> -----Original Message-----
> From: Liu, Yong
> Sent: Monday, September 21, 2020 2:27 PM
> To: 'Maxime Coquelin' <maxime.coque...@redhat.com>; Xia, Chenbo
> <chenbo....@intel.com>; Wang, Zhihong <zhihong.w...@intel.com>
> Cc: dev@dpdk.org
> Subject: RE: [PATCH v1 4/5] vhost: add packed ring vectorized dequeue
>
>
>
> > -----Original Message-----
> > From: Maxime Coquelin <maxime.coque...@redhat.com>
> > Sent: Friday, September 18, 2020 9:45 PM
> > To: Liu, Yong <yong....@intel.com>; Xia, Chenbo <chenbo....@intel.com>;
> > Wang, Zhihong <zhihong.w...@intel.com>
> > Cc: dev@dpdk.org
> > Subject: Re: [PATCH v1 4/5] vhost: add packed ring vectorized dequeue
> >
> >
> >
> > On 8/19/20 5:24 AM, Marvin Liu wrote:
> > > Optimize vhost packed ring dequeue path with SIMD instructions. Four
> > > descriptors status check and writeback are batched handled with
> AVX512
> > > instructions. Address translation operations are also accelerated by
> > > AVX512 instructions.
> > >
> > > If platform or compiler not support vectorization, will fallback to
> > > default path.
> > >
> > > Signed-off-by: Marvin Liu <yong....@intel.com>
> > >
> > > diff --git a/lib/librte_vhost/Makefile b/lib/librte_vhost/Makefile
> > > index 4f2f3e47da..c0cd7d498f 100644
> > > --- a/lib/librte_vhost/Makefile
> > > +++ b/lib/librte_vhost/Makefile
> > > @@ -31,6 +31,13 @@ CFLAGS += -DVHOST_ICC_UNROLL_PRAGMA
> > > endif
> > > endif
> > >
> > > +ifneq ($(FORCE_DISABLE_AVX512), y)
> > > + CC_AVX512_SUPPORT=\
> > > + $(shell $(CC) -march=native -dM -E - </dev/null 2>&1 | \
> > > + sed '/./{H;$$!d} ; x ; /AVX512F/!d; /AVX512BW/!d; /AVX512VL/!d' |
> \
> > > + grep -q AVX512 && echo 1)
> > > +endif
> > > +
> > > ifeq ($(CONFIG_RTE_LIBRTE_VHOST_NUMA),y)
> > > LDLIBS += -lnuma
> > > endif
> > > @@ -40,6 +47,12 @@ LDLIBS += -lrte_eal -lrte_mempool -lrte_mbuf -
> > lrte_ethdev -lrte_net
> > > SRCS-$(CONFIG_RTE_LIBRTE_VHOST) := fd_man.c iotlb.c socket.c vhost.c
> \
> > > vhost_user.c virtio_net.c vdpa.c
> > >
> > > +ifeq ($(CC_AVX512_SUPPORT), 1)
> > > +CFLAGS += -DCC_AVX512_SUPPORT
> > > +SRCS-$(CONFIG_RTE_LIBRTE_VHOST) += vhost_vec_avx.c
> > > +CFLAGS_vhost_vec_avx.o += -mavx512f -mavx512bw -mavx512vl
> > > +endif
> > > +
> > > # install includes
> > > SYMLINK-$(CONFIG_RTE_LIBRTE_VHOST)-include += rte_vhost.h
> > rte_vdpa.h \
> > > rte_vdpa_dev.h
> > rte_vhost_async.h
> > > diff --git a/lib/librte_vhost/meson.build b/lib/librte_vhost/meson.build
> > > index cc9aa65c67..c1481802d7 100644
> > > --- a/lib/librte_vhost/meson.build
> > > +++ b/lib/librte_vhost/meson.build
> > > @@ -8,6 +8,22 @@ endif
> > > if has_libnuma == 1
> > > dpdk_conf.set10('RTE_LIBRTE_VHOST_NUMA', true)
> > > endif
> > > +
> > > +if arch_subdir == 'x86'
> > > + if not machine_args.contains('-mno-avx512f')
> > > + if cc.has_argument('-mavx512f') and cc.has_argument('-
> > mavx512vl') and cc.has_argument('-mavx512bw')
> > > + cflags += ['-DCC_AVX512_SUPPORT']
> > > + vhost_avx512_lib =
> > > static_library('vhost_avx512_lib',
> > > + 'vhost_vec_avx.c',
> > > + dependencies:
> > > [static_rte_eal,
> > static_rte_mempool,
> > > + static_rte_mbuf,
> > > static_rte_ethdev,
> > static_rte_net],
> > > + include_directories:
> > > includes,
> > > + c_args: [cflags,
> > > '-mavx512f', '-mavx512bw', '-
> > mavx512vl'])
> > > + objs +=
> vhost_avx512_lib.extract_objects('vhost_vec_avx.c')
> > > + endif
> > > + endif
> > > +endif
> > > +
> > > if (toolchain == 'gcc' and cc.version().version_compare('>=8.3.0'))
> > > cflags += '-DVHOST_GCC_UNROLL_PRAGMA'
> > > elif (toolchain == 'clang' and cc.version().version_compare('>=3.7.0'))
> > > diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
> > > index 4a81f18f01..fc7daf2145 100644
> > > --- a/lib/librte_vhost/vhost.h
> > > +++ b/lib/librte_vhost/vhost.h
> > > @@ -1124,4 +1124,12 @@ virtio_dev_pktmbuf_alloc(struct virtio_net
> > *dev, struct rte_mempool *mp,
> > > return NULL;
> > > }
> > >
> > > +int
> > > +vhost_reserve_avail_batch_packed_avx(struct virtio_net *dev,
> > > + struct vhost_virtqueue *vq,
> > > + struct rte_mempool *mbuf_pool,
> > > + struct rte_mbuf **pkts,
> > > + uint16_t avail_idx,
> > > + uintptr_t *desc_addrs,
> > > + uint16_t *ids);
> > > #endif /* _VHOST_NET_CDEV_H_ */
> > > diff --git a/lib/librte_vhost/vhost_vec_avx.c
> > b/lib/librte_vhost/vhost_vec_avx.c
> > > new file mode 100644
> > > index 0000000000..e8361d18fa
> > > --- /dev/null
> > > +++ b/lib/librte_vhost/vhost_vec_avx.c
> > > @@ -0,0 +1,152 @@
> > > +/* SPDX-License-Identifier: BSD-3-Clause
> > > + * Copyright(c) 2010-2016 Intel Corporation
> > > + */
> > > +#include <stdint.h>
> > > +
> > > +#include "vhost.h"
> > > +
> > > +#define BYTE_SIZE 8
> > > +/* reference count offset in mbuf rearm data */
> > > +#define REFCNT_BITS_OFFSET ((offsetof(struct rte_mbuf, refcnt) - \
> > > + offsetof(struct rte_mbuf, rearm_data)) * BYTE_SIZE)
> > > +/* segment number offset in mbuf rearm data */
> > > +#define SEG_NUM_BITS_OFFSET ((offsetof(struct rte_mbuf, nb_segs) - \
> > > + offsetof(struct rte_mbuf, rearm_data)) * BYTE_SIZE)
> > > +
> > > +/* default rearm data */
> > > +#define DEFAULT_REARM_DATA (1ULL << SEG_NUM_BITS_OFFSET | \
> > > + 1ULL << REFCNT_BITS_OFFSET)
> > > +
> > > +#define DESC_FLAGS_SHORT_OFFSET (offsetof(struct
> vring_packed_desc,
> > flags) / \
> > > + sizeof(uint16_t))
> > > +
> > > +#define DESC_FLAGS_SHORT_SIZE (sizeof(struct vring_packed_desc) / \
> > > + sizeof(uint16_t))
> > > +#define BATCH_FLAGS_MASK (1 << DESC_FLAGS_SHORT_OFFSET | \
> > > + 1 << (DESC_FLAGS_SHORT_OFFSET + DESC_FLAGS_SHORT_SIZE) | \
> > > + 1 << (DESC_FLAGS_SHORT_OFFSET + DESC_FLAGS_SHORT_SIZE * 2) |
> > \
> > > + 1 << (DESC_FLAGS_SHORT_OFFSET + DESC_FLAGS_SHORT_SIZE * 3))
> > > +
> > > +#define FLAGS_BITS_OFFSET ((offsetof(struct vring_packed_desc, flags) -
> \
> > > + offsetof(struct vring_packed_desc, len)) * BYTE_SIZE)
> > > +
> > > +#define PACKED_FLAGS_MASK ((0ULL | VRING_DESC_F_AVAIL |
> > VRING_DESC_F_USED) \
> > > + << FLAGS_BITS_OFFSET)
> > > +#define PACKED_AVAIL_FLAG ((0ULL | VRING_DESC_F_AVAIL) <<
> > FLAGS_BITS_OFFSET)
> > > +#define PACKED_AVAIL_FLAG_WRAP ((0ULL | VRING_DESC_F_USED) <<
> \
> > > + FLAGS_BITS_OFFSET)
> > > +
> > > +#define DESC_FLAGS_POS 0xaa
> > > +#define MBUF_LENS_POS 0x6666
> > > +
> > > +int
> > > +vhost_reserve_avail_batch_packed_avx(struct virtio_net *dev,
> > > + struct vhost_virtqueue *vq,
> > > + struct rte_mempool *mbuf_pool,
> > > + struct rte_mbuf **pkts,
> > > + uint16_t avail_idx,
> > > + uintptr_t *desc_addrs,
> > > + uint16_t *ids)
> > > +{
> > > + struct vring_packed_desc *descs = vq->desc_packed;
> > > + uint32_t descs_status;
> > > + void *desc_addr;
> > > + uint16_t i;
> > > + uint8_t cmp_low, cmp_high, cmp_result;
> > > + uint64_t lens[PACKED_BATCH_SIZE];
> > > +
> > > + if (unlikely(avail_idx & PACKED_BATCH_MASK))
> > > + return -1;
> > > +
> > > + /* load 4 descs */
> > > + desc_addr = &vq->desc_packed[avail_idx];
> > > + __m512i desc_vec = _mm512_loadu_si512(desc_addr);
> > > +
> > > + /* burst check four status */
> > > + __m512i avail_flag_vec;
> > > + if (vq->avail_wrap_counter)
> > > +#if defined(RTE_ARCH_I686)
> > > + avail_flag_vec = _mm512_set4_epi64(PACKED_AVAIL_FLAG,
> > 0x0,
> > > + PACKED_FLAGS_MASK, 0x0);
> > > +#else
> > > + avail_flag_vec =
> > _mm512_maskz_set1_epi64(DESC_FLAGS_POS,
> > > + PACKED_AVAIL_FLAG);
> > > +
> > > +#endif
> > > + else
> > > +#if defined(RTE_ARCH_I686)
> > > + avail_flag_vec =
> > _mm512_set4_epi64(PACKED_AVAIL_FLAG_WRAP,
> > > + 0x0, PACKED_AVAIL_FLAG_WRAP,
> > 0x0);
> > > +#else
> > > + avail_flag_vec =
> > _mm512_maskz_set1_epi64(DESC_FLAGS_POS,
> > > + PACKED_AVAIL_FLAG_WRAP);
> > > +#endif
> > > +
> > > + descs_status = _mm512_cmp_epu16_mask(desc_vec, avail_flag_vec,
> > > + _MM_CMPINT_NE);
> > > + if (descs_status & BATCH_FLAGS_MASK)
> > > + return -1;
> > > +
> > > + /* check buffer fit into one region & translate address */
> > > + __m512i regions_low_addrs =
> > > + _mm512_loadu_si512((void *)&dev->regions_low_addrs);
> > > + __m512i regions_high_addrs =
> > > + _mm512_loadu_si512((void *)&dev->regions_high_addrs);
> > > + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) {
> > > + uint64_t addr_low = descs[avail_idx + i].addr;
> > > + uint64_t addr_high = addr_low + descs[avail_idx + i].len;
> > > + __m512i low_addr_vec = _mm512_set1_epi64(addr_low);
> > > + __m512i high_addr_vec = _mm512_set1_epi64(addr_high);
> > > +
> > > + cmp_low = _mm512_cmp_epi64_mask(low_addr_vec,
> > > + regions_low_addrs, _MM_CMPINT_NLT);
> > > + cmp_high = _mm512_cmp_epi64_mask(high_addr_vec,
> > > + regions_high_addrs, _MM_CMPINT_LT);
> > > + cmp_result = cmp_low & cmp_high;
> > > + int index = __builtin_ctz(cmp_result);
> > > + if (unlikely((uint32_t)index >= dev->mem->nregions))
> > > + goto free_buf;
> > > +
> > > + desc_addrs[i] = addr_low +
> > > + dev->mem->regions[index].host_user_addr -
> > > + dev->mem->regions[index].guest_phys_addr;
> > > + lens[i] = descs[avail_idx + i].len;
> > > + rte_prefetch0((void *)(uintptr_t)desc_addrs[i]);
> > > +
> > > + pkts[i] = virtio_dev_pktmbuf_alloc(dev, mbuf_pool, lens[i]);
> > > + if (!pkts[i])
> > > + goto free_buf;
> > > + }
> >
> > The above does not support vIOMMU, isn't it?
> >
> > The more the packed datapath evolves, the more it gets optimized for a
> > very specific configuration.
> >
> > In v19.11, indirect descriptors and chained buffers are handled as a
> > fallback. And now vIOMMU support is handled as a fallback.
> >
>
> Hi Maxime,
> Thanks for figuring out the feature miss. First version patch is lack of
> vIOMMU supporting.
> V2 patch will fix the feature gap between vectorized function and original
> batch function.
> So there will be no additional fallback introduced in vectorized patch set.
>
> IMHO, current packed optimization introduced complexity is for handling
> that gap between performance aimed frontend (like PMD) and normal
> network traffic (like TCP).
> Vectorized datapath is focusing in enhancing the performance of batched
> function. From function point of view, there will no difference between
> vectorized batched function and original batched function.
> Current packed ring path will remain the same if vectorized option is not
> enable. So I think the complexity won't increase too much. If there's any
> concern, please let me known.
>
> BTW, vectorized path can help performance a lot when vIOMMU enabled.
>
After double check, most performance difference came from runtime setting.
Performance gain of vectorized path is not so obvious.
> Regards,
> Marvin
>
> > I personnally don't like the path it is taking as it is adding a lot of
> > complexity on top of that.
> >
