[PATCH] sched/debug: Don't update sched_domain debug directories before sched_debug_init()

[PATCH] sched/debug: Don't update sched_domain debug directories before sched_debug_init()

[Valentin Schneider]

Since CPU capacity asymmetry can stem purely from maximum frequency
differences (e.g. Pixel 1), a rebuild of the scheduler topology can be
issued upon loading cpufreq, see:

  arch_topology.c::init_cpu_capacity_callback()

Turns out that if this rebuild happens *before* sched_debug_init() is
run (which is a late initcall), we end up messing up the sched_domain debug
directory: passing a NULL parent to debugfs_create_dir() ends up creating
the directory at the debugfs root, which in this case creates
/sys/kernel/debug/domains (instead of /sys/kernel/debug/sched/domains).

This currently doesn't happen on asymmetric systems which use cpufreq-scpi
or cpufreq-dt drivers, as those are loaded via
deferred_probe_initcall() (it is also a late initcall, but appears to be
ordered *after* sched_debug_init()).

Ionela has been working on detecting maximum frequency asymmetry via ACPI,
and that actually happens via a *device* initcall, thus before
sched_debug_init(), and causes the aforementionned debugfs mayhem.

One option would be to punt sched_debug_init() down to
fs_initcall_sync(). Preventing update_sched_domain_debugfs() from running
before sched_debug_init() appears to be the safer option.

Link: http://lore.kernel.org/r/20210514095339.12979-1-ionela.voinescu@arm.com
Fixes: 3b87f136f8fc ("sched,debug: Convert sysctl sched_domains to debugfs")
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
---
 kernel/sched/debug.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 3bdee5fd7d29..6b5d6c7612fd 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -388,6 +388,13 @@ void update_sched_domain_debugfs(void)
 {
 	int cpu, i;
 
+	/*
+	 * This can unfortunately be invoked before sched_debug_init() creates
+	 * the debug directory. Don't touch sd_sysctl_cpus until then.
+	 */
+	if (!debugfs_sched)
+		return;
+
 	if (!cpumask_available(sd_sysctl_cpus)) {
 		if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL))
 			return;
-- 
2.25.1

Re: [PATCH] sched/debug: Don't update sched_domain debug directories before sched_debug_init()

[Peter Zijlstra]

On Tue, May 18, 2021 at 02:07:25PM +0100, Valentin Schneider wrote:
> Since CPU capacity asymmetry can stem purely from maximum frequency
> differences (e.g. Pixel 1), a rebuild of the scheduler topology can be
> issued upon loading cpufreq, see:
> 
>   arch_topology.c::init_cpu_capacity_callback()
> 
> Turns out that if this rebuild happens *before* sched_debug_init() is
> run (which is a late initcall), we end up messing up the sched_domain debug
> directory: passing a NULL parent to debugfs_create_dir() ends up creating
> the directory at the debugfs root, which in this case creates
> /sys/kernel/debug/domains (instead of /sys/kernel/debug/sched/domains).
> 
> This currently doesn't happen on asymmetric systems which use cpufreq-scpi
> or cpufreq-dt drivers, as those are loaded via
> deferred_probe_initcall() (it is also a late initcall, but appears to be
> ordered *after* sched_debug_init()).
> 
> Ionela has been working on detecting maximum frequency asymmetry via ACPI,
> and that actually happens via a *device* initcall, thus before
> sched_debug_init(), and causes the aforementionned debugfs mayhem.
> 
> One option would be to punt sched_debug_init() down to
> fs_initcall_sync(). Preventing update_sched_domain_debugfs() from running
> before sched_debug_init() appears to be the safer option.
> 
> Link: http://lore.kernel.org/r/20210514095339.12979-1-ionela.voinescu@arm.com
> Fixes: 3b87f136f8fc ("sched,debug: Convert sysctl sched_domains to debugfs")
> Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>

Thanks!