
Ok, but perhaps rework this to have a:
Daniel Drake wrote:
Stephen Hemminger wrote:
Turn off TCP window scaling, your performance will be limited but about
as good as you can get with a corrupting firewall in between.
[snip]
For anyone else interested, the ISP is NTL (UK). The fix:
echo 409616384 131072
Daniel Drake wrote:
Stephen Hemminger wrote:
This is very familiar, and I just found the article I was thinking
of: http://lwn.net/Articles/92727/
I was also hit by that bug, on the same collection of websites, but
that particular problem was fixed for 2.6.8 or so. So I guess it is
From: Brent Cook [EMAIL PROTECTED]
Fix the tx interrupt handler to free completed tx descriptors even
when NAPI is enabled. Otherwise, the tx queue would fill up resulting
in poor performance and NETDEV WATCHDOG: iface: transmit timed out
messages.
Signed-off-by: Brent Cook [EMAIL PROTECTED]
From: Dale Farnsworth [EMAIL PROTECTED]
After resetting the hardware on a tx_timeout, call netif_wake_queue()
only if we have free tx descriptors.
Also, attempt to recover if mv643xx_eth_start_xmit() is called when
there are fewer free tx descriptors than expected.
The BUG_ON() call we are
David S. Miller [EMAIL PROTECTED] wrote:
I agree, it should behave just like ipv4.
AOL
That would've made quite a few races that we had to fix non-existant
by definition :)
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmVHI~} [EMAIL PROTECTED]
Home Page:
76 matches
Mail list logo