From: Jan Viktorin <viktorin@rehivetech.com>
To: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, Vlastimil Kosar <kosar@rehivetech.com>
Subject: Re: [dpdk-dev] [PATCH v2 15/16] lpm/arm: implement rte_lpm_lookupx4 using rte_lpm_lookup_bulk on for-x86
Date: Tue, 27 Oct 2015 16:38:52 +0100 [thread overview]
Message-ID: <20151027163852.1937e6ac@pcviktorin.fit.vutbr.cz> (raw)
In-Reply-To: <2601191342CEEE43887BDE71AB97725836AB58ED@irsmsx105.ger.corp.intel.com>
Hi Konstantin,
On Tue, 27 Oct 2015 15:31:44 +0000
"Ananyev, Konstantin" <konstantin.ananyev@intel.com> wrote:
> Hi Jan,
>
> > -----Original Message-----
> > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Jan Viktorin
> > Sent: Monday, October 26, 2015 4:38 PM
> > To: Thomas Monjalon; Hunt, David; dev@dpdk.org
> > Cc: Vlastimil Kosar
> > Subject: [dpdk-dev] [PATCH v2 15/16] lpm/arm: implement rte_lpm_lookupx4 using rte_lpm_lookup_bulk on for-x86
> >
> > From: Vlastimil Kosar <kosar@rehivetech.com>
> >
> > LPM function rte_lpm_lookupx4() uses i686/x86_64 SIMD intrinsics. Therefore,
> > the function is reimplemented using non-vector operations for non-x86
> > architectures. In the future, each architecture should have vectorized code.
> > This patch includes rudimentary emulation of intrinsic functions _mm_set_epi32(),
> > _mm_loadu_si128() and _mm_load_si128() for easy portability of existing
> > applications.
> >
> > LPM builds now when on ARM.
> >
> > FIXME: to be reworked
> >
> > Signed-off-by: Vlastimil Kosar <kosar@rehivetech.com>
> > Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
> > ---
> > config/defconfig_arm-armv7-a-linuxapp-gcc | 1 -
> > lib/librte_lpm/rte_lpm.h | 71 +++++++++++++++++++++++++++++++
> > 2 files changed, 71 insertions(+), 1 deletion(-)
> >
> > diff --git a/config/defconfig_arm-armv7-a-linuxapp-gcc b/config/defconfig_arm-armv7-a-linuxapp-gcc
> > index 5b582a8..33afb33 100644
> > --- a/config/defconfig_arm-armv7-a-linuxapp-gcc
> > +++ b/config/defconfig_arm-armv7-a-linuxapp-gcc
> > @@ -58,7 +58,6 @@ CONFIG_XMM_SIZE=16
> >
> > # fails to compile on ARM
> > CONFIG_RTE_LIBRTE_ACL=n
> > -CONFIG_RTE_LIBRTE_LPM=n
> >
> > # cannot use those on ARM
> > CONFIG_RTE_KNI_KMOD=n
> > diff --git a/lib/librte_lpm/rte_lpm.h b/lib/librte_lpm/rte_lpm.h
> > index c299ce2..4619992 100644
> > --- a/lib/librte_lpm/rte_lpm.h
> > +++ b/lib/librte_lpm/rte_lpm.h
> > @@ -47,7 +47,9 @@
> > #include <rte_byteorder.h>
> > #include <rte_memory.h>
> > #include <rte_common.h>
> > +#if defined(RTE_ARCH_X86_64) || defined(RTE_ARCH_I686)
> > #include <rte_vect.h>
> > +#endif
> >
> > #ifdef __cplusplus
> > extern "C" {
> > @@ -358,6 +360,7 @@ rte_lpm_lookup_bulk_func(const struct rte_lpm *lpm, const uint32_t * ips,
> > return 0;
> > }
> >
> > +#if defined(RTE_ARCH_X86_64) || defined(RTE_ARCH_I686)
> > /* Mask four results. */
> > #define RTE_LPM_MASKX4_RES UINT64_C(0x00ff00ff00ff00ff)
> >
> > @@ -472,6 +475,74 @@ rte_lpm_lookupx4(const struct rte_lpm *lpm, __m128i ip, uint16_t hop[4],
> > hop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)tbl[2] : defv;
> > hop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)tbl[3] : defv;
> > }
> > +#else
>
> Probably better to create an lib/librte_eal/common/include/arch/arm/rte_vect.h,
> and move all these x86 vector support emulation there?
> Konstantin
Sure. This patch is terribly wrong and it's not to be merged. It is a
question whether to make it this way (with the refactoring as you
suggested) or to make some general abstraction of the SSE calls in DPDK.
Jan
>
> > +// TODO: this code should be reworked.
> > +
> > +typedef struct {
> > + union uint128 {
> > + uint8_t uint8[16];
> > + uint32_t uint32[4];
> > + } val;
> > +} __m128i;
> > +
> > +static inline __m128i
> > +_mm_set_epi32(uint32_t v0, uint32_t v1, uint32_t v2, uint32_t v3)
> > +{
> > + __m128i res;
> > + res.val.uint32[0] = v0;
> > + res.val.uint32[1] = v1;
> > + res.val.uint32[2] = v2;
> > + res.val.uint32[3] = v3;
> > + return res;
> > +}
> > +
> > +static inline __m128i
> > +_mm_loadu_si128(__m128i * v)
> > +{
> > + __m128i res;
> > + res = *v;
> > + return res;
> > +}
> > +
> > +static inline __m128i
> > +_mm_load_si128(__m128i * v)
> > +{
> > + __m128i res;
> > + res = *v;
> > + return res;
> > +}
> > +
> > +/**
> > + * Lookup four IP addresses in an LPM table.
> > + *
> > + * @param lpm
> > + * LPM object handle
> > + * @param ip
> > + * Four IPs to be looked up in the LPM table
> > + * @param hop
> > + * Next hop of the most specific rule found for IP (valid on lookup hit only).
> > + * This is an 4 elements array of two byte values.
> > + * If the lookup was succesfull for the given IP, then least significant byte
> > + * of the corresponding element is the actual next hop and the most
> > + * significant byte is zero.
> > + * If the lookup for the given IP failed, then corresponding element would
> > + * contain default value, see description of then next parameter.
> > + * @param defv
> > + * Default value to populate into corresponding element of hop[] array,
> > + * if lookup would fail.
> > + */
> > +static inline void
> > +rte_lpm_lookupx4(const struct rte_lpm *lpm, __m128i ip, uint16_t hop[4],
> > + uint16_t defv)
> > +{
> > + rte_lpm_lookup_bulk(lpm, ip.val.uint32, hop, 4);
> > +
> > + hop[0] = (hop[0] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)hop[0] : defv;
> > + hop[1] = (hop[1] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)hop[1] : defv;
> > + hop[2] = (hop[2] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)hop[2] : defv;
> > + hop[3] = (hop[3] & RTE_LPM_LOOKUP_SUCCESS) ? (uint8_t)hop[3] : defv;
> > +}
> > +#endif
> >
> > #ifdef __cplusplus
> > }
> > --
> > 2.6.1
>
--
Jan Viktorin E-mail: Viktorin@RehiveTech.com
System Architect Web: www.RehiveTech.com
RehiveTech
Brno, Czech Republic
next prev parent reply other threads:[~2015-10-27 15:40 UTC|newest]
Thread overview: 72+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-26 16:37 [dpdk-dev] [PATCH v2 00/16] Support ARMv7 architecture Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 01/16] mk: Introduce " Jan Viktorin
2015-10-28 13:34 ` David Marchand
2015-10-28 17:32 ` Jan Viktorin
2015-10-28 17:36 ` Richardson, Bruce
2015-10-28 13:39 ` David Marchand
2015-10-28 17:32 ` Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 02/16] eal/arm: atomic operations for ARM Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 03/16] eal/arm: byte order " Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 04/16] eal/arm: cpu cycle " Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 05/16] eal/arm: implement rdtsc by PMU or clock_gettime Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 06/16] eal/arm: prefetch operations for ARM Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 07/16] eal/arm: spinlock operations for ARM (without HTM) Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 08/16] eal/arm: vector memcpy for ARM Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 09/16] eal/arm: use vector memcpy only when NEON is enabled Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 10/16] eal/arm: cpu flag checks for ARM Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 11/16] eal/arm: detect arm architecture in cpu flags Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 12/16] eal/arm: rwlock support for ARM Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 13/16] gcc/arm: avoid alignment errors to break build Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 14/16] maintainers: claim responsibility for ARMv7 Jan Viktorin
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 15/16] lpm/arm: implement rte_lpm_lookupx4 using rte_lpm_lookup_bulk on for-x86 Jan Viktorin
2015-10-27 15:31 ` Ananyev, Konstantin
2015-10-27 15:38 ` Jan Viktorin [this message]
2015-10-26 16:37 ` [dpdk-dev] [PATCH v2 16/16] acl: check for SSE 4.1 support Jan Viktorin
2015-10-27 15:55 ` Ananyev, Konstantin
2015-10-27 17:10 ` Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 00/17] Support ARMv7 architecture Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 01/17] mk: Introduce " Jan Viktorin
2015-10-28 10:09 ` David Marchand
2015-10-28 10:56 ` Jan Viktorin
2015-10-28 13:40 ` David Marchand
2015-10-28 13:44 ` Hunt, David
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 02/17] eal/arm: atomic operations for ARM Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 03/17] eal/arm: byte order " Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 04/17] eal/arm: cpu cycle " Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 05/17] eal/arm: implement rdtsc by PMU or clock_gettime Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 06/17] eal/arm: prefetch operations for ARM Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 07/17] eal/arm: spinlock operations for ARM (without HTM) Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 08/17] eal/arm: vector memcpy for ARM Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 09/17] eal/arm: use vector memcpy only when NEON is enabled Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 10/17] eal/arm: cpu flag checks for ARM Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 11/17] eal/arm: detect arm architecture in cpu flags Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 12/17] eal/arm: rwlock support for ARM Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 13/17] gcc/arm: avoid alignment errors to break build Jan Viktorin
2015-10-28 12:16 ` David Marchand
2015-10-28 17:34 ` Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 14/17] maintainers: claim responsibility for ARMv7 Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 15/17] eal/arm: add very incomplete rte_vect Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 16/17] lpm/arm: implement rte_lpm_lookupx4 using rte_lpm_lookup_bulk for non-x86 Jan Viktorin
2015-10-27 19:13 ` [dpdk-dev] [PATCH v3 17/17] acl: handle when SSE 4.1 is unsupported Jan Viktorin
2015-10-28 14:54 ` [dpdk-dev] [PATCH v3 00/17] Support ARMv7 architecture David Marchand
2015-10-28 17:38 ` Jan Viktorin
2015-10-28 17:58 ` David Marchand
2015-10-29 14:02 ` Thomas Monjalon
2015-10-29 14:09 ` Jan Viktorin
2015-10-29 15:02 ` Thomas Monjalon
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 00/15] " Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 01/15] eal/arm: atomic operations for ARM Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 02/15] eal/arm: byte order " Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 03/15] eal/arm: cpu cycle " Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 04/15] eal/arm: implement rdtsc by PMU or clock_gettime Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 05/15] eal/arm: prefetch operations for ARM Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 06/15] eal/arm: spinlock operations for ARM (without HTM) Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 07/15] eal/arm: vector memcpy for ARM Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 08/15] eal/arm: use vector memcpy only when NEON is enabled Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 09/15] eal/arm: cpu flag checks for ARM Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 10/15] eal/arm: detect arm architecture in cpu flags Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 11/15] eal/arm: rwlock support for ARM Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 12/15] eal/arm: add very incomplete rte_vect Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 13/15] gcc/arm: avoid alignment errors to break build Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 14/15] mk: Introduce ARMv7 architecture Jan Viktorin
2015-10-29 12:43 ` [dpdk-dev] [PATCH v4 15/15] maintainers: claim responsibility for ARMv7 Jan Viktorin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20151027163852.1937e6ac@pcviktorin.fit.vutbr.cz \
--to=viktorin@rehivetech.com \
--cc=dev@dpdk.org \
--cc=konstantin.ananyev@intel.com \
--cc=kosar@rehivetech.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).