On 04/05/2017 04:19 PM, Michal Hocko wrote: > On Wed 05-04-17 16:05:23, Anshuman Khandual wrote: >> On 04/04/2017 10:21 PM, Jérôme Glisse wrote: >>> Just a trivial fix. >>> >>> Signed-off-by: Jérôme Glisse <jglisse@xxxxxxxxxx> >>> Cc: Michal Hocko <mhocko@xxxxxxxx> >>> Cc: Dan Williams <dan.j.williams@xxxxxxxxx> >>> --- >>> kernel/memremap.c | 3 ++- >>> 1 file changed, 2 insertions(+), 1 deletion(-) >>> >>> diff --git a/kernel/memremap.c b/kernel/memremap.c >>> index faa9276..bbbe646 100644 >>> --- a/kernel/memremap.c >>> +++ b/kernel/memremap.c >>> @@ -366,7 +366,8 @@ void *devm_memremap_pages(struct device *dev, struct resource *res, >>> error = arch_add_memory(nid, align_start, align_size); >>> if (!error) >>> move_pfn_range_to_zone(&NODE_DATA(nid)->node_zones[ZONE_DEVICE], >>> - align_start, align_size); >>> + align_start >> PAGE_SHIFT, >>> + align_size >> PAGE_SHIFT); >> >> All this while it was taking up addresses instead of PFNs ? Then >> how it was working correctly before ? > > Because this code was embeded inside the arch_add_memory which did the > translation properly. See arch_add_memory implementations. Got your point. Checked both mainline kernel and mmotm branch v4.11-rc5-mmotm-2017-04-04-15-00, in both the places the code snippet seems to be different than here. For example arch_add _memory has the following signature instead. arch_add_memory(nid, align_start, align_size, true); and I dont see move_pfn_range_to_zone() at all. Which tree/ branch this patch is against ? -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx"> email@xxxxxxxxx </a>