Hi,
On ppc64 we want DMA writes (ie receive packets) to be aligned,
hopefully on a cacheline boundary. The NET_IP_ALIGN patch got us part of
the way, so that DMAs were 16 byte aligned.
The following patch addresses the other source of misalignment. Every
skb has 16 bytes of headroom reserved
Dear potential Speaker:
Based on your explicit interest expressed in the past and/or a recommendation
from a colleague who read your past publication(s), on behalf of the organizing
committee, we would like to extend a cordial
invitat on for you to submit a paper to the IPSI Transactions
Thomas Graf wrote:
Then how are you going to implement OSPF which has its own link state
detection?
OSPF doesn't have any link state detection. If carrier is up, hello
packets are sent out to the interface, if carrier is down, no hellos.
With hellos adjacencies can be formed to exchange routing
Yasuyuki KOZAKAI wrote:
I try to comment based on discussion with Yoshifuji-san and Miyazawa-san.
We think it's confusing for user to mix decrypted packets and pre-decrypted
ones to same hook. For example, if user want to accept packets encrypted by
transport mode ESP and drop others, he
Thomas Graf wrote:
* Thomas Graf [EMAIL PROTECTED] 2005-11-05 14:46
Assuming this is a separate bug, I'm not sure if this is the right
way to fix it. I think it would be better to rewrite the preferred
source address of all related local routes and only perform a
remove-and-add on the secondary
Hello,
Is it possible to use the sendfile system call with
UDP sockets in linux kernel ?
I saw in net/ipv4/udp.c a comment made by Hirokazu Takahashi,
which says sendfile() on UDP works now.
(this is in the linux 2.6.* kernels).
But when I am trying to write a user space program
which tries
Hello,
Thanks a lot for your answer !
Following your answer I did some reading in LDD3
I want to verify that I understood the interrupts mechanism right:
NF_IP_FORWARD
On 10/30/05, Harald Welte [EMAIL PROTECTED] wrote:
On Sat, Oct 29, 2005 at 03:39:30PM +0200, John Que wrote:
Hello,
I
Hi,
here an idea, how we might get rid of the nasty divide:
Stephen Hemminger wrote:
+/* RFC2861 Check whether we are limited by application or congestion
window + * This is the inverse of tcp_tso_should_defer
+ */
+static inline int tcp_cwnd_full(const struct tcp_sock *tp, u32 in_flight,
+
From: Stephen Hemminger [EMAIL PROTECTED]
Date: Mon, 7 Nov 2005 16:25:32 -0800
it could be limited to power of 2 and a shift?
That might not be a bad idea.
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at
From: Anton Blanchard [EMAIL PROTECTED]
Date: Tue, 8 Nov 2005 17:56:47 +1100
Any thoughts on this? The other option was to override dev_alloc_skb,
but I would prefer not to if possible.
There is still a lot of code that depends upon there being
a scratch of ~15 bytes there at the beginning of
From: Clemens Buchacher [EMAIL PROTECTED]
It's deprecated. Use %s, __FUNCTION__ instead.
Signed-off-by: Clemens Buchacher [EMAIL PROTECTED]
Signed-off-by: Maximilian Attems [EMAIL PROTECTED]
Signed-off-by: Domen Puncer [EMAIL PROTECTED]
Signed-off-by: Alexey Dobriyan [EMAIL PROTECTED]
Index:
On Tue, Nov 08, 2005 at 04:42:26AM +0100, Adrian Bunk wrote:
Looking at drivers/net/irda/Kconfig, it seems all DONGLE_OLD drivers
except EP7211_IR have more recent drivers.
My questions are:
Is there still any reason to keep the DONGLE_OLD drivers except
EP7211_IR?
The old
From: Anton Blanchard [EMAIL PROTECTED]
Date: Wed, 9 Nov 2005 03:12:20 +1100
Do you mean we expect exactly 16 bytes, or at least 16 bytes. Im happy
to ammend it so we always require the minimum 16 bytes, and can expand
it on architectures where it makes sense.
I meant at least 16 bytes.
To
Fix some of my bad spelling.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- skge-2.6.orig/drivers/net/skge.c
+++ skge-2.6/drivers/net/skge.c
@@ -130,7 +130,7 @@ static void skge_get_regs(struct net_dev
regs-len - B3_RI_WTO_R1);
}
-/* Wake on Lan only supported on
Go into power down mode on shutdown.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- skge-2.6.orig/drivers/net/skge.c
+++ skge-2.6/drivers/net/skge.c
@@ -1804,6 +1804,25 @@ static void yukon_mac_init(struct skge_h
skge_write16(hw, SK_REG(port, TX_GMF_CTRL_T), GMF_OPER_ON);
}
When skge is booted up, the PHY may be stuck in power down state
by the previous OS. So we may need to turn it on.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- skge-2.6.orig/drivers/net/skge.c
+++ skge-2.6/drivers/net/skge.c
@@ -2893,6 +2893,7 @@ static const char
Basic MII ioctl support for skge driver.
Signed-off-by: Stephen Hemminger [EMAIL PROTECTED]
--- skge-2.6.orig/drivers/net/skge.c
+++ skge-2.6/drivers/net/skge.c
@@ -37,6 +37,7 @@
#include linux/delay.h
#include linux/crc32.h
#include linux/dma-mapping.h
+#include linux/mii.h
#include
Minor update to skge driver:
* support MII tool interface.
handy for debugging, and some tools use it.
* fix some shutdown issues with some motherboards
--
Stephen Hemminger [EMAIL PROTECTED]
OSDL http://developer.osdl.org/~shemminger
-
To unsubscribe from this list:
Out of pseudo-morbid curiousity I was looking at inet_ehashfn, which looks
rather like this:
static inline int inet_ehashfn_orig(const __u32 laddr, const __u16 lport,
const __u32 faddr, const __u16 fport,
const int ehash_size)
{
int h = (laddr
The protocol field in ethernet headers is big-endian and should be
annotated as such. This patch allows detection of missing ntohs() calls
on the ethernet protocol field when sparse is run with __CHECK_ENDIAN__
defined.
This is a revised version that includes linux/types.h so that the
userspace
But when I am trying to write a user space program
which tries using sendfile with
UDP sockets I encounter erros. (with TCP sockets
sendfile works OK).
If you want help you need to supply more information. Please supply a
small sample program and also the error messages. You should also
David S. Miller wrote:
From: Rick Jones [EMAIL PROTECTED]
Date: Tue, 08 Nov 2005 11:11:17 -0800
What are the two lines xoring h with shifts of itself meant to be doing?
So that all the upper bits get mixed into the hash somehow, regardless
of the size of the hash table.
Is there a test
This patch contains code from Thomas Bogendoerfer which I have modified
somewhat. It works fine for me in my testing, but I would like
additional exposure to this code before submitting to mainline.
It should fix bugzilla bug #4219 as well.
The patch is based on 2.6.14-git11. The ethtool version
From: Rick Jones [EMAIL PROTECTED]
Date: Tue, 08 Nov 2005 11:33:16 -0800
Is there a test case I may have missed - otherwise it looks like it
was adding much randomness - particularly in the case of the IP's
being close to one another - the xor tends to knock-out the
upper bits of the IPs.
David S. Miller wrote:
From: Rick Jones [EMAIL PROTECTED]
Date: Tue, 08 Nov 2005 11:33:16 -0800
Is there a test case I may have missed - otherwise it looks like it
was adding much randomness - particularly in the case of the IP's
being close to one another - the xor tends to knock-out the
From: Rick Jones [EMAIL PROTECTED]
Date: Tue, 08 Nov 2005 12:12:19 -0800
Still seems to be doing as well - I'm assuming that when the
random IPs are generated that they do not have to be put into
network byte order, since they are randomly generated bits. When
the IPs aren't random, the
David S. Miller wrote:
From: Rick Jones [EMAIL PROTECTED]
Date: Tue, 08 Nov 2005 12:12:19 -0800
Still seems to be doing as well - I'm assuming that when the
random IPs are generated that they do not have to be put into
network byte order, since they are randomly generated bits. When
the IPs
Hi Yoshifuji,
/*
+ * find the first different bit between two addresses
+ * length of address must be a multiple of 32bits
+ */
+static inline int __ipv6_addr_diff(const void *token1, const void *token2, int
addrlen)
+{
+ const __u32 *a1 = token1, *a2 = token2;
+ int i;
+
+
Format output of ethtool -d for tg3 devices to make it more
readable. Each register is displayed as a 32-bit value and the
undefined regions in the register space are skipped.
Signed-off-by: Michael Chan [EMAIL PROTECTED]
diff -rup ethtool-3.a/ethtool.c ethtool-3.b/ethtool.c
---
Usual fix:
- b44_interrupt() does not schedule NAPI polling when the device is
going down;
- b44_close() waits for any scheduled NAPI polling before it starts
to release the private structures of the device.
Signed-off-by: Francois Romieu [EMAIL PROTECTED]
---
Signed-off-by: Francois Romieu [EMAIL PROTECTED]
---
74c2bceb5ef0b044b9c3206966494536f2c28fcf
diff --git a/drivers/net/b44.c b/drivers/net/b44.c
--- a/drivers/net/b44.c
+++ b/drivers/net/b44.c
@@ -29,8 +29,8 @@
#define DRV_MODULE_NAMEb44
#define PFX DRV_MODULE_NAME:
On Tuesday 08 November 2005 10:13, Greg KH wrote:
On Thu, Nov 03, 2005 at 07:55:38AM +1100, Herbert Xu wrote:
The recent rewrite of skb_copy_datagram_iovec broke the reception of
zero-size datagrams. This patch fixes it.
Signed-off-by: Herbert Xu [EMAIL PROTECTED]
Please apply it to
[NETFILTER] ctnetlink: check if protoinfo is present
This fixes an oops triggered from userspace. If we don't pass information
about the private protocol info, the reference to attr will be NULL. This is
likely to happen in update messages.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
[NETFILTER] nfnetlink: nfattr_parse() can never fail, make it void
nfattr_parse (and thus nfattr_parse_nested) always returns success. So we
can make them 'void' and remove all the checking at the caller side.
Based on original patch by Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald
Trying to build today's 2.6.14+git snapshot gives undefined references
to use_tempaddr
Looks like an ifdef got left out.
Signed-off-by: Peter Chubb [EMAIL PROTECTED]
net/ipv6/addrconf.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
Index: linux-2.6-import/net/ipv6/addrconf.c
* Hasso Tepper [EMAIL PROTECTED] 2005-11-08 17:14
OSPF doesn't care if there is connected route or not if !IFF_RUNNING. And
it doesn't matter if it's demand circuit or usual one. If kernel reports
link down (!IFF_RUNNING), no any packets are sent over this link -
hellos, LS updates etc. Hellos
[NETFILTER] ctnetlink: Add support to identify expectations by ID's
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit b98f5d02ab4cecbe4c3b038287e064973be27dcb
tree 135f77d58d8b18a19dc91c83e6f01c094f810be1
parent
[NETFILTER] PPTP helper: fix PNS-PAC expectation call id
The reply tuple of the PNS-PAC expectation was using the wrong call id.
So we had the following situation:
- PNS behind NAT firewall
- PNS call id requires NATing
- PNS-PAC gre packet arrives first
then the PNS-PAC expectation is matched,
[NETFILTER] ctnetlink: propagate error instaed of returning -EPERM
Propagate the error to userspace instead of returning -EPERM if the get
conntrack operation fails.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit
[NETFILTER] ctnetlink: kill unused includes
Kill some useless headers included in ctnetlink. They aren't used in any
way.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit 1f83ca8dd34407e20c17d0f172b8d9e59a326ba1
tree
[NETFILTER] ctnetlink: return -EINVAL if size is wrong
Return -EINVAL if the size isn't OK instead of -EPERM.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit 6ec2ee65ba91c23a2d36d11a9d92a435b30e56db
tree
[NETFILTER] ctnetlink: get_conntrack can use GFP_KERNEL
ctnetlink_get_conntrack is always called from user context, so GFP_KERNEL
is enough.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit 7b13da49573e77b6e122d9f9253728dac04d6658
tree
[NETFILTER] stop tracking ICMP error at early point
Currently connection tracking handles ICMP error like normal packets
if it failed to get related connection. But it fails that after all.
This makes connection tracking stop tracking ICMP error at early point.
Signed-off-by: Yasuyuki Kozakai
[NETFILTER] ctnetlink: add module alias to fix autoloading
Add missing module alias. This is a must to load ctnetlink on demand. For
example, the conntrack tool will fail if the module isn't loaded.
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
[NETFILTER] nfnetlink: only load subsystems if CAP_NET_ADMIN is set
Without this patch, any user can cause nfnetlink subsystems to be
autoloaded. Those subsystems however could add significant processing
overhead to packet processing, and would refuse any configuration messages
from
(seems like i can't even count, thus 16/15)
[NETFILTER] ctnetlink: ICMP_ID is u_int16_t not u_int8_t.
Signed-off-by: Krzysztof Piotr Oledzki [EMAIL PROTECTED]
Signed-off-by: Pablo Neira Ayuso [EMAIL PROTECTED]
Signed-off-by: Harald Welte [EMAIL PROTECTED]
---
commit
* Patrick McHardy [EMAIL PROTECTED] 2005-11-08 15:11
Yes, fixing it correctly looks very hard. Just changing the routes
doesn't seem right to me, someone might have added it with exactly
this prefsrc and doesn't want it to change, its also not clear how
to notify on this.
Right, OTOH this
Rick Jones [EMAIL PROTECTED] wrote:
Endianness often makes my head spin, so please be kind - are the addresses
being
converted to little endian before the hash (ntohl as it were) or is it just
happening as a consequence? If the former, why the conversion?
When you load a 32-bit integer
James Ketrenos wrote:
This series does the following:
[1/3] scripts/Lindent on ieee80211 subsystem.
[2/3] Fix problem with WEP unicast key index 0
[3/3] Stamp as 1.1.7
You can pull this series from:
rsync://bughost.org/repos/ieee80211-delta/.git/
Pulled, thanks.
Jeff
-
James Ketrenos wrote:
This series brings the ipw2100 and ipw2200 drivers up to be in sync w/
the current out of tree drivers (ipw2100 v1.1.3 and ipw2200 v1.0.8)
This series looks OK to me. However I appear to be missing patches #5
and #6, which causes the obvious cascade effect.
Can you
James Ketrenos wrote:
You can also pull this series from
rsync://bughost.org/repos/ipw-delta/.git/ (the GIT overlay is against
the ieee80211 overlay at rsync://bughost.org/repos/ieee80211-delta/.git/
vs. netdev-2.6#upstream)
Ignore my last email. I just pulled the quoted URL into netdev.
applied
-
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
applied 1-3
-
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
applied
-
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
applied all three ethtool-related patches
-
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
Please pull from 'upstream-linus' branch of
master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6.git
to receive the following updates, including a big ipw wireless update:
include/linux/eeprom.h| 136
Documentation/networking/README.ipw2100 | 119
Hello,
Thanks.
I use 2.6.11-1 kernel ,Fedora Core 4.
Here is the program I use and the error I get after
I send one UDP packet to this machine on the port it listens on ,9000.
(Note : input.txt is a nonempty small input file.)
#include sys/sendfile.h
#include sys/stat.h
#include sys/socket.h
Scott Feldman wrote:
I was going to say that eepro100's speedo_rx_link() does the same DMA
abuse as e100, but then I noticed one little detail: eepro100 sets both
EL (end of list) and S (suspend) bits in the RFD as it chains it to the
RFD list. e100 was only setting the EL bit. Hmmm,
58 matches
Mail list logo
