[LTP] Purpose of clock_nanosleep2 tests ?

[LTP] Purpose of clock_nanosleep2 tests ?

[Viresh Kumar]

Hi Cyril,

While working on the time64 variants I stumbled upon
testcases/kernel/syscalls/clock_nanosleep2/.

The commit  log says that we were trying to test clock_nanosleep2()
syscall, which I am unable to find, but still this ends up calling the kernel
variant directly for clock_nanosleep() only.

What am I missing ? Why is this stuff required ?

--
Viresh

[LTP] Purpose of clock_nanosleep2 tests ?

[Cyril Hrubis]

Hi!
> While working on the time64 variants I stumbled upon
> testcases/kernel/syscalls/clock_nanosleep2/.
> 
> The commit  log says that we were trying to test clock_nanosleep2()
> syscall, which I am unable to find, but still this ends up calling the kernel
> variant directly for clock_nanosleep() only.
> 
> What am I missing ? Why is this stuff required ?

Hmm, I guess that there is no clock_nanosleep2() and the test is a
result of a confusion of some kind. Maybe the author just confused
clock_nanosleep(2) with clock_nanosleep2().

Looking at clock_nanosleep() tests, there does not seem to be a test
for ABSTIME so the best solution would be to move the test to
clock_nanosleep directory. What do you think?

-- 
Cyril Hrubis
chrubis@suse.cz

[LTP] Purpose of clock_nanosleep2 tests ?

[Viresh Kumar]

On 21-04-20, 11:20, Cyril Hrubis wrote:
> Hi!
> > While working on the time64 variants I stumbled upon
> > testcases/kernel/syscalls/clock_nanosleep2/.
> > 
> > The commit  log says that we were trying to test clock_nanosleep2()
> > syscall, which I am unable to find, but still this ends up calling the kernel
> > variant directly for clock_nanosleep() only.
> > 
> > What am I missing ? Why is this stuff required ?
> 
> Hmm, I guess that there is no clock_nanosleep2() and the test is a
> result of a confusion of some kind. Maybe the author just confused
> clock_nanosleep(2) with clock_nanosleep2().
> 
> Looking at clock_nanosleep() tests, there does not seem to be a test
> for ABSTIME so the best solution would be to move the test to
> clock_nanosleep directory. What do you think?

clock_nanosleep03.c already tests ABSTIME. I think we can just drop
clock_nanosleep2 directory completely. I will send a patch if that looks okay.

-- 
viresh

[LTP] Purpose of clock_nanosleep2 tests ?

[Cyril Hrubis]

Hi!
> > > While working on the time64 variants I stumbled upon
> > > testcases/kernel/syscalls/clock_nanosleep2/.
> > > 
> > > The commit  log says that we were trying to test clock_nanosleep2()
> > > syscall, which I am unable to find, but still this ends up calling the kernel
> > > variant directly for clock_nanosleep() only.
> > > 
> > > What am I missing ? Why is this stuff required ?
> > 
> > Hmm, I guess that there is no clock_nanosleep2() and the test is a
> > result of a confusion of some kind. Maybe the author just confused
> > clock_nanosleep(2) with clock_nanosleep2().
> > 
> > Looking at clock_nanosleep() tests, there does not seem to be a test
> > for ABSTIME so the best solution would be to move the test to
> > clock_nanosleep directory. What do you think?
> 
> clock_nanosleep03.c already tests ABSTIME. I think we can just drop
> clock_nanosleep2 directory completely. I will send a patch if that looks okay.

clock_nanosleep03 is testing that inside of a time namespace, I think
that there is still value of having the simpler test as well.

-- 
Cyril Hrubis
chrubis@suse.cz

[LTP] Purpose of clock_nanosleep2 tests ?

[Cyril Hrubis]

Hi!
> > clock_nanosleep03.c already tests ABSTIME. I think we can just drop
> > clock_nanosleep2 directory completely. I will send a patch if that looks okay.
> 
> clock_nanosleep03 is testing that inside of a time namespace, I think
> that there is still value of having the simpler test as well.

Moreover the time namespaces does not support CLOCK_REALTIME so we
should implement clock_nanosleep() test with CLOCK_REALTIME and
TIMER_ABSTIME which is not covered at the moment anyways.

-- 
Cyril Hrubis
chrubis@suse.cz