search for: fuse_fill_dir_t

...en those rubbish values were out of the range allowed for microseconds. --- src/fuse.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/fuse.c b/src/fuse.c index 288c02a..dd4f139 100644 --- a/src/fuse.c +++ b/src/fuse.c @@ -175,6 +175,7 @@ mount_local_readdir (const char *path, void *buf, fuse_fill_dir_t filler, if (ss->val[i].ino >= 0) { struct stat statbuf; + memset (&statbuf, 0, sizeof statbuf); statbuf.st_dev = ss->val[i].dev; statbuf.st_ino = ss->val[i].ino; statbuf.st_mode = ss->val[i].mode; @@ -255,6 +256,7 @@...

It seems that FUSE can invoke flush to make sure the pending changes (e.g. to the attributes) of a file are set. Since a missing flush implementation is handled as if it were returning ENOSYS, this can cause issues later. To overcome this, just provide a stub implementation which does nothing, since we have nothing to do and don't want to have FUSE error out. Furthermore, uncomment the

...; +static char *fuse_options; +static struct fuse_chan *ch; +static struct fuse *fuse; +static struct timespec start_t; +static uint64_t size; + +static int nbdfuse_getattr (const char *path, struct stat *stbuf); +static int nbdfuse_readdir (const char *path, void *buf, + fuse_fill_dir_t filler, + off_t offset, struct fuse_file_info *fi); +static int nbdfuse_open (const char *path, struct fuse_file_info *fi); +static int nbdfuse_read (const char *path, char *buf, + size_t count, off_t offset, + struct fuse_f...

...wever, a google search for 'man fuse_main' pulls up https://man.openbsd.org/fuse_main.3 as its first hit, so I think FUSE has graduated to non-Linux systems, so we may have to revisit this later. > +static int > +nbdfuse_readdir (const char *path, void *buf, > + fuse_fill_dir_t filler, > + off_t offset, struct fuse_file_info *fi) > +{ > + if (strcmp (path, "/") != 0) > + return -ENOENT; > + > + filler (buf, ".", NULL, 0); > + filler (buf, "..", NULL, 0); > + filler (buf, filename, NULL, 0); >...

...); + + for (i = 0; i < r->len; ++i) + statns_to_old_stat (&r->val[i], &ret->val[i]); + + return ret; +} diff --git a/src/fuse.c b/src/fuse.c index 00f9092..08a8784 100644 --- a/src/fuse.c +++ b/src/fuse.c @@ -165,7 +165,7 @@ mount_local_readdir (const char *path, void *buf, fuse_fill_dir_t filler, */ names = malloc ((ents->len + 1) * sizeof (char *)); if (names) { - CLEANUP_FREE_STAT_LIST struct guestfs_stat_list *ss = NULL; + CLEANUP_FREE_STATNS_LIST struct guestfs_statns_list *ss = NULL; CLEANUP_FREE_XATTR_LIST struct guestfs_xattr_list *xattrs = NULL;...

This version doesn't crash or cause hung processes or stuck mountpoints, so that's an improvement. Rich.

This patch is just for review. It enables FUSE support in the API via two new calls, 'guestfs_mount_local' and 'guestfs_umount_local'. FUSE turns out to be very easy to deadlock (necessitating that the machine be rebooted). Running the test from the third patch is usually an effective way to demonstrate this. However I have not yet managed to produce a simple reproducer that

This changes the proposed API slightly. Previously 'mount-local' generating a 'mounted' event when the filesystem was ready, and from the 'mounted' event you had to effectively do a fork. Now, 'mount-local' just initializes the mountpoint and you have to call 'mount-local-run' to enter the FUSE main loop. Between these calls you can do a fork or whatever