Messages in this thread | | | Date | Wed, 8 Feb 2023 08:26:19 -0500 | From | Phil Auld <> | Subject | Re: [PATCH v2] sched/debug: Put sched/domains files under the verbose flag |
| |
Hi Ingo, Peter,
On Fri, Jan 20, 2023 at 11:33:30AM -0500 Phil Auld wrote: > The debug files under sched/domains can take a long time to regenerate, > especially when updates are done one at a time. Move these files under > the sched verbose debug flag. Allow changes to verbose to trigger > generation of the files. This lets a user batch the updates but still > have the information available. The detailed topology printk messages > are also under verbose. > > Discussion that lead to this approach can be found in the link below. > > Simplified code to maintain use of debugfs bool routines suggested by > Michael Ellerman <mpe@ellerman.id.au>. > > Signed-off-by: Phil Auld <pauld@redhat.com> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> > Cc: Ingo Molnar <mingo@redhat.com> > Cc: Peter Zijlstra (Intel) <peterz@infradead.org> > Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com> > Cc: Valentin Schneider <vschneid@redhat.com> > Cc: Vishal Chourasia <vishalc@linux.vnet.ibm.com> > Cc: Vincent Guittot <vincent.guittot@linaro.org> > Link: https://lore.kernel.org/all/Y01UWQL2y2r69sBX@li-05afa54c-330e-11b2-a85c-e3f3aa0db1e9.ibm.com/ > ---
Thoughts on this one?
Thanks, Phil
> > v2: fix comment typo and use cpumask_empty() > > kernel/sched/debug.c | 52 +++++++++++++++++++++++++++++++++++++++++--- > 1 file changed, 49 insertions(+), 3 deletions(-) > > diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c > index 1637b65ba07a..0b2340a79b65 100644 > --- a/kernel/sched/debug.c > +++ b/kernel/sched/debug.c > @@ -280,6 +280,45 @@ static const struct file_operations sched_dynamic_fops = { > > __read_mostly bool sched_debug_verbose; > > +#ifdef CONFIG_SMP > +static struct dentry *sd_dentry; > + > + > +static ssize_t sched_verbose_write(struct file *filp, const char __user *ubuf, > + size_t cnt, loff_t *ppos) > +{ > + ssize_t result; > + bool orig; > + > + cpus_read_lock(); > + mutex_lock(&sched_domains_mutex); > + > + orig = sched_debug_verbose; > + result = debugfs_write_file_bool(filp, ubuf, cnt, ppos); > + > + if (sched_debug_verbose && !orig) > + update_sched_domain_debugfs(); > + else if (!sched_debug_verbose && orig) { > + debugfs_remove(sd_dentry); > + sd_dentry = NULL; > + } > + > + mutex_unlock(&sched_domains_mutex); > + cpus_read_unlock(); > + > + return result; > +} > +#else > +#define sched_verbose_write debugfs_write_file_bool > +#endif > + > +static const struct file_operations sched_verbose_fops = { > + .read = debugfs_read_file_bool, > + .write = sched_verbose_write, > + .open = simple_open, > + .llseek = default_llseek, > +}; > + > static const struct seq_operations sched_debug_sops; > > static int sched_debug_open(struct inode *inode, struct file *filp) > @@ -303,7 +342,7 @@ static __init int sched_init_debug(void) > debugfs_sched = debugfs_create_dir("sched", NULL); > > debugfs_create_file("features", 0644, debugfs_sched, NULL, &sched_feat_fops); > - debugfs_create_bool("verbose", 0644, debugfs_sched, &sched_debug_verbose); > + debugfs_create_file_unsafe("verbose", 0644, debugfs_sched, &sched_debug_verbose, &sched_verbose_fops); > #ifdef CONFIG_PREEMPT_DYNAMIC > debugfs_create_file("preempt", 0644, debugfs_sched, NULL, &sched_dynamic_fops); > #endif > @@ -345,7 +384,6 @@ late_initcall(sched_init_debug); > #ifdef CONFIG_SMP > > static cpumask_var_t sd_sysctl_cpus; > -static struct dentry *sd_dentry; > > static int sd_flags_show(struct seq_file *m, void *v) > { > @@ -402,15 +440,23 @@ void update_sched_domain_debugfs(void) > if (!debugfs_sched) > return; > > + if (!sched_debug_verbose) > + return; > + > if (!cpumask_available(sd_sysctl_cpus)) { > if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL)) > return; > cpumask_copy(sd_sysctl_cpus, cpu_possible_mask); > } > > - if (!sd_dentry) > + if (!sd_dentry) { > sd_dentry = debugfs_create_dir("domains", debugfs_sched); > > + /* rebuild sd_sysctl_cpus if empty since it gets cleared below */ > + if (cpumask_empty(sd_sysctl_cpus)) > + cpumask_copy(sd_sysctl_cpus, cpu_online_mask); > + } > + > for_each_cpu(cpu, sd_sysctl_cpus) { > struct sched_domain *sd; > struct dentry *d_cpu; > -- > 2.31.1 >
--
| |