I'm running an Intel 82598EB 10G AF DA on a Dell PowerEdge 1950 (Xeon
5430 + 5000X chipset). To enable DCA on my machine, I enabled I/OAT in
BIOS and applied the patch(http://patchwork.kernel.org/patch/15273/)
to the I/OAT driver. I confirmed that ixgbe_setup_dca() (in ixgbe
2.0.34.3) is done succe
Ronciak, John wrote:
Are you running out of CPU cycles? What does the interrupt rate look
like?
How do I get the rate? I am still instrumenting things.
Are the interrupts being shared at all? Can you 'cat
/proc/interrupts' to see how they are mapped? Please provide this output.
Our CPUs lo
And lastly, please try without bonding when performance tuning to eliminate
bonding from being the issue.
Also, with the slower cpu you're likely hurting your throughput with the high
interrupt rates.
Using TSO on the transmit side may be precluded by the bonding driver in F9
(not sure) and th
Are you running out of CPU cycles? What does the interrupt rate look like?
Are the interrupts being shared at all? Can you 'cat /proc/interrupts' to see
how they are mapped? Please provide this output.
How is the realtec device being used in this testing?
Have you tried the stand-alone e100
Ronciak, John wrote:
Are the e1000 NICs in 32 bit PCI slots? What is the exact e1000 HW being used?
64-bit slots - I have a combination of MT and XT boards have tried MT -
MT, XT - XT and MT - XT, also RTL8111 to both MT and XT cards. I get
identical results in all combinations.
I settled
Are the e1000 NICs in 32 bit PCI slots? What is the exact e1000 HW being used?
Cheers,
John
---
"...that your people will judge you on what you can build, not what you
destroy.", B. Obama, 2009
>-Original Message-
>From: Howard
The problem has been solved.
1. If the traffic is generated through raw sockets, then both packet
constructing code for ethX as well as the ethX interrupt are assigned
to the same processor.
2. I discovered this by:
- looking on htop, /proc/interrupts and then proper entries in /proc/irq
3. When I
This may not be right forum to ask this question but I am hoping
somebody will know the right place for me to go with this.
I have a network built around some Cisco switches. There is a 3550-12T
and a pair of 2950-48's connected via gigabit links to this switch. I
have 7 server boxes connected on
Hi, Jesse
I have solved the issue. The problem is PBA setting.
In the original driver, TXA is set to zero, which works for pci, but not
pcie.
Now my 82572EI card works well.
Thanks.
Roy
> -Original Message-
> From: Zang Roy-R61911
> Sent: Friday, June 19, 2009 9:57 AM
> To: Zang Roy-R6191
Hi,
On Sun, Jun 14, 2009 at 07:09:45PM +0200, Rafael J. Wysocki wrote:
> On Sunday 14 June 2009, Andreas Mohr wrote:
> > - why do we call netif_device_detach() _after_ doing hardware shutdown
> > of the network controller? I'd guess this can cause huge issues?
> > Someone told me he had rtnl l
I also tried the pktgen driver, and it works very well too, just as
the raw sockets approach. The problem is therefore with my driver. I'm
looking on pktgen code and my driver code, and the differences are:
1. The pktgen driver is selecting queues.. I can't say anything more
about the code, becaus
11 matches
Mail list logo