hw/9pfs: Add new virtfs option writeout=immediate skip host page cache

writeout=immediate implies the after pwritev we do a sync_file_range.

Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

4 files changed, 48 insertions, 10 deletions

diff --git a/hw/9pfs/virtio-9p-device.c b/hw/9pfs/virtio-9p-device.c
index e5b68dad5a..403eed089e 100644
--- a/hw/9pfs/virtio-9p-device.c
+++ b/hw/9pfs/virtio-9p-device.c
@@ -115,6 +115,7 @@ VirtIODevice *virtio_9p_init(DeviceState *dev, V9fsConf *conf)
         exit(1);
     }
 
+    s->ctx.export_flags = fse->export_flags;
     s->ctx.fs_root = g_strdup(fse->path);
     len = strlen(conf->tag);
     if (len > MAX_TAG_LEN) {
diff --git a/hw/9pfs/virtio-9p-handle.c b/hw/9pfs/virtio-9p-handle.c
index 5c8b5ed471..b248e44916 100644
--- a/hw/9pfs/virtio-9p-handle.c
+++ b/hw/9pfs/virtio-9p-handle.c
@@ -192,16 +192,29 @@ static ssize_t handle_preadv(FsContext *ctx, int fd, const struct iovec *iov,
 static ssize_t handle_pwritev(FsContext *ctx, int fd, const struct iovec *iov,
                               int iovcnt, off_t offset)
 {
+    ssize_t ret;
 #ifdef CONFIG_PREADV
-    return pwritev(fd, iov, iovcnt, offset);
+    ret = pwritev(fd, iov, iovcnt, offset);
 #else
     int err = lseek(fd, offset, SEEK_SET);
     if (err == -1) {
         return err;
     } else {
-        return writev(fd, iov, iovcnt);
+        ret = writev(fd, iov, iovcnt);
     }
 #endif
+#ifdef CONFIG_SYNC_FILE_RANGE
+    if (ret > 0 && ctx->export_flags & V9FS_IMMEDIATE_WRITEOUT) {
+        /*
+         * Initiate a writeback. This is not a data integrity sync.
+         * We want to ensure that we don't leave dirty pages in the cache
+         * after write when writeout=immediate is sepcified.
+         */
+        sync_file_range(fd, offset, ret,
+                        SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE);
+    }
+#endif
+    return ret;
 }
 
 static int handle_chmod(FsContext *fs_ctx, V9fsPath *fs_path, FsCred *credp)
diff --git a/hw/9pfs/virtio-9p-local.c b/hw/9pfs/virtio-9p-local.c
index 9559ff6550..47295b7e27 100644
--- a/hw/9pfs/virtio-9p-local.c
+++ b/hw/9pfs/virtio-9p-local.c
@@ -203,16 +203,30 @@ static ssize_t local_preadv(FsContext *ctx, int fd, const struct iovec *iov,
 static ssize_t local_pwritev(FsContext *ctx, int fd, const struct iovec *iov,
                              int iovcnt, off_t offset)
 {
+    ssize_t ret
+;
 #ifdef CONFIG_PREADV
-    return pwritev(fd, iov, iovcnt, offset);
+    ret = pwritev(fd, iov, iovcnt, offset);
 #else
     int err = lseek(fd, offset, SEEK_SET);
     if (err == -1) {
         return err;
     } else {
-        return writev(fd, iov, iovcnt);
+        ret = writev(fd, iov, iovcnt);
     }
 #endif
+#ifdef CONFIG_SYNC_FILE_RANGE
+    if (ret > 0 && ctx->export_flags & V9FS_IMMEDIATE_WRITEOUT) {
+        /*
+         * Initiate a writeback. This is not a data integrity sync.
+         * We want to ensure that we don't leave dirty pages in the cache
+         * after write when writeout=immediate is sepcified.
+         */
+        sync_file_range(fd, offset, ret,
+                        SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE);
+    }
+#endif
+    return ret;
 }
 
 static int local_chmod(FsContext *fs_ctx, V9fsPath *fs_path, FsCred *credp)
diff --git a/hw/9pfs/virtio-9p.c b/hw/9pfs/virtio-9p.c
index c01c31aa25..3958788248 100644
--- a/hw/9pfs/virtio-9p.c
+++ b/hw/9pfs/virtio-9p.c
@@ -80,6 +80,20 @@ void cred_init(FsCred *credp)
     credp->fc_rdev = -1;
 }
 
+static int get_dotl_openflags(V9fsState *s, int oflags)
+{
+    int flags;
+    /*
+     * Filter the client open flags
+     */
+    flags = oflags & ~(O_NOCTTY | O_ASYNC | O_CREAT);
+    /*
+     * Ignore direct disk access hint until the server supports it.
+     */
+    flags &= ~O_DIRECT;
+    return flags;
+}
+
 void v9fs_string_init(V9fsString *str)
 {
     str->data = NULL;
@@ -1598,10 +1612,7 @@ static void v9fs_open(void *opaque)
         err = offset;
     } else {
         if (s->proto_version == V9FS_PROTO_2000L) {
-            flags = mode;
-            flags &= ~(O_NOCTTY | O_ASYNC | O_CREAT);
-            /* Ignore direct disk access hint until the server supports it. */
-            flags &= ~O_DIRECT;
+            flags = get_dotl_openflags(s, mode);
         } else {
             flags = omode_to_uflags(mode);
         }
@@ -1650,8 +1661,7 @@ static void v9fs_lcreate(void *opaque)
         goto out_nofid;
     }
 
-    /* Ignore direct disk access hint until the server supports it. */
-    flags &= ~O_DIRECT;
+    flags = get_dotl_openflags(pdu->s, flags);
     err = v9fs_co_open2(pdu, fidp, &name, gid,
                         flags | O_CREAT, mode, &stbuf);
     if (err < 0) {