Re: [PATCH v6 1/4] rcu: Make call_rcu() lazy to save power

[Joel Fernandes <joel@joelfernandes.org>]

On Mon, Sep 26, 2022 at 04:59:44PM -0700, Paul E. McKenney wrote:
> On Mon, Sep 26, 2022 at 07:47:50PM -0400, Joel Fernandes wrote:
> > 
> > 
> > > On Sep 26, 2022, at 6:32 PM, Paul E. McKenney <paulmck@kernel.org> wrote:
> > > 
> > > On Mon, Sep 26, 2022 at 09:02:21PM +0000, Joel Fernandes wrote:
> > >> On Mon, Sep 26, 2022 at 09:32:44PM +0200, Uladzislau Rezki wrote:
> > >> [...]
> > >>>>>> On my KVM machine the boot time is affected:
> > >>>>>> 
> > >>>>>> <snip>
> > >>>>>> [    2.273406] e1000 0000:00:03.0 eth0: Intel(R) PRO/1000 Network Connection
> > >>>>>> [   11.945283] e1000 0000:00:03.0 ens3: renamed from eth0
> > >>>>>> [   22.165198] sr 1:0:0:0: [sr0] scsi3-mmc drive: 4x/4x cd/rw xa/form2 tray
> > >>>>>> [   22.165206] cdrom: Uniform CD-ROM driver Revision: 3.20
> > >>>>>> [   32.406981] sr 1:0:0:0: Attached scsi CD-ROM sr0
> > >>>>>> [  104.115418] process '/usr/bin/fstype' started with executable stack
> > >>>>>> [  104.170142] EXT4-fs (sda1): mounted filesystem with ordered data mode. Quota mode: none.
> > >>>>>> [  104.340125] systemd[1]: systemd 241 running in system mode. (+PAM +AUDIT +SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD -IDN2 +IDN -PCRE2 default-hierarchy=hybrid)
> > >>>>>> [  104.340193] systemd[1]: Detected virtualization kvm.
> > >>>>>> [  104.340196] systemd[1]: Detected architecture x86-64.
> > >>>>>> [  104.359032] systemd[1]: Set hostname to <pc638>.
> > >>>>>> [  105.740109] random: crng init done
> > >>>>>> [  105.741267] systemd[1]: Reached target Remote File Systems.
> > >>>>>> <snip>
> > >>>>>> 
> > >>>>>> 2 - 11 and second delay is between 32 - 104. So there are still users which must
> > >>>>>> be waiting for "RCU" in a sync way.
> > >>>>> 
> > >>>>> I was wondering if you can compare boot logs and see which timestamp does the
> > >>>>> slow down start from. That way, we can narrow down the callback. Also another
> > >>>>> idea is, add "trace_event=rcu:rcu_callback,rcu:rcu_invoke_callback
> > >>>>> ftrace_dump_on_oops" to the boot params, and then manually call
> > >>>>> "tracing_off(); panic();" from the code at the first printk that seems off in
> > >>>>> your comparison of good vs bad. For example, if "crng init done" timestamp is
> > >>>>> off, put the "tracing_off(); panic();" there. Then grab the serial console
> > >>>>> output to see what were the last callbacks that was queued/invoked.
> > >>>> 
> > >>>> We do seem to be in need of some way to quickly and easily locate the
> > >>>> callback that needed to be _flush() due to a wakeup.
> > >>>> 
> > >>> <snip>
> > >>> diff --git a/kernel/workqueue.c b/kernel/workqueue.c
> > >>> index aeea9731ef80..fe1146d97f1a 100644
> > >>> --- a/kernel/workqueue.c
> > >>> +++ b/kernel/workqueue.c
> > >>> @@ -1771,7 +1771,7 @@ bool queue_rcu_work(struct workqueue_struct *wq, struct rcu_work *rwork)
> > >>> 
> > >>>        if (!test_and_set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(work))) {
> > >>>                rwork->wq = wq;
> > >>> -               call_rcu(&rwork->rcu, rcu_work_rcufn);
> > >>> +               call_rcu_flush(&rwork->rcu, rcu_work_rcufn);
> > >>>                return true;
> > >>>        }
> > >>> 
> > >>> <snip>
> > >>> 
> > >>> ?
> > >>> 
> > >>> But it does not fully solve my boot-up issue. Will debug tomorrow further.
> > >> 
> > >> Ah, but at least its progress, thanks. Could you send me a patch to include
> > >> in the next revision with details of this?
> > >> 
> > >>>> Might one more proactive approach be to use Coccinelle to locate such
> > >>>> callback functions?  We might not want -all- callbacks that do wakeups
> > >>>> to use call_rcu_flush(), but knowing which are which should speed up
> > >>>> slow-boot debugging by quite a bit.
> > >>>> 
> > >>>> Or is there a better way to do this?
> > >>>> 
> > >>> I am not sure what Coccinelle is. If we had something automated that measures
> > >>> a boot time and if needed does some profiling it would be good. Otherwise it
> > >>> is a manual debugging mainly, IMHO.
> > >> 
> > >> Paul, What about using a default-off kernel CONFIG that splats on all lazy
> > >> call_rcu() callbacks that do a wake up. We could use the trace hooks to do it
> > >> in kernel I think. I can talk to Steve to get ideas on how to do that but I
> > >> think it can be done purely from trace events (we might need a new
> > >> trace_end_invoke_callback to fire after the callback is invoked). Thoughts?
> > > 
> > > Could you look for wakeups invoked between trace_rcu_batch_start() and
> > > trace_rcu_batch_end() that are not from interrupt context?  This would
> > > of course need to be associated with a task rather than a CPU.
> > 
> > Yes this sounds good, but we also need to know if the callbacks are lazy or not since wake-up is ok from a non lazy one. I think I’ll need a table to track that at queuing time.
> 
> Agreed.
> 
> > > Note that you would need to check for wakeups from interrupt handlers
> > > even with the extra trace_end_invoke_callback().  The window where an
> > > interrupt handler could do a wakeup would be reduced, but not eliminated.
> > 
> > True! Since this is a  debugging option, can we not just disable interrupts across callback invocation?
> 
> Not without terminally annoying lockdep, at least for any RCU callbacks
> doing things like spin_lock_bh().
> 

Sorry if my last email bounced. Looks like my iPhone betrayed me this once ;)

I was thinking something like this:
1. Put a flag in rcu_head to mark CBs as lazy.
2. Add a trace_rcu_invoke_callback_end() trace point.

Both #1 and #2 can be a debug CONFIG option. #2 can be a tracepoint and not
exposed if needed.

3. Put an in-kernel probe on both trace_rcu_invoke_callback_start() and
trace_rcu_invoke_callback_end(). In the start probe, set a per-task flag if
the current CB is lazy. In the end probe, clear it.

4. Put an in-kernel probe on trace_rcu_sched_wakeup().

Splat in the wake up probe if:
1. Hard IRQs are on.
2. The per-cpu flag is set.

#3 actually does not even need probes if we can directly call the functions
from the rcu_do_batch() function.

I'll work on it in the morning and also look into Vlad's config.

thanks,

 - Joel