On 06/16/2017 08:58 AM, Max Reitz wrote: > Signed-off-by: Max Reitz <mre...@redhat.com> > --- > In order to pass, this depends on "fix: avoid an infinite loop or a > dangling pointer problem in img_commit" > (http://lists.nongnu.org/archive/html/qemu-block/2017-06/msg00443.html)
Looks like that is now 4172a003 > and on the "block: Don't compare strings in bdrv_reopen_prepare()" > series > (http://lists.nongnu.org/archive/html/qemu-block/2017-06/msg00424.html). Still pending. > --- > tests/qemu-iotests/020 | 27 +++++++++++++++++++++++++++ > tests/qemu-iotests/020.out | 17 +++++++++++++++++ > 2 files changed, 44 insertions(+) > > diff --git a/tests/qemu-iotests/020 b/tests/qemu-iotests/020 > index 7a11110..83d5ef0 100755 > --- a/tests/qemu-iotests/020 > +++ b/tests/qemu-iotests/020 > @@ -110,6 +110,33 @@ for offset in $TEST_OFFSETS; do > io_zero readv $(( offset + 64 * 1024 + 65536 * 4 )) 65536 65536 1 > done > _check_test_img > +_cleanup > +TEST_IMG=$TEST_IMG_SAVE > + > +echo > +echo 'Testing failing commit' > +echo > + > +# Create an image with a null backing file to which committing will fail > (with > +# ENOSPC so we can distinguish the result from some generic EIO which may be > +# generated anywhere in the block layer) > +_make_test_img -b "json:{'driver': 'raw', > + 'file': { > + 'driver': 'blkdebug', > + 'inject-error': [{ > + 'event': 'write_aio', > + 'errno': 28, Still nasty that we encoded an errno value rather than using a stringified enum, but that's not your problem. > + 'once': true > + }], > + 'image': { > + 'driver': 'null-co' > + }}}" The comment does wonders for explaining how it all works! > + > +# Just write anything so comitting will not be a no-op s/comitting/committing/ With the typo fix, Reviewed-by: Eric Blake <ebl...@redhat.com> -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org
signature.asc
Description: OpenPGP digital signature
