Re: [Libguestfs] [PATCH v3.1 1/9] uuid: add support to change uuid of btrfs partition

-----Original Message----- From: libguestfs-bounces(a)redhat.com [mailto:libguestfs-bounces@redhat.com] On Behalf Of Pino Toscano Sent: Tuesday, June 30, 2015 12:07 AM To: libguestfs(a)redhat.com Subject: Re: [Libguestfs] [PATCH v3.1 1/9] uuid: add support to change uuid of btrfs partition In data lunedì 29 giugno 2015 02:36:57, Chen, Hanxiao ha scritto: > > > -----Original Message----- > > From: libguestfs-bounces(a)redhat.com [mailto:libguestfs-bounces@redhat.com] On > > Behalf Of Pino Toscano > > Sent: Friday, June 26, 2015 8:12 PM > > To: libguestfs(a)redhat.com > > Subject: Re: [Libguestfs] [PATCH v3.1 1/9] uuid: add support to change uuid of btrfs > > partition > > > > In data venerdì 26 giugno 2015 17:35:36, Chen Hanxiao ha scritto: > > > btrfs-progs v4.1 add support to change uuid of btrfs fs. > > > > > > Signed-off-by: Chen Hanxiao <chenhanxiao(a)cn.fujitsu.com&gt; > > > --- > > > v3.1: fix a typo > > > v3: set errno as ENOTSUP when btrfstune -u is not available > > > v2: put btrfs operation back to daemon/btrfs.c > > > move tests to tests/btrfs > > > > > > daemon/btrfs.c | 60 ++++++++++++++++++++++++++++++++++++++++++ > > > daemon/daemon.h | 1 + > > > daemon/uuids.c | 6 ++--- > > > tests/btrfs/test-btrfs-misc.pl | 17 ++++++++++++ > > > 4 files changed, 80 insertions(+), 4 deletions(-) > > > > > > diff --git a/daemon/btrfs.c b/daemon/btrfs.c > > > index 20e5e6b..a69c512 100644 > > > --- a/daemon/btrfs.c > > > +++ b/daemon/btrfs.c > > > @@ -790,6 +790,44 @@ do_btrfs_device_delete (char *const *devices, const char > > *fs) > > > return 0; > > > } > > > > > > + > > > +/* btrfstune command add two new options > > > + * -U UUID change fsid to UUID > > > + * -u change fsid, use a random one > > > + * since v4.1 > > > + * We could check wheter 'btrfstune' support > > > + * '-u' and '-U UUID' option by checking the output of > > > + * 'btrfstune' command. > > > + */ > > > +static int > > > +test_btrfstune_uuid_opt (void) > > > +{ > > > + static int result = -1; > > > + if (result != -1) > > > + return result; > > > + > > > + CLEANUP_FREE char *err = NULL; > > > + > > > + int r = commandr (NULL, &err, str_btrfstune, NULL); > > > + > > > + if (r == -1) { > > > + reply_with_error ("btrfstune: %s", err); > > > + return -1; > > > + } > > > + > > > + /* FIXME currently btrfstune do not support '--help'. > > > > So --help should be passed anyway to the btrfstune invocation, so with > > newer versions hopefully there will not be complaints about it anymore. > > Fine. > > > > [snip] > > > diff --git a/tests/btrfs/test-btrfs-misc.pl b/tests/btrfs/test-btrfs-misc.pl > > > index 2fe7c59..b47caab 100755 > > > --- a/tests/btrfs/test-btrfs-misc.pl > > > +++ b/tests/btrfs/test-btrfs-misc.pl > > > @@ -20,6 +20,7 @@ > > > > > > use strict; > > > use warnings; > > > +use Errno; > > > > > > use Sys::Guestfs; > > > > > > @@ -47,5 +48,21 @@ my $label = $g->vfs_label ("/dev/sda1"); > > > die "unexpected label: expecting 'newlabel' but got '$label'" > > > unless $label eq "newlabel"; > > > > > > +# Setting btrfs UUID > > > +eval { > > > + $g->set_uuid ("/dev/sda1", "12345678-1234-1234-1234-123456789012"); > > > +}; > > > +# FIXME: ignore ESRCH > > > > What is ESRCH about? > > The problem is every time I got a ESRCH here. > I have no idea about this. > > Can we just ignore it by: > } elsif ($err == Errno::ESRCH()) { > ; > } else { > die $@; > } No, please investigate why you are getting such errno. One guess is that a leftover from a failing operation, and that set_uuid actually succeeded. In that case, checking first for $@ might do the trick: eval { ... } if ($@) { # failed, check last_errno } else { # good, compare with vfs_uuid }

> > > +my $err = $g->last_errno (); > > > + > > > +if ($err == 0) { > > > + my $uuid = $g->vfs_uuid ("/dev/sda1"); > > > + die "unexpected uuid expecting > > > + '12345678-1234-1234-1234-123456789012' but got '$uuid'" > > > + unless $uuid eq "12345678-1234-1234-1234-123456789012"; > > > +} elsif ($err == Errno::ENOTSUP()) { > > > + warn "$0: skipping test for btrfs UUID change feature is not available\n"; > > > +} > > > > IIRC 'warn' prints the newline at the end already. > > > > This is missing to check when $err is a valid error. > > > > Can we just: > } else { > die $@; > } Yes, that should be fine. Thanks, -- Pino Toscano _______________________________________________ Libguestfs mailing list Libguestfs(a)redhat.com https://www.redhat.com/mailman/listinfo/libguestfs