Re: rq lock contention due to commit af7f588d8f73

[Peter Zijlstra <peterz@infradead.org>]

On Mon, Mar 27, 2023 at 04:05:02PM +0800, Aaron Lu wrote:
> Hi Mathieu,
> 
> I was doing some optimization work[1] for kernel scheduler using a
> database workload: sysbench+postgres and before I submit my work, I
> rebased my patch on top of latest v6.3-rc kernels to see if everything
> still works expected and then I found rq's lock became very heavily
> contended as compared to v6.2 based kernels.
> 
> Using the above mentioned workload, before commit af7f588d8f73("sched:
> Introduce per-memory-map concurrency ID"), the profile looked like:
> 
>      7.30%     0.71%  [kernel.vmlinux]            [k] __schedule
>      0.03%     0.03%  [kernel.vmlinux]            [k] native_queued_spin_lock_slowpath
> 
> After that commit:
> 
>     49.01%     0.87%  [kernel.vmlinux]            [k] __schedule
>     43.20%    43.18%  [kernel.vmlinux]            [k] native_queued_spin_lock_slowpath

Unlike what the subject says, if it is due to that commit, then it's not
rq lock but that new cid_lock thing.

Can you frob init/Kconfig and make SCHED_MM_CID user configurable and
disable it to confirm?

(also, mathieu, when you do the below, you'll see it is in a weird spot)

diff --git a/init/Kconfig b/init/Kconfig
index 1fb5f313d18f..f2661f73f3dd 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -1042,7 +1042,8 @@ config RT_GROUP_SCHED
 endif #CGROUP_SCHED
 
 config SCHED_MM_CID
-	def_bool y
+	bool "RSEQ Concurrency ID"
+	default y
 	depends on SMP && RSEQ
 
 config UCLAMP_TASK_GROUP