CIFS module build verification system

[Pavel Shilovsky]

Hi,

During the second part of the GSoC program of this year I has been
working on build verification system for CIFS kernel module.

Now, I got the system ready to test module with Cthon and XFS test
suites. Both suites has some tests (that are not working with CIFS)
disabled. I was trying to make it easier to add new test suites
further.

The system is based on bash scripts that can be run from cron. The
scripts automatically pull the recent changes from the predefined git
repository, build the module and run tests against it.

The git repo for build verification system scripts:
http://git.altlinux.org/people/piastry/public/?p=cifs-tests.git;a=summary

The git repo for patched Cthon tests (merged by Steve Dickson into his tree):
http://git.altlinux.org/people/piastry/public/?p=cthon04.git;a=summary

The git repo for patched XFS tests:
http://git.altlinux.org/people/piastry/public/?p=xfstests.git;a=summary

[Steve French]

This looks promising ... note the typo ("NFS" instead of "CIFS" in the
error message in the xfstest patch)

+ cifs)
+ # Don't know how to check an NFS filesystem, yet.

Post by Pavel Shilovsky
Hi,
During the second part of the GSoC program of this year I has been
working on build verification system for CIFS kernel module.
Now, I got the system ready to test module with Cthon and XFS test
suites. Both suites has some tests (that are not working with CIFS)
disabled. I was trying to make it easier to add new test suites
further.
The system is based on bash scripts that can be run from cron. The
scripts automatically pull the recent changes from the predefined git
repository, build the module and run tests against it.
http://git.altlinux.org/people/piastry/public/?p=cifs-tests.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=cthon04.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=xfstests.git;a=summary
--
Best regards,
Pavel Shilovsky.

[David Disseldorp]

Hi Pavel,

Post by Pavel Shilovsky
Hi,
During the second part of the GSoC program of this year I has been
working on build verification system for CIFS kernel module.
Now, I got the system ready to test module with Cthon and XFS test
suites. Both suites has some tests (that are not working with CIFS)
disabled. I was trying to make it easier to add new test suites
further.
The system is based on bash scripts that can be run from cron. The
scripts automatically pull the recent changes from the predefined git
repository, build the module and run tests against it.

Thanks for putting this together!

Post by Pavel Shilovsky
http://git.altlinux.org/people/piastry/public/?p=cifs-tests.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=cthon04.git;a=summary

I haven't looked at your cifs-tests or cthon04 tests yet.

Post by Pavel Shilovsky
http://git.altlinux.org/people/piastry/public/?p=xfstests.git;a=summary

nfs)
export MKFS_OPTIONS=$NFS_MKFS_OPTIONS
;;
+ nfs)
+ export MKFS_OPTIONS=$CIFS_MKFS_OPTIONS
+ ;;

Should be "cifs)".

Would be good to add an auto group to tests/cifs/group, which could be
added to as existing tests are made to pass on cifs.

Also, a brief description in README of how to run against a cifs mount
would be useful. E.g.
http://git.samba.org/?p=ddiss/fstests.git;a=commitdiff;h=b99c5b12799b4db3edcb6430e000db6afb5db9a9

[Pavel Shilovsky]

Post by Pavel Shilovsky
Hi,
During the second part of the GSoC program of this year I has been
working on build verification system for CIFS kernel module.
Now, I got the system ready to test module with Cthon and XFS test
suites. Both suites has some tests (that are not working with CIFS)
disabled. I was trying to make it easier to add new test suites
further.
The system is based on bash scripts that can be run from cron. The
scripts automatically pull the recent changes from the predefined git
repository, build the module and run tests against it.
http://git.altlinux.org/people/piastry/public/?p=cifs-tests.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=cthon04.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=xfstests.git;a=summary

Hi Steve, David,

Thank you for your notes and suggestions!

Do you understand right that I can squash David's commit (started with
"SQUASH") into my patch?

Not sure about group file. CIFS (as NFS) doesn't have it's own tests
now. Do you mean to add "test/cifs" directory with an empty (except a
commented header) group file inside (to let us add cifs specific tests
in future)?

[David Disseldorp]

On Wed, 20 Aug 2014 11:32:59 +0400, Pavel Shilovsky wrote:

...

Post by Pavel Shilovsky

Post by Pavel Shilovsky
http://git.altlinux.org/people/piastry/public/?p=cifs-tests.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=cthon04.git;a=summary
http://git.altlinux.org/people/piastry/public/?p=xfstests.git;a=summary

Hi Steve, David,
Thank you for your notes and suggestions!
Do you understand right that I can squash David's commit (started with
"SQUASH") into my patch?

Sure, go for it.

Post by Pavel Shilovsky
Not sure about group file. CIFS (as NFS) doesn't have it's own tests
now. Do you mean to add "test/cifs" directory with an empty (except a
commented header) group file inside (to let us add cifs specific tests
in future)?

My mistake, I thought the generic tests could be regrouped
per-filesystem. An empty template would be fine.

Cheers, David

[David Disseldorp]

Post by David Disseldorp

Post by Pavel Shilovsky
Do you understand right that I can squash David's commit (started with
"SQUASH") into my patch?

Sure, go for it.

Also, please post your next round to the fstests mailing list
(fstests-***@public.gmane.org) for feedback. I'd like to get this work
upstream.

Cheers, David

[Pavel Shilovsky]

Post by David Disseldorp
=20
=20

Post by David Disseldorp

Do you understand right that I can squash David's commit (started w=

ith

Post by David Disseldorp

Post by David Disseldorp

"SQUASH") into my patch? =20

=20
Sure, go for it.

=20
Also, please post your next round to the fstests mailing list
upstream.

Good idea, thanks. So, I will send two patches: my main patch (with you=
r squash commit inside) and your Readme patch. Also I can squash the Re=
adme one too and post one big patch -- what is more preferable?

Steve, David, do you want me to add your reviewed-by/acked-by to the ma=
in patch?

[David Disseldorp]

Good idea, thanks. So, I will send two patches: my main patch (with your squash commit inside) and your Readme patch. Also I can squash the Readme one too and post one big patch -- what is more preferable?

One patch with the two commits from my repo squashed in should be fine.

Steve, David, do you want me to add your reviewed-by/acked-by to the main patch?

If it's just your original change with the other two patches squashed,
then please add Reviewed-by: David Disseldorp <ddiss-***@public.gmane.org>

Cheers, David