[dm-devel] use regular gendisk registration in device mapper

[Mike Snitzer]

On Tue, Jul 27 2021 at  4:38P -0400,
Milan Broz <gmazyland at gmail.com> wrote:

> On 27/07/2021 18:18, Mike Snitzer wrote:
> > On Tue, Jul 27 2021 at 12:02P -0400,
> > Christoph Hellwig <hch at lst.de> wrote:
> > 
> >> On Tue, Jul 27, 2021 at 11:58:46AM -0400, Mike Snitzer wrote:
> >>>> This did not make a different to my testing
> >>>> using dmsetup and the lvm2 tools.
> >>>
> >>> I'll try these changes running through the lvm2 testsuite.
> >>
> >> Btw, is ther documentation on how to run it somewhere?  I noticed
> >> tests, old-tests and unit-tests directories, but no obvious way
> >> to run them.
> > 
> > I haven't tracked how it has changed in a while, but I always run:
> > make check_local
> > 
> > (but to do that you first need to ./configure how your distro does
> > it... so that all targets are enabled, etc. Then: make).
> > 
> > Will revisit this shortly and let you know if my process needed to
> > change at all due to lvm2 changes.
> 
> BTW it would be also nice to run cryptsetup testsuite as root - we do a lot
> of DM operations there (and we depend on sysfs on some places).
> 
> You can just run configure, make and then make check.

Once I installed all deps, I got all but one passing with Christoph's changes:

Block_size: 512, Data_size: 256000B, FEC_roots: 9, Corrupted_bytes: 4 [no-superblock][one_device_test]Usage: lt-veritysetup [-?Vv] [-?|--help] [--usage] [-V|--version]
        [--cancel-deferred] [--check-at-most-once] [--data-block-size=bytes]
        [--data-blocks=blocks] [--debug] [--deferred] [--fec-device=path]
        [--fec-offset=bytes] [--fec-roots=bytes] [--format=number]
        [-h|--hash string] [--hash-block-size=bytes] [--hash-offset=bytes]
        [--ignore-corruption] [--ignore-zero-blocks] [--no-superblock]
        [--panic-on-corruption] [--restart-on-corruption]
        [--root-hash-file=STRING] [--root-hash-signature=STRING]
        [-s|--salt hex string] [--uuid=STRING] [-v|--verbose]
        [OPTION...] <action> <action-specific>
-s=e48da609055204e89ae53b655ca2216dd983cf3cb829f34f63a297d106d53e2d: unknown option
[N/A, test skipped]
FEC repair failed
FAILED backtrace:
500 ./verity-compat-test
FAIL: verity-compat-test

Seems like a test bug.

Mike