[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 026/104] virtiofsd: Fast path for virtio read
From: |
Masayoshi Mizuma |
Subject: |
Re: [PATCH 026/104] virtiofsd: Fast path for virtio read |
Date: |
Fri, 17 Jan 2020 13:54:12 -0500 |
On Thu, Dec 12, 2019 at 04:37:46PM +0000, Dr. David Alan Gilbert (git) wrote:
> From: "Dr. David Alan Gilbert" <address@hidden>
>
> Readv the data straight into the guests buffer.
>
> Signed-off-by: Dr. David Alan Gilbert <address@hidden>
> With fix by:
> Signed-off-by: Eryu Guan <address@hidden>
> ---
> tools/virtiofsd/fuse_lowlevel.c | 5 +
> tools/virtiofsd/fuse_virtio.c | 159 ++++++++++++++++++++++++++++++++
> tools/virtiofsd/fuse_virtio.h | 4 +
> 3 files changed, 168 insertions(+)
>
> diff --git a/tools/virtiofsd/fuse_lowlevel.c b/tools/virtiofsd/fuse_lowlevel.c
> index c2b114cf5b..5f80625652 100644
> --- a/tools/virtiofsd/fuse_lowlevel.c
> +++ b/tools/virtiofsd/fuse_lowlevel.c
> @@ -475,6 +475,11 @@ static int fuse_send_data_iov_fallback(struct
> fuse_session *se,
> return fuse_send_msg(se, ch, iov, iov_count);
> }
>
> + if (fuse_lowlevel_is_virtio(se) && buf->count == 1 &&
> + buf->buf[0].flags == (FUSE_BUF_IS_FD | FUSE_BUF_FD_SEEK)) {
> + return virtio_send_data_iov(se, ch, iov, iov_count, buf, len);
> + }
> +
> abort(); /* Will have taken vhost path */
> return 0;
> }
> diff --git a/tools/virtiofsd/fuse_virtio.c b/tools/virtiofsd/fuse_virtio.c
> index c33e0f7e8c..146cd3f702 100644
> --- a/tools/virtiofsd/fuse_virtio.c
> +++ b/tools/virtiofsd/fuse_virtio.c
> @@ -230,6 +230,165 @@ err:
> return ret;
> }
>
> +/*
> + * Callback from fuse_send_data_iov_* when it's virtio and the buffer
> + * is a single FD with FUSE_BUF_IS_FD | FUSE_BUF_FD_SEEK
> + * We need send the iov and then the buffer.
> + * Return 0 on success
> + */
> +int virtio_send_data_iov(struct fuse_session *se, struct fuse_chan *ch,
> + struct iovec *iov, int count, struct fuse_bufvec
> *buf,
> + size_t len)
> +{
> + int ret = 0;
> + VuVirtqElement *elem;
> + VuVirtq *q;
> +
> + assert(count >= 1);
> + assert(iov[0].iov_len >= sizeof(struct fuse_out_header));
> +
> + struct fuse_out_header *out = iov[0].iov_base;
> + /* TODO: Endianness! */
> +
> + size_t iov_len = iov_size(iov, count);
> + size_t tosend_len = iov_len + len;
> +
> + out->len = tosend_len;
> +
> + fuse_log(FUSE_LOG_DEBUG, "%s: count=%d len=%zd iov_len=%zd\n", __func__,
> + count, len, iov_len);
> +
> + /* unique == 0 is notification which we don't support */
> + assert(out->unique);
> +
> + /* For virtio we always have ch */
> + assert(ch);
> + assert(!ch->qi->reply_sent);
> + elem = ch->qi->qe;
> + q = &ch->qi->virtio_dev->dev.vq[ch->qi->qidx];
> +
> + /* The 'in' part of the elem is to qemu */
> + unsigned int in_num = elem->in_num;
> + struct iovec *in_sg = elem->in_sg;
> + size_t in_len = iov_size(in_sg, in_num);
> + fuse_log(FUSE_LOG_DEBUG, "%s: elem %d: with %d in desc of length %zd\n",
> + __func__, elem->index, in_num, in_len);
> +
> + /*
> + * The elem should have room for a 'fuse_out_header' (out from fuse)
> + * plus the data based on the len in the header.
> + */
> + if (in_len < sizeof(struct fuse_out_header)) {
> + fuse_log(FUSE_LOG_ERR, "%s: elem %d too short for out_header\n",
> + __func__, elem->index);
> + ret = -E2BIG;
The ret should be positive value, right?
ret = E2BIG;
> + goto err;
> + }
> + if (in_len < tosend_len) {
> + fuse_log(FUSE_LOG_ERR, "%s: elem %d too small for data len %zd\n",
> + __func__, elem->index, tosend_len);
> + ret = -E2BIG;
ret = E2BIG;
> + goto err;
> + }
> +
> + /* TODO: Limit to 'len' */
> +
> + /* First copy the header data from iov->in_sg */
> + copy_iov(iov, count, in_sg, in_num, iov_len);
> +
> + /*
> + * Build a copy of the the in_sg iov so we can skip bits in it,
> + * including changing the offsets
> + */
> + struct iovec *in_sg_cpy = calloc(sizeof(struct iovec), in_num);
assert(in_sg_cpy) should be here? in case calloc() fails...
> + memcpy(in_sg_cpy, in_sg, sizeof(struct iovec) * in_num);
> + /* These get updated as we skip */
> + struct iovec *in_sg_ptr = in_sg_cpy;
> + int in_sg_cpy_count = in_num;
> +
> + /* skip over parts of in_sg that contained the header iov */
> + size_t skip_size = iov_len;
> +
> + size_t in_sg_left = 0;
> + do {
> + while (skip_size != 0 && in_sg_cpy_count) {
> + if (skip_size >= in_sg_ptr[0].iov_len) {
> + skip_size -= in_sg_ptr[0].iov_len;
> + in_sg_ptr++;
> + in_sg_cpy_count--;
> + } else {
> + in_sg_ptr[0].iov_len -= skip_size;
> + in_sg_ptr[0].iov_base += skip_size;
> + break;
> + }
> + }
> +
> + int i;
> + for (i = 0, in_sg_left = 0; i < in_sg_cpy_count; i++) {
> + in_sg_left += in_sg_ptr[i].iov_len;
> + }
> + fuse_log(FUSE_LOG_DEBUG,
> + "%s: after skip skip_size=%zd in_sg_cpy_count=%d "
> + "in_sg_left=%zd\n",
> + __func__, skip_size, in_sg_cpy_count, in_sg_left);
> + ret = preadv(buf->buf[0].fd, in_sg_ptr, in_sg_cpy_count,
> + buf->buf[0].pos);
> +
> + fuse_log(FUSE_LOG_DEBUG, "%s: preadv_res=%d(%m) len=%zd\n",
> + __func__, ret, len);
"%m" should be removed? because it may show the previous errno even if preadv()
is succsess. Like as:
[ID: 00000079] virtio_send_data_iov: after skip skip_size=0 in_sg_cpy_count=1
in_sg_left=65536
[ID: 00000079] virtio_send_data_iov: preadv_res=16000(No such file or
directory) len=65536
Otherwise, looks good to me:
Reviewed-by: Masayoshi Mizuma <address@hidden>
Thanks,
Masa
> + if (ret == -1) {
> + ret = errno;
> + free(in_sg_cpy);
> + goto err;
> + }
> + if (ret < len && ret) {
> + fuse_log(FUSE_LOG_DEBUG, "%s: ret < len\n", __func__);
> + /* Skip over this much next time around */
> + skip_size = ret;
> + buf->buf[0].pos += ret;
> + len -= ret;
> +
> + /* Lets do another read */
> + continue;
> + }
> + if (!ret) {
> + /* EOF case? */
> + fuse_log(FUSE_LOG_DEBUG, "%s: !ret in_sg_left=%zd\n", __func__,
> + in_sg_left);
> + break;
> + }
> + if (ret != len) {
> + fuse_log(FUSE_LOG_DEBUG, "%s: ret!=len\n", __func__);
> + ret = EIO;
> + free(in_sg_cpy);
> + goto err;
> + }
> + in_sg_left -= ret;
> + len -= ret;
> + } while (in_sg_left);
> + free(in_sg_cpy);
> +
> + /* Need to fix out->len on EOF */
> + if (len) {
> + struct fuse_out_header *out_sg = in_sg[0].iov_base;
> +
> + tosend_len -= len;
> + out_sg->len = tosend_len;
> + }
> +
> + ret = 0;
> +
> + vu_queue_push(&se->virtio_dev->dev, q, elem, tosend_len);
> + vu_queue_notify(&se->virtio_dev->dev, q);
> +
> +err:
> + if (ret == 0) {
> + ch->qi->reply_sent = true;
> + }
> +
> + return ret;
> +}
> +
> /* Thread function for individual queues, created when a queue is 'started'
> */
> static void *fv_queue_thread(void *opaque)
> {
> diff --git a/tools/virtiofsd/fuse_virtio.h b/tools/virtiofsd/fuse_virtio.h
> index 135a14875a..cc676b9193 100644
> --- a/tools/virtiofsd/fuse_virtio.h
> +++ b/tools/virtiofsd/fuse_virtio.h
> @@ -26,4 +26,8 @@ int virtio_loop(struct fuse_session *se);
> int virtio_send_msg(struct fuse_session *se, struct fuse_chan *ch,
> struct iovec *iov, int count);
>
> +int virtio_send_data_iov(struct fuse_session *se, struct fuse_chan *ch,
> + struct iovec *iov, int count,
> + struct fuse_bufvec *buf, size_t len);
> +
> #endif
> --
> 2.23.0
>
>
- Re: [PATCH 026/104] virtiofsd: Fast path for virtio read,
Masayoshi Mizuma <=