On Tue, Jun 03, 2025 at 04:51:00PM -0400, Joel Fernandes wrote:
> Hi Boqun and Paul,
>
> On 6/3/2025 4:44 PM, Boqun Feng wrote:
> > On Tue, Jun 03, 2025 at 01:01:19PM -0700, Paul E. McKenney wrote:
> >> Hello!
> >>
> >> From what I can see, when I start testing v6.16-rc1, I will need to pull
> >> in the following commits and patches into my pile in my -rcu "dev" branch:
> >>
> >> 6c59a7c789f0 ("rcutorture: Add tests for SRCU up/down reader primitives")
> >> 0a67dd132d1d ("rcutorture: Pull rcu_torture_updown() loop body into new
> >> function")
> >> c27d0d38f2ca ("rcutorture: Complain if an ->up_read() is delayed more than
> >> 10 seconds")
> >> 3b4863a42139 ("rcutorture: Check for ->up_read() without matching
> >> ->down_read()")
>
> Paul and this one too? I dropped it from both my PR and my rcu/dev branch:
> Subject: [PATCH 5/5] rcutorture: Fix rcutorture_one_extend_check() splat in RT
Good catch, but I do have that one in my -rcu "dev" branch. Mostly
because I was lazy and never rebased on top of your final pull request.
Which reminds me, it is time to get Joel shared-RCU privileges if he
does not already have them, correct?
> >> The above commits must be ordered after this -rcu "dev" commit:
> >>
> >> db950fccd45a ("rcutorture: Make rcutorture_one_extend_check() account for
> >> hardirq")
> >> Tested-by: kernel test robot <[email protected]>
> >>
> >> There might be a few conflicts to resolve. ;-)
> >>
> >> I need to drop the following commits because they are already in mainline
> >> (thank you, Joel!):
> >>
> >> 604d70035eaf ("torture: Check for "Call trace:" as well as "Call Trace:")
> >> 2e5158390089 ("rcutorture: Reduce TREE01 CPU overcommit")
> >> 90c0254a6d48 ("rcutorture: Remove MAXSMP and CPUMASK_OFFSTACK from TREE01")
> >>
> >> Anything else?
> >>
> >
> > Hmm... we didn't sync Joel's rcu/dev to the rcu repo, did we? I believe
> > Joel had a few patches queued in the rcu/dev of his repo:
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/jfern/linux.git rcu/dev
> >
> > , which should be handed over to the next guy.
> >
> > These are the commmits:
> >
> > ac05b397cb2b ("rcutorture: Start rcu_torture_writer() after
> > rcu_torture_reader()")
> > 006bc5216bd0 ("rcutorture: Print only one rtort_pipe_count splat")
> > af97b7685824 ("rcutorture: Check for no up/down readers at task level")
> > 2baf873c6c2b ("torture: Suppress torture.sh "Zero time" messages for
> > disabled tests")
I have these.
> > b0fb3ada8fb0 ("rcu: Robustify rcu_is_cpu_rrupt_from_idle()")
I have this as an "EXP" experimental commit at the moment.
> > e2cf1ccc99a3 ("rcu/exp: Warn on CPU lagging for too long within hotplug
> > IPI's blindspot")
> > e8b1ec78cf81 ("rcu/exp: Warn on QS requested on dying CPU")
> > b4783fc5ab1f ("rcu/exp: Remove needless CPU up quiescent state report")
> > a7e008d32033 ("rcu/exp: Remove confusing needless full barrier on task
> > unblock")
> > 6e8a70ce57b6 ("rcu/exp: Protect against early QS report")
But I do not have these!
> > @Joel, do you want me to sync this rcu/dev branch to the rcu repo then?
>
> Yes Boqun, go ahead and sync it with rcu's /dev. Thanks.
>
> Also how do I get added to access the RCU tree?
Boqun, would you like to send the kernel.org request? Just to get them
used to requests not coming from me?
Thanx, Paul
