Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> wrote: > On Tue, Aug 01, 2017 at 02:56:14PM +0900, Minchan Kim wrote: >> This patch is ready for solving race problems caused by TLB batch. > > s/is ready/is a preparatory patch/ > >> For that, we will increase/decrease TLB flush pending count of >> mm_struct whenever tlb_[gather|finish]_mmu is called. >> >> Before making it simple, this patch separates architecture specific >> part and rename it to arch_tlb_[gather|finish]_mmu and generic part >> just calls it. >> >> It shouldn't change any behavior. >> >> Cc: Ingo Molnar <mingo@xxxxxxxxxx> >> Cc: Russell King <linux@xxxxxxxxxxxxxxx> >> Cc: Tony Luck <tony.luck@xxxxxxxxx> >> Cc: Martin Schwidefsky <schwidefsky@xxxxxxxxxx> >> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx> >> Cc: Heiko Carstens <heiko.carstens@xxxxxxxxxx> >> Cc: Yoshinori Sato <ysato@xxxxxxxxxxxxxxxxxxxx> >> Cc: Jeff Dike <jdike@xxxxxxxxxxx> >> Cc: linux-arch@xxxxxxxxxxxxxxx >> Cc: Nadav Amit <nadav.amit@xxxxxxxxx> >> Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx> >> Signed-off-by: Minchan Kim <minchan@xxxxxxxxxx> > > You could alias arch_generic_tlb_finish_mmu as arch_tlb_gather_mmu > simiilar to how other arch-generic helpers are done to avoid some > #ifdeffery but otherwise Minchan, Andrew wishes me to send one series that combines both series. What about this comment from Mel? It seems you intentionally did not want to alias them... BTW: patch 4 should add “#include <asm/tlb.h>" - I’ll add it. Thanks, Nadav
