Am Dienstag 15 November 2005 03:43 schrieb jamal:
I'll have more time for a comment this evening, but let me ask one question
until then:
> 1) I dont think operstare_useroverride is needed.
You said the same to Thomas on IFF_WAIT. Both operstate_useroverride and
IFF_WAIT exist to allow userspa
We are seeing approx. 2x throughput increase for most workloads.
Absolute numbers for UFO throughput vary, the numbers below are for
pci-x 1.0 slots. In pci-x 2.0 slots we are seeing UFO performance close
to 10Gbps.
> -Original Message-
> From: Ananda Raju [mailto:[EMAIL PROTECTED]
> Sent
From: Matt Mackall <[EMAIL PROTECTED]>
Date: Mon, 14 Nov 2005 22:29:47 -0800
> Can we make any assumptions about the size and position of fragments.
> For instance, will the first N data bytes of a UDP packet all be in
> the same fragment?
Nope, they can be fragmented any way possible.
For packe
On Mon, Nov 14, 2005 at 09:41:30PM -0800, David S. Miller wrote:
> From: "David S. Miller" <[EMAIL PROTECTED]>
> Date: Mon, 14 Nov 2005 21:39:22 -0800 (PST)
>
> > From: Matt Mackall <[EMAIL PROTECTED]>
> > Date: Mon, 14 Nov 2005 21:23:58 -0800
> >
> > > What is "packet split" in this context?
> >
From: David Monro <[EMAIL PROTECTED]>
Date: Mon, 14 Nov 2005 13:02:42 +1030
> I don't know if it just needs to be told not to calulate the checksums
> for v6 packets, or what.
It generates a generic checksum calculation on all packets,
similarly to the Sun GEM chip, which ought to be totally
prot
Whitespace fixes
Signed-off-by: Jeff Kirsher <[EMAIL PROTECTED]>
Signed-off-by: John Ronciak <[EMAIL PROTECTED]>
Signed-off-by: Jesse Brandeburg <[EMAIL PROTECTED]>
diff -up linux-2.6/drivers/net/ixgb/ixgb.h linux-2.6.new/drivers/net/ixgb/ixgb.h
--- linux-2.6/drivers/net/ixgb/ixgb.h 2005-11-14
TSO fixes
- fix rare early completion when using TSO
- extra descriptor for the sentinel descriptor
Signed-off-by: Jeff Kirsher <[EMAIL PROTECTED]>
Signed-off-by: John Ronciak <[EMAIL PROTECTED]>
Signed-off-by: Jesse Brandeburg <[EMAIL PROTECTED]>
diff -up linux-2.6/drivers/net/ixgb/ixgb_main.c
Comments/Driver Version
Signed-off-by: Jeff Kirsher <[EMAIL PROTECTED]>
Signed-off-by: John Ronciak <[EMAIL PROTECTED]>
Signed-off-by: Jesse Brandeburg <[EMAIL PROTECTED]>
diff -up linux-2.6/drivers/net/ixgb/ixgb_main.c
linux-2.6.new/drivers/net/ixgb/ixgb_main.c
--- linux-2.6/drivers/net/ixgb/ix
ixgb driver update
Signed-off-by: Jeff Kirsher <[EMAIL PROTECTED]>
Signed-off-by: John Ronciak <[EMAIL PROTECTED]>
Signed-off-by: Jesse Brandeburg <[EMAIL PROTECTED]>
1. TSO fixes
- fix rare early completion when using TSO
- extra descriptor for the sentinel descriptor
2. Comments/Driver Version
From: YOSHIFUJI Hideaki <[EMAIL PROTECTED]>
Date: Tue, 15 Nov 2005 10:22:37 +0900 (JST)
> In article <[EMAIL PROTECTED]> (at Mon, 14 Nov 2005 18:42:30 +0800), Yan
> Zheng <[EMAIL PROTECTED]> says:
>
> > The "score.rule++" doesn't make any sense for me.
> > According to codes above, I think it s
From: "David S. Miller" <[EMAIL PROTECTED]>
Date: Mon, 14 Nov 2005 21:39:22 -0800 (PST)
> From: Matt Mackall <[EMAIL PROTECTED]>
> Date: Mon, 14 Nov 2005 21:23:58 -0800
>
> > What is "packet split" in this context?
>
> It's a mode of buffering used by the e1000 driver.
BTW, the issue is that in
From: Matt Mackall <[EMAIL PROTECTED]>
Date: Mon, 14 Nov 2005 21:23:58 -0800
> What is "packet split" in this context?
It's a mode of buffering used by the e1000 driver.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordom
On Mon, Nov 14, 2005 at 01:15:38PM -0800, Jeff Kirsher wrote:
> When using packet split, netpoll times out when doing a netdump.
What is "packet split" in this context? You ought to cc: the netdump
people as well, as it's not part of the mainline kernel.
--
Mathematics is the supreme nostalgia o
As in the case of Stefan, only the negative comments:
1) I think we need to separate the oper state from the rest; so
no need to add dormant to be in netdev_state_t.
2) Events need only be generated from/to down state
3) IFF_WAIT is not needed. A device goes from NOTPRESENT
to DOWN; and may go
-- Forwarded message --
Date: Mon, 14 Nov 2005 23:30:11 +0100
From: Charles-Edouard Ruault <[EMAIL PROTECTED]>
To: linux-kernel@vger.kernel.org
Subject: [BUG] kernel 2.6.14.2 breaks IPSEC
Hi All,
i've found out that since i've upgraded to kernel 2.6.14.2 ( problem also
applies to
On Mon, 2005-14-11 at 16:53 +0100, Krzysztof Halasa wrote:
> jamal <[EMAIL PROTECTED]> writes:
>
> "Dormant" is not a flag (1 bit) but rather a specific set of flags
> (exactly as there is no "running" nor "l2down" bits). So set_dormant*()
> (with bit flags) doesn't make much sense unless it chan
On Mon, 2005-14-11 at 16:48 +0100, Stefan Rompf wrote:
> Am Montag 14 November 2005 14:57 schrieb jamal:
>
> > My suggestion is at this point to ignore any L3 issues and have people
> > post their patches. RFC 2863 states MUST be taken into consideration.
> > Proper naming must be taken into accou
Mitch is several thousand miles away and may not get a chance to ACK to
this in time.
This patch looks fine. It functions properly. We have
started testing with it and no problems have been reported so far.
I would request that it be applied upstream, we have lost this window
many a times be
Andrew Morton <[EMAIL PROTECTED]> wrote:
>
> So far, i've looked into net/ipv4/esp4.c and i can confirm that the
> correct spi has been selected and inserted into the packet in function
> esp_output
> esph->spi = x->id.spi; ( line 97 ).
> It looks as if the corruption happens later down the stac
In article <[EMAIL PROTECTED]> (at Mon, 14 Nov 2005 18:42:30 +0800), Yan Zheng
<[EMAIL PROTECTED]> says:
> The "score.rule++" doesn't make any sense for me.
> According to codes above, I think it should be "hiscore.rule++;" .
Oops, you're right.
> Signed-off-by: Yan Zheng<[EMAIL PROTECTED]>
Ac
Begin forwarded message:
Date: Mon, 14 Nov 2005 23:30:11 +0100
From: Charles-Edouard Ruault <[EMAIL PROTECTED]>
To: linux-kernel@vger.kernel.org
Subject: [BUG] kernel 2.6.14.2 breaks IPSEC
Hi All,
i've found out that since i've upgraded to kernel 2.6.14.2 ( problem
also applies to 2.6.14 ),
I think this got fixed, didn't it?
If so, should we backport the fix into 2.6.14.x?
Begin forwarded message:
Date: Mon, 14 Nov 2005 20:30:21 +0100
From: Frank van Maarseveen <[EMAIL PROTECTED]>
To: linux-kernel@vger.kernel.org
Subject: [2.6.14.2] Debug: sleeping function called from invalid co
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] fix type of sysctl variables in nf_conntrack_ipv6
>
> These variables should be unsigned. This fixes sysctl handler for
> nf_ct_frag6_{low,high}_thresh.
>
> Signed-off-by: Yasuyuki Kozakai <[EMAIL PROTECT
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] nf_conntrack: fix possibility of infinite loop while evicting
> nf_ct_frag6_queue
>
> This synchronizes nf_ct_reasm with ipv6 reassembly, and fixes a possibility
> of an infinite loop if CPUs evict and cre
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] fix leak of fragment queue at unloading nf_conntrack_ipv6
>
> This patch makes nf_conntrack_ipv6 free all IPv6 fragment queues at module
> unloading time. Also introduce a BUG_ON if we ever again have leak
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER]: cleanup IPv6 Netfilter Kconfig
>
> This removes linux 2.4 configs in comments as TODO lists.
> And this also move the entry of nf_conntrack to top like IPv4 Netfilter
> Kconfig.
>
> Based on original patc
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] nfnetlink: unconditionally require CAP_NET_ADMIN
>
> This patch unconditionally requires CAP_NET_ADMIN for all nfnetlink
> mesages. It also removes the per-message cap_required field, since all
> existing
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER]: link 'netfilter' before ipv4
>
> Staticaly linked nf_conntrack_ipv4 requires nf_conntrack. but currently
> nf_conntrack is linked after it. This changes the order of ipv4 and netfilter
> to fix this.
>
>
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] nf_conntrack: Add missing code to TCP conntrack module
>
> Looks like the nf_conntrack TCP code was slightly mismerged: it does
> not contain an else branch present in the IPv4 version. Let's add that
> cod
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] ctnetlink: More thorough size checking of attributes
>
> Add missing size checks. Thanks Patrick McHardy for the hint.
>
> Signed-off-by: Pablo Neira Ayuso <[EMAIL PROTECTED]>
> Signed-off-by: Harald Welte
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] nfnetlink: skip size check if size not specified (== 0)
>
> Skip sizecheck if the size of the attribute wasn't specified, ie. zero.
>
> Signed-off-by: Pablo Neira Ayuso <[EMAIL PROTECTED]>
> Signed-off-by:
From: Harald Welte <[EMAIL PROTECTED]>
Date: Sun, 13 Nov 2005 22:51:59 +0100
> [NETFILTER] ctnetlink: use size_t to make gcc-4.x happy
>
> Make gcc-4.x happy. Use size_t instead of int. Thanks to Patrick McHardy
> for the hint.
>
> Signed-off-by: Pablo Neira Ayuso <[EMAIL PROTECTED]>
> Signed-of
Hi,
Will send you updated patch which uses __cpu_to_be64().
Here is quick test I did with 60k packet buffer. With UFO we reach upto 5.87
Gbits/sec compared to 3.55Gbits/sec non UFO case
With two threads we can reach upto 7.25 Gbits/sec where as non UFO still
remains at 3.5 Gbits/sec
linux:/hom
Unfortunately, I am currently several thousand miles away from my
Linux boxes, so I can't apply and test the patch. A quick
eyeball of the patch looks OK, though.
-Mitch
>-Original Message-
>From: John W. Linville [mailto:[EMAIL PROTECTED]
>These are the sysfs changes for the bonding
On Monday 14 November 2005 21:25, Ananda Raju wrote:
> Hi,
> This patch implements the UFO support in S2io driver. This patch uses the UFO
> interface available in linux-2.6.15 kernel.
Can you share some numbers on how much difference it makes vs non UFO?
> +#ifdef __BIG_ENDIAN
> +
> Signed-off-by: Jesse Brandeburg <[EMAIL PROTECTED]>
Signed-off-by: Ian McDonald <[EMAIL PROTECTED]>
Just a note to people given my history in flagging this one earlier -
I have tested this for latency and throughput. With the previous patch
I was getting 20 msec ping and 25 Mbit throughput. Now
Hi,
Kernel version 2.6.15 onwards provides UDP Fragmentation Offload (UFO)
This patch to ethtool provides UFO (UDP Fragmentation Offload) on/off support
using -K option similar to feature TSO.
To find out whether UFO is enabled or not use
#ethtool -k eth3
Offload parameters for eth3:
rx-checksu
On Mon, Nov 14, 2005 at 04:26:56PM -0500, Jeff Garzik wrote:
> John W. Linville wrote:
> > fec_8xx: make CONFIG_FEC_8XX depend on CONFIG_8xx
>
> unneeded, now? the fix to require CONFIG_FEC already went in.
Yeah, you might think that... :-)
Apparently CONFIG_FEC depends on one of several
John W. Linville wrote:
A collection of minor changes, mostly bug fixes. This also includes
Adrian's hostap.c rename to hostap_main.c (acked by Jouni).
I went ahead and pulled-in my e1000 changes and my fec_8xx Kconfig
change. I don't think they are controversial. (I copied the e1000
guys on
When using packet split, netpoll times out when doing a netdump.
Server logs:
--netdump[14973]: Got too many timeouts in handshaking, ignoring
client 172.0.2.250
--netdump[14973]: Got too many timeouts waiting for SHOW_STATUS for
client 172.0.2.250, rebooting it.
When packet split is not used, ne
For the four versions of hardware that we (currently) support microcode
download on, the default configuration of our receive interrupt mitigation
microcode was too aggressive, and caused unnecessary delays when pinging,
and low(er) throughput on single connection latency sensitive performance
On Tue, Nov 15, 2005 at 07:46:58AM +1100, Herbert Xu wrote:
> John W. Linville <[EMAIL PROTECTED]> wrote:
> >
> > diff --git a/drivers/net/wan/pc300_tty.c b/drivers/net/wan/pc300_tty.c
> > index 52f26b9..931cbdf 100644
> > --- a/drivers/net/wan/pc300_tty.c
> > +++ b/drivers/net/wan/pc300_tty.c
> >
Fedora-netdev!
This message is to announce the availability of a new Fedora-based
kernel repository. The kernels available there are based upon
the standard Fedora kernels, with the addition of current upstream
networking patches which are more recent than the Fedora kernel's
upstream base. More
John W. Linville <[EMAIL PROTECTED]> wrote:
>
> diff --git a/drivers/net/wan/pc300_tty.c b/drivers/net/wan/pc300_tty.c
> index 52f26b9..931cbdf 100644
> --- a/drivers/net/wan/pc300_tty.c
> +++ b/drivers/net/wan/pc300_tty.c
> @@ -689,7 +689,7 @@ static void cpc_tty_rx_work(void * data)
>
Hi,
This patch implements the UFO support in S2io driver. This patch uses the UFO
interface available in linux-2.6.15 kernel.
Please review the patch.
Signed-off-by: Ananda Raju <[EMAIL PROTECTED]>
---
diff -upNr old/drivers/net/s2io.c new/drivers/net/s2io.c
--- old/drivers/net/s2io.c 2005-
> My suggestion is at this point to ignore any L3 issues and have people
> post their patches. RFC 2863 states MUST be taken into consideration.
> Proper naming must be taken into account.
Split up in 3 patches, not implementing the bits allowing userspace to
trigger leaving dormant state.
[NET]
Ok..not really a bug report since I don't know exactly what
all went wrong, or how...but here's what I saw. If anyone
else has seen anything wierd with 2.6.14 networking, especially
related to NFS and/or e100, please let me know.
I left 2.6.14 (plus my hacks, could be the problem as always)
runn
Am Montag 14 November 2005 17:44 schrieb Krzysztof Halasa:
[As Jamal answers later, I'll just comment on the technical questions]
> Why do you do that instead of atomic write?
as described in net/core/dev.c:
* The @dev_base list is protected by @dev_base_lock and the rtln
* semaphore.
*
* P
John W. Linville wrote:
On Sun, Nov 13, 2005 at 10:24:15PM -0800, Ben Greear wrote:
This been dealt with already?
/home/greear/git/linux-2.6/drivers/net/wan/pc300_tty.c: In function
`cpc_tty_rx_work':
/home/greear/git/linux-2.6/drivers/net/wan/pc300_tty.c:692: warning:
passing arg 1 of `kfre
[EMAIL PROTECTED] a écrit :
Use hits to speed up the SACK processing. Various forms
of this have been used by TCP developers (Web100, STCP, BIC)
to avoid the 2x linear search of outstanding segments.
--- net-2.6.orig/include/linux/tcp.h
+++ net-2.6/include/linux/tcp.h
@@ -307,6 +307,21 @@ struc
Stefan Rompf <[EMAIL PROTECTED]> writes:
> +++ linux-2.6.14/net/core/link_watch.c
> @@ -75,7 +75,14 @@ void linkwatch_run_queue(void)
> clear_bit(__LINK_STATE_LINKWATCH_PENDING, &dev->state);
>
> if (dev->flags & IFF_UP) {
> - if (netif_carrier_ok(
I want to apologize in advance - I wont be able to comment on anything
for another 10 hours or so. I will comment tonight.
cheers,
jamal
-
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.o
jamal <[EMAIL PROTECTED]> writes:
> My suggestion is at this point to ignore any L3 issues and have people
> post their patches. RFC 2863 states MUST be taken into consideration.
> Proper naming must be taken into account.
I'll not repost my patch because you already have it and it lacks only
co
Am Montag 14 November 2005 14:57 schrieb jamal:
> My suggestion is at this point to ignore any L3 issues and have people
> post their patches. RFC 2863 states MUST be taken into consideration.
> Proper naming must be taken into account.
here we go. I've removed OPER_DORMANTL3* as requested and ch
jamal <[EMAIL PROTECTED]> writes:
> Thats the idea. But for now - I suggest we take L3 out of the equation
> and we revisit after the first agreeable patch is out.
L3 as in DORMANT_L3{DOWN,UP}? Sure.
>> Come on, making a patch showing the general idea as well as the
>> problematic details (locki
Stefan,
After talking to Thomas it seems to me there are no conceptual
differences - at least he and Krzysztof agree on the principle of
of the dormant state.
So what you mention below is more of implementation issues. All
implementation issues are resolvable.
My suggestion is at this point to i
On Sun, Nov 13, 2005 at 10:24:15PM -0800, Ben Greear wrote:
> This been dealt with already?
>
> /home/greear/git/linux-2.6/drivers/net/wan/pc300_tty.c: In function
> `cpc_tty_rx_work':
> /home/greear/git/linux-2.6/drivers/net/wan/pc300_tty.c:692: warning:
> passing arg 1 of `kfree' discards qual
The "score.rule++" doesn't make any sense for me.
According to codes above, I think it should be "hiscore.rule++;" .
Signed-off-by: Yan Zheng<[EMAIL PROTECTED]>
Index: net/ipv6/addrconf.c
--- a/net/ipv6/addrconf.c 2005-11-13 12
Hi.
I'm doing a graduation research where the goal is create a new protocol
in the linux kernel. The protocol choosen was TCPXM, an hybrid reliable
sender-initiated multicast/unicast aimed for small environment such as
grids.
In the last two months I studied the network subsystem, thinking a
In isl_38xx.c
In routine isl38xx_trigger-device
Move unnecessary udelay/register read.
This is only required when hand-compiling the driver and
setting VERBOSE > SHOW_ERROR_MESSAGES
Signed-off-by: Roger While <[EMAIL PROTECTED]>
diff -Naur linux-2.6.15rc1orig/drivers/net/wireless/prism5
Hi Harald,
would you mind merging the prink()s ...
Harald Welte wrote:
> diff --git a/net/bridge/netfilter/ebt_log.c b/net/bridge/netfilter/ebt_log.c
> --- a/net/bridge/netfilter/ebt_log.c
> +++ b/net/bridge/netfilter/ebt_log.c
> @@ -55,17 +57,19 @@ static void print_MAC(unsigned char *p)
> }
>
Ben Greear <[EMAIL PROTECTED]> wrote:
>
> I don't see how this code (from 2.6.14, skbuff.h) can be correct. Shouldn't
> we be doing some subtraction against the base time value somewhere? Or maybe
> just the
> comment is incorrect?
The base is now zero :)
Treat this as a reminder for us to fi
On Fri, Nov 11, 2005 at 03:00:10PM -0800, David S. Miller wrote:
> There is another bug I see there in the cleanup path too.
> We don't unregister the inet6 protosw on the failure path.
> But this appears totally harmless since you can't reference
> the inet6 proto switch table once you unload the
David S. Miller wrote:
From: Ben Greear <[EMAIL PROTECTED]>
Date: Wed, 09 Nov 2005 13:41:02 -0800
So, I am asking the TSO folks: Will you pay any attention to
a bug report against 2.6.13.2, or would I just be wasting my
time?
If the answer is yes, then I'll get on it..otherwise, I'll wait un
64 matches
Mail list logo
