There was a request in this thread for a little more detail on how the
bonding sysfs interface works. I have been running this set of patches
in a test environment so am familiar enough with the interface that I
should be able to give a brief explanation. Here goes...
-
This version of bon
> Linux is a general purpose operating system.
>
> Even as a dedicated router, a router daemon still has to execute
> in userspace to do BGP etc. signaling with routing peers. The
> administrator also might want to run diagnostic tools to monitor
> the network.
>
> You cannot spin polling on the d
On Wed, Jul 06, 2005 at 04:24:53PM -0400, jaegert wrote:
> The patch has been revised to attach the security context
> to the xfrm_state and xfrm_policy rather than the selector.
Thanks a lot for making the change Trent.
> On input, the policy retrieved must also be authorized for the socket
>
Stuart Zilm wrote:
# this works - the source selected is $SRC
ip route add $DST dev eth0 src $SRC # implicit table main
# this fails - the source selected is chosen from main
ip route del $DST dev eth0 src $SRC # implicit table main - NOTE: if
this route remains, this source address will be ch
> >
> > > In my SMP platform, there is no other processes running. The usage of
> CPUs
> > > are 100% and 0%. How could I make Nic interrupts not arrive at only
one
> CPU,
> > > or balance the interrupt between two CPUs?
> >
> > This doesn't work. If you try to split up the work for one network
>
>
> > In my SMP platform, there is no other processes running. The usage of
CPUs
> > are 100% and 0%. How could I make Nic interrupts not arrive at only one
CPU,
> > or balance the interrupt between two CPUs?
>
> This doesn't work. If you try to split up the work for one network
> card amongst mul
From: "Qinghua(Kevin) Ye" <[EMAIL PROTECTED]>
Date: Wed, 6 Jul 2005 15:57:00 -0600
> In my SMP platform, there is no other processes running. The usage of CPUs
> are 100% and 0%. How could I make Nic interrupts not arrive at only one CPU,
> or balance the interrupt between two CPUs?
This doesn't
> It is not Click's right to make this kind of decision, that is what
> we have the process scheduler for.
>
Click's scheduler is aim to make packet processing tasks with highest
priority. It's just for the dedicated use. Since Linux is a general OS, and
it is free and fatastic, so it is possible
Patches for SELinux. Note that the patch applies to 2.6.13-rc2.
Regards,
Trent.
==
This patch series implements per packet access control via the
extension of the Linux Security Modules (LSM) interface by hooks in
the XFRM and pfkey subsystems
The patch has been revised to attach the security context
to the xfrm_state and xfrm_policy rather than the selector.
The function xfrm_policy_bysel was only used to match policies
for addition/removal from SPD which now requires matching security
context as well. The function has been changed t
On Wed, 6 Jul 2005 19:37:21 +0200
Andi Kleen <[EMAIL PROTECTED]> wrote:
> Hallo,
>
> I did a recent patch to make PCI driver ->probe run on the local node
> of a PCI device. This gives better NUMA locality. One problem is that
> most network drivers do a lot of initialization and memory allocatio
From: "Qinghua(Kevin) Ye" <[EMAIL PROTECTED]>
Date: Wed, 6 Jul 2005 14:12:29 -0600
> Yes, you are right. Click acturally will release the CPU to OS at interval.
> Other processes will be responded at this interval.
It is not Click's right to make this kind of decision, that is what
we have the pr
From: "Qinghua(Kevin) Ye" <[EMAIL PROTECTED]>
Date: Wed, 6 Jul 2005 13:15:40 -0600
> Yes, It wastes CPU cycles if there is other process running. However, as it
> being a dedicated router, it should not be a problem. The process of packets
> is the only task it is supposed to do.
Linux is a gener
net: move protocol specific config options out of net/Kconfig
Move the protocol specific config options out to the specific protocols.
With this change net/Kconfig now starts to become readable and serve as a
good basis for further rearrangings.
The menu structure is left almost intact, except tha
From: Andi Kleen <[EMAIL PROTECTED]>
Date: Wed, 6 Jul 2005 19:37:21 +0200
> I would prefer a way that does not require driver patching.
You're going to have to hack the drivers to do this, I think.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [E
Based on patch originally submitted by Randy Dunlap I have made two
patches that modify the Kconfig bits for net/.
First patch create a new top-level menu just above "Device Drivers"
named "Networking". Hereby one no longer has to look deep in Device
Drivers to find IPv6 for example.
Sample menu
net: restructure kconfig for net/
Restructure Networking menu structure creating a separate menu
just before Device Drivers.
Device drivers for ordinary NIC's are still to be found
in the Device Drivers section, but Bluetooth, IrDA and ax25
are located with their corresponding menu entries.
This
From: "Qinghua(Kevin) Ye" <[EMAIL PROTECTED]>
Date: Wed, 6 Jul 2005 11:36:40 -0600
> > > Compared to NAPI, click polling will disable Nic interrupts during its
> > > operation, even there is no any packets in the rx buffer.
> >
> > This destroys latency if you only recheck the RX buffer using
> >
Yes, It wastes CPU cycles if there is other process running. However, as it
being a dedicated router, it should not be a problem. The process of packets
is the only task it is supposed to do.
>
> > > > Compared to NAPI, click polling will disable Nic interrupts during
its
> > > > operation, even th
Andi Kleen <[EMAIL PROTECTED]> :
[...]
> I would prefer a way that does not require driver patching.
Why ?
You need a link which is not available (assuming one refuses to play
ugly games with SET_NETDEV_DEV or bus_info/pci_name). Please make it
explicit and save everybody's pain. Imho it is not a
This patchs adds support for the Olitec ISDN PCI card in the hisax
gazel driver. The gazel driver supports this card, but wasn't aware of
its PCI ids. Users used to modify the PCI ids of a supported card in
include/linux/pci_ids.h and recompile their kernel to get this card
running, as said in most
> > Compared to NAPI, click polling will disable Nic interrupts during its
> > operation, even there is no any packets in the rx buffer.
>
> This destroys latency if you only recheck the RX buffer using
> timer interrupts. Even with HZ=1000, on gigabit links your packet
> latency will be terrible.
Hallo,
I did a recent patch to make PCI driver ->probe run on the local node
of a PCI device. This gives better NUMA locality. One problem is that
most network drivers do a lot of initialization and memory allocation
not in ->probe, but in ->open. One way to fix them would be to do
the same net/co
Hello
here is requested information
ethtool -S eth0
NIC statistics:
rx_packets: 7595954
tx_packets: 3592596
rx_bytes: 2608713255
tx_bytes: 251613192
rx_errors: 10
tx_errors: 0
rx_dropped: 10
tx_dropped: 0
multicast: 0
collisions: 0
rx_length_errors: 0
Hi.
>> I wonder if it would be better to add a dev->priv_flags for,
>> e.g., IFF_SLAVE_INACTIVE or the like, set that when a slave is marked
>> inactive, and clear it when the slave becomes active (similarly to how
>> NOARP is done today). The rationale for this is that NOARP can be
>> chang
Hi
This is with the 20050630-2.6.13-rc1-sis190-test.patch driver
scp ... same as always
dmesg gives "eth0: pad error. status = 012005eb"
scp text [EMAIL PROTECTED]:[OK]
scp [EMAIL PROTECTED]:text . [STALL]
scp text blah@ [OK]
scp blah@:text . [OK]
Hmm, this might help
26 matches
Mail list logo
