Only use the "avx512vbmi" compiler target when it is actually supported
by the compiler.
The order of mfex_impls and the 'dpif_miniflow_extract_impl_idx' enum
have to be changed to keep the start index and size of the impl list
correct in both VBMI and non VBMI cases.
Signed-off-by: Cian Ferriter <cian.ferri...@intel.com>
---
v2:
* Don't register vbmi specialized mfex impls unless VBMI is actually
available.
* This required some re-ordering of the mfex impl lists.
---
acinclude.m4 | 14 +++++++++++
configure.ac | 1 +
lib/dpif-netdev-extract-avx512.c | 23 ++++++++++++++---
lib/dpif-netdev-private-extract.c | 41 ++++++++++++++++---------------
lib/dpif-netdev-private-extract.h | 12 +++++----
5 files changed, 63 insertions(+), 28 deletions(-)
diff --git a/acinclude.m4 b/acinclude.m4
index 932ff4693..31033edca 100644
--- a/acinclude.m4
+++ b/acinclude.m4
@@ -100,6 +100,20 @@ AC_DEFUN([OVS_CHECK_AVX512BW_DQ], [
fi
])
+dnl OVS_CHECK_AVX512VBMI
+dnl
+dnl Checks if compiler supports AVX512VBMI instructions.
+AC_DEFUN([OVS_CHECK_AVX512VBMI], [
+ OVS_CHECK_CC_OPTION(
+ [-mavx512vbmi], [ovs_have_cc_mavx512vbmi=yes],
+ [ovs_have_cc_mavx512vbmi=no])
+ AM_CONDITIONAL([HAVE_AVX512VBMI], [test $ovs_have_cc_mavx512vbmi = yes])
+ if test "$ovs_have_cc_mavx512vbmi" = yes; then
+ AC_DEFINE([HAVE_AVX512VBMI], [1],
+ [Define to 1 if compiler supports AVX512VBMI.])
+ fi
+])
+
dnl OVS_ENABLE_WERROR
AC_DEFUN([OVS_ENABLE_WERROR],
[AC_ARG_ENABLE(
diff --git a/configure.ac b/configure.ac
index f849837e0..12b4010e5 100644
--- a/configure.ac
+++ b/configure.ac
@@ -188,6 +188,7 @@ OVS_CHECK_DPIF_AVX512_DEFAULT
OVS_CHECK_MFEX_AUTOVALIDATOR
OVS_CHECK_AVX512
OVS_CHECK_AVX512BW_DQ
+OVS_CHECK_AVX512VBMI
AC_ARG_VAR(KARCH, [Kernel Architecture String])
AC_SUBST(KARCH)
diff --git a/lib/dpif-netdev-extract-avx512.c b/lib/dpif-netdev-extract-avx512.c
index f36a6eab0..5d62f3607 100644
--- a/lib/dpif-netdev-extract-avx512.c
+++ b/lib/dpif-netdev-extract-avx512.c
@@ -110,7 +110,9 @@ _mm512_maskz_permutex2var_epi8_skx(__mmask64 k_mask,
/* Wrapper function required to enable ISA. */
static inline __m512i
+#if HAVE_AVX512VBMI
__attribute__((__target__("avx512vbmi")))
+#endif
_mm512_maskz_permutexvar_epi8_wrap(__mmask64 kmask, __m512i idx, __m512i a)
{
return _mm512_maskz_permutexvar_epi8(kmask, idx, a);
@@ -481,7 +483,7 @@ mfex_avx512_process(struct dp_packet_batch *packets,
odp_port_t in_port,
void *pmd_handle OVS_UNUSED,
const enum MFEX_PROFILES profile_id,
- const uint32_t use_vbmi)
+ const uint32_t use_vbmi OVS_UNUSED)
{
uint32_t hitmask = 0;
struct dp_packet *packet;
@@ -544,7 +546,11 @@ mfex_avx512_process(struct dp_packet_batch *packets,
*/
__m512i v512_zeros = _mm512_setzero_si512();
__m512i v_blk0;
+#if HAVE_AVX512VBMI
if (__builtin_constant_p(use_vbmi) && use_vbmi) {
+#else
+ if (0) {
+#endif
v_blk0 = _mm512_maskz_permutexvar_epi8_wrap(k_shuf, v_shuf,
v_pkt0);
} else {
@@ -626,7 +632,9 @@ mfex_avx512_process(struct dp_packet_batch *packets,
}
-#define DECLARE_MFEX_FUNC(name, profile) \
+#if HAVE_AVX512VBMI
+#define VBMI_MFEX_FUNC(name, profile) \
+ \
uint32_t \
__attribute__((__target__("avx512f"))) \
__attribute__((__target__("avx512vbmi"))) \
@@ -637,7 +645,12 @@ mfex_avx512_vbmi_##name(struct dp_packet_batch *packets,
\
{ \
return mfex_avx512_process(packets, keys, keys_size, in_port, \
pmd_handle, profile, 1); \
-} \
+}
+#else
+#define VBMI_MFEX_FUNC(name, profile)
+#endif
+
+#define BASIC_MFEX_FUNC(name, profile) \
\
uint32_t \
__attribute__((__target__("avx512f"))) \
@@ -650,6 +663,10 @@ mfex_avx512_##name(struct dp_packet_batch *packets,
\
pmd_handle, profile, 0); \
}
+#define DECLARE_MFEX_FUNC(name, profile) \
+VBMI_MFEX_FUNC(name, profile) \
+BASIC_MFEX_FUNC(name, profile) \
+
/* Each profile gets a single declare here, which specializes the function
* as required.
*/
diff --git a/lib/dpif-netdev-private-extract.c
b/lib/dpif-netdev-private-extract.c
index 43b8b824e..ea2b03e5c 100644
--- a/lib/dpif-netdev-private-extract.c
+++ b/lib/dpif-netdev-private-extract.c
@@ -56,45 +56,46 @@ static struct dpif_miniflow_extract_impl mfex_impls[] = {
/* Compile in implementations only if the compiler ISA checks pass. */
#if (__x86_64__ && HAVE_AVX512F && HAVE_LD_AVX512_GOOD && HAVE_AVX512BW_DQ \
&& __SSE4_2__)
- [MFEX_IMPL_VBMI_IPv4_UDP] = {
- .probe = mfex_avx512_vbmi_probe,
- .extract_func = mfex_avx512_vbmi_ip_udp,
- .name = "avx512_vbmi_ipv4_udp", },
-
[MFEX_IMPL_IPv4_UDP] = {
.probe = mfex_avx512_probe,
.extract_func = mfex_avx512_ip_udp,
.name = "avx512_ipv4_udp", },
- [MFEX_IMPL_VBMI_IPv4_TCP] = {
- .probe = mfex_avx512_vbmi_probe,
- .extract_func = mfex_avx512_vbmi_ip_tcp,
- .name = "avx512_vbmi_ipv4_tcp", },
-
[MFEX_IMPL_IPv4_TCP] = {
.probe = mfex_avx512_probe,
.extract_func = mfex_avx512_ip_tcp,
.name = "avx512_ipv4_tcp", },
- [MFEX_IMPL_VBMI_DOT1Q_IPv4_UDP] = {
- .probe = mfex_avx512_vbmi_probe,
- .extract_func = mfex_avx512_vbmi_dot1q_ip_udp,
- .name = "avx512_vbmi_dot1q_ipv4_udp", },
-
[MFEX_IMPL_DOT1Q_IPv4_UDP] = {
.probe = mfex_avx512_probe,
.extract_func = mfex_avx512_dot1q_ip_udp,
.name = "avx512_dot1q_ipv4_udp", },
- [MFEX_IMPL_VBMI_DOT1Q_IPv4_TCP] = {
- .probe = mfex_avx512_vbmi_probe,
- .extract_func = mfex_avx512_vbmi_dot1q_ip_tcp,
- .name = "avx512_vbmi_dot1q_ipv4_tcp", },
-
[MFEX_IMPL_DOT1Q_IPv4_TCP] = {
.probe = mfex_avx512_probe,
.extract_func = mfex_avx512_dot1q_ip_tcp,
.name = "avx512_dot1q_ipv4_tcp", },
+#if HAVE_AVX512VBMI
+ [MFEX_IMPL_VBMI_IPv4_UDP] = {
+ .probe = mfex_avx512_vbmi_probe,
+ .extract_func = mfex_avx512_vbmi_ip_udp,
+ .name = "avx512_vbmi_ipv4_udp", },
+
+ [MFEX_IMPL_VBMI_IPv4_TCP] = {
+ .probe = mfex_avx512_vbmi_probe,
+ .extract_func = mfex_avx512_vbmi_ip_tcp,
+ .name = "avx512_vbmi_ipv4_tcp", },
+
+ [MFEX_IMPL_VBMI_DOT1Q_IPv4_UDP] = {
+ .probe = mfex_avx512_vbmi_probe,
+ .extract_func = mfex_avx512_vbmi_dot1q_ip_udp,
+ .name = "avx512_vbmi_dot1q_ipv4_udp", },
+
+ [MFEX_IMPL_VBMI_DOT1Q_IPv4_TCP] = {
+ .probe = mfex_avx512_vbmi_probe,
+ .extract_func = mfex_avx512_vbmi_dot1q_ip_tcp,
+ .name = "avx512_vbmi_dot1q_ipv4_tcp", },
+#endif
#endif
};
diff --git a/lib/dpif-netdev-private-extract.h
b/lib/dpif-netdev-private-extract.h
index bbb80ddca..ea1d03918 100644
--- a/lib/dpif-netdev-private-extract.h
+++ b/lib/dpif-netdev-private-extract.h
@@ -83,14 +83,16 @@ enum dpif_miniflow_extract_impl_idx {
MFEX_IMPL_STUDY,
#if (__x86_64__ && HAVE_AVX512F && HAVE_LD_AVX512_GOOD && HAVE_AVX512BW_DQ \
&& __SSE4_2__)
- MFEX_IMPL_VBMI_IPv4_UDP,
MFEX_IMPL_IPv4_UDP,
- MFEX_IMPL_VBMI_IPv4_TCP,
MFEX_IMPL_IPv4_TCP,
- MFEX_IMPL_VBMI_DOT1Q_IPv4_UDP,
MFEX_IMPL_DOT1Q_IPv4_UDP,
- MFEX_IMPL_VBMI_DOT1Q_IPv4_TCP,
MFEX_IMPL_DOT1Q_IPv4_TCP,
+#if HAVE_AVX512VBMI
+ MFEX_IMPL_VBMI_IPv4_UDP,
+ MFEX_IMPL_VBMI_IPv4_TCP,
+ MFEX_IMPL_VBMI_DOT1Q_IPv4_UDP,
+ MFEX_IMPL_VBMI_DOT1Q_IPv4_TCP,
+#endif
#endif
MFEX_IMPL_MAX
};
@@ -103,7 +105,7 @@ extern struct ovs_mutex dp_netdev_mutex;
#if (__x86_64__ && HAVE_AVX512F && HAVE_LD_AVX512_GOOD && HAVE_AVX512BW_DQ \
&& __SSE4_2__)
-#define MFEX_IMPL_START_IDX MFEX_IMPL_VBMI_IPv4_UDP
+#define MFEX_IMPL_START_IDX MFEX_IMPL_IPv4_UDP
#else
#define MFEX_IMPL_START_IDX MFEX_IMPL_MAX
--
2.25.1
_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
