Hello Stephen,
as posted to LKLM and netdev recently, I'm seeing reproducable hangs with the
new sky2 of 2.6.16-rc3 on my Marvel 88E8053.
I was advised to mail you directly, and I've filed a bugreport #6084 on kernel
bugzilla, my config, lspci output and dmesg are attached there. Please tell
m
Joy,
Thanks for your comment and sorry for the delay. Did you mean a
separate error code for 'null' context? The current code catches the
case when the sid is SECSID_NULL, and returns ENOPROTOOPT. The
question is whether we want to create a different error code for this
case. Any suggestions?
removed magic number 33 as suggested by Arnaldo
32 bit modular socket ioctl emulation for 64 bit kernel
This patch is a re-submit of an earlier patch submitted by Andrew Hendry
that did not make it into the Linux Kernel. Here is some more information
about this patch.
This patch allows use of
32 - 64 converstion for patch 5
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.16-rc3-vanilla/net/x25/af_x25.c
linux-2.6.16-rc3/net/x25/af_x25.c
--- linux-2.6.16-rc3-vanilla/net/x25/af_x25.c 2006-02-16 15:33:48.0
+1100
+++ linux-2.6.16-rc3/net/x25/af_
32 bit modular socket ioctl emulation for 64 bit kernel
This patch allows 32 bit x25 module structures to be passed to
a 64 bit kernel via ioctl using the new compat_sock_ioctl registration
mechanism instead of the obsolete 'register_ioctl32_conversion into hash table'
mechanism.
Signed-off-b
32 bit modular socket ioctl emulation for 64 bit kernel
This patch allows an x25 server application to run on a 64 bit kernel, by
fixing the following error message from the kernel.
T2 kernel: schedule_timeout:
wrong timeout value from 88164796
Signed-off-by:Shaun Pe
Includes correction from Arnaldo's suggestions.
32 bit modular socket ioctl emulation for 64 bit kernel
The following patch provides 32 bit userland ioctl support for modular (x.25
type)
socket ioctls in a 64 bit kernel. Since the the register_ioctl32_conversion()
is now obsolete, this patch pr
32 bit modular socket ioctl emulation for 64 bit kernel
This patch is the first step towards migration of the 'handler functions'
for 32-64 bit userspace-kernel conversion, away from the ioctl32_hash_table.
It will be used by the x25 socket layer.
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED
Thanks for that Aranaldo
I have corrected this so as to be compliant with OSI Network services
for DTE facilities and rebuilt and retested the patches
Corrected patches follow.
On Wed, 2006-02-15 at 20:35 -0200, Arnaldo Carvalho de Melo wrote:
> On 2/15/06, Shaun Pereira <[EMAIL PROTECTED]> wro
Let's try this again. How does this look, Robert?
There's a race in pktgen which can lead to a double
free of a pktgen_dev's skb. If a worker thread is in
the midst of doing fill_packet(), and the controlling
thread gets a "stop" message, the already freed skb
can be freed once again in pktgen_st
This is a simplified version of the netlink_has_listeners patch
to let netlink multicast senders check for listeners before
generating messages. The last version also had a bug that made
the mask only contain the subscribed groups of the socket used
for the last bind operation. This version should
(Sorry, resent due to wrong message in CC, sending again to keep consistency)
I have been having problems for some while with sky2 on my machine even
with sky2 0.15 (as in 2.6.16-rc3-git4)
But, i found a patch from Stephen that noone seemed to have commented
on, so i thought i'd better try it out
From: Patrick McHardy <[EMAIL PROTECTED]>
Date: Wed, 15 Feb 2006 18:25:44 +0100
> Another fix related to the netfilter IPsec patches.
Applied, thanks a lot.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at ht
From: Patrick McHardy <[EMAIL PROTECTED]>
Date: Wed, 15 Feb 2006 18:25:31 +0100
> This patch fixes the crash in xfrm4_output_finish reported by
> multiple people.
Looks good, applied.
Thanks Patrick.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to
32 bit modular socket ioctl emulation for 64 bit kernel
This patch is a re-submit of an earlier patch submitted by Andrew Hendry
that did not make it into the Linux Kernel. Here is some more information
about this patch.
This patch allows use of the optional user facility to insert ITU-T
(http
Allow patch 5 to use 32-64 bit conversion mechanism
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.16-rc3-vanilla/net/x25/af_x25.c
linux-2.6.16-rc3/net/x25/af_x25.c
--- linux-2.6.16-rc3-vanilla/net/x25/af_x25.c 2006-02-15 11:17:03.0
+1100
+++ linux-2.
32 bit modular socket ioctl emulation for 64 bit kernel
This patch is the first step towards migration of the 'handler functions'
for 32-64 bit userspace-kernel conversion, away from the ioctl32_hash_table.
It will be used by the x25 socket layer.
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED
32 bit modular socket ioctl emulation for 64 bit kernel
This patch allows 32 bit x25 module structures to be passed to
a 64 bit kernel via ioctl using the new compat_sock_ioctl registration
mechanism instead of the obsolete 'register_ioctl32_conversion into hash table'
mechanism.
Signed-off-b
32 bit modular socket ioctl emulation for 64 bit kernel
This patch allows an x25 server application to run on a 64 bit kernel, by
fixing the following error message from the kernel.
T2 kernel: schedule_timeout:
wrong timeout value from 88164796
Signed-off-by:Shaun Pe
Sorry Dave,
My mail client line wrapped the patch. I will resend the lot
Please ignore the previous set.
The following patch provides 32 bit userland ioctl support for modular (x.25
type)
socket ioctls in a 64 bit kernel. Since the the register_ioctl32_conversion()
is now obsolete, this patch
On 2/15/06, Shaun Pereira <[EMAIL PROTECTED]> wrote:
> + switch (*p) {
> + case X25_FAC_CALLING_AE:
> + if (p[1] > 33)
> + break;
> + dte_facs->calling_len = p[2];
> +
Allow patch 5 to use 32-64 bit conversion mechanism
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.16-rc3-vanilla/net/x25/af_x25.c
linux-2.6.16-rc3/net/x25/af_x25.c
--- linux-2.6.16-rc3-vanilla/net/x25/af_x25.c 2006-02-15
11:17:03.0 +1100
+++ linux-2.6.
Hi
The following patch provides 32 bit userland ioctl support for modular
(x.25 type) socket ioctls in a 64 bit kernel. Since the the
register_ioctl32_conversion() is now obsolete, this patch provides a
mechanism to allow 32 bit user space ioctls to reach the kernel.
Signed-off-by:Shaun Pereira
This patch is the first step towards migration of the 'handler
functions' for 32-64 bit userspace-kernel conversion, away from the
ioctl32_hash_table. It will be used by the x25 socket layer.
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
Acked-by: Arnd Bergmann <[EMAIL PROTECTED]>
diff -uprN
This patch allows 32 bit x25 module structures to be passed to
a 64 bit kernel via ioctl using the new compat_sock_ioctl registration
mechanism instead of the obsolete 'register_ioctl32_conversion into hash
table' mechanism.
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
Acked-by: Arnd Bergma
This patch is a re-submit of an earlier patch submitted by Andrew Hendry
that did not make it into the Linux Kernel. Here is some more
information about this patch. Hope this is helpful.
This patch allows use of the optional user facility to insert ITU-T
(http://www.itu.int/ITU-T/) specified DTE
On Wednesday 15 February 2006 18:12, you wrote:
> >>I need the IPW card to monitor the traffic of my bcm43xx cards.
> >>I really need working monitor mode in the ipw2200 at the moment.
> >>If someone knows a way to get it working, _please_ tell me.
> >>(hwcrypto=0 does not work. Already tried that)
This patch allows an x25 server application to run on a 64 bit kernel,
by fixing the following error message from the kernel.
T2 kernel: schedule_timeout:
wrong timeout value from 88164796
Signed-off-by:Shaun Pereira <[EMAIL PROTECTED]>
Acked-by: Arnd Bergmann <[EMAIL
On Wed, Feb 15, 2006 at 11:02:23PM +0100, Michael Buesch wrote:
> On Wednesday 15 February 2006 18:12, you wrote:
> > The problem is a version conflict between the boot firmware provided in
> > the 2.4 firmware package and the monitor mode firmware file. We're
> > working to correct that problem.
On Wednesday 15 February 2006 18:12, you wrote:
> Michael Buesch wrote:
>
> >On Wednesday 15 February 2006 01:18, you wrote:
> >
> >
> >>Hi,
> >>
> >>When I put my ipw2200 into monitor mode, I get my dmesg spammed with this:
> >>ipw2200: Firmware error detected. Restarting.
> >>The card looses
On Wednesday 15 February 2006 20:21, you wrote:
> This patch fixes an oops when bcm43xx-d80211 module is unloaded.
This is already fixed in my tree.
> Signed-off-by: Jiri Benc <[EMAIL PROTECTED]>
>
> --- dscape.testing.orig/drivers/net/wireless/bcm43xx-d80211/bcm43xx_main.c
> 2006-02-15 20:0
On Wed, Feb 15, 2006 at 03:21:50PM +0200, Ilia Sotnikov wrote:
>
> Totally agree but perhaps we should ask the confirmation from someone?
That's what this list is for :) Send a patch and if there are no objections
it should be go in.
> The removal of routing decisions by TOS field would made the
TCPV4 inside an IPSEC tunnel doesn't react to PMTU Discovery on first
connection attempt. After that connection hangs. Subsequent TCP
connections to the same peer however use the correct MTU. ICMP
Fragmentation Needed messages are also fall into the IPSEC tunnel.
As far as I can tell the problem a
On Wed, Feb 08, 2006 at 06:35:27PM +0900, Horms wrote:
> > [PATCH] Better fixup for the orinoco driver
> >
> > The latest kernel added a pretty ugly fix for the orinoco etherleak bug
> > which contains bogus skb->len checks already done by the caller and causes
> > copies of all odd sized frames (
This patch fixes an oops when bcm43xx-d80211 module is unloaded.
Signed-off-by: Jiri Benc <[EMAIL PROTECTED]>
--- dscape.testing.orig/drivers/net/wireless/bcm43xx-d80211/bcm43xx_main.c
2006-02-15 20:06:07.0 +0100
+++ dscape.testing/drivers/net/wireless/bcm43xx-d80211/bcm43xx_main.c
On Wed, 15 Feb 2006, David S. Miller wrote:
>
> I don't see how this fixes the race
>
Well, of course. I was just checking that you were
paying attention.
It looks like the right thing to do is to let the
worker thread free up the skb in all cases. New
patch soon.
--
Arthur
-
To un
Arthur Kepner writes:
> There's a race in pktgen which can lead to a double
> free of a pktgen_dev's skb. If a worker thread is in
> the midst of doing fill_packet(), and the controlling
> thread gets a "stop" message, the already freed skb
> can be freed once again in pktgen_stop_device().
On Wed, Feb 15, 2006 at 01:23:19PM +0200, Pekka J Enberg wrote:
> > Hmm. Either I am totally confused or we don't even attempt suspend/resume
> > for eisa and mca bus devices. Care to try this patch?
On Wed, 2006-02-15 at 12:40 +, Russell King wrote:
> Please don't use struct device_driver su
This patch fixes the crash in xfrm4_output_finish reported by
multiple people.
[XFRM]: Fix SNAT-related crash in xfrm4_output_finish
When a packet matching an IPsec policy is SNATed so it doesn't match any
policy anymore it looses its xfrm bundle, which makes xfrm4_output_finish
crash because of
Another fix related to the netfilter IPsec patches.
[NETFILTER]: Don't invoke okfn in CONFIG_NETFILTER=n variant of nf_hook()
nf_hook() is supposed to call the netfilter hook and return control of the
packet back to the caller in case it may pass, the okfn is only used for
queueing.
Signed-off-b
Michael Buesch wrote:
>On Wednesday 15 February 2006 01:18, you wrote:
>
>
>>Hi,
>>
>>When I put my ipw2200 into monitor mode, I get my dmesg spammed with this:
>>ipw2200: Firmware error detected. Restarting.
>>The card looses packets, obviously.
>>
>>I googled, and I found a website [1], stati
On 2/15/06, Herbert Xu <[EMAIL PROTECTED]> wrote:
> How about dropping TOS from the hash code instead? How many people
> out there actually route using the TOS and use more than two TOS values?
Totally agree but perhaps we should ask the confirmation from someone?
The removal of routing decisions
Pekka J Enberg <[EMAIL PROTECTED]> wrote:
>
> Hi Andrew,
>
> On Wed, 15 Feb 2006, Andrew Morton wrote:
> > Problem is, it doesn't resume correctly either with or without the patch:
> > it needs rmmod+modprobe to get it going again. (Which is better than the
> > aic7xxx driver, which has a co
On Wed, Feb 15, 2006 at 01:23:19PM +0200, Pekka J Enberg wrote:
> Hmm. Either I am totally confused or we don't even attempt suspend/resume
> for eisa and mca bus devices. Care to try this patch?
Please don't use struct device_driver suspend/resume methods.
--
Russell King
Linux kernel2.6
Pekka J Enberg <[EMAIL PROTECTED]> wrote:
>
> On Wed, 15 Feb 2006, Andrew Morton wrote:
> > I have a 3c509, and I'm not afraid to use it!
> >
> > Problem is, it doesn't resume correctly either with or without the patch:
> > it needs rmmod+modprobe to get it going again. (Which is better than the
On Wed, Feb 15, 2006 at 09:34:23AM +0200, Ilia Sotnikov wrote:
>
> In some cases the patch could led to a large number of lookups on
> rt_hash_code over all possible TOS values (additional 8 passes). In
> complex configurations with many routing tables it could be
> computationally expensive. Tha
Hi Andrew,
On Wed, 15 Feb 2006, Andrew Morton wrote:
> Problem is, it doesn't resume correctly either with or without the patch:
> it needs rmmod+modprobe to get it going again. (Which is better than the
> aic7xxx driver, which has a coronary and panics the kernel on post-resume
> reboot).
Hmm.
On Wed, 15 Feb 2006, Andrew Morton wrote:
> I have a 3c509, and I'm not afraid to use it!
>
> Problem is, it doesn't resume correctly either with or without the patch:
> it needs rmmod+modprobe to get it going again. (Which is better than the
> aic7xxx driver, which has a coronary and panics the
Pekka Enberg <[EMAIL PROTECTED]> wrote:
>
> I am looking for someone with 3c509 netword card that can do
> suspend/resume to test this patch.
I have a 3c509, and I'm not afraid to use it!
Problem is, it doesn't resume correctly either with or without the patch:
it needs rmmod+modprobe to get it
From: Stephen Hemminger <[EMAIL PROTECTED]>
Date: Tue, 14 Feb 2006 15:41:42 -0800
> Looks like somebody forgot to use the _bh spin_lock variant. We ran into a
> deadlock where br->hello_timer expired while br_stp_disable_br() walked
> br->port_list.
>
> Signed-off-by: Adrian Drzewiecki <[EMAIL
From: Arthur Kepner <[EMAIL PROTECTED]>
Date: Tue, 14 Feb 2006 13:55:58 -0800 (PST)
>
> There's a race in pktgen which can lead to a double
> free of a pktgen_dev's skb. If a worker thread is in
> the midst of doing fill_packet(), and the controlling
> thread gets a "stop" message, the already fr
From: Herbert Xu <[EMAIL PROTECTED]>
Date: Tue, 14 Feb 2006 12:49:24 +1100
> On Mon, Feb 13, 2006 at 06:25:01PM +0100, Patrick McHardy wrote:
> > I finally got around to fixing the "ip_finish_output2: No header cache
> > and no neighbour!" problem reported by Andi Kleen. Instead of rerouting
> > t
[PKT_SCHED]: Keep backlog counter in sch_sfq
Keep backlog counter in SFQ qdisc to make it usable as child qdisc with
RED.
Signed-off-by: Patrick McHardy <[EMAIL PROTECTED]>
---
commit 78564b29480707a727ef83e5186036d2a280dd57
tree a87b1f211fa627c0d8da240d5e042ffac19324ad
parent 7a2ef66a3ef9321be0
[PKT_SCHED]: Dump child qdisc handle in sch_{atm,dsmark}
A qdisc should set tcm_info to the child qdisc handle in its class dump
function.
Signed-off-by: Patrick McHardy <[EMAIL PROTECTED]>
---
commit 3394462061e134f6f4e0d5d4910b63227d810ab7
tree 14348d5f7e8fad88f1d66d7a1a0cef86c147b1ee
parent a
[PKT_SCHED]: Qdisc drop operation is optional
The drop operation is optional and qdiscs must check if childs support it.
Signed-off-by: Patrick McHardy <[EMAIL PROTECTED]>
---
commit a69d87ea9357e7f1e8979328dd4878a5c79e6f02
tree f68f812188bc0cec40e6b674d9bc5ddd24f6835c
parent 25bf368b3d98668c5d5
[PKT_SCHED]: Convert sch_red to a classful qdisc
Convert sch_red to a classful qdisc. All qdiscs that maintain accurate
backlog counters are eligible as child qdiscs. When a queue limit larger
than zero is given, a bfifo qdisc is used for backwards compatibility.
Current versions of tc enforce a l
Hi Dave,
following are a couple of net/sched patches for 2.6.17. Please apply if Jamal
doesn't have objections.
net/sched/sch_atm.c|1
net/sched/sch_dsmark.c |1
net/sched/sch_netem.c |4 -
net/sched/sch_prio.c |2
net/sched/sch_red.c| 179
[PKT_SCHED]: Restore TBF change semantic
When TBF was converted to a classful qdisc, the semantic of the limit
parameter was broken. On initilization an inner bfifo qdisc is created
for backwards compatibility, when changing parameters however the new
limit is ignored and the current child qdisc r
58 matches
Mail list logo
