After CloudNetEngine vswitch technical preview is launched, we received quite
a few queries on vswitches performance comparison, but we cannot simply give a
test result on our platform because performance varies on different H/Ws and
different workloads, and that's why we encourage you to try the
I'd like to hope that's my methodology problem, but I just follow the
installation guide without any customization.
Hi Mark, do you have any performance data share with us? Maybe we are using
different type of workloads, like I mentioned I am using typical data center
workload, I guess you are
> On Jul 22, 2015, at 2:36 AM, Gray, Mark D wrote:
>
>
>>
>> I'd like to hope that's my methodology problem, but I just follow the
>> installation guide without any customization.
>>
>> Hi Mark, do you have any performance data share with us? Maybe we are
>> using different type of workloads
Hi Bruce,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Bruce Richardson
> Sent: Tuesday, July 21, 2015 11:25 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] ixgbe: fix check for split packets
>
> The check for split packets to be reassembled in the
In ieee1588fwd.c, timestamp.tv_sec is a size_t variable,
which is a long int, but it was being printed with PRIu64,
causing an issue on 32 bits.
Signed-off-by: Pablo de Lara
---
app/test-pmd/ieee1588fwd.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/app/test-pmd/ieee15
Hi Pablo,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Pablo de Lara
> Sent: Wednesday, July 22, 2015 2:39 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] testpmd: fix wrong variable type in ieee1588fwd
> for
> 32 bits
>
> In ieee1588fwd.c, timest
Hi, Vladimir
When compiling thash for no-X86 arches, it fails with the following errors.
I wonder if
it is possible to make the thash library arch-independent?
== Build app/test
CC test_thash.o
In file included from /u/zlu.bjg/git/dpdk.org/app/test/test_thash.c:40:
/u/zlu.bjg/git/dpdk.org/tile-
Hi,
> -Original Message-
> From: Lu, Wenzhuo
> Sent: Wednesday, July 22, 2015 8:26 AM
> To: De Lara Guarch, Pablo; dev at dpdk.org
> Subject: RE: [dpdk-dev] [PATCH] testpmd: fix wrong variable type in
> ieee1588fwd for 32 bits
>
> Hi Pablo,
>
> > -Original Message-
> > From: dev
> >>
> >> I'd like to hope that's my methodology problem, but I just follow the
> >> installation guide without any customization.
> >>
> >> Hi Mark, do you have any performance data share with us? Maybe we are
> >> using different type of workloads, like I mentioned I am using
> >> typical data ce
In ieee1588fwd.c, timestamp.tv_sec is a time_t variable,
which is a long int, but it was being printed with PRIu64,
causing an issue on 32 bits.
Signed-off-by: Pablo de Lara
---
Changes in v2:
- Corrected wrong commit message
app/test-pmd/ieee1588fwd.c | 4 ++--
1 file changed, 2 insertions(+)
> "Jun Xiao" wrote:
>
> > After CloudNetEngine vswitch technical preview is launched, we
> > received quite a few queries on vswitches performance comparison, but
> > we cannot simply give a test result on our platform because
> > performance varies on different H/Ws and different workloads, and
Hi,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Pablo de Lara
> Sent: Wednesday, July 22, 2015 4:07 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2] testpmd: fix wrong variable type in ieee1588fwd
> for 32 bits
>
> In ieee1588fwd.c, timestamp.tv
> Malloc was moved to the EAL and dummy malloc library was left
> to not break apps that had a librte_malloc.so dependency.
> Note that the dummy library will be removed in the next release.
>
> When building a combined library, all objects are copied to the same
> directory before creating the li
On Wed, Jul 22, 2015 at 01:44:14AM +0100, Lu, Wenzhuo wrote:
> Hi Bruce,
>
> > -Original Message-
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Bruce Richardson
> > Sent: Tuesday, July 21, 2015 11:25 PM
> > To: dev at dpdk.org
> > Subject: [dpdk-dev] [PATCH] ixgbe: fix check
2015-07-20 11:40, Stephen Hemminger:
> The number of descriptors can be either zero to use the whole
> available ring, or some value smaller. This is used to limit
> the number of mbufs allocated for the receive ring. If more
> descriptors are requested than available the size is silently
> truncat
> This integrates my change and earlier change by Ouyang Changchun
> into one fix. And second patch is minor stuff found while reviewing.
>
> Stephen Hemminger (2):
> virtio: fix queue size and number of descriptors
> virtio: small cleanups
Applied, thanks
Please take care of patch numbering
This patch enables jumbo frame support for the 82583V.
It has been tested ( rx and tx ) with real HW.
Signed-off-by: Klaus Degner
---
drivers/net/e1000/em_ethdev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/e1000/em_ethdev.c b/drivers/net/e1000/em_ethdev.c
in
> In order to keep the ABI consistent with the old hash library,
> hash_func_init_val field has been moved, so it remains
> at the same offset as previously, since hash_func and
> hash_func_init_val are fields accesed by the public function
> rte_hash_hash and must keep the same offset as older ver
The check for split packets to be reassembled in the vector ixgbe PMD
was incorrectly only checking the first 16 elements of the array instead
of all 32. This is fixed by changing the uint32_t values to be uint64_t
instead.
Fixes: cf4b4708a88a ("ixgbe: improve slow-path perf with vector scattered
Hi,
And what happens if someone changes RTE_IXGBE_VPMD_RX_BURST to something
else than 32? I guess this bug were introduced when someone raised it
from 16 to 32
I think you are better off with a for loop which uses this value. Or at
least make a comment around RTE_IXGBE_VPMD_RX_BURST that if yo
ixgbe_recv_pkts_lro uses field rx_free_trigger
in structure ixgbe_rx_queue, but that field is only defined
if IXGBE_RX_ALLOW_BULK_ALLOC is enabled, so even though
that field is not used when it is disabled,
compiler complains about it.
Therefore, the lines of code that use that field
have been ifde
2015-07-22 10:50, Pablo de Lara:
> ixgbe_recv_pkts_lro uses field rx_free_trigger
> in structure ixgbe_rx_queue, but that field is only defined
> if IXGBE_RX_ALLOW_BULK_ALLOC is enabled, so even though
> that field is not used when it is disabled,
> compiler complains about it.
Why this option is
2015-07-22 10:50, Pablo de Lara:
> ixgbe_recv_pkts_lro uses field rx_free_trigger
> in structure ixgbe_rx_queue, but that field is only defined
> if IXGBE_RX_ALLOW_BULK_ALLOC is enabled, so even though
> that field is not used when it is disabled,
> compiler complains about it.
> Therefore, the lin
On Wed, Jul 22, 2015 at 10:47:34AM +0100, Zoltan Kiss wrote:
> Hi,
>
> And what happens if someone changes RTE_IXGBE_VPMD_RX_BURST to something
> else than 32? I guess this bug were introduced when someone raised it from
> 16 to 32
Actually, no, this bug is purely due to me getting my maths wrong
> > In ieee1588fwd.c, timestamp.tv_sec is a time_t variable, which is a long
> > int, but
> > it was being printed with PRIu64, causing an issue on 32 bits.
> >
> > Signed-off-by: Pablo de Lara
> Acked-by: Wenzhuo Lu
Applied, thanks
2015-07-14 14:50, Zhang, Helin:
> From: Wu, Jingjing
> > Because VF multi-queues can be supported, l3fwd can run on vf.
> > Suggest to remove the l3fwd-vf example.
> Totally agree with this!
> But we need the confirmation from validation guys of that l3fwd works
> quite well on VF with all NICs (e.
Sergio,
As the maintainer of memory allocation, would you consider using
libhugetlbfs in DPDK for Linux?
It may simplify a part of our memory allocator and avoid some potential
bugs which would be already fixed in the dedicated lib.
> > If FM10K_DEBUG_DRIVER is enabled, then the log messages about
> > function entry are missing newline causing extremely long lines.
> >
> > Signed-off-by: Stephen Hemminger
>
> Acked-by : Jing Chen
Applied, thanks
> > The return in fm10k_dev_handle_fault has useless conditional
> > since it always returns 0 here.
> >
> > Signed-off-by: Stephen Hemminger
>
> Acked-by: Jing Chen
Applied, thanks
> > The fm10k driver was reading the interrupt cause register but then
> > using the interrupt mask register defines to look at the bits.
> > The result is that if a fault happens, the driver would never clear
> > the fault and would get into an infinite cycle of interrupts.
> >
> > Note: I don't
2015-07-21 10:40, Bruce Richardson:
> On Mon, Jul 20, 2015 at 11:01:34PM +0530, Rahul Lakkireddy wrote:
> > This series of patches fix compilation and enable CXGBE poll mode driver for
> > FreeBSD. The first patch fixes a limitation of nic_uio that only binds to
> > devices present in rte_pci_dev_
On 22/07/15 10:59, Bruce Richardson wrote:
> On Wed, Jul 22, 2015 at 10:47:34AM +0100, Zoltan Kiss wrote:
>> Hi,
>>
>> And what happens if someone changes RTE_IXGBE_VPMD_RX_BURST to something
>> else than 32? I guess this bug were introduced when someone raised it from
>> 16 to 32
>
> Actually, n
On 22/07/15 14:19, Zoltan Kiss wrote:
> Btw. vPMD was a bit misleading abbreviation for me, it took me a while
> until I realized 'v' stands for 'vector', not 'virtualization' as in
> most cases nowadays.
>
Though that's mostly my fault to not to check the documentation :)
> -Original Message-
> From: Zoltan Kiss [mailto:zoltan.kiss at linaro.org]
> Sent: Wednesday, July 22, 2015 2:20 PM
> To: Richardson, Bruce
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] ixgbe: fix check for split packets
>
>
>
> On 22/07/15 10:59, Bruce Richardson wrote:
2015-07-22 13:35, Richardson, Bruce:
> From: Zoltan Kiss [mailto:zoltan.kiss at linaro.org]
> > On 22/07/15 10:59, Bruce Richardson wrote:
> > > The vector PMD always works off a fixed 32 burst size. Any change to
> > > that will lead to many changes in the code, so I don't believe a loop is
> > ne
2015-07-17 09:16, Bruce Richardson:
> On Thu, Jul 16, 2015 at 04:47:23PM -0700, Stephen Hemminger wrote:
> > - if (user_cb == NULL &&
> > - (user_cb = rte_zmalloc("INTR_USER_CALLBACK",
> > - sizeof(struct rte_eth_dev_callback), 0))) {
> > + if (!user_cb)
>
>
__mempool_get_trailer() calculated header's address.
The address of trailer should set after element area.
This patch fixes this calculating.
This issue was mixed with the following commit:
Fixes: 97e7e685bfcd ("mempool: add structure for object trailers")
Signed-off-by: Yuichi Nakai
---
lib/l
2015-07-16 16:47, Stephen Hemminger:
> These were found by running the Linux kernel coccinelle scripts
> on the DPDK source.
>
> Stephen Hemminger (2):
> rte_ethdev: fix crash if malloc fails in rte_eth_dev_callback_register
> kni: fix coccinelle warnings
Applied, thanks
Marvin/Waterman
Could you help to check if l3fwd is good enough for all cases (1g/10/40g, PF
and VF, single queue/multiple queue)?
We aim to remove l3fwd-vf to reduce an example application which is not so
necessary.
Thank you!
Regards,
Helin
> -Original Message-
> From: Thomas Monjalo
As I remember, the problem is that inside l3fwd each I/O lcore tries to claim a
TX queue on each port in use for itself
(to avoid any synchronisation overhead).
Obviously on some legacy (and virtual) devices this is not possible.
On l3fwd-vf, several lcores share the same TX queue.
(synchronisatio
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev
> Sent: Monday, July 20, 2015 5:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCHv3 1/5] ethdev: add new API to retrieve RX/TX queue
> information
>
> Add the ability for the upper la
> -Original Message-
> From: Zhang, Helin
> Sent: Wednesday, July 22, 2015 5:51 PM
> To: Ananyev, Konstantin
> Cc: dev at dpdk.org
> Subject: RE: [dpdk-dev] [PATCHv3 1/5] ethdev: add new API to retrieve RX/TX
> queue information
>
>
>
> > -Original Message-
> > From: dev [mail
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev
> Sent: Monday, July 20, 2015 5:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCHv3 2/5] i40e: add support for eth_(rxq|txq)_info_get
>
> Signed-off-by: Konstantin Ananyev
Acked-by:
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev
> Sent: Monday, July 20, 2015 5:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCHv3 3/5] ixgbe: add support for
> eth_(rxq|txq)_info_get
>
> Signed-off-by: Konstantin Ananyev
Acked-b
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev
> Sent: Monday, July 20, 2015 5:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCHv3 4/5] e1000: add support for
> eth_(rxq|txq)_info_get
>
> Signed-off-by: Konstantin Ananyev
Acked-b
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Konstantin Ananyev
> Sent: Monday, July 20, 2015 5:19 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCHv3 5/5] testpmd: add new command to display RX/TX
> queue information
>
> Signed-off-by: Konstantin An
Add the ability for the upper layer to query RX/TX queue information.
Right now supported for:
ixgbe, i40e, e1000 PMDs.
Konstantin Ananyev (5):
ethdev: add new API to retrieve RX/TX queue information
i40e: add support for eth_(rxq|txq)_info_get
ixgbe: add support for eth_(rxq|txq)_info_get
Signed-off-by: Konstantin Ananyev
---
drivers/net/ixgbe/ixgbe_ethdev.c | 4
drivers/net/ixgbe/ixgbe_ethdev.h | 6 ++
drivers/net/ixgbe/ixgbe_rxtx.c | 42
3 files changed, 52 insertions(+)
diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/driver
Signed-off-by: Konstantin Ananyev
---
app/test-pmd/cmdline.c | 48
app/test-pmd/config.c | 67 ++
app/test-pmd/testpmd.h | 2 ++
3 files changed, 117 insertions(+)
diff --git a/app/test-pmd/cmdline.c b/app/tes
Add the ability for the upper layer to query RX/TX queue information.
Add new structures:
struct rte_eth_rxq_info
struct rte_eth_txq_info
new functions:
rte_eth_rx_queue_info_get
rte_eth_tx_queue_info_get
into rte_etdev API.
Left extra free space in the queue info structures,
so extra fields co
Signed-off-by: Konstantin Ananyev
---
drivers/net/i40e/i40e_ethdev.c | 2 ++
drivers/net/i40e/i40e_ethdev.h | 5 +
drivers/net/i40e/i40e_rxtx.c | 42 ++
3 files changed, 49 insertions(+)
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40
Signed-off-by: Konstantin Ananyev
---
drivers/net/e1000/e1000_ethdev.h | 12
drivers/net/e1000/em_ethdev.c| 2 ++
drivers/net/e1000/em_rxtx.c | 38 ++
drivers/net/e1000/igb_ethdev.c | 4
drivers/net/e1000/igb_rxtx.c | 36 +
On Wed, 22 Jul 2015 19:28:51 +0100
Konstantin Ananyev wrote:
> Add the ability for the upper layer to query RX/TX queue information.
>
> Add new structures:
> struct rte_eth_rxq_info
> struct rte_eth_txq_info
>
> new functions:
> rte_eth_rx_queue_info_get
> rte_eth_tx_queue_info_get
>
> into r
2015-07-22 17:00, Ananyev, Konstantin:
> From: Zhang, Helin
> > > -#ifdef __cplusplus
> > > -}
> > > -#endif
> > > -
> > > /**
> > > * Set the list of multicast addresses to filter on an Ethernet device.
> > > *
> > > @@ -3882,4 +3952,9 @@ extern int
> > > rte_eth_timesync_read_rx_timestamp(ui
On Thu, 23 Jul 2015 09:48:55 +0800
mac_leehk at yahoo.com.hk wrote:
> From: marco
Thank you for addressing a real bug.
But there are several issues with the patch as submitted:
* the standard way to handle allocation failure in network drivers is to drop
the
received packet and reuse the
55 matches
Mail list logo
