On Mon, Jan 27, 2014 at 8:49 PM, Ben Pfaff wrote:
> On Mon, Jan 27, 2014 at 05:48:35PM -0800, pshelar at nicira.com wrote:
>> From: Pravin B Shelar
>>
>> Following patch adds DPDK netdev-class to userspace datapath.
>> Approach taken in this patch differs from Intel?? DPDK vSwitch
>> where DPDK d
From: Pravin B Shelar
Following patch adds DPDK netdev-class to userspace datapath.
Approach taken in this patch differs from Intel? DPDK vSwitch
where DPDK datapath switching is done in saparate process. This
patch adds support for DPDK type port and uses OVS userspace
datapath for switching.
Why LFENCE, rather than CPUID? I guess LFENCE does not prevent
out-of-order execution for non-load instructions across it.
This link has detailed information on RDTSC, RDTSCP, and CPUID:
http://www.intel.com/content/dam/www/public/us/en/documents/white-papers/ia-32-ia-64-benchmark-code-execution-p
Hi,
>> It would be nice if I can bypass set_tsc_freq_from_cpuinfo() in
>> set_tsc_freq().
>
> I think it would not solve the problem because your clock is varying and the
> TSC calibration must be updated accordingly with different values by core.
Reasonably new Intel CPUs (including Nehalem) has
> LLVM clang requires an explicitly sized "cmp" assembly instruction.
>
> Signed-off-by: Olivier Matz
Acked and applied.
--
Thomas
> From: Mats Liljegren
>
> Use command line parameters to get the name of the interface.
> This name is converted into if_index, which is provided as
> device info.
>
> Signed-off-by: Mats Liljegren
Acked and applied
Thank you
--
Thomas
24/01/2014 14:24, liljegren.mats2 at gmail.com:
> From: Mats Liljegren
>
> This field is intended for pcap to describe the name of the interface
> as known to Linux. It is an interface index, but can be translated into
> an interface name using if_indextoname() function.
>
> When using pcap, int
Hello,
16/01/2014 07:59, neeraj kumar:
> It is in regard to this commit *d92c66f5ac53a7b803efcb4782e601ea99c38a7e*
>
> By default in 32 bit kernel *High Memory Support* option of make menuconfig
> is selected with size 4 GB, therefore it gives compilation error.
>
> Thus to use KNI on 32 bit sy
Dmitry,
I cannot thank you enough for this information. This too was my main
problem. I put a "small" unmeasured delay before the call to
rte_eth_rx_burst() and suddenly it starts returning bursts of 512 packets
vs. 4!!
Best Regards,
Mike
On Wed, Jan 22, 2014 at 9:52 AM, Dmitry Vyal wrote:
>
Yes, i will add a new function that includes the lfence;
for the performance penalty, we did not see noticable performance impact
on our full software, so we did not see any reason to use 2 functions,
but it's certainly because we make a very limited number of calls to
rdtsc and it's true that
Hi Banashankar,
For proper TCP checksum calculation you have to calculate checksum over
pseudo header (see app/test-pmd/cmdline.c) and put result in tcp_hdr->cksum
2014-01-26 Wiles, Roger Keith
> Hi Banashankar,
>
> The tx_conf is used in the pktgen_config_ports() with the
> rte_eth_tx_queue_s
Hi Vladimir,
Once I added the pseudo header I was able to offload the TCP checksum.
Thanks a lot.
Thanks
Banashankar
On Mon, Jan 27, 2014 at 12:02 AM, Vladimir Medvedkin
wrote:
> Hi Banashankar,
>
> For proper TCP checksum calculation you have to calculate checksum over
> pseudo header (see ap
Hi,
In pktgen.c, tx_conf have a flag txq_flags which is set to IXGBE_SIMPLE_FLAGS
which inturn has value ETH_TXQ_FLAGS_NOOFFLOADS which is defined in
rte_ethdev.h and has values which disables all kind of chksum offloads. So
once I commented off that particular flag, offload started working.
Yea I
24/01/2014 12:42, Fran?ois-Fr?d?ric Ozog:
> IMHO, adding the lfence for all cases is introducing an un-necessary
> performance penalty.
>
> What about adding rte_rdtsc_sync() or rte_rdtsc_serial() with the comment
> about the rdtsc instruction behavior so that developers can choose which
> form th
Hello,
27/01/2014 03:56, Sangjin Han:
> Cause: No probed ethernet devices - check that
> CONFIG_RTE_LIBRTE_IGB_PMD=y and that CONFIG_RTE_LIBRTE_EM_PMD=y and
> that CONFIG_RTE_LIBRTE_IXGBE_PMD=y in your configuration file
[...]
> I googled for possible causes and came across a similar case:
> http
Hey,
As I remember, for TCP and UDP checksum HW offload, SW is responsible for
generating and setting psudo-header checksum.
>From 82599 datasheet, 7.2.5.2:
Note: For non-TSO, software still needs to calculate a full checksum for the
TCP/
UDP pseudo-header. This checksum of the pseudo-header sho
16 matches
Mail list logo
