On Mon, Apr 08, 2024 at 11:27:02PM -0700, Jane Chu wrote: > On 4/8/2024 12:42 PM, Matthew Wilcox (Oracle) wrote: > > +++ b/include/linux/ksm.h > > @@ -81,15 +81,9 @@ struct folio *ksm_might_need_to_copy(struct folio *folio, > > void rmap_walk_ksm(struct folio *folio, struct rmap_walk_control *rwc); > > void folio_migrate_ksm(struct folio *newfolio, struct folio *folio); > > - > > -#ifdef CONFIG_MEMORY_FAILURE > > -void collect_procs_ksm(struct page *page, struct list_head *to_kill, > > - int force_early); > > -#endif > > - > > -#ifdef CONFIG_PROC_FS > > +void collect_procs_ksm(struct folio *folio, struct page *page, > > + struct list_head *to_kill, int force_early); > > long ksm_process_profit(struct mm_struct *); > > -#endif /* CONFIG_PROC_FS */ > Why is the #ifdef-#endif CONFIG_PROC_FS removed? In ksm.c, > ksm_process_profit() is defined within the config switch. Yes, but there's no need to put ifdefs around function declarations. All it does is add a rather bogus dependency on the CONFIG symbol. There's no harm in having a declaration for a function which doesn't exist.
