All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert (git)" <dgilbert@redhat.com>
To: qemu-devel@nongnu.org, renzhen@linux.alibaba.com,
	eguan@linux.alibaba.com, ganesh.mahalingam@intel.com,
	m.mizuma@jp.fujitsu.com, mszeredi@redhat.com,
	misono.tomohiro@jp.fujitsu.com, tao.peng@linux.alibaba.com,
	piaojun@huawei.com, stefanha@redhat.com, vgoyal@redhat.com,
	mst@redhat.com, berrange@redhat.com
Subject: [PATCH 09/25] virtiofsd: Pass write iov's all the way through
Date: Thu, 24 Oct 2019 12:27:02 +0100	[thread overview]
Message-ID: <20191024112718.34657-10-dgilbert@redhat.com> (raw)
In-Reply-To: <20191024112718.34657-1-dgilbert@redhat.com>

From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>

Pass the write iov pointing to guest RAM all the way through rather
than copying the data.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
 contrib/virtiofsd/fuse_virtio.c | 73 ++++++++++++++++++++++++++++++---
 1 file changed, 67 insertions(+), 6 deletions(-)

diff --git a/contrib/virtiofsd/fuse_virtio.c b/contrib/virtiofsd/fuse_virtio.c
index 68036e2171..98f77d338d 100644
--- a/contrib/virtiofsd/fuse_virtio.c
+++ b/contrib/virtiofsd/fuse_virtio.c
@@ -452,6 +452,10 @@ static void *fv_queue_thread(void *opaque)
                  __func__, qi->qidx, (size_t)evalue, in_bytes, out_bytes);
 
         while (1) {
+            bool allocated_bufv = false;
+            struct fuse_bufvec bufv;
+            struct fuse_bufvec *pbufv;
+
             /*
              * An element contains one request and the space to send our
              * response They're spread over multiple descriptors in a
@@ -493,14 +497,71 @@ static void *fv_queue_thread(void *opaque)
                          __func__, elem->index);
                 assert(0); /* TODO */
             }
-            copy_from_iov(&fbuf, out_num, out_sg);
-            fbuf.size = out_len;
+            /* Copy just the first element and look at it */
+            copy_from_iov(&fbuf, 1, out_sg);
+
+            if (out_num > 2 &&
+                out_sg[0].iov_len == sizeof(struct fuse_in_header) &&
+                ((struct fuse_in_header *)fbuf.mem)->opcode == FUSE_WRITE &&
+                out_sg[1].iov_len == sizeof(struct fuse_write_in)) {
+                /*
+                 * For a write we don't actually need to copy the
+                 * data, we can just do it straight out of guest memory
+                 * but we must sitll copy the headers in case the guest
+                 * was nasty and changed them while we were using them.
+                 */
+                if (se->debug) {
+                    fprintf(stderr, "%s: Write special case\n", __func__);
+                }
+
+                /* copy the fuse_write_in header after the fuse_in_header */
+                fbuf.mem += out_sg->iov_len;
+                copy_from_iov(&fbuf, 1, out_sg + 1);
+                fbuf.mem -= out_sg->iov_len;
+                fbuf.size = out_sg[0].iov_len + out_sg[1].iov_len;
+
+                /* Allocate the bufv, with space for the rest of the iov */
+                allocated_bufv = true;
+                pbufv = malloc(sizeof(struct fuse_bufvec) +
+                               sizeof(struct fuse_buf) * (out_num - 2));
+
+                pbufv->count = 1;
+                pbufv->buf[0] = fbuf;
+
+                size_t iovindex, pbufvindex;
+                iovindex = 2; /* 2 headers, separate iovs */
+                pbufvindex = 1; /* 2 headers, 1 fusebuf */
+
+                for (; iovindex < out_num; iovindex++, pbufvindex++) {
+                    pbufv->count++;
+                    pbufv->buf[pbufvindex].pos = ~0; /* Dummy */
+                    pbufv->buf[pbufvindex].flags = 0;
+                    pbufv->buf[pbufvindex].mem = out_sg[iovindex].iov_base;
+                    pbufv->buf[pbufvindex].size = out_sg[iovindex].iov_len;
+                }
+            } else {
+                /* Normal (non fast write) path */
+
+                /* Copy the rest of the buffer */
+                fbuf.mem += out_sg->iov_len;
+                copy_from_iov(&fbuf, out_num - 1, out_sg + 1);
+                fbuf.mem -= out_sg->iov_len;
+                fbuf.size = out_len;
 
-            /* TODO! Endianness of header */
+                /* TODO! Endianness of header */
 
-            /* TODO: Add checks for fuse_session_exited */
-            struct fuse_bufvec bufv = { .buf[0] = fbuf, .count = 1 };
-            fuse_session_process_buf_int(se, &bufv, &ch);
+                /* TODO: Add checks for fuse_session_exited */
+                bufv.buf[0] = fbuf;
+                bufv.count = 1;
+                pbufv = &bufv;
+            }
+            pbufv->idx = 0;
+            pbufv->off = 0;
+            fuse_session_process_buf_int(se, pbufv, &ch);
+
+            if (allocated_bufv) {
+                free(pbufv);
+            }
 
             if (!qi->reply_sent) {
                 fuse_log(FUSE_LOG_DEBUG, "%s: elem %d no reply sent\n",
-- 
2.23.0



  parent reply	other threads:[~2019-10-24 11:52 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-24 11:26 [PATCH 00/25] virtiofs daemon (security) Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 01/25] virtiofsd: passthrough_ll: create new files in caller's context Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 02/25] virtiofsd: passthrough_ll: add lo_map for ino/fh indirection Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 03/25] virtiofsd: passthrough_ll: add ino_map to hide lo_inode pointers Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 04/25] virtiofsd: passthrough_ll: add dirp_map to hide lo_dirp pointers Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 05/25] virtiofsd: passthrough_ll: add fd_map to hide file descriptors Dr. David Alan Gilbert (git)
2019-10-24 11:26 ` [PATCH 06/25] virtiofsd: passthrough_ll: add fallback for racy ops Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 07/25] virtiofsd: validate path components Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 08/25] virtiofsd: Plumb fuse_bufvec through to do_write_buf Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` Dr. David Alan Gilbert (git) [this message]
2019-10-24 11:27 ` [PATCH 10/25] virtiofsd: add fuse_mbuf_iter API Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 11/25] virtiofsd: validate input buffer sizes in do_write_buf() Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 12/25] virtiofsd: check input buffer size in fuse_lowlevel.c ops Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 13/25] virtiofsd: prevent ".." escape in lo_do_lookup() Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 14/25] virtiofsd: prevent ".." escape in lo_do_readdir() Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 15/25] virtiofsd: use /proc/self/fd/ O_PATH file descriptor Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 16/25] virtiofsd: sandbox mount namespace Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 17/25] virtiofsd: move to an empty network namespace Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 18/25] virtiofsd: move to a new pid namespace Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 19/25] virtiofsd: add seccomp whitelist Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 20/25] virtiofsd: Parse flag FUSE_WRITE_KILL_PRIV Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 21/25] virtiofsd: Drop CAP_FSETID if client asked for it Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 22/25] virtiofsd: set maximum RLIMIT_NOFILE limit Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 23/25] virtiofsd: add security guide document Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 24/25] virtiofsd: add --syslog command-line option Dr. David Alan Gilbert (git)
2019-10-24 11:27 ` [PATCH 25/25] virtiofsd: print log only when priority is high enough Dr. David Alan Gilbert (git)

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=20191024112718.34657-10-dgilbert@redhat.com \
    --to=dgilbert@redhat.com \
    --cc=berrange@redhat.com \
    --cc=eguan@linux.alibaba.com \
    --cc=ganesh.mahalingam@intel.com \
    --cc=m.mizuma@jp.fujitsu.com \
    --cc=misono.tomohiro@jp.fujitsu.com \
    --cc=mst@redhat.com \
    --cc=mszeredi@redhat.com \
    --cc=piaojun@huawei.com \
    --cc=qemu-devel@nongnu.org \
    --cc=renzhen@linux.alibaba.com \
    --cc=stefanha@redhat.com \
    --cc=tao.peng@linux.alibaba.com \
    --cc=vgoyal@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.