On 09-01-2007 17:08, [EMAIL PROTECTED] wrote:
Hello,
I have been trying to get the RED qdisc and ECN to work for the past few weeks
and all my experiments have failed. Here is the setup I am using.
Src -- R1 -- R2 -- Dst
Between Src and R1 is a 100Mbps link and between R1 and R2 a
Hello,
I have a machine with 2 dual core CPUs. This machine runs Fedora Core 6.
I have two Intel e1000 GigaBit network cards on this machine; I use bonding so
that the machine assigns the same IP address to both NICs ;
It seems to me that bonding is configured OK, bacuse when running:
cat
Hi,
During the holiday season, I posted a patch that fixed this problem without
using spinlocks nor disabling interrupts.
http://marc.theaimsgroup.com/?l=linux-netdevm=116649413613845w=2
With this patch applied, I confirmed that the system doesn't panic.
I think this patch can fix this
Hi Dave,
Since we stop using dev_alloc_skb on the IrDA TX frame, we constantly run
into the case of the skb headroom being 0, and thus we call skb_cow for
every IrDA TX frame.
This patch uses a local buffer and memcpy the skb to it, saving us a
kmalloc for each of those IrDA TX frames.
With USB2.0 bulk out MTU can be 512 bytes, so checking it only for 64 bytes is
incorrect.
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
---
drivers/net/irda/irda-usb.c |2 --
1 files changed, 0 insertions(+), 2 deletions(-)
diff --git a/drivers/net/irda/irda-usb.c
Michael Tokarev [EMAIL PROTECTED] wrote:
Note there's no funny/interesting hardware involved, like network cards with
tcp checksumming offload capabilities (this is plain dumb 8139 card).
The 8139 card might be dumb, but the driver isn't :) It emulates
checksum offload in software, meaning
Hi Mark,
On 1/15/07, Mark Ryden [EMAIL PROTECTED] wrote:
I have a machine with 2 dual core CPUs. This machine runs Fedora Core 6.
I have two Intel e1000 GigaBit network cards on this machine; I use bonding so
that the machine assigns the same IP address to both NICs ;
cat /proc/interrupts
On 10-01-2007 11:01, Patrick McHardy wrote:
[IPROUTE]: Replace usec by time in function names
Rename functions containing usec since they don't necessarily return
usec units anymore.
Signed-off-by: Patrick McHardy [EMAIL PROTECTED]
---
...
diff --git a/tc/q_cbq.c b/tc/q_cbq.c
index
On Sat, 2007-01-13 at 18:17 +0100, Maarten Lankhorst wrote:
Remove CONFIG_NET_WIRELESS
Nothing uses this, and it breaks the kernel build if a wireless device is
used with a unsupported type of bus.
Verified this with a grep.
I don't really care about the symbol and I'm in favour of removing
On 10-01-2007 11:01, Patrick McHardy wrote:
[IPROUTE]: Introduce tc_calc_xmitsize and use where appropriate
Add tc_calc_xmitsize() as complement to tc_calc_xmittime(), which calculates
the size that can be transmitted at a given rate during a given time.
Replace all expressions of the form
Johannes Berg schreef:
On Sat, 2007-01-13 at 18:17 +0100, Maarten Lankhorst wrote:
Remove CONFIG_NET_WIRELESS
Nothing uses this, and it breaks the kernel build if a wireless device is
used with a unsupported type of bus.
Verified this with a grep.
I don't really care about the
On Mon, 2007-01-15 at 13:55 +0100, Maarten Lankhorst wrote:
Johannes Berg schreef:
On Sat, 2007-01-13 at 18:17 +0100, Maarten Lankhorst wrote:
Remove CONFIG_NET_WIRELESS
Nothing uses this, and it breaks the kernel build if a wireless device is
used with a unsupported type of bus.
Herbert Xu wrote:
Michael Tokarev [EMAIL PROTECTED] wrote:
Note there's no funny/interesting hardware involved, like network cards with
tcp checksumming offload capabilities (this is plain dumb 8139 card).
The 8139 card might be dumb, but the driver isn't :) It emulates
checksum offload in
On Mon, 15 Jan 2007 13:31:06 +, Johannes Berg wrote:
On Mon, 2007-01-15 at 13:55 +0100, Maarten Lankhorst wrote:
Enabling this doesn't cause anything to fail, but my wireless router
doesn't have a pci bus, but instead a native SSB, so CONFIG_NET_WIRELESS
isn't selected. This in turn
Hello.
I am trying to get a RTL8111 (RealTek ethernet controller) running w.
the r8169 kernel module. I am using kernel 2.6.19.2 on a
LinuxFromScratch system; the motherboard on which said RTL8111 sits is
an Asus P5B.
lspci says (regarding the ethernet chip):
Jarek Poplawski wrote:
On 10-01-2007 11:01, Patrick McHardy wrote:
[IPROUTE]: Introduce tc_calc_xmitsize and use where appropriate
Add tc_calc_xmitsize() as complement to tc_calc_xmittime(), which calculates
the size that can be transmitted at a given rate during a given time.
Replace all
The 3c59x.c in kernel 2.6.18 (and as I see later ones too) attempts
to enable PME from the D0 state. The PME config space on Dell Optiplexs
for this chip has a zero in the capabilities as it doesn't 'wake from d0'.
So the pci_wake call fails, its result is not tested, so no error is reported.
Hello all.
The 3c59x.c in kernel 2.6.18 (and as I see later ones too) attempts
to enable PME from the already awake D0 state. The PME config space on Dell
Optiplexs
for this chip has a zero in the capabilities for this bit-- no 'wake from d0'.
The pci_enable_wake in 2.6.18 tests the
On Mon, Jan 15, 2007 at 03:38:57PM +0100, Jens Stroebel wrote:
During the use of network connections, we experience network transfer
stops during which a transfer seems to stall completely for many
seconds, after which the transfer runs as if nothing happened.
Addition:
Trying to debug the
Kenzo Iwami wrote:
With this patch applied, I confirmed that the system doesn't panic.
I think this patch can fix this problem.
Does this patch have problems.
Kenzo,
thanks for staying patient while most of us were out or busy. Apart from acknowledging
that you might have fixed a problem
(trying again, this time to the correct maintainer)
All,
Similar to this commit:
http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=d15e9c4d9a75702b30e00cdf95c71c88e3f3f51e
It's not safe in cp_start_xmit to blindly call spin_lock_irq and then
spin_unlock_irq,
From: Auke Kok [EMAIL PROTECTED]
ixgb: Don't stop queue unnecesarily
We don't need to stop twice in ixgb_xmit_frame.
Signed-off-by: Auke Kok [EMAIL PROTECTED]
diff --git a/drivers/net/ixgb/ixgb_main.c b/drivers/net/ixgb/ixgb_main.c
index 51bd7e8..83f4d67 100644
---
Mark Ryden wrote:
Hello,
I have a machine with 2 dual core CPUs. This machine runs Fedora Core 6.
I have two Intel e1000 GigaBit network cards on this machine; I use
bonding so
that the machine assigns the same IP address to both NICs ;
It seems to me that bonding is configured OK, bacuse
Stephen,
After some days of uptime, I've been seeing 'transmit timed out'
messages [1]. Let me know if there is any useful debugging you'd like.
--- [1]
sky2 v1.10 addr 0xdfb0 irq 16 Yukon-EC (0xb6) rev 1
sky2 eth1: addr 00:03:2d:05:9c:27
sky2 lan0: enabling interface
sky2 lan0: Link is up
At 11:00 AM 1/15/2007 -0500, Dan Williams wrote:
On Mon, 2007-01-15 at 09:12 -0600, Harry Coin wrote:
Hello all.
The 3c59x.c in kernel 2.6.18 (and as I see later ones too) attempts
to enable PME from the already awake D0 state. The PME config space on
Dell
Optiplexs
for this chip has a
Hello from Iowa.
Below please find a fix to the Wake On Lan function in the e100.c (intel
10/100) driver. With the original driver distributed in kernel 2.6.18 in
debian etch, wake on lan did not work. This was tested on 14 dell
optiplexes with built-in ethernet chips in a totally
Harry Coin wrote:
Hello from Iowa.
Below please find a fix to the Wake On Lan function in the e100.c (intel
10/100) driver. With the original driver distributed in kernel 2.6.18
in debian etch, wake on lan did not work. This was tested on 14 dell
optiplexes with built-in ethernet chips
Please reproduce problem with this patch, then do:
cat /proc/sys/net/sky2/lan0
This patch (which shouldn't go into the mainline driver), adds a debug
interface to sky2 driver to dump the receive and transmit rings.
The file /proc/net/sky2/ethX will show the status of transmits in process,
On Sat, 13 Jan 2007 14:03:29 +0100
Tino Keitel [EMAIL PROTECTED] wrote:
On Tue, Jan 02, 2007 at 20:10:15 +0100, Tino Keitel wrote:
[...]
Btw., I just built 2.6.20-rc3 with patches 4 and 5 and wake on LAN now
works. Thanks for your work.
Hi,
I had some failures during resume from
Michael Tokarev a e'crit :
Any idea how to force sending FIN-with-data?
int flag_on = 1;
setsockopt(fd, SOL_TCP, TCP_CORK, flag_on, sizeof(int));
send(fd, data, datalen, 0);
close(fd);
Eric Dumazet
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to
On Mon, Jan 15, 2007 at 10:21:49 -0800, Stephen Hemminger wrote:
On Sat, 13 Jan 2007 14:03:29 +0100
Tino Keitel [EMAIL PROTECTED] wrote:
On Tue, Jan 02, 2007 at 20:10:15 +0100, Tino Keitel wrote:
[...]
Btw., I just built 2.6.20-rc3 with patches 4 and 5 and wake on LAN now
At 10:19 AM 1/15/2007 -0800, Auke Kok wrote:
Have you tried the version in 2.6.19?
I even tried copying and pasting the e100_down and the latest PM stuff from
the newest e100.c version on sourceforge. I admit to being defeated as to
how to join a sourceforge group. Too many hours writing
Jiri Benc schreef:
On Mon, 15 Jan 2007 13:31:06 +, Johannes Berg wrote:
On Mon, 2007-01-15 at 13:55 +0100, Maarten Lankhorst wrote:
Enabling this doesn't cause anything to fail, but my wireless router
doesn't have a pci bus, but instead a native SSB, so CONFIG_NET_WIRELESS
isn't
Harry Coin wrote:
At 10:19 AM 1/15/2007 -0800, Auke Kok wrote:
Have you tried the version in 2.6.19?
I even tried copying and pasting the e100_down and the latest PM stuff
from the newest e100.c version on sourceforge. I admit to being
defeated as to how to join a sourceforge group. Too
Chris Lalancette [EMAIL PROTECTED] :
[...]
Similar to this commit:
http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=d15e9c4d9a75702b30e00cdf95c71c88e3f3f51e
It's not safe in cp_start_xmit to blindly call spin_lock_irq and then
spin_unlock_irq, since it
On Mon, Jan 15, 2007 at 04:34:41PM +0300, Michael Tokarev wrote:
# ethtool -k eth0
Offload parameters for eth0:
Cannot get device rx csum settings: Operation not supported
Cannot get device tx csum settings: Operation not supported
Cannot get device scatter-gather settings: Operation not
Herbert Xu wrote:
On Mon, Jan 15, 2007 at 04:34:41PM +0300, Michael Tokarev wrote:
[]
So I guess the problem is not related to hw checksumming offloading.
Nope, it just means that 8139too doesn't provide ethtool handlers to
disable checksum offloading.
So I suggest that you try doing the
On Mon, Jan 15, 2007 at 08:56:35PM +0100, Francois Romieu wrote:
As I understand http://lkml.org/lkml/2006/12/12/239, something like the
patch below should had been sent instead. Herbert, ack/nak ?
Sorry, what I said in that thread is in error. Netpoll may
unfortunately call the transmit
On Tue, Jan 16, 2007 at 12:46:08AM +0300, Michael Tokarev wrote:
I'm doing the capture on an intermediate host - the whole day today ;)
Cool, I was just trying to make sure :)
The trace is here: http://www.corpit.ru/mjt/bad-tcp-cksum-dmp.bin
I'll take a look.
Are you using anything extra
Michael Tokarev a écrit :
Eric Dumazet wrote:
Michael Tokarev a e'crit :
Any idea how to force sending FIN-with-data?
int flag_on = 1;
setsockopt(fd, SOL_TCP, TCP_CORK, flag_on, sizeof(int));
send(fd, data, datalen, 0);
close(fd);
That produces two packets - one (or more - depending on the
Christoph Hellwig wrote:
On Wed, Jan 10, 2007 at 06:41:37PM -0600, Jay Cliburn wrote:
+struct csum_param {
+ unsigned buf_len:14;
+ unsigned dma_int:1;
+ unsigned pkt_int:1;
+ u16 valan_tag;
+ unsigned eop:1;
+ /* command */
+ unsigned coalese:1;
+
Jay Cliburn [EMAIL PROTECTED] :
[...]
I welcome any comments on the rationality of this approach.
An URL for the current version of the patch would be welcome too :o)
--
Ueimor
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More
Francois Romieu wrote:
Jay Cliburn [EMAIL PROTECTED] :
[...]
I welcome any comments on the rationality of this approach.
An URL for the current version of the patch would be welcome too :o)
Sorry. Forgot to do that. The current version may be found here:
From: Sridhar Samudrala [EMAIL PROTECTED]
Date: Thu, 11 Jan 2007 11:41:25 -0800
[SCTP]: Set correct error cause value for missing parameters
sctp_process_missing_param() needs to use the SCTP_ERROR_MISS_PARAM
error cause value.
Signed-off-by: Vlad Yasevich [EMAIL PROTECTED]
From: Sridhar Samudrala [EMAIL PROTECTED]
Date: Thu, 11 Jan 2007 11:41:27 -0800
[SCTP]: Verify some mandatory parameters.
Verify init_tag and a_rwnd mandatory parameters in INIT and
INIT-ACK chunks.
Signed-off-by: Vlad Yasevich [EMAIL PROTECTED]
Signed-off-by: Sridhar Samudrala [EMAIL
From: Sridhar Samudrala [EMAIL PROTECTED]
Date: Thu, 11 Jan 2007 11:41:29 -0800
[SCTP]: Correctly handle unexpected INIT-ACK chunk.
Consider the chunk as Out-of-the-Blue if we don't have
an endpoint. Otherwise discard it as before.
Signed-off-by: Vlad Yasevich [EMAIL PROTECTED]
From: Sridhar Samudrala [EMAIL PROTECTED]
Date: Thu, 11 Jan 2007 11:41:32 -0800
[SCTP]: Fix SACK sequence during shutdown
Currently, when association enters SHUTDOWN state,the
implementation will SACK any DATA first and then transmit
the SHUTDOWN chunk. This is against the order required
On Tue, Jan 16, 2007 at 12:46:08AM +0300, Michael Tokarev wrote:
The trace is here: http://www.corpit.ru/mjt/bad-tcp-cksum-dmp.bin
I'm sorry but this dump does NOT look like it was taken from an
intermediate box. I verified two bad checksums (chosen randomly)
and they were both correct but
From: Samuel Ortiz [EMAIL PROTECTED]
Date: Mon, 15 Jan 2007 11:15:11 +0200
Since we stop using dev_alloc_skb on the IrDA TX frame, we constantly run
into the case of the skb headroom being 0, and thus we call skb_cow for
every IrDA TX frame.
This patch uses a local buffer and memcpy the skb
On Tue, Jan 16, 2007 at 02:27:39PM +1100, Herbert Xu wrote:
I'm sorry but this dump does NOT look like it was taken from an
intermediate box. I verified two bad checksums (chosen randomly)
and they were both correct but partial checksums. This means that
this dump was most likely taken
From: Samuel Ortiz [EMAIL PROTECTED]
Date: Mon, 15 Jan 2007 11:15:42 +0200
With USB2.0 bulk out MTU can be 512 bytes, so checking it only for 64 bytes is
incorrect.
Signed-off-by: Samuel Ortiz [EMAIL PROTECTED]
Applied, thanks a lot.
-
To unsubscribe from this list: send the line
David Miller [EMAIL PROTECTED] wrote:
Technically this is a bug fix too because once an SKB hits the
transmit function it should essentially be immutable, ie. you
shouldn't be writing to it. tcpdump sniffers could be looking
at the SKB, as one example.
We do have a way around that with
Hi,
I think the return value of rt6_nlmsg_size() should includes the
amount of RTA_METRICS.
Regards,
---
net/ipv6/route.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 8c3d568..5f0043c 100644
--- a/net/ipv6/route.c
+++
Hi,
I'm sorry to re-send...
I think the return value of rt6_nlmsg_size() should includes the
amount of RTA_METRICS.
Regards,
Signed-off-by: Noriaki TAKAMIYA [EMAIL PROTECTED]
---
net/ipv6/route.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/net/ipv6/route.c
This series is an attempt to generalize the async I/O paths to be
implementation agnostic. It completely eliminates knowledge of
the kiocb structure in the generic code and makes it private within the
current aio code. Things get noticeably cleaner without that layering
violation.
The new
Define a new function typedef for I/O completion at the file/iovec level --
typedef void (file_endio_t)(void *endio_data, ssize_t count, int err);
and convert aio_complete and all its callers to this new prototype.
---
drivers/usb/gadget/inode.c | 24 +++---
fs/aio.c
Convert the internals of blkdev_direct_IO to use a generic endio function,
instead of directly calling aio_complete. This may also fix some bugs/races
in this code, for instance it checks bio-bi_size instead of assuming it's
zero, and it atomically accumulates the bytes_done counter (assuming
This converts the iternals of nfs's directIO support to use a generic endio
function, instead of directly calling aio_complete. It's pretty easy
because it already has a pretty abstracted completion path.
---
diff -urpN -X dontdiff a/fs/nfs/direct.c b/fs/nfs/direct.c
--- a/fs/nfs/direct.c
Convert code using iocb-ki_left to use the more generic iov_length() call.
---
diff -urpN -X dontdiff a/fs/ocfs2/file.c b/fs/ocfs2/file.c
--- a/fs/ocfs2/file.c 2007-01-10 11:50:26.0 -0800
+++ b/fs/ocfs2/file.c 2007-01-10 12:42:09.0 -0800
@@ -1157,7 +1157,7 @@ static ssize_t
This removes the aio implementation from the usb gadget file system. Aside
from making very creative (!) use of the aio retry path, it can't be of any
use performance-wise because it always kmalloc()s a bounce buffer for the
*whole* I/O size. Perhaps the only reason to keep it around is the
This converts the internals of __blockdev_direct_IO in fs/direct-io.c to use
a generic endio function, instead of directly calling aio_complete. It also
changes the semantics of dio_iodone to be more friendly to its only users,
xfs and ocfs2. This allows the caller to know how to release locks
On Mon, Jan 15, 2007 at 05:54:50PM -0800, Nate Diller wrote:
Convert code using iocb-ki_left to use the more generic iov_length() call.
No way. We need to reduce the numer of iovec traversals, not adding
more of them.
-
To unsubscribe from this list: send the line unsubscribe netdev in
the
This converts the _locking variant of blockdev_direct_IO to use a generic
endio function, and updates all the FS callsites.
---
Documentation/filesystems/Locking |5 +++--
Documentation/filesystems/vfs.txt |5 +++--
fs/block_dev.c|9 -
fs/ext2/inode.c
Remove unused arg from socket operations
The sendmsg and recvmsg socket operations take a kiocb pointer, but none of
the functions actually use it. There's really no need even theoretically,
it's really quite ugly having it there at all. Also, removing it will pave
the way for a more generic
64 matches
Mail list logo
