On Wed, Mar 10, 2021 at 09:06:48AM -0800, Minchan Kim wrote:
> On Wed, Mar 10, 2021 at 05:46:56PM +0100, Michal Hocko wrote:
> > On Wed 10-03-21 08:05:36, Minchan Kim wrote:
> > > On Wed, Mar 10, 2021 at 02:07:05PM +0100, Michal Hocko wrote:
> > [...]
> > > > The is a lot of churn indeed. Have you considered adding $FOO_lglvl
> > > > variants for those so that you can use them for your particular case
> > > > without affecting most of existing users? Something similar we have
> > > > discussed in other email thread regarding lru_add_drain_all?
> > >
> > > I thought that way but didn't try since it couldn't make them
> > > atomic(For example, other printk place in other context will
> > > affect by the $FOO_lglvl).
> >
> > I do not follow. I meant something like the following (likely incomplete
> > but you should get an idea).
>
> Oh, I thought you wanted to override loglevel temporally.
>
> old_lvl = save_printk_lvl(new level);
> dump_page();
> restore_printk_lvl(old_lvl);
>
> >
> > diff --git a/include/linux/page_owner.h b/include/linux/page_owner.h
> > index 3468794f83d2..71b402eb8f78 100644
> > --- a/include/linux/page_owner.h
> > +++ b/include/linux/page_owner.h
> > @@ -14,7 +14,7 @@ extern void __set_page_owner(struct page *page,
> > extern void __split_page_owner(struct page *page, unsigned int nr);
> > extern void __copy_page_owner(struct page *oldpage, struct page *newpage);
> > extern void __set_page_owner_migrate_reason(struct page *page, int reason);
> > -extern void __dump_page_owner(struct page *page);
> > +extern void __dump_page_owner(struct page *page, const char *loglvl);
> > extern void pagetypeinfo_showmixedcount_print(struct seq_file *m,
> > pg_data_t *pgdat, struct zone *zone);
> >
> > @@ -46,10 +46,10 @@ static inline void set_page_owner_migrate_reason(struct page *page, int reason)
> > if (static_branch_unlikely(&page_owner_inited))
> > __set_page_owner_migrate_reason(page, reason);
> > }
> > -static inline void dump_page_owner(struct page *page)
> > +static inline void dump_page_owner(struct page *page, const char *loglvl)
> > {
> > if (static_branch_unlikely(&page_owner_inited))
> > - __dump_page_owner(page);
> > + __dump_page_owner(page, loglvl);
> > }
> > #else
> > static inline void reset_page_owner(struct page *page, unsigned int order)
> > @@ -69,7 +69,7 @@ static inline void copy_page_owner(struct page *oldpage, struct page *newpage)
> > static inline void set_page_owner_migrate_reason(struct page *page, int reason)
> > {
> > }
> > -static inline void dump_page_owner(struct page *page)
> > +static inline void dump_page_owner(struct page *page, const char *loglvl)
> > {
> > }
> > #endif /* CONFIG_PAGE_OWNER */
> > diff --git a/kernel/stacktrace.c b/kernel/stacktrace.c
> > index 9f8117c7cfdd..1b13135d9916 100644
> > --- a/kernel/stacktrace.c
> > +++ b/kernel/stacktrace.c
> > @@ -14,6 +14,18 @@
> > #include <linux/kallsyms.h>
> > #include <linux/stacktrace.h>
> >
> > +void __stack_trace_print(const unsigned long *entries, unsigned int nr_entries,
> > + int spacesconst, char *loglvl)
> > +{
> > + unsigned int i;
> > +
> > + if (WARN_ON(!entries))
> > + return;
> > +
> > + for (i = 0; i < nr_entries; i++)
> > + printk("%s%*c%pS\n", loglvl, 1 + spaces, ' ', (void *)entries[i]);
> > +}
>
> That's exactly I did with introducing pr_loglevel. I wanted to address
> *all places* to use dump_page and stack_trace_print since some folks
> might ask me to fix all the broken place all at once. I'm getting tired
> with such hassle.
>
> void dump_page(const char *log_lvl, struct page *page, const char *reason)
> {
> __dump_page(log_lvl, page, reason);
> dump_page_owner(log_lvl, page);
> }
> EXPORT_SYMBOL(dump_page);
Since it's good to have regardless of the patch, I posted next revision
with removeing ratelimit and put something more in description to
proceed the work.
Posted v3 - https://lore.kernel.org/linux-mm/20210310180104.517886-1-minchan@xxxxxxxxxx/
