[LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Petr Vorel]

Kernel commit ab89060fbc92e ("pidfs: return -EREMOTE when PIDFD_GET_INFO
is called on another ns") from v7.0 was backported into stable/LTS
kernels v6.18.14 (7446125afb6d9) and v6.19.10 (2920abaf56c09).

Reported-by: Wei Gao <wegao@suse.com>
Signed-off-by: Petr Vorel <pvorel@suse.cz>
---
 testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
index 0f24d4741f..cc66d54cf4 100644
--- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
+++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
@@ -54,7 +54,7 @@ static void setup(void)
 	if (!ioctl_pidfd_info_exit_supported())
 		tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");
 
-	if (tst_kvercmp(7, 0, 0) >= 0)
+	if (tst_kvercmp(6, 18, 14) >= 0)
 		err_nr = EREMOTE;
 }
 
-- 
2.53.0


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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Li Wang]

Hi Petr, Wei,

Petr Vorel wrote:

> Kernel commit ab89060fbc92e ("pidfs: return -EREMOTE when PIDFD_GET_INFO
> is called on another ns") from v7.0 was backported into stable/LTS
> kernels v6.18.14 (7446125afb6d9) and v6.19.10 (2920abaf56c09).

Is this backport just a single stable kernel or a range of version?

> Reported-by: Wei Gao <wegao@suse.com>
> Signed-off-by: Petr Vorel <pvorel@suse.cz>
> ---
>  testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> index 0f24d4741f..cc66d54cf4 100644
> --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> @@ -54,7 +54,7 @@ static void setup(void)
>  	if (!ioctl_pidfd_info_exit_supported())
>  		tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");
>  
> -	if (tst_kvercmp(7, 0, 0) >= 0)
> +	if (tst_kvercmp(6, 18, 14) >= 0)

What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?


-- 
Regards,
Li Wang

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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Petr Vorel]

Hi Li,

[ Discussion => Cc Cyril and Jan ]

> Hi Petr, Wei,

> Petr Vorel wrote:

> > Kernel commit ab89060fbc92e ("pidfs: return -EREMOTE when PIDFD_GET_INFO
> > is called on another ns") from v7.0 was backported into stable/LTS
> > kernels v6.18.14 (7446125afb6d9) and v6.19.10 (2920abaf56c09).

> Is this backport just a single stable kernel or a range of version?

It was backported into 6.18 stable (in v6.18.14) and in 6.19 (in v6.19.10).

> > Reported-by: Wei Gao <wegao@suse.com>
> > Signed-off-by: Petr Vorel <pvorel@suse.cz>
> > ---
> >  testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)

> > diff --git a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > index 0f24d4741f..cc66d54cf4 100644
> > --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > @@ -54,7 +54,7 @@ static void setup(void)
> >  	if (!ioctl_pidfd_info_exit_supported())
> >  		tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");

> > -	if (tst_kvercmp(7, 0, 0) >= 0)
> > +	if (tst_kvercmp(6, 18, 14) >= 0)

> What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?

FYI I haven't tested that, but it should fail on 6.18.0..v6.18.13 and on
v6.19.0..v6.19.9 due missing backport.

I'd assume that's correct, because on stable/LTS we should IMHO be more strict
to ensure it behaves as expected. WDYT?

Kind regards,
Petr

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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Li Wang]

> > > --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > @@ -54,7 +54,7 @@ static void setup(void)
> > >  	if (!ioctl_pidfd_info_exit_supported())
> > >  		tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");
> 
> > > -	if (tst_kvercmp(7, 0, 0) >= 0)
> > > +	if (tst_kvercmp(6, 18, 14) >= 0)
> 
> > What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?
> 
> FYI I haven't tested that, but it should fail on 6.18.0..v6.18.13 and on
> v6.19.0..v6.19.9 due missing backport.
> 
> I'd assume that's correct, because on stable/LTS we should IMHO be more strict
> to ensure it behaves as expected. WDYT?

Hmm, maybe that's acceptable. Or just:

  if ((tst_kvercmp(6, 18, 14) == 0 || 
       tst_kvercmp(6, 19, 10) == 0) ||
       tst_kvercmp(7, 0, 0) >= 0)

And it'd be great to have Jan's comments, he is an expert in
the kernel maintenance area:).

-- 
Regards,
Li Wang

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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Jan Stancek via ltp]

On Fri, Apr 24, 2026 at 9:13 AM Li Wang <li.wang@linux.dev> wrote:
>
> > > > --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > @@ -54,7 +54,7 @@ static void setup(void)
> > > >   if (!ioctl_pidfd_info_exit_supported())
> > > >           tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");
> >
> > > > - if (tst_kvercmp(7, 0, 0) >= 0)
> > > > + if (tst_kvercmp(6, 18, 14) >= 0)
> >
> > > What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?
> >
> > FYI I haven't tested that, but it should fail on 6.18.0..v6.18.13 and on
> > v6.19.0..v6.19.9 due missing backport.
> >
> > I'd assume that's correct, because on stable/LTS we should IMHO be more strict
> > to ensure it behaves as expected. WDYT?
>
> Hmm, maybe that's acceptable. Or just:
>
>   if ((tst_kvercmp(6, 18, 14) == 0 ||
>        tst_kvercmp(6, 19, 10) == 0) ||
>        tst_kvercmp(7, 0, 0) >= 0)
>
> And it'd be great to have Jan's comments, he is an expert in
> the kernel maintenance area:).

ehm, not sure about that :-).

This is change in error code, we didn't treat previous value as error for older
releases, so why do that now? I don't see it covered in man pages (yet),
so why be strict?

I think the patch should either use more specific version ranges or
allow both errnos on versions <= 7.0.0.


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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Wei Gao via ltp]

On Fri, Apr 24, 2026 at 09:33:55AM +0200, Jan Stancek wrote:
> On Fri, Apr 24, 2026 at 9:13 AM Li Wang <li.wang@linux.dev> wrote:
> >
> > > > > --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > > +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > > @@ -54,7 +54,7 @@ static void setup(void)
> > > > >   if (!ioctl_pidfd_info_exit_supported())
> > > > >           tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");
> > >
> > > > > - if (tst_kvercmp(7, 0, 0) >= 0)
> > > > > + if (tst_kvercmp(6, 18, 14) >= 0)
> > >
> > > > What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?
> > >
> > > FYI I haven't tested that, but it should fail on 6.18.0..v6.18.13 and on
> > > v6.19.0..v6.19.9 due missing backport.
> > >
> > > I'd assume that's correct, because on stable/LTS we should IMHO be more strict
> > > to ensure it behaves as expected. WDYT?
> >
> > Hmm, maybe that's acceptable. Or just:
> >
> >   if ((tst_kvercmp(6, 18, 14) == 0 ||
> >        tst_kvercmp(6, 19, 10) == 0) ||
> >        tst_kvercmp(7, 0, 0) >= 0)
> >
> > And it'd be great to have Jan's comments, he is an expert in
> > the kernel maintenance area:).
> 
> ehm, not sure about that :-).
> 
> This is change in error code, we didn't treat previous value as error for older
> releases, so why do that now? I don't see it covered in man pages (yet),
> so why be strict?
> 
> I think the patch should either use more specific version ranges or
> allow both errnos on versions <= 7.0.0.
> 

I agree, i prefer use my patch :)
https://patchwork.ozlabs.org/project/ltp/patch/20260406023538.14739-1-wegao@suse.com/

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

Re: [LTP] ioctl_pidfd06: Update kernel version

[linuxtestproject.agent]

Hi Petr,

On 2026-04-24, Petr Vorel wrote:
> ioctl_pidfd06: Update kernel version

> -	if (tst_kvercmp(7, 0, 0) >= 0)
> +	if (tst_kvercmp(6, 18, 14) >= 0)
>  		err_nr = EREMOTE;

tst_kvercmp(6, 18, 14) >= 0 is too broad: because tst_kvercmp compares
(major, minor, patch) as ordered integers, it evaluates true for all
v6.19.x, including v6.19.0-v6.19.9 which do not carry the backport
(which landed in v6.19.10). Those kernels would see a false EREMOTE
expectation and TFAIL.

The check needs to handle both stable branches independently, e.g.:

	if (tst_kvercmp(7, 0, 0) >= 0 ||
	    (tst_kvercmp(6, 18, 14) >= 0 && tst_kvercmp(6, 19, 0) < 0) ||
	    tst_kvercmp(6, 19, 10) >= 0)
		err_nr = EREMOTE;

Also, since this corrects a bug (TFAIL on kernels with the backport),
please add a Fixes: tag referencing the commit that introduced the
original tst_kvercmp(7, 0, 0) check.

---
Note:

Our agent completed the review of the patch. The full review can be
found at: https://github.com/linux-test-project/ltp-agent/actions/runs/24879016987

The agent can sometimes produce false positives although often its
findings are genuine. If you find issues with the review, please
comment this email or ignore the suggestions.

Regards,
LTP AI Reviewer

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

Re: [LTP] [PATCH 1/1] ioctl_pidfd06: Update kernel version

[Petr Vorel]

Hi all,

first, thanks for a review!

> On Fri, Apr 24, 2026 at 9:13 AM Li Wang <li.wang@linux.dev> wrote:

> > > > > --- a/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > > +++ b/testcases/kernel/syscalls/ioctl/ioctl_pidfd06.c
> > > > > @@ -54,7 +54,7 @@ static void setup(void)
> > > > >   if (!ioctl_pidfd_info_exit_supported())
> > > > >           tst_brk(TCONF, "PIDFD_INFO_EXIT is not supported by ioctl()");

> > > > > - if (tst_kvercmp(7, 0, 0) >= 0)
> > > > > + if (tst_kvercmp(6, 18, 14) >= 0)

> > > > What will happen if this test is run on kernel range: v6.19.0 ~ v6.19.9?

> > > FYI I haven't tested that, but it should fail on 6.18.0..v6.18.13 and on
> > > v6.19.0..v6.19.9 due missing backport.

> > > I'd assume that's correct, because on stable/LTS we should IMHO be more strict
> > > to ensure it behaves as expected. WDYT?

> > Hmm, maybe that's acceptable. Or just:

> >   if ((tst_kvercmp(6, 18, 14) == 0 ||
'==' I guess you mean '>=', this would not work on v6.18.15..v6.18.24

> >        tst_kvercmp(6, 19, 10) == 0) ||
'==': although 6.19.14 is EOL, that might change.

> >        tst_kvercmp(7, 0, 0) >= 0)

Before sending I was thinking about:

if ((tst_kvercmp(6, 19, 10) >= 0) ||
    (tst_kvercmp(6, 18, 14) >= 0 && tst_kvercmp(6, 19, 0) < 0)

Because the only version which should be avoided is v6.19.0..v6.19.9.
I can send v2.

> > And it'd be great to have Jan's comments, he is an expert in
> > the kernel maintenance area:).

> ehm, not sure about that :-).

> This is change in error code, we didn't treat previous value as error for older
> releases, so why do that now? I don't see it covered in man pages (yet),
> so why be strict?

> I think the patch should either use more specific version ranges or
> allow both errnos on versions <= 7.0.0.

FYI Wei's approach was not to be strict:
https://patchwork.ozlabs.org/project/ltp/patch/20260424060543.462475-1-pvorel@suse.cz/

I did not supersede it (somebody did), feel free to ack this approach (and I
will not send v2).

Kind regards,
Petr

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

Re: [LTP] ioctl_pidfd06: Update kernel version

[Petr Vorel]

> Hi Petr,

> On 2026-04-24, Petr Vorel wrote:
> > ioctl_pidfd06: Update kernel version

> > -	if (tst_kvercmp(7, 0, 0) >= 0)
> > +	if (tst_kvercmp(6, 18, 14) >= 0)
> >  		err_nr = EREMOTE;

> tst_kvercmp(6, 18, 14) >= 0 is too broad: because tst_kvercmp compares
> (major, minor, patch) as ordered integers, it evaluates true for all
> v6.19.x, including v6.19.0-v6.19.9 which do not carry the backport
> (which landed in v6.19.10). Those kernels would see a false EREMOTE
> expectation and TFAIL.

> The check needs to handle both stable branches independently, e.g.:

> 	if (tst_kvercmp(7, 0, 0) >= 0 ||
> 	    (tst_kvercmp(6, 18, 14) >= 0 && tst_kvercmp(6, 19, 0) < 0) ||
> 	    tst_kvercmp(6, 19, 10) >= 0)
> 		err_nr = EREMOTE;

Obviously tst_kvercmp(6, 19, 10) >= 0 will match also tst_kvercmp(7, 0, 0) >= 0,
therefore "tst_kvercmp(7, 0, 0) >= 0" is not needed.

Kind regards,
Petr

> Also, since this corrects a bug (TFAIL on kernels with the backport),
> please add a Fixes: tag referencing the commit that introduced the
> original tst_kvercmp(7, 0, 0) check.

Do we want to add Fixes when original commit was correct?

Kind regards,
Petr

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

Re: [LTP] ioctl_pidfd06: Update kernel version

[Petr Vorel]

...
> > Also, since this corrects a bug (TFAIL on kernels with the backport),
> > please add a Fixes: tag referencing the commit that introduced the
> > original tst_kvercmp(7, 0, 0) check.

> Do we want to add Fixes when original commit was correct?

Shouldn't be Follow-up: used?

Kind regards,
Petr

> Kind regards,
> Petr

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