Hi,Fengguang and Yuanhan,thank you for your test, I find the reason which causes that warnings. set_page_block_order() is called by free_area_init_core() and sparse_init(), But free_area_init_core() is defined as __paginginit, sparse_init() and set_page_block_order() are defined as __init. #ifdef CONFIG_SPARSEMEM #define __paginginit __meminit #else #define __paginginit __init #endif So if we config sparse memory and memory hotplug, when hot add a new pgdat, free_area_init_core() will call set_page_block_order(), but set_page_block_order() already has been released. Here is the new patch, just use __paginginit instead of __init. Thanks Xishi Qiu Signed-off-by: Xishi Qiu <qiuxishi@xxxxxxxxxx> --- mm/internal.h | 2 ++ mm/page_alloc.c | 4 ++-- mm/sparse.c | 3 +++ 3 files changed, 7 insertions(+), 2 deletions(-) diff -puN mm/internal.h~mm-setup-pageblock_order-before-its-used-by-sparsemem mm/internal.h --- a/mm/internal.h~mm-setup-pageblock_order-before-its-used-by-sparsemem +++ a/mm/internal.h @@ -349,3 +349,5 @@ extern u32 hwpoison_filter_enable; extern unsigned long vm_mmap_pgoff(struct file *, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); + +extern void set_pageblock_order(void); diff -puN mm/page_alloc.c~mm-setup-pageblock_order-before-its-used-by-sparsemem mm/page_alloc.c --- a/mm/page_alloc.c~mm-setup-pageblock_order-before-its-used-by-sparsemem +++ a/mm/page_alloc.c @@ -4303,7 +4303,7 @@ static inline void setup_usemap(struct p #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE /* Initialise the number of pages represented by NR_PAGEBLOCK_BITS */ -static inline void __init set_pageblock_order(void) +void __paginginit set_pageblock_order(void) { unsigned int order; @@ -4331,7 +4331,7 @@ static inline void __init set_pageblock_ * include/linux/pageblock-flags.h for the values of pageblock_order based on * the kernel config */ -static inline void set_pageblock_order(void) +void __paginginit set_pageblock_order(void) { } diff -puN mm/sparse.c~mm-setup-pageblock_order-before-its-used-by-sparsemem mm/sparse.c --- a/mm/sparse.c~mm-setup-pageblock_order-before-its-used-by-sparsemem +++ a/mm/sparse.c @@ -495,6 +495,9 @@ void __init sparse_init(void) struct page **map_map; #endif + /* Setup pageblock_order for HUGETLB_PAGE_SIZE_VARIABLE */ + set_pageblock_order(); + /* * map is using big page (aka 2M in x86 64 bit) * usemap is less one page (aka 24 bytes) On 2012/10/14 18:42, Fengguang Wu wrote: > Hi Xishi, > > FYI, there are new compile warnings show up in > > commit: ca57df79d4f64e1a4886606af4289d40636189c5 mm: setup pageblock_order before it's used by sparsemem > config: powerpc-allyesconfig # make ARCH=powerpc allyesconfig > > All warnings: > > WARNING: vmlinux.o(.meminit.text+0x1228): Section mismatch in reference from the function .free_area_init_core.isra.34() to the function .init.text:.set_pageblock_order() > The function __meminit .free_area_init_core.isra.34() references > a function __init .set_pageblock_order(). > If .set_pageblock_order is only used by .free_area_init_core.isra.34 then > annotate .set_pageblock_order with a matching annotation. > > --- > 0-DAY kernel build testing backend Open Source Technology Center > Fengguang Wu, Yuanhan Liu Intel Corporation > > -- To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html