Re: [PATCH 1/2] file-posix: Use OFD lock only if the filesystem supports the lock

[Masayoshi Mizuma]

On Wed, Nov 18, 2020 at 03:16:53PM +0000, Daniel P. Berrangé wrote:
> On Thu, Nov 05, 2020 at 11:01:01PM -0500, Masayoshi Mizuma wrote:
> > From: Masayoshi Mizuma <m.mizuma@jp.fujitsu.com>
> > 
> > locking=auto doesn't work if the filesystem doesn't support OFD lock.
> > In that situation, following error happens:
> > 
> >   qemu-system-x86_64: -blockdev 
> > driver=qcow2,node-name=disk,file.driver=file,file.filename=/mnt/guest.qcow2,file.locking=auto:
> >  Failed to lock byte 100
> > 
> > qemu_probe_lock_ops() judges whether qemu can use OFD lock
> > or not with doing fcntl(F_OFD_GETLK) to /dev/null. So the
> > error happens if /dev/null supports OFD lock, but the filesystem
> > doesn't support the lock.
> > 
> > Lock the actual file, not /dev/null, using F_OFD_SETLK and if that
> > fails, then fallback to F_SETLK.
> > 
> > Signed-off-by: Masayoshi Mizuma <m.mizuma@jp.fujitsu.com>
> > ---
> >  block/file-posix.c   |  56 ++++++++--------
> >  include/qemu/osdep.h |   2 +-
> >  util/osdep.c         | 149 ++++++++++++++++++++++++++++---------------
> >  3 files changed, 125 insertions(+), 82 deletions(-)
> 
> 
> > diff --git a/util/osdep.c b/util/osdep.c
> > index 66d01b9160..454e8ef9f4 100644
> > --- a/util/osdep.c
> > +++ b/util/osdep.c
> > @@ -117,9 +117,6 @@ int qemu_mprotect_none(void *addr, size_t size)
> >  
> >  #ifndef _WIN32
> >  
> > -static int fcntl_op_setlk = -1;
> > -static int fcntl_op_getlk = -1;
> > -
> >  /*
> >   * Dups an fd and sets the flags
> >   */
> > @@ -187,68 +184,87 @@ static int qemu_parse_fdset(const char *param)
> >      return qemu_parse_fd(param);
> >  }
> >  
> > -static void qemu_probe_lock_ops(void)
> > +bool qemu_has_ofd_lock(int orig_fd)
> >  {
> > -    if (fcntl_op_setlk == -1) {
> >  #ifdef F_OFD_SETLK
> > -        int fd;
> > -        int ret;
> > -        struct flock fl = {
> > -            .l_whence = SEEK_SET,
> > -            .l_start  = 0,
> > -            .l_len    = 0,
> > -            .l_type   = F_WRLCK,
> > -        };
> > -
> > -        fd = open("/dev/null", O_RDWR);
> > -        if (fd < 0) {
> > +    int fd;
> > +    int ret;
> > +    struct flock fl = {
> > +        .l_whence = SEEK_SET,
> > +        .l_start  = 0,
> > +        .l_len    = 0,
> > +        .l_type   = F_RDLCK,
> > +    };
> > +
> > +    fd = qemu_dup(orig_fd);
> 
> Consider that we're *not* using  OFD locks, and QEMU already
> has 'foo.qcow2' open for an existing disk backend, and it is
> locked.
> 
> Now someone tries to hot-add 'foo.qcow2' for a second disk
> by mistake.  Doing this qemu_dup + qemu_close will cause
> the existing locks to be removed AFAICT.

Thank you for pointing it out. I'll remove this qemu_dup() and
check orig_fd directly.

> 
> > +    if (fd >= 0) {
> > +        ret = fcntl_setfl(fd, O_RDONLY);
> > +        if (ret) {
> >              fprintf(stderr,
> > -                    "Failed to open /dev/null for OFD lock probing: %s\n",
> > -                    strerror(errno));
> > -            fcntl_op_setlk = F_SETLK;
> > -            fcntl_op_getlk = F_GETLK;
> > -            return;
> > -        }
> > -        ret = fcntl(fd, F_OFD_GETLK, &fl);
> > -        close(fd);
> > -        if (!ret) {
> > -            fcntl_op_setlk = F_OFD_SETLK;
> > -            fcntl_op_getlk = F_OFD_GETLK;
> > -        } else {
> > -            fcntl_op_setlk = F_SETLK;
> > -            fcntl_op_getlk = F_GETLK;
> > +                    "Failed to fcntl for OFD lock probing.\n");
> > +            qemu_close(fd);
> > +            return false;
> >          }
> > +    }
> > +
> > +    ret = fcntl(fd, F_OFD_GETLK, &fl);
> > +    qemu_close(fd);
> > +
> > +    if (ret == 0) {
> > +        return true;
> > +    } else {
> > +        return false;
> > +    }
> >  #else
> > -        fcntl_op_setlk = F_SETLK;
> > -        fcntl_op_getlk = F_GETLK;
> > +    return false;
> >  #endif
> > -    }
> >  }
> >  
> > -bool qemu_has_ofd_lock(void)
> > -{
> > -    qemu_probe_lock_ops();
> >  #ifdef F_OFD_SETLK
> > -    return fcntl_op_setlk == F_OFD_SETLK;
> > +static int _qemu_lock_fcntl(int fd, struct flock *fl)
> > +{
> > +    int ret;
> > +    bool ofd_lock = true;
> > +
> > +    do {
> > +        if (ofd_lock) {
> > +            ret = fcntl(fd, F_OFD_SETLK, fl);
> > +            if ((ret == -1) && (errno == EINVAL)) {
> > +                ofd_lock = false;
> > +            }
> > +        }
> > +
> > +        if (!ofd_lock) {
> > +            /* Fallback to POSIX lock */
> > +            ret = fcntl(fd, F_SETLK, fl);
> > +        }
> > +    } while (ret == -1 && errno == EINTR);
> 
> THis loop is confusing to read. I'd suggest creating a
> wrapper
> 
>   qemu_fcntl()
> 
> that does the while (ret == -1 && errno == EINTR) loop,
> so that this locking code can be clearer without the
> loop.

Great idea. I'll make qemu_fcntl().

Thanks!
Masa

> 
> > +
> > +    return ret == -1 ? -errno : 0;
> > +}
> >  #else
> > -    return false;
> > -#endif
> > +static int _qemu_lock_fcntl(int fd, struct flock *fl)
> > +{
> > +    int ret;
> > +
> > +    do {
> > +        ret = fcntl(fd, F_SETLK, fl);
> > +    } while (ret == -1 && errno == EINTR);
> > +
> > +    return ret == -1 ? -errno : 0;
> >  }
> > +#endif
> 
> 
> Regards,
> Daniel
> -- 
> |: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org         -o-            https://fstop138.berrange.com :|
> |: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
>