Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

[Joel Fernandes]

[-- Attachment #1: Type: text/plain, Size: 838 bytes --]

Hello,

On Wed, 2 Apr 2025 16:17:06 GMT, Sebastian Andrzej Siewior wrote:
> On 2025-03-31 14:03:06 [-0700], Paul E. McKenney wrote:
> > The torture.sh --do-rt command-line parameter is intended to mimic -rt
> > kernels.  Now that CONFIG_PREEMPT_RT is upstream, this commit makes this
> > mimicking more precise.
> > 
> > Note that testing of RCU priority boosting is disabled in favor
> > of forward-progress testing of RCU callbacks.  If it turns out to be
> > possible to make kernels built with CONFIG_PREEMPT_RT=y to tolerate
> > testing of both, both will be enabled.
> 
> Not sure what you point at here: You can build a PREEMPT_RT kernel and
> RCU boosting is enabled by default. You could disable it if needed.

Yeah, RCU_BOOST has default y if PREEMPT_RT.

Paul, should we be disabling it in the --do-rt script?

thanks,

 - Joel

Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

[Paul E. McKenney]

On Wed, Apr 02, 2025 at 12:19:13PM -0400, Joel Fernandes wrote:
> Hello,
> 
> On Wed, 2 Apr 2025 16:17:06 GMT, Sebastian Andrzej Siewior wrote:
> > On 2025-03-31 14:03:06 [-0700], Paul E. McKenney wrote:
> > > The torture.sh --do-rt command-line parameter is intended to mimic -rt
> > > kernels.  Now that CONFIG_PREEMPT_RT is upstream, this commit makes this
> > > mimicking more precise.
> > > 
> > > Note that testing of RCU priority boosting is disabled in favor
> > > of forward-progress testing of RCU callbacks.  If it turns out to be
> > > possible to make kernels built with CONFIG_PREEMPT_RT=y to tolerate
> > > testing of both, both will be enabled.
> > 
> > Not sure what you point at here: You can build a PREEMPT_RT kernel and
> > RCU boosting is enabled by default. You could disable it if needed.
> 
> Yeah, RCU_BOOST has default y if PREEMPT_RT.
> 
> Paul, should we be disabling it in the --do-rt script?

You should have a "rcutorture.test_boost=0" from f2ac55968df2
("rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT").

Plus I just now made additional adjustments based on Sebastian Siewior's
feedback.  He is likely to have additional feedback, so this is currently
a "squash!" commit in my tree.

							Thanx, Paul

Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

[Joel Fernandes]

On 4/2/2025 3:17 PM, Paul E. McKenney wrote:
> On Wed, Apr 02, 2025 at 12:19:13PM -0400, Joel Fernandes wrote:
>> Hello,
>>
>> On Wed, 2 Apr 2025 16:17:06 GMT, Sebastian Andrzej Siewior wrote:
>>> On 2025-03-31 14:03:06 [-0700], Paul E. McKenney wrote:
>>>> The torture.sh --do-rt command-line parameter is intended to mimic -rt
>>>> kernels.  Now that CONFIG_PREEMPT_RT is upstream, this commit makes this
>>>> mimicking more precise.
>>>>
>>>> Note that testing of RCU priority boosting is disabled in favor
>>>> of forward-progress testing of RCU callbacks.  If it turns out to be
>>>> possible to make kernels built with CONFIG_PREEMPT_RT=y to tolerate
>>>> testing of both, both will be enabled.
>>>
>>> Not sure what you point at here: You can build a PREEMPT_RT kernel and
>>> RCU boosting is enabled by default. You could disable it if needed.
>>
>> Yeah, RCU_BOOST has default y if PREEMPT_RT.
>>
>> Paul, should we be disabling it in the --do-rt script?
> 
> You should have a "rcutorture.test_boost=0" from f2ac55968df2
> ("rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT").
> 
> Plus I just now made additional adjustments based on Sebastian Siewior's
> feedback.  He is likely to have additional feedback, so this is currently
> a "squash!" commit in my tree.
Ok, thanks. If you could repost whenever it is ready to the list, I'll pull it.

Also I'm guessing that checkpatch one does not need any changes, but we've to
add a reminder to revert it in the future.

thanks,

- Joel

Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

[Paul E. McKenney]

On Sat, Apr 05, 2025 at 07:01:13AM -0400, Joel Fernandes wrote:
> 
> 
> On 4/2/2025 3:17 PM, Paul E. McKenney wrote:
> > On Wed, Apr 02, 2025 at 12:19:13PM -0400, Joel Fernandes wrote:
> >> Hello,
> >>
> >> On Wed, 2 Apr 2025 16:17:06 GMT, Sebastian Andrzej Siewior wrote:
> >>> On 2025-03-31 14:03:06 [-0700], Paul E. McKenney wrote:
> >>>> The torture.sh --do-rt command-line parameter is intended to mimic -rt
> >>>> kernels.  Now that CONFIG_PREEMPT_RT is upstream, this commit makes this
> >>>> mimicking more precise.
> >>>>
> >>>> Note that testing of RCU priority boosting is disabled in favor
> >>>> of forward-progress testing of RCU callbacks.  If it turns out to be
> >>>> possible to make kernels built with CONFIG_PREEMPT_RT=y to tolerate
> >>>> testing of both, both will be enabled.
> >>>
> >>> Not sure what you point at here: You can build a PREEMPT_RT kernel and
> >>> RCU boosting is enabled by default. You could disable it if needed.
> >>
> >> Yeah, RCU_BOOST has default y if PREEMPT_RT.
> >>
> >> Paul, should we be disabling it in the --do-rt script?
> > 
> > You should have a "rcutorture.test_boost=0" from f2ac55968df2
> > ("rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT").
> > 
> > Plus I just now made additional adjustments based on Sebastian Siewior's
> > feedback.  He is likely to have additional feedback, so this is currently
> > a "squash!" commit in my tree.
> Ok, thanks. If you could repost whenever it is ready to the list, I'll pull it.

If there are no complaints by Monday, Pacific Time, I will resend it.

> Also I'm guessing that checkpatch one does not need any changes, but we've to
> add a reminder to revert it in the future.

I have added a reminder in my usual todo.txt file.  I have no objection
to it being added to a more visible location.  On possible place would
be the RCU Design Documents gdoc, but I will let you guys choose.

							Thanx, Paul

Re: [v2,04/12] rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT

[Joel Fernandes]

> On Apr 5, 2025, at 1:10 PM, Paul E. McKenney <paulmck@kernel.org> wrote:
> 
> On Sat, Apr 05, 2025 at 07:01:13AM -0400, Joel Fernandes wrote:
>> 
>> 
>>> On 4/2/2025 3:17 PM, Paul E. McKenney wrote:
>>> On Wed, Apr 02, 2025 at 12:19:13PM -0400, Joel Fernandes wrote:
>>>> Hello,
>>>> 
>>>> On Wed, 2 Apr 2025 16:17:06 GMT, Sebastian Andrzej Siewior wrote:
>>>>> On 2025-03-31 14:03:06 [-0700], Paul E. McKenney wrote:
>>>>>> The torture.sh --do-rt command-line parameter is intended to mimic -rt
>>>>>> kernels.  Now that CONFIG_PREEMPT_RT is upstream, this commit makes this
>>>>>> mimicking more precise.
>>>>>> 
>>>>>> Note that testing of RCU priority boosting is disabled in favor
>>>>>> of forward-progress testing of RCU callbacks.  If it turns out to be
>>>>>> possible to make kernels built with CONFIG_PREEMPT_RT=y to tolerate
>>>>>> testing of both, both will be enabled.
>>>>> 
>>>>> Not sure what you point at here: You can build a PREEMPT_RT kernel and
>>>>> RCU boosting is enabled by default. You could disable it if needed.
>>>> 
>>>> Yeah, RCU_BOOST has default y if PREEMPT_RT.
>>>> 
>>>> Paul, should we be disabling it in the --do-rt script?
>>> 
>>> You should have a "rcutorture.test_boost=0" from f2ac55968df2
>>> ("rcutorture: Make torture.sh --do-rt use CONFIG_PREEMPT_RT").
>>> 
>>> Plus I just now made additional adjustments based on Sebastian Siewior's
>>> feedback.  He is likely to have additional feedback, so this is currently
>>> a "squash!" commit in my tree.
>> Ok, thanks. If you could repost whenever it is ready to the list, I'll pull it.
> 
> If there are no complaints by Monday, Pacific Time, I will resend it.
> 
>> Also I'm guessing that checkpatch one does not need any changes, but we've to
>> add a reminder to revert it in the future.
> 
> I have added a reminder in my usual todo.txt file.  I have no objection
> to it being added to a more visible location.  On possible place would
> be the RCU Design Documents gdoc, but I will let you guys choose.

Makes sense, I think we should add this and other todo items somewhere - the location you suggest sounds reasonable to me, I've noted it. I am also wondering Boqun - are you working on that PR handbook doc patch? If so, please send for review and inclusion in 6.16 :)

Thanks,

 - Joel 


> 
>                            Thanx, Paul