Skip to contentSkip to conversation

LTP release preprations

LTP Cyril Hrubis 5 participants 21 messages Jan 15 – Jan 30

lore
Share & tools
lore
#0 lore raw

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

Signature · 5 lines
Cyril Hrubis
[email protected]

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#1 lore raw
Quoted 3 lines · Cyril Hrubis

It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

I hope to promote the standardization of the LTP API namespace. After some discussion, we still haven't found a satisfactory prefix. e.g.

tst_: core LTP API (lib/)
etst_: non‑core extended library in (libs/)
otst_: legacy/compat library (no new features)

But now, I'm also considering using tst_*, est_*, and ost_* directly.

That would great if we can finalize this before this release.

Quoted 2 lines

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

And, I will start the pre-release testing (against CentOS) soon.

--
Regards,
Li Wang

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#2 lore raw

Hi!

Quoted 5 lines

tst_: core LTP API (lib/)
etst_: non‑core extended library in (libs/)
otst_: legacy/compat library (no new features)

But now, I'm also considering using tst_*, est_*, and ost_* directly.

Maybe tse_ (test extended) and tso_ (test old).

Quoted 1 line

That would great if we can finalize this before this release.

I do not have a strong opinion here, whatever is short enough would do.

Signature · 5 lines
Cyril Hrubis
[email protected]

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#3 lore raw

Hi Cyril,

tse_ and tso_ sound good. Short and distinct enough. I will start to prepare my patch V2 based on this.

Signature · 2 lines
Regards,
Li Wang
Quoted 18 lines · Cyril Hrubis

Hi!

Quoted 5 lines

tst_: core LTP API (lib/)
etst_: non‑core extended library in (libs/)
otst_: legacy/compat library (no new features)

But now, I'm also considering using tst_*, est_*, and ost_* directly.

Maybe tse_ (test extended) and tso_ (test old).

Quoted 1 line

That would great if we can finalize this before this release.

I do not have a strong opinion here, whatever is short enough would do.

--
Cyril Hrubis
[email protected]

Signature · 6 lines
Regards,
Li Wang


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#4 lore raw
In reply to Cyril Hrubis

Hi Cyril, all,

Quoted 7 lines

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

I plan to merge today Martin's fix for lio_listio/2-1.c:
https://lore.kernel.org/ltp/[email protected]/

I hope we could merge userfaultfd rewrite:
https://github.com/linux-test-project/ltp/pull/1280

Hopefully v7 "new cmd support option for needs_cmds" is ready now:
https://lore.kernel.org/ltp/[email protected]/

ioctl_pidfd02-06: Add CONFIG_USER_NS and CONFIG_PID_NS to needs_kconfigs
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/
(I somehow agree with Jan that sometimes just check /proc or /sys path is better than imply kconfig, but majority agreed to use .needs_kconfigs).

Maybe fix clone10.c on 32 bit, if we find a root cause. This is just an investigation:
https://lore.kernel.org/ltp/[email protected]/

We got no further response for some RPC fix, I'm not sure if it's ready:
https://lore.kernel.org/ltp/[email protected]/

If we have time:
tst_filesystems01.c: Add test for .filesystems
https://patchwork.ozlabs.org/project/ltp/list/?series=481881&state=*

checkpatch.pl update
https://patchwork.ozlabs.org/project/ltp/list/?series=487179&state=*

Kind regards,
Petr

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#5 lore raw
Quoted 20 lines · Petr Vorel

Hi Cyril, all,

Quoted 7 lines

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

I plan to merge today Martin's fix for lio_listio/2-1.c:
https://lore.kernel.org/ltp/[email protected]/

I hope we could merge userfaultfd rewrite:
https://github.com/linux-test-project/ltp/pull/1280

Hopefully v7 "new cmd support option for needs_cmds" is ready now:
https://lore.kernel.org/ltp/[email protected]/

Quoted 8 lines

ioctl_pidfd02-06: Add CONFIG_USER_NS and CONFIG_PID_NS to needs_kconfigs
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/
(I somehow agree with Jan that sometimes just check /proc or /sys path is better than imply kconfig, but majority agreed to use .needs_kconfigs).

Maybe fix clone10.c on 32 bit, if we find a root cause. This is just an investigation:
https://lore.kernel.org/ltp/[email protected]/

CLONE_SETTLS expects different parameters on x86 vs x86_64, I'd skip 32bit temporarily until it's resolved.

Quoted 17 lines

We got no further response for some RPC fix, I'm not sure if it's ready:
https://lore.kernel.org/ltp/[email protected]/

If we have time:
tst_filesystems01.c: Add test for .filesystems
https://patchwork.ozlabs.org/project/ltp/list/?series=481881&state=*

checkpatch.pl update
https://patchwork.ozlabs.org/project/ltp/list/?series=487179&state=*

Kind regards,
Petr

--
Mailing list info: https://lists.linux.it/listinfo/ltp

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#9 lore raw
In reply to Cyril Hrubis
Quoted 4 lines · Cyril Hrubis

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

Hello,
I'm finishing a partial revert of the ioctl_sg01 USB workaround which I'd like to get merged before release. I'll try to send the patch today.

Signature · 10 lines
Martin Doucha   [email protected]
SW Quality Engineer
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#12 lore raw
Quoted 3 lines

Hi!
Are there any other fixes that should be considered before the git is tagged?

* ioctl_pidfd02-06: Add CONFIG_USER_NS and CONFIG_PID_NS to needs_kconfigs
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/
(Or I can prepare an alternative fix to do a runtime check, I also plan to send a patch with formalising the rules)

* tst_net.sh fixes would be nice to get merged:
https://patchwork.ozlabs.org/project/ltp/list/?series=489902&state=*
https://lore.kernel.org/ltp/[email protected]/
https://lore.kernel.org/ltp/[email protected]/

https://patchwork.ozlabs.org/project/ltp/list/?series=489469&state=*

It should be safe to merge these:

* v5 tst_filesystems01.c: Add test for .filesystems
https://patchwork.ozlabs.org/project/ltp/list/?series=489469&state=*

* checkpatch.pl enhancement to check for typos
https://patchwork.ozlabs.org/project/ltp/list/?series=487179&state=*
(I did not merged it as I hoped for a discussion what else to enable for warnings, but that could be done as a separate effort after the release).

FYI I dared to merge IMA fix:
https://github.com/linux-test-project/ltp/commit/1ba58805ea5466df5202ff2e7b5b3d0cd5c81c37

Kind regards,
Petr

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#13 lore raw
Quoted 4 lines · Petr Vorel

* ioctl_pidfd02-06: Add CONFIG_USER_NS and CONFIG_PID_NS to needs_kconfigs
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/
(Or I can prepare an alternative fix to do a runtime check, I also plan to send a patch with formalising the rules)

The original patch should be safe to be merged.

Because we can disable or limit the creation of new namespaces only by turning off the Kconfig options:

- CONFIG_USER_NS
- CONFIG_PID_NS

If those options are unset, the namespace type simply doesn’t exist (no /proc/self/ns/user).

If they’re set, there isn’t a runtime knob (sysctl, proc, or sysfs) that can disables them in general-purpose kernels.

Signature · 6 lines
Regards,
Li Wang


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#14 lore raw
Quoted 12 lines
Quoted 4 lines · Petr Vorel

* ioctl_pidfd02-06: Add CONFIG_USER_NS and CONFIG_PID_NS to needs_kconfigs
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/
(Or I can prepare an alternative fix to do a runtime check, I also plan to send a patch with formalising the rules)

The original patch should be safe to be merged.

Because we can disable or limit the creation of new namespaces only by turning off the Kconfig options:

Quoted 8 lines

- CONFIG_USER_NS
- CONFIG_PID_NS

If those options are unset, the namespace type simply doesn’t exist (no /proc/self/ns/user).

If they’re set, there isn’t a runtime knob (sysctl, proc, or sysfs) that can disables them in general-purpose kernels.

Yeah, it's bool, let's merge it.

Kind regards,
Petr

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#15 lore raw
In reply to Cyril Hrubis

Cyril Hrubis <[email protected]> wrote:

Hi!

Quoted 3 lines

Are there any other fixes that should be considered before the git is tagged?

No from my side, and LTP pre-release test looks good on our latest RHEL 9/10 series.

I think with the merge of some extra fixes suggested by Petr, we could make it a release. Better on this Friday.

Signature · 5 lines
Regards,
Li Wang

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#16 lore raw
In reply to Cyril Hrubis
Quoted 3 lines · Cyril Hrubis

Hi!
Are there any other fixes that should be considered before the git is tagged?

Hi,
our pre-release testing has found a failure in userfaultfd03 on kernel v5.14 because the test can't find /dev/userfaultfd. According to kernel git, the file was added in this commit:

https://github.com/torvalds/linux/commit/2d5de004e009add27db76c5cdc9f1f7f7dc087e7

which was first merged into kernel v6.1. But the test has .min_kver="5.11".

Ricardo, how did you choose the minimal kernel version for this test?

I recommend removing the .min_kver attribute and checking for the presence of /dev/userfaultfd instead. If the file is missing and the kernel is older than v6.1, then TCONF, otherwise TBROK.

Signature · 10 lines
Martin Doucha   [email protected]
SW Quality Engineer
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#17 lore raw
Quoted 8 lines · Martin Doucha
Quoted 3 lines · Cyril Hrubis

Hi!
Are there any other fixes that should be considered before the git is tagged?

Hi,
our pre-release testing has found a failure in userfaultfd03 on kernel v5.14 because the test can't find /dev/userfaultfd. According to kernel git, the file was added in this commit:

Quoted 6 lines

https://github.com/torvalds/linux/commit/2d5de004e009add27db76c5cdc9f1f7f7dc087e7

which was first merged into kernel v6.1. But the test has .min_kver="5.11".

Ricardo, how did you choose the minimal kernel version for this test?

https://lwn.net/Articles/897260/

Quoted 2 lines

I recommend removing the .min_kver attribute and checking for the presence of /dev/userfaultfd instead. If the file is missing and the kernel is older than v6.1, then TCONF, otherwise TBROK.

Ok.  Will try to submit a fix ASAP.

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#18 lore raw

Hi!

Quoted 5 lines
Quoted 3 lines

I recommend removing the .min_kver attribute and checking for the presence of /dev/userfaultfd instead. If the file is missing and the kernel is older than v6.1, then TCONF, otherwise TBROK.

Ok.  Will try to submit a fix ASAP.

Please do so till the end of today, so that we can include that in the release.

Signature · 5 lines
Cyril Hrubis
[email protected]

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#19 lore raw
In reply to Cyril Hrubis

Hi!

Quoted 7 lines · Cyril Hrubis

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

There's one more thing to do and it's to push a new version of kirk a minor fix. LTP should update it accordingly.

I didn't see any issue with kirk so far, so we should do this now before the release, WDYT?

Signature · 7 lines
Andrea Cervesato
SUSE QE Automation Engineer Linux
[email protected]


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
#20 lore raw
Quoted 8 lines · Andrea Cervesato via ltp

Hi!

Quoted 5 lines · Cyril Hrubis

Hi!
It's about we start to prepare for the January release. We have roughly a week for getting in all patches that should go in. So if there is anything that should be reviewed please point it out ASAP.

Quoted 6 lines
Quoted 2 lines

As for me I'm going to (hopefully) finalize the ground rules patch and send another version today.

There's one more thing to do and it's to push a new version of kirk a minor fix. LTP should update it accordingly.

+1

Quoted 3 lines

I didn't see any issue with kirk so far, so we should do this now before the release, WDYT?

Yes, better to have that v3.2.1 version together, thanks!

Signature · 5 lines
Regards,
Li Wang

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp