[dpdk-dev] Performance hit - NICs on different CPU sockets

On Thu, Jun 16, 2016 at 10:19 PM, Wiles, Keith wrote: > > On 6/16/16, 3:16 PM, "dev on behalf of Wiles, Keith" on behalf of keith.wiles at intel.com> wrote: > >> >>On 6/16/16, 3:00 PM, "Take Ceara" wrote: >> >>>On Thu, Jun 16, 2016 at 9:33 PM, Wiles, Keith >>>wrote: On 6/16/16, 1:20 PM,

[dpdk-dev] [PATCH v10 4/7] ethdev: make get port by name and get name by port public

2016-06-15 15:06, Reshma Pattan: > Converted rte_eth_dev_get_port_by_name to a public API. > Converted rte_eth_dev_get_name_by_port to a public API. > Updated the release notes with the changes. It is not an API change, just a new API, so no need to reference it in the release notes.

[dpdk-dev] [PATCH 4/4] doc: add MTU update to feature matrix for enic

Signed-off-by: John Daley --- doc/guides/nics/overview.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/doc/guides/nics/overview.rst b/doc/guides/nics/overview.rst index 29a6163..6b30085 100644 --- a/doc/guides/nics/overview.rst +++ b/doc/guides/nics/overview.rst @@ -92,7

[dpdk-dev] [PATCH 3/4] enic: add an update MTU function for non-Rx scatter mode

Provide an update MTU callbaack. The function returns -ENOTSUP if Rx scatter is enabled. Updating the MTU to be greater than the value configured via the Cisco CIMC/UCSM management interface is allowed provided it is still less than the maximum egress packet size allowed by the NIC.

[dpdk-dev] [PATCH 2/4] enic: set the max allowed MTU for the NIC

The max MTU is set to the max egress packet size allowed by the VIC minus the size of a an IPv4 L2 header with .1Q (18 bytes). Signed-off-by: John Daley --- drivers/net/enic/enic.h| 1 + drivers/net/enic/enic_ethdev.c | 3 ++- drivers/net/enic/enic_res.c| 25

[dpdk-dev] [PATCH 1/4] enic: enable NIC max packet size discovery

Pull in common VNIC code which enables querying for max egress packet size. Signed-off-by: John Daley --- There are some non-related fields and defines in this file because it is shared with other drivers and interfaces to the VIC. drivers/net/enic/base/vnic_enet.h | 17 - 1

[dpdk-dev] [PATCH 0/4] enic: enable MTU update callback

This patchset determines the max egress packet size allowed on the NIC and uses it to set an upper limit for MTU. An MTU update function is added, but only works if Rx scatter is disabled. If Rx scatter is enabled, -ENOSUP is returned. Another patch with Rx scatter support will come later. These

[dpdk-dev] Performance hit - NICs on different CPU sockets

On Thu, Jun 16, 2016 at 9:33 PM, Wiles, Keith wrote: > On 6/16/16, 1:20 PM, "Take Ceara" wrote: > >>On Thu, Jun 16, 2016 at 6:59 PM, Wiles, Keith >>wrote: >>> >>> On 6/16/16, 11:56 AM, "dev on behalf of Wiles, Keith" >> dpdk.org on behalf of keith.wiles at intel.com> wrote: >>> On

[dpdk-dev] [PATCH v4] e1000: configure VLAN TPID

This patch enables configuring the outer TPID for double VLAN. Note that all other TPID values are read only. Signed-off-by: Beilei Xing --- v4 changes: Optimize the code to be more readable. v3 changes: Update commit log and comments. v2 changes: Modify return value. Cause inner TPID is not

[dpdk-dev] [PATCH v10 3/7] ethdev: add new fields to ethdev info struct

2016-06-15 15:06, Reshma Pattan: > The new fields nb_rx_queues and nb_tx_queues are added to the > rte_eth_dev_info structure. > Changes to API rte_eth_dev_info_get() are done to update these new fields > to the rte_eth_dev_info object. The ABI is changed, not the API. > Release notes is updated

[dpdk-dev] [PATCH 4/4] app/test: typo fixing

Fixing typo in the performance tests for example preftest to perftest. Signed-off-by: Jain, Deepak K --- app/test/test_cryptodev_perf.c | 26 +- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/app/test/test_cryptodev_perf.c

[dpdk-dev] [PATCH 3/4] app/test: updating AES SHA performance test

From: Fiona Trahe Updating the AES performance test in line with snow3g peformance test. Output format has been updated so as to get better understanding of numbers. Signed-off-by: Fiona Trahe Signed-off-by: Jain, Deepak K --- app/test/test_cryptodev.h | 2 +

[dpdk-dev] [PATCH 2/4] app/test: adding Snow3g performance test

From: Fiona Trahe Adding performance test for snow3g wireless algorithm. Performance test can run over both software and hardware. Signed-off-by: Fiona Trahe Signed-off-by: Jain, Deepak K Signed-off-by: Declan Doherty --- app/test/test_cryptodev.h | 2 +-

[dpdk-dev] [PATCH 1/4] cryptodev: add rte_crypto_op_bulk_free function

From: Declan Doherty Adding rte_crypto_op_bulk_free to free up the ops in bulk so as to expect improvement in performance. Signed-off-by: Declan Doherty --- lib/librte_cryptodev/rte_crypto.h | 15 +++ 1 file changed, 15 insertions(+) diff --git

[dpdk-dev] [PATCH 0/4] Extending cryptodev Performance tests

Performance tests haven been extended in this patchset. Patchset consists of 4 patches: Patch 1 adds new function rte_crypto_op_bulk_free to be used in patch 2 and patch 3. Patch 2 add snow3g performance tests. Patch 3 updates the existing aes performanc test Patch 4 fixes the typo in names of

[dpdk-dev] [PATCH v3 3/4] bonding: take queue spinlock in rx/tx burst functions

2016-06-16 16:41, Iremonger, Bernard: > Hi Thomas, > > > 2016-06-16 15:32, Bruce Richardson: > > > On Mon, Jun 13, 2016 at 01:28:08PM +0100, Iremonger, Bernard wrote: > > > > > Why does this particular PMD need spinlocks when doing RX and TX, > > > > > while other device types do not? How is

[dpdk-dev] Performance hit - NICs on different CPU sockets

On Thu, Jun 16, 2016 at 6:59 PM, Wiles, Keith wrote: > > On 6/16/16, 11:56 AM, "dev on behalf of Wiles, Keith" dpdk.org on behalf of keith.wiles at intel.com> wrote: > >> >>On 6/16/16, 11:20 AM, "Take Ceara" wrote: >> >>>On Thu, Jun 16, 2016 at 5:29 PM, Wiles, Keith >>>wrote: >>>

[dpdk-dev] Performance hit - NICs on different CPU sockets

On 6/16/16, 3:16 PM, "dev on behalf of Wiles, Keith" wrote: > >On 6/16/16, 3:00 PM, "Take Ceara" wrote: > >>On Thu, Jun 16, 2016 at 9:33 PM, Wiles, Keith >>wrote: >>> On 6/16/16, 1:20 PM, "Take Ceara" wrote: >>> On Thu, Jun 16, 2016 at 6:59 PM, Wiles, Keith wrote: > > On

[dpdk-dev] Performance hit - NICs on different CPU sockets

On 6/16/16, 3:00 PM, "Take Ceara" wrote: >On Thu, Jun 16, 2016 at 9:33 PM, Wiles, Keith wrote: >> On 6/16/16, 1:20 PM, "Take Ceara" wrote: >> >>>On Thu, Jun 16, 2016 at 6:59 PM, Wiles, Keith >>>wrote: On 6/16/16, 11:56 AM, "dev on behalf of Wiles, Keith" >>> dpdk.org on behalf of

[dpdk-dev] [PATCH] ena: Update PMD to cooperate with latest ENA firmware

This patch includes: * Update of ENA communication layer * Fixed memory management issue After allocating memzone it's required to zeroize it as well as freeing memzone with dedicated function. * Added debug area and host information * Disabling readless communication regarding

[dpdk-dev] [PATCH v3 17/17] ethdev: get rid of device type

From: David Marchand Now that hotplug has been moved to eal, there is no reason to keep the device type in this layer. Signed-off-by: David Marchand --- app/test/virtual_pmd.c| 2 +- drivers/net/af_packet/rte_eth_af_packet.c | 2 +-

[dpdk-dev] [PATCH v3 16/17] ethdev: convert to eal hotplug

From: David Marchand Remove bus logic from ethdev hotplug by using eal for this. Current api is preserved: - the last port that has been created is tracked to return it to the application when attaching, - the internal device name is reused when detaching. We can

[dpdk-dev] [PATCH v3 15/17] eal: add hotplug operations for pci and vdev

From: David Marchand hotplug which deals with resources should come from the layer that already handles them, i.e. eal. For both attach and detach operations, 'name' is used to select the bus that will handle the request. Signed-off-by: David Marchand ---

[dpdk-dev] [PATCH v3 14/17] ethdev: do not scan all pci devices on attach

From: David Marchand No need to scan all devices, we only need to update the device being attached. Signed-off-by: David Marchand --- lib/librte_eal/common/eal_common_pci.c | 11 --- lib/librte_ether/rte_ethdev.c | 3 --- 2 files changed, 8

[dpdk-dev] [PATCH v3 12/17] pci: add a helper for device name

From: David Marchand eal is a better place than crypto / ethdev for naming resources. Add a helper in eal and make use of it in crypto / ethdev. Signed-off-by: David Marchand --- lib/librte_cryptodev/rte_cryptodev.c| 27 ---

[dpdk-dev] [PATCH v3 11/17] eal/linux: move back interrupt thread init before setting affinity

From: David Marchand Now that virtio pci driver is initialized in a constructor, iopl() stuff happens early enough so that interrupt thread can be created right after plugin loading. This way, chelsio driver should be happy again [1]. [1]

[dpdk-dev] [PATCH v3 10/17] ethdev: get rid of eth driver register callback

From: David Marchand Now that all pdev are pci drivers, we don't need to register ethdev drivers through a dedicated channel. Signed-off-by: David Marchand --- lib/librte_ether/rte_ethdev.c | 22 -- lib/librte_ether/rte_ethdev.h

[dpdk-dev] [PATCH v3 09/17] crypto: get rid of crypto driver register callback

From: David Marchand Now that all pdev are pci drivers, we don't need to register crypto drivers through a dedicated channel. Signed-off-by: David Marchand --- lib/librte_cryptodev/rte_cryptodev.c | 22 ---

[dpdk-dev] [PATCH v3 08/17] drivers: convert all pdev drivers as pci drivers

From: David Marchand Simplify crypto and ethdev pci drivers init by using newly introduced init macros and helpers. Those drivers then don't need to register as "rte_driver"s anymore. virtio and mlx* drivers use the general purpose RTE_INIT macro, as they both need

[dpdk-dev] [PATCH v3 07/17] ethdev: export init/uninit common wrappers for pci drivers

From: David Marchand Preparing for getting rid of eth_drv, here are two wrappers that can be used by pci drivers that assume a 1 to 1 association between pci resource and upper interface. Signed-off-by: David Marchand --- lib/librte_ether/rte_ethdev.c | 14

[dpdk-dev] [PATCH v3 06/17] crypto: export init/uninit common wrappers for pci drivers

From: David Marchand Preparing for getting rid of rte_cryptodev_driver, here are two wrappers that can be used by pci drivers that assume a 1 to 1 association between pci resource and upper interface. Signed-off-by: David Marchand ---

[dpdk-dev] [PATCH v3 05/17] eal: introduce init macros

From: David Marchand Introduce a RTE_INIT macro used to mark an init function as a constructor. Current eal macros have been converted to use this (no functional impact). RTE_EAL_PCI_REGISTER is added as a helper for pci drivers. Suggested-by: Jan Viktorin

[dpdk-dev] [PATCH v3 04/17] eal: remove duplicate function declaration

From: David Marchand rte_eal_dev_init is declared in both eal_private.h and rte_dev.h since its introduction. This function has been exported in ABI, so remove it from eal_private.h Fixes: e57f20e05177 ("eal: make vdev init path generic for both virtual and pci

[dpdk-dev] [PATCH v3 03/17] drivers: align pci driver definitions

From: David Marchand Pure coding style, but it might make it easier later if we want to move fields in rte_cryptodev_driver and eth_driver structures. Signed-off-by: David Marchand --- drivers/crypto/qat/rte_qat_cryptodev.c | 2 +- drivers/net/ena/ena_ethdev.c

[dpdk-dev] [PATCH v3 02/17] crypto: no need for a crypto pmd type

From: David Marchand This information is not used and just adds noise. Signed-off-by: David Marchand --- lib/librte_cryptodev/rte_cryptodev.c | 8 +++- lib/librte_cryptodev/rte_cryptodev.h | 2 -- lib/librte_cryptodev/rte_cryptodev_pmd.h | 3 +-- 3 files

[dpdk-dev] [PATCH v3 01/17] pci: no need for dynamic tailq init

From: David Marchand These lists can be initialized once and for all at build time. With this, those lists are only manipulated in a common place (and we could even make them private). A nice side effect is that pci drivers can now register in constructors.

[dpdk-dev] [PATCH v3 00/17] prepare for rte_device / rte_driver

From: David Marchand * Original patch series is from David Marchand. This is just a rebase over master (d76c19309) * Following discussions with Jan [1] and some cleanup I started on pci code, here is a patchset that reworks pdev drivers registration and hotplug api.

[dpdk-dev] [PATCH] port: add kni interface support

Hi Cristian, The new patch has been submitted just now. Please note that I do ignore some check patch errors this time. B.R. Ethan 2016-06-13 21:18 GMT+08:00 Dumitrescu, Cristian < cristian.dumitrescu at intel.com>: > Hi Ethan, > > > > Great, we?ll wait for your patch later this week then. I

[dpdk-dev] [PATCH v3 3/3] port: document update

add kni configurations into the document of ip pipeline sample application Signed-off-by: WeiJie Zhuang --- doc/guides/sample_app_ug/ip_pipeline.rst | 112 +++ 1 file changed, 83 insertions(+), 29 deletions(-) diff --git a/doc/guides/sample_app_ug/ip_pipeline.rst

[dpdk-dev] [PATCH v3 2/3] port: add kni nodrop writer

1. add no drop writing operations to the kni port 2. support dropless kni config in the ip pipeline sample application Signed-off-by: WeiJie Zhuang --- examples/ip_pipeline/app.h | 2 + examples/ip_pipeline/config_parse.c | 31 - examples/ip_pipeline/init.c | 26

[dpdk-dev] [PATCH v3 1/3] port: add kni interface support

1. add KNI port type to the packet framework 2. add KNI support to the IP Pipeline sample Application 3. some bug fix Signed-off-by: WeiJie Zhuang --- v2: * Fix check patch error. v3: * Fix code review comments. --- doc/api/doxy-api-index.md | 1 +

[dpdk-dev] [PATCH v4 0/3] Keep-alive enhancements

> Remy Horton (3): > eal: export keepalive state enumerations > eal: add additional keepalive callbacks > examples/l2fwd-keepalive: add IPC liveness reporting Applied, thanks Just a last comment: the agent in the example should not appear in examples/Makefile.

[dpdk-dev] Performance hit - NICs on different CPU sockets

On Thu, Jun 16, 2016 at 5:29 PM, Wiles, Keith wrote: > > Right now I do not know what the issue is with the system. Could be too many > Rx/Tx ring pairs per port and limiting the memory in the NICs, which is why > you get better performance when you have 8 core per port. I am not really >

[dpdk-dev] [PATCH] qat: fix for VFs not getting recognized

2016-06-16 16:29, Jain, Deepak K: > Due to addition of CLASS_ID in EAL, class_id is > amended into the code. Why the VF is not recognized? The class id should not be mandatory.

[dpdk-dev] [PATCH v5 0/7] Remove string operations from xstats

> Remy Horton (7): > rte: change xstats to use integer ids > drivers/net/ixgbe: change xstats to use integer ids > drivers/net/e1000: change xstats to use integer ids > drivers/net/fm10k: change xstats to use integer ids > drivers/net/i40e: change xstats to use integer ids >

[dpdk-dev] [PATCH v3] i40e: configure MTU

On 6/16/16, 10:40 AM, "dev on behalf of Yong Wang" wrote: >On 5/16/16, 5:27 AM, "dev on behalf of Olivier Matz" on behalf of olivier.matz at 6wind.com> wrote: > >>Hi Beilei, >> >>On 05/13/2016 10:15 AM, Beilei Xing wrote: >>> This patch enables configuring MTU for i40e. >>> Since changing MTU

[dpdk-dev] [PATCH v3] i40e: configure MTU

On 5/16/16, 5:27 AM, "dev on behalf of Olivier Matz" wrote: >Hi Beilei, > >On 05/13/2016 10:15 AM, Beilei Xing wrote: >> This patch enables configuring MTU for i40e. >> Since changing MTU needs to reconfigure queue, stop port first >> before configuring MTU. >> >> Signed-off-by: Beilei Xing >>

[dpdk-dev] [PATCH v5 1/4] lib/librte_ether: support device reset

2016-06-15 11:03, Wenzhuo Lu: > +/** > + * Reset an Ethernet device. > + * > + * @param port_id > + * The port identifier of the Ethernet device. > + */ > +int > +rte_eth_dev_reset(uint8_t port_id); Please explain in the doxygen comment what means a reset. We must understand why and when an

[dpdk-dev] [PATCH v3 1/2] ethdev: add tunnel and port RSS offload types

On Fri, Apr 01, 2016 at 07:59:33PM +0530, Jerin Jacob wrote: > On Fri, Apr 01, 2016 at 04:04:13PM +0200, Thomas Monjalon wrote: > > 2016-03-31 02:21, Jerin Jacob: > > > - added VXLAN, GENEVE and NVGRE tunnel flow types > > > - added PORT flow type for accounting physical/virtual > > > port or

[dpdk-dev] Performance hit - NICs on different CPU sockets

On Thu, Jun 16, 2016 at 4:58 PM, Wiles, Keith wrote: > > From the output below it appears the x710 devices 01:00.[0-3] are on socket 0 > And the x710 devices 02:00.[0-3] sit on socket 1. > I assume there's a mistake here. The x710 devices on socket 0 are: $ lspci | grep -ie "01:.*x710" 01:00.0

[dpdk-dev] [PATCH v5 1/1] eal: fix resource leak of mapped memory

Patch fixes resource leak in rte_eal_hugepage_attach() where mapped files were not freed back to the OS in case of failure. Patch uses the behavior of Linux munmap: "It is not an error if the indicated range does not contain any mapped pages". Coverity issue: 13295, 13296, 13303 Fixes:

[dpdk-dev] [PATCHv7 1/6] pmdinfogen: Add buildtools and pmdinfogen utility

On 06/16/2016 04:33 PM, Neil Horman wrote: > On Thu, Jun 16, 2016 at 03:29:57PM +0300, Panu Matilainen wrote: >> On 06/09/2016 08:46 PM, Neil Horman wrote: >>> pmdinfogen is a tool used to parse object files and build json strings for >>> use in later determining hardware support in a dso or

[dpdk-dev] [PATCH v3 3/4] bonding: take queue spinlock in rx/tx burst functions

2016-06-16 15:32, Bruce Richardson: > On Mon, Jun 13, 2016 at 01:28:08PM +0100, Iremonger, Bernard wrote: > > > Why does this particular PMD need spinlocks when doing RX and TX, while > > > other device types do not? How is adding/removing devices from a bonded > > > device different to other

[dpdk-dev] Performance hit - NICs on different CPU sockets

On 6/16/16, 11:56 AM, "dev on behalf of Wiles, Keith" wrote: > >On 6/16/16, 11:20 AM, "Take Ceara" wrote: > >>On Thu, Jun 16, 2016 at 5:29 PM, Wiles, Keith >>wrote: >> >>> >>> Right now I do not know what the issue is with the system. Could be too >>> many Rx/Tx ring pairs per port and

[dpdk-dev] Performance hit - NICs on different CPU sockets

On 6/16/16, 11:20 AM, "Take Ceara" wrote: >On Thu, Jun 16, 2016 at 5:29 PM, Wiles, Keith wrote: > >> >> Right now I do not know what the issue is with the system. Could be too many >> Rx/Tx ring pairs per port and limiting the memory in the NICs, which is why >> you get better performance

[dpdk-dev] [PATCH v5] eal: out-of-bounds write

Overrunning array mcfg->memseg of 256 44-byte elements at element index 257 using index j. Fixed by add condition with message information. Fixes: af75078fece3 ("first public release") Coverity ID 13282 Signed-off-by: Slawomir Mrozowicz --- v5: - update message v4: - remove check condition from

[dpdk-dev] [PATCH v2] xenvirt: fix compilation after mempool changes

On Mon, Jun 13, 2016 at 01:54:29PM +0200, Christian Ehrhardt wrote: > Yeah, working now - thanks for the fast update! > > Kind Regards, > Christian > > Christian Ehrhardt > Software Engineer, Ubuntu Server > Canonical Ltd > Applied to dpdk-next-net/rel_16_07 /Bruce

[dpdk-dev] [PATCH] app/testpmd: unchecked return value

> > Calling rte_eth_dev_rss_hash_update without checking return value. > > Fixed by handle return value and print out error status. > > > > Fixes: ce8d561418d4 ("app/testpmd: add port configuration settings") > > Coverity ID 119251 > > > > Signed-off-by: Slawomir Mrozowicz > > Acked-by: Pablo

[dpdk-dev] [PATCH v5 00/25] DPDK PMD for ThunderX NIC device

On Thu, Jun 16, 2016 at 11:58:27AM +0100, Bruce Richardson wrote: > On Thu, Jun 16, 2016 at 03:01:02PM +0530, Jerin Jacob wrote: > > On Wed, Jun 15, 2016 at 03:39:25PM +0100, Bruce Richardson wrote: > > > On Wed, Jun 15, 2016 at 12:36:15AM +0530, Jerin Jacob wrote: > > > > This patch set provides

[dpdk-dev] [PATCH 0/2] vhost: Fix leaks on migration.

Thanks for fixing them! Would you please resend them, with a rebase based on master branch of following tree: http://dpdk.org/browse/next/dpdk-next-virtio/ --yliu On Thu, Jun 16, 2016 at 11:32:03AM +0300, Ilya Maximets wrote: > Ilya Maximets (2): > vhost: fix leak of file

[dpdk-dev] [PATCH v3 3/4] bonding: take queue spinlock in rx/tx burst functions

Hi Thomas, > 2016-06-16 15:32, Bruce Richardson: > > On Mon, Jun 13, 2016 at 01:28:08PM +0100, Iremonger, Bernard wrote: > > > > Why does this particular PMD need spinlocks when doing RX and TX, > > > > while other device types do not? How is adding/removing devices > > > > from a bonded device

[dpdk-dev] [PATCH v3 0/5] vhost/virtio performance loopback utility

> > Zhihong Wang (5): > > testpmd: add retry option > > testpmd: configurable tx_first burst number > > testpmd: show throughput in port stats > > testpmd: handle all rxqs in rss setup > > testpmd: show topology at forwarding start > > Series-acked-by: Pablo de Lara Applied, thanks

[dpdk-dev] [PATCH v3 5/5] testpmd: show topology at forwarding start

2016-06-16 11:09, De Lara Guarch, Pablo: > > --- a/app/test-pmd/testpmd.c > > +++ b/app/test-pmd/testpmd.c > > @@ -1016,6 +1016,7 @@ start_packet_forwarding(int with_tx_first) > > flush_fwd_rx_queues(); > > > > fwd_config_setup(); > > + fwd_config_display(); > >

[dpdk-dev] [PATCH v5 00/25] DPDK PMD for ThunderX NIC device

On Thu, Jun 16, 2016 at 04:47:39PM +0530, Jerin Jacob wrote: > On Thu, Jun 16, 2016 at 11:58:27AM +0100, Bruce Richardson wrote: > > On Thu, Jun 16, 2016 at 03:01:02PM +0530, Jerin Jacob wrote: > > > On Wed, Jun 15, 2016 at 03:39:25PM +0100, Bruce Richardson wrote: > > > > On Wed, Jun 15, 2016 at

[dpdk-dev] [PATCHv7 5/6] pmdinfo.py: Add tool to query binaries for hw and other support information

On 06/09/2016 08:47 PM, Neil Horman wrote: > This tool searches for the primer sting PMD_DRIVER_INFO= in any ELF binary, > and, if found parses the remainder of the string as a json encoded string, > outputting the results in either a human readable or raw, script parseable > format > > Note that,

[dpdk-dev] [PATCH v3 3/4] bonding: take queue spinlock in rx/tx burst functions

On Mon, Jun 13, 2016 at 01:28:08PM +0100, Iremonger, Bernard wrote: > Hi Bruce, > > > > > Subject: Re: [dpdk-dev] [PATCH v3 3/4] bonding: take queue spinlock in rx/tx > > burst functions > > > > On Sun, Jun 12, 2016 at 06:11:28PM +0100, Bernard Iremonger wrote: > > > Use rte_spinlock_trylock()

[dpdk-dev] [PATCHv7 1/6] pmdinfogen: Add buildtools and pmdinfogen utility

On 06/09/2016 08:46 PM, Neil Horman wrote: > pmdinfogen is a tool used to parse object files and build json strings for > use in later determining hardware support in a dso or application binary. > pmdinfo looks for the non-exported symbol names this_pmd_name and > this_pmd_tbl (where n is a

[dpdk-dev] [PATCH v6 00/38] new bnxt poll mode driver library

On Wed, Jun 15, 2016 at 02:23:00PM -0700, Stephen Hurd wrote: > The bnxt poll mode library (librte_pmd_bnxt) implements support for > Broadcom NetXtreme C-Series. These adapters support Standards- > compliant 10/25/50Gbps 30MPPS full-duplex throughput. > > Information about this family of

[dpdk-dev] [PATCH v3] rte_hash: add scalable multi-writer insertion w/ Intel TSX

This patch introduced scalable multi-writer Cuckoo Hash insertion based on a split Cuckoo Search and Move operation using Intel TSX. It can do scalable hash insertion with 22 cores with little performance loss and negligible TSX abortion rate. * Added an extra rte_hash flag definition to switch

[dpdk-dev] [PATCH v3] rte_hash: add scalable multi-writer insertion w/ Intel TSX

Here's the latest version of the rte_hash multi-writer patch. It's re-based on top of the latest head as of Jun 16, 2016. http://dpdk.org/dev/patchwork/patch/13886/ http://dpdk.org/dev/patchwork/patch/12589/ v3 changes: * Made spinlock as fall back behavior when developer choose to use

[dpdk-dev] [PATCH v5 00/25] DPDK PMD for ThunderX NIC device

On Wed, Jun 15, 2016 at 03:39:25PM +0100, Bruce Richardson wrote: > On Wed, Jun 15, 2016 at 12:36:15AM +0530, Jerin Jacob wrote: > > This patch set provides the initial version of DPDK PMD for the > > built-in NIC device in Cavium ThunderX SoC family. > > > > Implemented features and ThunderX

[dpdk-dev] Performance hit - NICs on different CPU sockets

On 6/16/16, 9:36 AM, "Take Ceara" wrote: >Hi Keith, > >On Tue, Jun 14, 2016 at 3:47 PM, Wiles, Keith wrote: Normally the limitation is in the hardware, basically how the PCI bus is connected to the CPUs (or sockets). How the PCI buses are connected to the system depends on the

[dpdk-dev] [PATCH v4] eal: out-of-bounds write

On 06/15/2016 04:25 PM, Slawomir Mrozowicz wrote: > Overrunning array mcfg->memseg of 256 44-byte elements > at element index 257 using index j. > Fixed by add condition with message information. > > Fixes: af75078fece3 ("first public release") > Coverity ID 13282 > > Signed-off-by: Slawomir

[dpdk-dev] [PATCH v4] e1000: configure VLAN TPID

> -Original Message- > From: Xing, Beilei > Sent: Thursday, June 16, 2016 9:36 PM > To: Zhang, Helin > Cc: dev at dpdk.org; Xing, Beilei > Subject: [PATCH v4] e1000: configure VLAN TPID > > This patch enables configuring the outer TPID for double VLAN. > Note that all other TPID

[dpdk-dev] [PATCH v4 1/1] eal: fix resource leak of mapped memory

On 15/06/2016 13:25, Marcin Kerlin wrote: > Patch fixes resource leak in rte_eal_hugepage_attach() where mapped files > were not freed back to the OS in case of failure. Patch uses the behavior > of Linux munmap: "It is not an error if the indicated range does not > contain any mapped pages". > >

[dpdk-dev] [PATCH v6 00/38] new bnxt poll mode driver library

On Thu, Jun 16, 2016 at 9:24 AM, Bruce Richardson < bruce.richardson at intel.com> wrote: > On Wed, Jun 15, 2016 at 02:23:00PM -0700, Stephen Hurd wrote: > > The bnxt poll mode library (librte_pmd_bnxt) implements support for > > Broadcom NetXtreme C-Series. These adapters support Standards- > >

[dpdk-dev] [PATCH] examples/ip_pipeline: fix build error for gcc 4.8

> -Original Message- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Thomas Monjalon > Sent: Tuesday, June 14, 2016 9:04 PM > To: Mrzyglod, DanielX T > Cc: dev at dpdk.org; Singh, Jasvinder ; > Dumitrescu, Cristian > Subject: Re: [dpdk-dev] [PATCH] examples/ip_pipeline: fix

[dpdk-dev] [PATCH v13 2/3] app/test: test external mempool manager

Use a minimal custom mempool external ops and check that it also passes basic mempool autotests. Signed-off-by: Olivier Matz Signed-off-by: David Hunt Acked-by: Shreyansh Jain Acked-by: Olivier Matz --- app/test/test_mempool.c | 122 +++- 1 file

[dpdk-dev] [PATCH v13 1/3] mempool: support external mempool operations

Until now, the objects stored in a mempool were internally stored in a ring. This patch introduces the possibility to register external handlers replacing the ring. The default behavior remains unchanged, but calling the new function rte_mempool_set_ops_byname() right after

[dpdk-dev] [PATCH v13 0/3] mempool: add external mempool manager

Here's the latest version of the External Mempool Manager patchset. It's re-based on top of the latest head as of 15/6/2016, including Olivier's 35-part patch series on mempool re-org [1] [1] http://dpdk.org/ml/archives/dev/2016-May/039229.html v13 changes: * Added in extra opaque data

[dpdk-dev] enic in passhtrough mode tx drops

Hi all, I'm running a vm attached to 2 cisco Virtual Card Interfaces in passthrough mode in a cisco UCS. The vNICs are configured on access mode without VLAN ID. The incoming packets are arriving with 802.1q header containing vlan priority bit according to the class of service configured on

[dpdk-dev] [PATCH v2 00/17] prepare for rte_device / rte_driver

On Thu, 16 Jun 2016 08:42:29 + Shreyansh Jain wrote: > Hi, > > > -Original Message- > > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com] > > Sent: Thursday, June 16, 2016 1:04 PM > > To: Shreyansh Jain > > Cc: David Marchand ; viktorin at > > rehivetech.com; > > dev at

[dpdk-dev] [PATCH v2 00/17] prepare for rte_device / rte_driver

On Thu, 16 Jun 2016 11:19:59 +0200 Thomas Monjalon wrote: > 2016-06-16 10:23, Jan Viktorin: > > I think, we should consider to move it to somebody else. I would work on > > it, however, I don't see all the tasks that are to be done. That's why I > > was waiting to finalize those patchs by

[dpdk-dev] [PATCH v12 0/3] mempool: add external mempool manager

On 16/6/2016 9:58 AM, Olivier MATZ wrote: >>> >>> So I don't think we should have more cache misses whether it's >>> placed at the beginning or at the end. Maybe I'm missing something... >>> >>> I still believe it's better to group the 2 fields as they are >>> tightly linked together. It could

[dpdk-dev] [PATCH] hash: new function to retrieve a key given its position

> -Original Message- > From: Yari Adan Petralanda [mailto:yari.adan.petralanda at ericsson.com] > Sent: Thursday, June 16, 2016 9:23 AM > To: Richardson, Bruce; De Lara Guarch, Pablo; Juan Antonio Montesinos > Delgado > Cc: dev at dpdk.org > Subject: [PATCH] hash: new function to

[dpdk-dev] [PATCH v4] eal: out-of-bounds write

On 15/06/2016 14:25, Slawomir Mrozowicz wrote: > Overrunning array mcfg->memseg of 256 44-byte elements > at element index 257 using index j. > Fixed by add condition with message information. > > Fixes: af75078fece3 ("first public release") > Coverity ID 13282 > > Signed-off-by: Slawomir

[dpdk-dev] random pkt generator PMD

On 15.06.2016 19:02, Neil Horman wrote: > On Wed, Jun 15, 2016 at 03:43:56PM +0600, Yerden Zhumabekov wrote: >> Hello everybody, >> >> DPDK already got a number of PMDs for various eth devices, it even has PMD >> emulations for backends such as pcap, sw rings etc. >> >> I've been thinking about

[dpdk-dev] [PATCH v2] enic: scattered Rx

For performance reasons, this patch uses 2 VIC RQs per RQ presented to DPDK. The VIC requires that each descriptor be marked as either a start of packet (SOP) descriptor or a non-SOP descriptor. A one RQ solution requires skipping descriptors when receiving small packets and results in bad

[dpdk-dev] [PATCH v2 2/2] vhost: unmap log memory on cleanup.

Fixes memory leak on QEMU migration. Fixes: 54f9e32305d4 ("vhost: handle dirty pages logging request") Signed-off-by: Ilya Maximets --- lib/librte_vhost/vhost-net.h | 1 + lib/librte_vhost/vhost_user/virtio-net-user.c | 15 +-- 2 files changed, 14 insertions(+), 2

[dpdk-dev] [PATCH v2 1/2] vhost: fix leak of file descriptors.

While migration of vhost-user device QEMU allocates memfd to store information about dirty pages and sends fd to vhost-user process. File descriptor for this memory should be closed to prevent "Too many open files" error for vhost-user process after some amount of migrations. Ex.: # ls

[dpdk-dev] [PATCH v2 0/2] vhost: Fix leaks on migration.

v2: * rebased on top of dpdk-next-virtio/master Ilya Maximets (2): vhost: fix leak of file descriptors. vhost: unmap log memory on cleanup. lib/librte_vhost/vhost-net.h | 1 + lib/librte_vhost/vhost_user/virtio-net-user.c | 16 ++-- 2 files changed, 15

[dpdk-dev] [PATCH v2] rte_hash: add scalable multi-writer insertion w/ Intel TSX

Hi Wei, > -Original Message- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Wei Shen > Sent: Thursday, June 16, 2016 5:53 AM > To: dev at dpdk.org > Cc: De Lara Guarch, Pablo; stephen at networkplumber.org; Tai, Charlie; > Maciocco, Christian; Gobriel, Sameh; Shen, Wei1 >

[dpdk-dev] [PATCH v3 3/3] mempool: allow for user-owned mempool caches

The mempool cache is only available to EAL threads as a per-lcore resource. Change this so that the user can create and provide their own cache on mempool get and put operations. This works with non-EAL threads too. This commit introduces the new API calls: rte_mempool_cache_create(size,

[dpdk-dev] [PATCH v3 2/3] mempool: use bit flags instead of is_mp and is_mc

Pass the same flags as in rte_mempool_create(). Changes API calls: rte_mempool_generic_put(mp, obj_table, n, flags) rte_mempool_generic_get(mp, obj_table, n, flags) Signed-off-by: Lazaros Koromilas --- lib/librte_mempool/rte_mempool.h | 58 +--- 1

[dpdk-dev] [PATCH v3 1/3] mempool: deprecate specific get/put functions

This commit introduces the API calls: rte_mempool_generic_put(mp, obj_table, n, is_mp) rte_mempool_generic_get(mp, obj_table, n, is_mc) Deprecates the API calls: rte_mempool_mp_put_bulk(mp, obj_table, n) rte_mempool_sp_put_bulk(mp, obj_table, n) rte_mempool_mp_put(mp, obj)

[dpdk-dev] [PATCH v3 0/3] mempool: user-owned mempool caches

Updated version of the user-owned cache patchset. It applies on top of the latest external mempool manager patches from David Hunt [1]. [1] http://dpdk.org/ml/archives/dev/2016-June/041479.html v3 changes: * Deprecate specific mempool API calls instead of removing them. * Split deprecation

[dpdk-dev] [PATCH v5 00/25] DPDK PMD for ThunderX NIC device

On Thu, Jun 16, 2016 at 03:01:02PM +0530, Jerin Jacob wrote: > On Wed, Jun 15, 2016 at 03:39:25PM +0100, Bruce Richardson wrote: > > On Wed, Jun 15, 2016 at 12:36:15AM +0530, Jerin Jacob wrote: > > > This patch set provides the initial version of DPDK PMD for the > > > built-in NIC device in

[dpdk-dev] [PATCH] hash: new function to retrieve a key given its position

On Thu, Jun 16, 2016 at 10:23:42AM +, Juan Antonio Montesinos Delgado wrote: > Hi, > > As I understand it, the hash table entry can change position in the first > hash table but the index in the second hash table remains the same. So, > regardless the bucket the entry is in, the index (of

[dpdk-dev] [PATCH v1 02/28] eal: extract function eal_parse_sysfs_valuef

Sorry, didn't notice this email earlier... Comments inline > -Original Message- > From: Jan Viktorin [mailto:viktorin at rehivetech.com] > Sent: Wednesday, June 15, 2016 3:26 PM > To: Shreyansh Jain > Cc: dev at dpdk.org; David Marchand ; Thomas > Monjalon > ; Bruce Richardson

[dpdk-dev] [PATCH] igb_uio: fix build with backported kernel

On 15 June 2016 at 11:59, Ferruh Yigit wrote: > On 6/15/2016 4:57 PM, Ferruh Yigit wrote: > > Following compile error observed with CentOS 6.8, which uses kernel > > kernel-devel-2.6.32-642.el6.x86_64: > > > > CC eal_thread.o > > .../build/lib/librte_eal/linuxapp/igb_uio/igb_uio.c: > > In

  1   2   >