Skip to contentSkip to conversation

Patch "fsnotify: consistent behavior for parent not watching children" has been added to the 5.15-stable tree

LTP Petr Vorel 2 participants 2 messages Aug 30, 2024 3 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): CAOQ4uxjxSQESOZVCV_yL=uwG5+Y89sYRozVSgHPpaudG4CwZqw@mail.gmail.com

A message that is not in the archive (sent off-list or not yet received): [email protected]

Petr Vorel

author
#3 lore raw

Hi Chuck,

Quoted 21 lines

Hi Amir-

Quoted 11 lines · Amir Goldstein

Please follow up in adjusting the LTP test fanotify09 to update this comment with the appropriate 5.15.y version:

* Test cases #6-#7 are regression tests for commit:
* (from v5.19, unlikely to be backported thus not in .tags):
*
* e730558adffb fanotify: consistent behavior for parent not
watching children

Quoted 7 lines

As well as this hackish test condition:

       if (tc->ignore && tst_kvercmp(5, 19, 0) < 0) {
               tst_res(TCONF, "ignored mask on parent dir has undefined "
                               "behavior on kernel < 5.19");
               return;
       }
Quoted 5 lines

I'm not sure how this check should be altered.

The fix in question now appears in both 5.10.220 and 5.15.154. Which means the commit is in, for instance, 5.10.222 and 5.15.160, but not in 5.13.0 or 5.17.0.

Thank you for noticing that. The comment on the test patch [1] was:

* Test cases #6-#7 are regression tests for commit:
* (from v5.19, unlikely to be backported thus not in .tags):

I see in the end it got backported. Now we have to decide whether it's important enough that we should put commit in the tag and run it unconditionally (preferred way for fixes) or from the latest stable kernel version which got the backport.

AFAIK LTP should not try to hide bugs, report missing fix. And it's up to testers and the projects if they consider the problem as WONTFIX or the kernel they test will backport the fix.

Kind regards,
Petr

[1] https://lore.kernel.org/ltp/[email protected]/

Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp
#4 lore raw
Quoted 27 lines · Petr Vorel

Hi Chuck,

Quoted 3 lines

Hi Amir-

Quoted 1 line

On Mar 20, 2024, at 2:57 AM, Amir Goldstein <[email protected]>

Quoted 9 lines · wrote:
Quoted 8 lines

Please follow up in adjusting the LTP test fanotify09 to update this comment with the appropriate 5.15.y version:

* Test cases #6-#7 are regression tests for commit:
* (from v5.19, unlikely to be backported thus not in .tags):
*
* e730558adffb fanotify: consistent behavior for parent not
watching children

Quoted 4 lines
Quoted 4 lines

As well as this hackish test condition:

if (tc->ignore && tst_kvercmp(5, 19, 0) < 0) {
tst_res(TCONF, "ignored mask on parent dir has

undefined "

Quoted 5 lines
Quoted 3 lines
                               "behavior on kernel < 5.19");
               return;
       }

I'm not sure how this check should be altered.

Quoted 9 lines
Quoted 3 lines

The fix in question now appears in both 5.10.220 and 5.15.154. Which means the commit is in, for instance, 5.10.222 and 5.15.160, but not in 5.13.0 or 5.17.0.

Thank you for noticing that. The comment on the test patch [1] was:

* Test cases #6-#7 are regression tests for commit:
* (from v5.19, unlikely to be backported thus not in .tags):

I suggest to add the fix commit to .tags and change the minimal kernel for testing tc->ignore to v5.10.0

Testing on v5.13 v5.17 or older v5.10 and v5.15 will just see the bug as Petr said and could decide to upgrade to newer LTS.

Still I prefer not to add unneeded noise when testing kernels older that v5.10 but Im fine with whatever you chose to do.

Thanks,
Amir.

Quoted 18 lines

I see in the end it got backported. Now we have to decide whether it's important
enough that we should put commit in the tag and run it unconditionally (preferred way for fixes) or from the latest stable kernel version which got the
backport.

AFAIK LTP should not try to hide bugs, report missing fix. And it's up to testers and the projects if they consider the problem as WONTFIX or the kernel
they test will backport the fix.

Kind regards,
Petr

[1]
https://lore.kernel.org/ltp/[email protected]/

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