Re: [dpdk-dev] DPDK 18.11.2 LTS update

On Thu, 16 May 2019 23:12:45 +0200 Thomas Monjalon wrote: > 16/05/2019 20:57, Kevin Traynor: > > Plan is to have a 18.11.2 RC1 ready early next week for validation > > teams. There are a small few outstanding backports that need authors > > help - please send by next Monday so I can put into RC1.

Re: [dpdk-dev] [PATCH v2 2/7] ether: do not mark ethernet address and header as packed

On Thu, 16 May 2019 21:40:05 +0100 Bruce Richardson wrote: > On Thu, May 16, 2019 at 11:04:22AM -0700, Stephen Hemminger wrote: > > Since the Ethernet address and header are naturally aligned > > correctly on all architectures, there is no need to mark them > > as packed. This results in faster c

Re: [dpdk-dev] [PATCH] kvargs: trim spaces at the beginning and end of key and values

On Thu, 16 May 2019 20:55:10 + "Ergin, Mesut A" wrote: > > -Original Message- > > From: Stephen Hemminger [mailto:step...@networkplumber.org] > > Sent: Thursday, May 16, 2019 9:37 AM > > To: Ergin, Mesut A > > Cc: olivier.m...@6wind.com; dev@dpdk.org > > Subject: Re: [dpdk-dev] [PATC

[dpdk-dev] [PATCH v2 0/1] net/e1000: set/clear GO_LINKD bit only if PHY reset is not blocked

From: Shweta Choudaha Hello, This patch fixes an issue where GO_LINKD bit is incorrectly set/cleared even when PHY reset is blocked and is updated to fix the checkpatch warning Fixes f34dec0b41 ("igb: force phy power up/down") Thanks, Shweta Shweta Choudaha (1): net/e1000: set/clear GO_LIN

[dpdk-dev] [PATCH v2 1/1] net/e1000: set/clear GO_LINKD bit only if PHY reset is not blocked

From: Shweta Choudaha When PHY reset is blocked as is the case when BMC is connected via NC-SI do not set GO_LINKD bit in PHY power management register in dev_stop as this will disconnect the PHY. Also, in dev_close clear the GO_LINKD bit only if PHY reset is not blocked Fixes: 3af34dec0b41 ("ig

[dpdk-dev] [PATCH 1/1] net/e1000: set/clear GO_LINKD bit only if PHY reset is not blocked

From: Shweta Choudaha When PHY reset is blocked as is the case when BMC is connected via NC-SI do not set GO_LINKD bit in PHY power management register in dev_stop as this will disconnect the PHY. Also, in dev_close clear the GO_LINKD bit only if PHY reset is not blocked Fixes: 3af34dec0b41 ("ig

[dpdk-dev] [PATCH 0/1] net/e1000: set/clear GO_LINKD bit only if PHY reset is not blocked

From: Shweta Choudaha Hello, This fixes an issue where GO_LINKD bit is incorrectly set/cleared even when PHY reset is blocked. Fixes f34dec0b41 ("igb: force phy power up/down") Thanks, Shweta Shweta Choudaha (1): net/e1000: set/clear GO_LINKD bit only if PHY reset is not blocked drivers/

Re: [dpdk-dev] DPDK Release Status Meeting 16/5/2019

On Thu, 2019-05-16 at 18:02 +0100, Ferruh Yigit wrote: > * Can be good to build contact information of the involved testers > from > various companies, for faster/better communication, if you are > doing LTS > testing please reach out to Kevin or Luca Shall we add a list to the MAINTAINE

Re: [dpdk-dev] [PATCH 2/3] net/i40e: add runtime option to disable vector rx

> -Original Message- > From: Maxime Coquelin [mailto:maxime.coque...@redhat.com] > Sent: Thursday, May 16, 2019 1:17 AM > To: Ergin, Mesut A ; Xing, Beilei > ; Zhang, Qi Z > Cc: dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH 2/3] net/i40e: add runtime option to disable > vector rx > > >

Re: [dpdk-dev] DPDK 18.11.2 LTS update

16/05/2019 20:57, Kevin Traynor: > Plan is to have a 18.11.2 RC1 ready early next week for validation > teams. There are a small few outstanding backports that need authors > help - please send by next Monday so I can put into RC1. Please could you share the list of patches which need some help? S

Re: [dpdk-dev] [PATCH 3/3] net/i40e: fix inadvertent override of vector RX allowance

> -Original Message- > From: Maxime Coquelin [mailto:maxime.coque...@redhat.com] > Sent: Thursday, May 16, 2019 1:18 AM > To: Ergin, Mesut A ; Xing, Beilei > ; Zhang, Qi Z > Cc: dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH 3/3] net/i40e: fix inadvertent override of > vector > RX allowa

Re: [dpdk-dev] [PATCH] kvargs: trim spaces at the beginning and end of key and values

> -Original Message- > From: Stephen Hemminger [mailto:step...@networkplumber.org] > Sent: Thursday, May 16, 2019 9:37 AM > To: Ergin, Mesut A > Cc: olivier.m...@6wind.com; dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH] kvargs: trim spaces at the beginning and end > of > key and values

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, May 16, 2019 at 09:37:42PM +0100, Bruce Richardson wrote: > On Thu, May 16, 2019 at 10:04:54AM -0700, Stephen Hemminger wrote: > > Agree if structure is naturally packed, adding packed attribute is a > > mistake. > > Do you want to also include such a change in your patchset, or will I do

Re: [dpdk-dev] [PATCH v2 2/7] ether: do not mark ethernet address and header as packed

On Thu, May 16, 2019 at 11:04:22AM -0700, Stephen Hemminger wrote: > Since the Ethernet address and header are naturally aligned > correctly on all architectures, there is no need to mark them > as packed. This results in faster code, and also gets rid of > warnings with gcc-9 about getting address

Re: [dpdk-dev] [PATCH] ether: mark ethernet addresses as being 2-byte aligned

On Thu, May 16, 2019 at 07:04:03PM +0100, Kevin Traynor wrote: > On 16/05/2019 16:54, Bruce Richardson wrote: > > When including the rte_ether.h header in applications with warnings > > enabled, a warning was given because of the assumption of 2-byte alignment > > of ethernet addresses when process

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, May 16, 2019 at 10:04:54AM -0700, Stephen Hemminger wrote: > On Thu, 16 May 2019 17:36:43 +0100 > Bruce Richardson wrote: > > > On Thu, May 16, 2019 at 05:07:45PM +0100, Bruce Richardson wrote: > > > On Thu, May 16, 2019 at 09:06:52AM -0700, Stephen Hemminger wrote: > > > > On Thu, 16 M

Re: [dpdk-dev] [PATCH v2 0/6] Pseudo-random number generation improvements

On Thu, May 16, 2019 at 07:55:04PM +0200, Mattias Rönnblom wrote: > Make DPDK pseudo-random number generation multi-thread safe, go faster > and produce better-quality pseudo-random numbers. > > This patchset requires the yet-to-be-merged rdseed build detection > patch for rdseed support to be ena

[dpdk-dev] DPDK 18.11.2 LTS update

Hi All, DPDK 18.11.2 will contain backports of relevant fixes from DPDK 19.05. Plan is to have a 18.11.2 RC1 ready early next week for validation teams. There are a small few outstanding backports that need authors help - please send by next Monday so I can put into RC1. Depending on what is fou

Re: [dpdk-dev] [PATCH v2 3/7] ether: deinline non-critical functions

Acked-by: Rami Rosen

[dpdk-dev] [PATCH v3] net/i40e: Eliminate weak symbols in i40e_rxtx.c

Use of weak symbols can hide makefile errors especially when custom makefiles are used. Removing the use of weak symbols to avoid a stub function being linked in production code. Signed-off-by: David Harton --- v3 - added CONFIG_RTE_LIBRTE_I40E_INC_VECTOR to makefile v2 - added CC_AVX2_SUPPORT

[dpdk-dev] [PATCH v2 7/7] cmdline: use ether_unformat_addr

Now that there is a version of ether_aton in rte_ether, it can be used by the cmdline ethernet address parser. Signed-off-by: Stephen Hemminger --- lib/Makefile | 2 +- lib/librte_cmdline/Makefile | 3 +- lib/librte_cmdline/cmdline_parse_etherad

[dpdk-dev] [PATCH v2 5/7] ethdev: use eth_unformat_addr

Use eth_unformat_addr, so that ethdev can be built and work without the cmdline library. The dependency on cmdline was an arrangement of convenience anyway. Signed-off-by: Stephen Hemminger --- lib/Makefile | 1 - lib/librte_ethdev/Makefile| 2 +- lib/librte_ethdev/m

[dpdk-dev] [PATCH v2 6/7] ether: use bitops to speedup comparison

Using bit operations like or and xor is faster than a loop on all architectures. Really just explicit unrolling. Similar cast to uint16 unaligned is already done in other functions here. Signed-off-by: Stephen Hemminger --- lib/librte_net/rte_ether.h | 17 +++-- 1 file changed, 7 in

[dpdk-dev] [PATCH v2 4/7] ether: add eth_unformat_addr

Make a function that coresponds with eth_aton_r which can be used to convert string to ether_addr. This also allows rte_ethdev to no longer depend on the cmdline library. Signed-off-by: Stephen Hemminger --- lib/librte_net/rte_ether.c | 47 ++ lib/librte_net/

[dpdk-dev] [PATCH v2 3/7] ether: deinline non-critical functions

Formatting ethernet address and getting a random value are not in critical path so they should not be inlined. Signed-off-by: Stephen Hemminger --- lib/librte_net/Makefile| 1 + lib/librte_net/rte_ether.c | 29 + lib/librte_net/rte_ether.h

[dpdk-dev] [PATCH v2 0/7] ether: improvements and optimizations

This is a collection of patches around the ethernet address manipulation routines in librte_net/rte_ether. This does not include the prefixing of the ether_address which is still under review. v2 incorporates feedback and Bruce's alignement patch Bruce Richardson (1): ether: mark ethernet addr

[dpdk-dev] [PATCH v2 1/7] ether: mark ethernet addresses as being 2-byte aligned

From: Bruce Richardson When including the rte_ether.h header in applications with warnings enabled, a warning was given because of the assumption of 2-byte alignment of ethernet addresses when processing them. .../include/rte_ether.h:149:2: warning: converting a packed ‘const struct ether_addr

[dpdk-dev] [PATCH v2 2/7] ether: do not mark ethernet address and header as packed

Since the Ethernet address and header are naturally aligned correctly on all architectures, there is no need to mark them as packed. This results in faster code, and also gets rid of warnings with gcc-9 about getting address of packed structure. Suggested-by: Bruce Richardson Signed-off-by: Steph

Re: [dpdk-dev] [PATCH] ether: mark ethernet addresses as being 2-byte aligned

On 16/05/2019 16:54, Bruce Richardson wrote: > When including the rte_ether.h header in applications with warnings > enabled, a warning was given because of the assumption of 2-byte alignment > of ethernet addresses when processing them. > > .../include/rte_ether.h:149:2: warning: converting a pac

[dpdk-dev] [PATCH v2 6/6] eal: add PRNG to MAINTAINERS and release notes

Add a section on PRNG in MAINTAINERS, and information on the PRNG improvements in the 19.08 release notes. Signed-off-by: Mattias Rönnblom --- MAINTAINERS| 6 ++ doc/guides/rel_notes/release_19_08.rst | 13 + 2 files changed, 19 insertions(+) diff --

[dpdk-dev] [PATCH v2 4/6] eal: introduce random generator function with upper bound

Add a function rte_rand_max() which generates an uniformly distributed pseudo-random number less than a user-specified upper bound. The commonly used pattern rte_rand() % SOME_VALUE creates biased results (as in some values in the range are more frequently occurring than others) if SOME_VALUE is n

[dpdk-dev] [PATCH v2 5/6] eal: add bounded PRNG performance tests

Add best- and worst-case performance tests for rte_rand_max(). Signed-off-by: Mattias Rönnblom --- app/test/test_rand_perf.c | 19 ++- 1 file changed, 18 insertions(+), 1 deletion(-) diff --git a/app/test/test_rand_perf.c b/app/test/test_rand_perf.c index 771713757..fe797ebfa 10

[dpdk-dev] [PATCH v2 3/6] eal: improve entropy for initial PRNG seed

Replace the use of rte_get_timer_cycles() with getentropy() for seeding the pseudo-random number generator. getentropy() provides a more truly random value. getentropy() requires glibc 2.25 and Linux kernel 3.17. In case getentropy() is not found at compile time, or the relevant syscall fails in r

[dpdk-dev] [PATCH v2 1/6] eal: replace libc-based random number generation with LFSR

This commit replaces rte_rand()'s use of lrand48() with a DPDK-native combined Linear Feedback Shift Register (LFSR) (also known as Tausworthe) pseudo-random number generator. This generator is faster and produces better-quality random numbers than the linear congruential generator (LCG) of lib's

[dpdk-dev] [PATCH v2 0/6] Pseudo-random number generation improvements

Make DPDK pseudo-random number generation multi-thread safe, go faster and produce better-quality pseudo-random numbers. This patchset requires the yet-to-be-merged rdseed build detection patch for rdseed support to be enabled. http://patchwork.dpdk.org/patch/53404/ Thanks to Stephen Hemminger, B

[dpdk-dev] [PATCH v2 2/6] eal: add pseudo-random number generation performance test

Add performance test for rte_rand(). Signed-off-by: Mattias Rönnblom --- app/test/Makefile | 1 + app/test/test_rand_perf.c | 75 +++ 2 files changed, 76 insertions(+) create mode 100644 app/test/test_rand_perf.c diff --git a/app/test/Makefile b/app

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, 16 May 2019 17:36:43 +0100 Bruce Richardson wrote: > On Thu, May 16, 2019 at 05:07:45PM +0100, Bruce Richardson wrote: > > On Thu, May 16, 2019 at 09:06:52AM -0700, Stephen Hemminger wrote: > > > On Thu, 16 May 2019 17:03:37 +0100 > > > Bruce Richardson wrote: > > > > > > > On Wed,

[dpdk-dev] DPDK Release Status Meeting 16/5/2019

Minutes 16 May 2019 --- Agenda: * Release Dates * Subtrees * OvS * Conferences * Opens Participants: * Debian/Microsoft * Intel * Marvell * Mellanox * NXP * RedHat Release Dates - * v19.05 is released on Tuesday, 14 May * https://mails.dpdk.org/archives/dev/2019-

Re: [dpdk-dev] [PATCH] kvargs: trim spaces at the beginning and end of key and values

On Wed, 15 May 2019 22:41:07 -0700 Mesut Ali Ergin wrote: > +/* trim leading and trailing spaces */ > +static char * > +trim_space(char *str) > +{ > + char *start, *end; > + > + for (start = str; *start; start++) { > + if (!isspace((unsigned char) start[0])) > +

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, May 16, 2019 at 05:07:45PM +0100, Bruce Richardson wrote: > On Thu, May 16, 2019 at 09:06:52AM -0700, Stephen Hemminger wrote: > > On Thu, 16 May 2019 17:03:37 +0100 > > Bruce Richardson wrote: > > > > > On Wed, May 15, 2019 at 03:19:52PM -0700, Stephen Hemminger wrote: > > > > Using bit

Re: [dpdk-dev] [PATCH v1] net/ice: change RSS RETA size to meet with ETH_RSS_RETA_SIZE_x

> -Original Message- > From: Maxime Coquelin [mailto:maxime.coque...@redhat.com] > Sent: Thursday, May 16, 2019 19:36 > To: Wang, Haiyue ; Zhang, Qi Z > ; dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH v1] net/ice: change RSS RETA size to meet with > ETH_RSS_RETA_SIZE_x > > > > On 5/14/19

Re: [dpdk-dev] [PATCH] net/af_xdp: add libelf for pmd_af_xdp.

On Thu, May 16, 2019 at 1:59 AM Luca Boccassi wrote: > > On Wed, 2019-05-15 at 13:24 -0700, William Tu wrote: > > The libbpf requires linking elf library. Fix it by > > adding -lelf. > > > > Fixes: f1debd77efaf ("net/af_xdp: introduce AF_XDP PMD") > > Signed-off-by: William Tu < > > u9012...@gmail

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, May 16, 2019 at 09:06:52AM -0700, Stephen Hemminger wrote: > On Thu, 16 May 2019 17:03:37 +0100 > Bruce Richardson wrote: > > > On Wed, May 15, 2019 at 03:19:52PM -0700, Stephen Hemminger wrote: > > > Using bit operations like or and xor is faster than a loop > > > on all architectures. R

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, 16 May 2019 17:03:37 +0100 Bruce Richardson wrote: > On Wed, May 15, 2019 at 03:19:52PM -0700, Stephen Hemminger wrote: > > Using bit operations like or and xor is faster than a loop > > on all architectures. Really just explicit unrolling. > > > > Similar cast to uint16 unaligned is alr

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Wed, May 15, 2019 at 03:19:52PM -0700, Stephen Hemminger wrote: > Using bit operations like or and xor is faster than a loop > on all architectures. Really just explicit unrolling. > > Similar cast to uint16 unaligned is already done in > other functions here. > > Signed-off-by: Stephen Hemmin

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On Thu, 16 May 2019 11:03:10 +0200 Mattias Rönnblom wrote: > On 2019-05-16 00:19, Stephen Hemminger wrote: > > Using bit operations like or and xor is faster than a loop > > on all architectures. Really just explicit unrolling. > > > > Similar cast to uint16 unaligned is already done in > > othe

[dpdk-dev] [PATCH] ether: mark ethernet addresses as being 2-byte aligned

When including the rte_ether.h header in applications with warnings enabled, a warning was given because of the assumption of 2-byte alignment of ethernet addresses when processing them. .../include/rte_ether.h:149:2: warning: converting a packed ‘const struct ether_addr’ pointer (alignment 1) t

Re: [dpdk-dev] [RFC v8] /net: memory interface (memif)

On Thu, 16 May 2019 13:46:58 +0200 Jakub Grajciar wrote: > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c > index f0061d99f..b05f1462d 100644 > --- a/app/test-pmd/testpmd.c > +++ b/app/test-pmd/testpmd.c > @@ -2493,6 +2493,8 @@ pmd_test_exit(void) > device = rt

[dpdk-dev] [RFC] ethdev: enhance metadata as flow rule criteria

Current implementation of rte_flow metadata match sets metadata only for egress traffic leaving the need for such feature in ingress flows unsatisfied and the lack of metadata index or ID limits it to a single metadata action and match per flow which is a drawback in several scenarios we would like

Re: [dpdk-dev] [RFC v8] /net: memory interface (memif)

On Thu, 16 May 2019 13:46:58 +0200 Jakub Grajciar wrote: > + /* remote info */ > + char remote_name[64]; /**< remote app name */ > + char remote_if_name[64]; Hard coding magic string sizes has future potential for disaster. Could you at least add a #define. > +type

Re: [dpdk-dev] [RFC v8] /net: memory interface (memif)

On Thu, 16 May 2019 13:46:58 +0200 Jakub Grajciar wrote: > +enum memif_role_t { > + MEMIF_ROLE_MASTER, > + MEMIF_ROLE_SLAVE, > +}; Because master/slave terminology is potentially culturally offensive it is flagged by many corporate source scanning tools. Could you use primary/secondary

[dpdk-dev] [RFC] net/mlx5: support for flow action on VLAN header

VLAN actions support is implemented in librte_ethdev, and in test-pmd application, based on [1] Generic flow API. These actions conform to the VLAN actions defined in [2] the OpenFlow Switch Specification. rte_flow defines the following VLAN actions: 1. OF_POP_VLAN Pop the outer-most VLAN hea

Re: [dpdk-dev] [RFC v8] /net: memory interface (memif)

On Thu, 16 May 2019 13:46:58 +0200 Jakub Grajciar wrote: > + > +#define ETH_MEMIF_DEFAULT_SOCKET_FILENAME"/tmp/memif.sock" > +#define ETH_MEMIF_DEFAULT_RING_SIZE 10 > +#define ETH_MEMIF_DEFAULT_PKT_BUFFER_SIZE2048 The current best practice on Linux is to use /run for sockets. It

Re: [dpdk-dev] [RFC v8] /net: memory interface (memif)

On Thu, 16 May 2019 13:46:58 +0200 Jakub Grajciar wrote: > +struct memif_queue { > + struct rte_mempool *mempool;/**< mempool for RX packets */ > + uint16_t in_port; /**< port id */ > + > + struct pmd_internals *pmd; /**< device internals

Re: [dpdk-dev] [PATCH v2] net/i40e: Eliminate weak symbols in i40e_rxtx.c

On Wed, May 15, 2019 at 12:13:46PM -0400, David Harton wrote: > Use of weak symbols can hide makefile errors especially when > custom makefiles are used. Removing the use of weak symbols > to avoid a stub function being linked in production code. > > Signed-off-by: David Harton > --- > > v2 - a

Re: [dpdk-dev] DPDK Windows Community call - 02 May 2019

sending again with fixed Cc list 16/05/2019 11:41, Thomas Monjalon: > 14/05/2019 17:48, Thomas Monjalon: > > 10/05/2019 01:34, Ranjit Menon: > > > Contact Cathal to be added to this meeting invite. > > > > Would it be possible to move this bi-weekly meeting > > to the Tuesday (same time), please?

[dpdk-dev] [RFC] net/mlx5: add support of LRO in MLX5 PMD

Introduction: LRO (Large Receive Offload) is intended to reduce host CPU overhead when processing Rx TCP packets. LRO works by aggregating multiple incoming packets from a single stream into a larger buffer, before they are passed higher up the networking stack. Thus reducing the number of packet

[dpdk-dev] [RFC v8] /net: memory interface (memif)

Memory interface (memif), provides high performance packet transfer over shared memory. Signed-off-by: Jakub Grajciar --- MAINTAINERS |6 + app/test-pmd/testpmd.c |2 + config/common_base |5 + config/commo

[dpdk-dev] [PATCH v2 2/3] examples/ip_fragmentation: Enabling IP checksum offload in mbuf

As per the documentation to use any IP offload features, application must set required offload flags into mbuf->ol_flags. Signed-off-by: Sunil Kumar Kori --- examples/ip_fragmentation/main.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/examples/ip_fragmentation/main

[dpdk-dev] [PATCH v2 3/3] examples/ip_reassembly: Enabling IP checksum offload in mbuf

As per the documentation to use any IP offload features, application must set required offload flags into mbuf->ol_flags. Signed-off-by: Sunil Kumar Kori --- examples/ip_reassembly/main.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/examples/ip_reassembly/main.c b/examples/ip_reasse

[dpdk-dev] [PATCH v2 1/3] lib/librte_ip_frag: Remove PKT_TX_IP_CKSUM offload flags

Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use checksum offload whether it is supported by platform or not. Also documentation does not provide any expected value of ol_flags in retu

[dpdk-dev] [PATCH v2 0/3] Removing PKT_TX_IP_CKSUM from fragmentation/reassembly lib

v2: Checkpath warning are removed Patchset contains below changes: 1. Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use chec

Re: [dpdk-dev] [PATCH v1] net/ice: change RSS RETA size to meet with ETH_RSS_RETA_SIZE_x

On 5/14/19 3:19 PM, Haiyue Wang wrote: Since ice can support 128, 512, 2K three RSS RETA size, and if set 2K by default, this will make it not compatible with ETH_RSS_RETA_SIZE_x value definition, limit it to 512 if the cap.rss_table_size exceeds the max value 512. It looks like a fix, so "F

[dpdk-dev] [PATCH 3/3] examples/ip_reassembly: Enabling IP checksum offload in mbuf

Signed-off-by: Sunil Kumar Kori --- examples/ip_reassembly/main.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c index 17b55d4..686e364 100644 --- a/examples/ip_reassembly/main.c +++ b/examples/ip_reassembly/main.c @@ -353,6

[dpdk-dev] [PATCH 2/3] examples/ip_fragmentation: Enabling IP checksum offload in mbuf

Signed-off-by: Sunil Kumar Kori --- examples/ip_fragmentation/main.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/examples/ip_fragmentation/main.c b/examples/ip_fragmentation/main.c index e90a61e..d821967 100644 --- a/examples/ip_fragmentation/main.c +++ b/examples/i

[dpdk-dev] [PATCH 0/3] Removing PKT_TX_IP_CKSUM from fragmentation/reassembly lib

Patchset contains below changes: 1. Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use checksum offload whether it is supporte

[dpdk-dev] [PATCH 1/3] lib/librte_ip_frag: Remove PKT_TX_IP_CKSUM offload flags

Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use checksum offload whether it is supported by platform or not. Also documentation does not provide any expected value of ol_flags in retu

[dpdk-dev] [PATCH 3/3] examples/ip_reassembly: Enabling IP checksum offload in mbuf

Change-Id: Idce2db6b3e9437bcbe1ff42281fde42f36902c24 Signed-off-by: Sunil Kumar Kori --- examples/ip_reassembly/main.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c index 17b55d4..686e364 100644 --- a/examples/ip_reassembly

[dpdk-dev] [PATCH 2/3] examples/ip_fragmentation: Enabling IP checksum offload in mbuf

Change-Id: I616dc217a96f0bd4c77dd70959dff3c208524f84 Signed-off-by: Sunil Kumar Kori --- examples/ip_fragmentation/main.c | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/examples/ip_fragmentation/main.c b/examples/ip_fragmentation/main.c index e90a61e..d821967 100644 --

[dpdk-dev] [PATCH 0/3] Removing PKT_TX_IP_CKSUM from fragmentation/reassembly lib

Patchset contains below changes: 1. Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use checksum offload whether it is supporte

[dpdk-dev] [PATCH 1/3] lib/librte_ip_frag: Remove PKT_TX_IP_CKSUM offload flags

Currently PKT_TX_IP_CKSUM is being set into mbuf->ol_flags during fragmentation and reassemble operation implicitly. Because of this, application is forced to use checksum offload whether it is supported by platform or not. Also documentation does not provide any expected value of ol_flags in retu

[dpdk-dev] [Bug 280] X710 PF Reset Issue with DPDK VF Driver

https://bugs.dpdk.org/show_bug.cgi?id=280 Bug ID: 280 Summary: X710 PF Reset Issue with DPDK VF Driver Product: DPDK Version: 19.05 Hardware: Other OS: Linux Status: CONFIRMED Severity: major Pri

[dpdk-dev] [Bug 279] Unaligned memory access when reading ipv6 header

https://bugs.dpdk.org/show_bug.cgi?id=279 Bug ID: 279 Summary: Unaligned memory access when reading ipv6 header Product: DPDK Version: 19.02 Hardware: x86 OS: Linux Status: CONFIRMED Severity: normal

Re: [dpdk-dev] [RFC 3/4] ethdev: use eth_unformat_addr

On Wed, May 15, 2019 at 03:19:51PM -0700, Stephen Hemminger wrote: > Use eth_unformat_addr, so that ethdev can be built and work > without the cmdline library. The dependency on cmdline was > an arrangement of convenience anyway. > > Signed-off-by: Stephen Hemminger > --- > lib/librte_ethdev/rte

Re: [dpdk-dev] DPDK Windows Community call - 02 May 2019

14/05/2019 17:48, Thomas Monjalon: > 10/05/2019 01:34, Ranjit Menon: > > Contact Cathal to be added to this meeting invite. > > Would it be possible to move this bi-weekly meeting > to the Tuesday (same time), please? > > [...] > > Encourage other dpdk.org Windows partners to join this call. > >

Re: [dpdk-dev] [PATCH 2/4] net/bonding: fix LACP fast queue Rx handler

Hello Chas, On Fri, Apr 19, 2019 at 12:50 AM Chas Williams <3ch...@gmail.com> wrote: > On 4/18/19 3:11 AM, David Marchand wrote: > > Hello Chas, > > > > On Fri, Apr 12, 2019 at 4:02 PM Chas Williams <3ch...@gmail.com > > > wrote: > > > > I should have some time this w

Re: [dpdk-dev] [RFC 4/4] net/ether: use bitops to speedup comparison

On 2019-05-16 00:19, Stephen Hemminger wrote: Using bit operations like or and xor is faster than a loop on all architectures. Really just explicit unrolling. Similar cast to uint16 unaligned is already done in other functions here. Signed-off-by: Stephen Hemminger --- lib/librte_net/rte_eth

Re: [dpdk-dev] [PATCH] net/af_xdp: add libelf for pmd_af_xdp.

On Wed, 2019-05-15 at 13:24 -0700, William Tu wrote: > The libbpf requires linking elf library. Fix it by > adding -lelf. > > Fixes: f1debd77efaf ("net/af_xdp: introduce AF_XDP PMD") > Signed-off-by: William Tu < > u9012...@gmail.com > > > --- > mk/rte.app.mk | 2 +- > 1 file changed, 1 insertion

Re: [dpdk-dev] [PATCH] net/vmxnet3: uninitialized variable fix

On Thu, May 16, 2019 at 9:10 AM Yogev Chaimovich wrote: I am not sure this is the right fix but I don't know this driver so I will let Yong comment on this. On the form, please, have a look at the recommandations for submitting patches: http://doc.dpdk.org/guides/contributing/patches.html#commit

Re: [dpdk-dev] [PATCH 3/3] net/i40e: fix inadvertent override of vector RX allowance

On 5/16/19 6:28 AM, Mesut Ali Ergin wrote: When i40e_rx_vec_dev_conf_condition_check_default() determines whether vector receive functions would be allowed during initialization phase, it should honor previously recorded disallowance during configuration phase, and not override simply because

Re: [dpdk-dev] [PATCH 2/3] net/i40e: add runtime option to disable vector rx

On 5/16/19 6:28 AM, Mesut Ali Ergin wrote: Vector RX functions are not at feature parity with non-vector ones and currently the vector RX path is enabled by default. Hence, the only option to force selection of non-vector variants and be able to retain functionality is to disable vector PMD gl

Re: [dpdk-dev] [PATCH] examples/vhost_crypto: fix unchecked return value

On Wed, May 15, 2019 at 05:25:48PM +0100, Fan Zhang wrote: This patch tries to fix the coverity issue of unchecked return value. Since the function that causes the problem is unused, it is removed completely. Coverity issue: 336816 Fixes: f5188211c721 ("examples/vhost_crypto: add sample applicat

Re: [dpdk-dev] [PATCH] ring: remove unnecessary forward declaration

On Thu, May 16, 2019 at 03:28:56PM +0800, Xiaolong Ye wrote: > As memzone.h is introduced by commit 38c9817ee1d8 ("mempool: adjust name > size in related data types"), forward declaration for rte_memzone is no > longer needed. > > Signed-off-by: Xiaolong Ye Acked-by: Olivier Matz Thanks

[dpdk-dev] [PATCH] ring: remove unnecessary forward declaration

As memzone.h is introduced by commit 38c9817ee1d8 ("mempool: adjust name size in related data types"), forward declaration for rte_memzone is no longer needed. Signed-off-by: Xiaolong Ye --- lib/librte_ring/rte_ring.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/lib/librte_ring/rte_ring.

Re: [dpdk-dev] [RFC 3/4] ethdev: use eth_unformat_addr

On Thu, May 16, 2019 at 12:20 AM Stephen Hemminger < step...@networkplumber.org> wrote: > Use eth_unformat_addr, so that ethdev can be built and work > without the cmdline library. The dependency on cmdline was > an arrangement of convenience anyway. > > Signed-off-by: Stephen Hemminger > --- >

Re: [dpdk-dev] [RFC 2/4] net/ether: add eth_unformat_addr

On Thu, May 16, 2019 at 12:20 AM Stephen Hemminger < step...@networkplumber.org> wrote: > Make a function that coresponds with eth_aton_r which can > nit: corresponds be used to convert string to ether_addr. > > This also allows rte_ethdev to no longer depend on the > cmdline library. > > Signed

Re: [dpdk-dev] [RFC 1/4] net/ether: deinline non-critical functions

On Thu, May 16, 2019 at 12:20 AM Stephen Hemminger < step...@networkplumber.org> wrote: > Formatting ethernet address and getting a random value are > not in critical path so they should not be inlined. > > Signed-off-by: Stephen Hemminger > --- > lib/librte_net/Makefile| 1 + > lib

[dpdk-dev] [PATCH] net/vmxnet3: uninitialized variable fix

Coverity issue: 323479 Fixes: 30f77abe net/vmxnet3: support stats reset Cc: sta...@dpdk.org Signed-off-by: yogev ch --- drivers/net/vmxnet3/vmxnet3_ethdev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/vmxnet3/vmxnet3_ethdev.c b/drivers/net/vmxnet3/vmxnet3