Message ID | 1459844572-53069-2-git-send-email-puck.chen@hisilicon.com |
---|---|
State | New |
Headers | show |
add Mel Gorman On 2016/4/5 16:22, Chen Feng wrote: > Make the pfn always valid when using flat memory. > If the reserved memory is not align to memblock-size, > there will be holes in zone. > > This patch makes the memory in buddy always in the > array of mem-map. > > Signed-off-by: Chen Feng <puck.chen@hisilicon.com> > Signed-off-by: Fu Jun <oliver.fu@hisilicon.com> > --- > arch/arm64/mm/init.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c > index ea989d8..0e1d5b7 100644 > --- a/arch/arm64/mm/init.c > +++ b/arch/arm64/mm/init.c > @@ -306,7 +306,8 @@ static void __init free_unused_memmap(void) > struct memblock_region *reg; > > for_each_memblock(memory, reg) { > - start = __phys_to_pfn(reg->base); > + start = round_down(__phys_to_pfn(reg->base), > + MAX_ORDER_NR_PAGES); > > #ifdef CONFIG_SPARSEMEM > /* > @@ -327,8 +328,8 @@ static void __init free_unused_memmap(void) > * memmap entries are valid from the bank end aligned to > * MAX_ORDER_NR_PAGES. > */ > - prev_end = ALIGN(__phys_to_pfn(reg->base + reg->size), > - MAX_ORDER_NR_PAGES); > + prev_end = round_up(__phys_to_pfn(reg->base + reg->size), > + MAX_ORDER_NR_PAGES); > } > > #ifdef CONFIG_SPARSEMEM >
On Mon, Apr 11, 2016 at 07:08:38PM +0800, Xishi Qiu wrote: > On 2016/4/5 16:22, Chen Feng wrote: > > > Make the pfn always valid when using flat memory. > > If the reserved memory is not align to memblock-size, > > there will be holes in zone. > > > > This patch makes the memory in buddy always in the > > array of mem-map. > > > > Signed-off-by: Chen Feng <puck.chen@hisilicon.com> > > Signed-off-by: Fu Jun <oliver.fu@hisilicon.com> > > --- > > arch/arm64/mm/init.c | 7 ++++--- > > 1 file changed, 4 insertions(+), 3 deletions(-) > > > > diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c > > index ea989d8..0e1d5b7 100644 > > --- a/arch/arm64/mm/init.c > > +++ b/arch/arm64/mm/init.c > > @@ -306,7 +306,8 @@ static void __init free_unused_memmap(void) > > How about let free_unused_memmap() support for CONFIG_SPARSEMEM_VMEMMAP? We would need extra care to check that the memmap was actually allocated in the first place. -- Catalin
diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index ea989d8..0e1d5b7 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -306,7 +306,8 @@ static void __init free_unused_memmap(void) struct memblock_region *reg; for_each_memblock(memory, reg) { - start = __phys_to_pfn(reg->base); + start = round_down(__phys_to_pfn(reg->base), + MAX_ORDER_NR_PAGES); #ifdef CONFIG_SPARSEMEM /* @@ -327,8 +328,8 @@ static void __init free_unused_memmap(void) * memmap entries are valid from the bank end aligned to * MAX_ORDER_NR_PAGES. */ - prev_end = ALIGN(__phys_to_pfn(reg->base + reg->size), - MAX_ORDER_NR_PAGES); + prev_end = round_up(__phys_to_pfn(reg->base + reg->size), + MAX_ORDER_NR_PAGES); } #ifdef CONFIG_SPARSEMEM