From: Michal Hocko <mhocko@xxxxxxxx> Subject: mm: move mm_percpu_wq initialization earlier Yang Li has reported that drain_all_pages triggers a WARN_ON which means that this function is called earlier than the mm_percpu_wq is initialized on arm64 with CMA configured: [ 0.276449] WARNING: CPU: 2 PID: 1 at mm/page_alloc.c:2423 drain_all_pages+0x244/0x25c [ 0.276537] Modules linked in: [ 0.276594] CPU: 2 PID: 1 Comm: swapper/0 Not tainted 4.11.0-rc1-next-20170310-00027-g64dfbc5 #127 [ 0.276693] Hardware name: Freescale Layerscape 2088A RDB Board (DT) [ 0.276764] task: ffffffc07c4a6d00 task.stack: ffffffc07c4a8000 [ 0.276831] PC is at drain_all_pages+0x244/0x25c [ 0.276886] LR is at start_isolate_page_range+0x14c/0x1f0 [...] [ 0.279000] [<ffffff80081636bc>] drain_all_pages+0x244/0x25c [ 0.279065] [<ffffff80081c675c>] start_isolate_page_range+0x14c/0x1f0 [ 0.279137] [<ffffff8008166a48>] alloc_contig_range+0xec/0x354 [ 0.279203] [<ffffff80081c6c5c>] cma_alloc+0x100/0x1fc [ 0.279263] [<ffffff8008481714>] dma_alloc_from_contiguous+0x3c/0x44 [ 0.279336] [<ffffff8008b25720>] atomic_pool_init+0x7c/0x208 [ 0.279399] [<ffffff8008b258f0>] arm64_dma_init+0x44/0x4c [ 0.279461] [<ffffff8008083144>] do_one_initcall+0x38/0x128 [ 0.279525] [<ffffff8008b20d30>] kernel_init_freeable+0x1a0/0x240 [ 0.279596] [<ffffff8008807778>] kernel_init+0x10/0xfc [ 0.279654] [<ffffff8008082b70>] ret_from_fork+0x10/0x20 Fix this by moving the whole setup_vmstat which is an initcall right now to init_mm_internals which will be called right after the WQ subsystem is initialized. Link: http://lkml.kernel.org/r/20170315164021.28532-1-mhocko@xxxxxxxxxx Signed-off-by: Michal Hocko <mhocko@xxxxxxxx> Reported-by: Yang Li <pku.leo@xxxxxxxxx> Tested-by: Yang Li <pku.leo@xxxxxxxxx> Tested-by: Xiaolong Ye <xiaolong.ye@xxxxxxxxx> Cc: Mel Gorman <mgorman@xxxxxxx> Cc: Vlastimil Babka <vbabka@xxxxxxx> Cc: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- include/linux/mm.h | 2 ++ init/main.c | 2 ++ mm/vmstat.c | 4 +--- 3 files changed, 5 insertions(+), 3 deletions(-) diff -puN include/linux/mm.h~mm-move-mm_percpu_wq-initialization-earlier include/linux/mm.h --- a/include/linux/mm.h~mm-move-mm_percpu_wq-initialization-earlier +++ a/include/linux/mm.h @@ -32,6 +32,8 @@ struct user_struct; struct writeback_control; struct bdi_writeback; +void init_mm_internals(void); + #ifndef CONFIG_NEED_MULTIPLE_NODES /* Don't use mapnrs, do it properly */ extern unsigned long max_mapnr; diff -puN init/main.c~mm-move-mm_percpu_wq-initialization-earlier init/main.c --- a/init/main.c~mm-move-mm_percpu_wq-initialization-earlier +++ a/init/main.c @@ -1022,6 +1022,8 @@ static noinline void __init kernel_init_ workqueue_init(); + init_mm_internals(); + do_pre_smp_initcalls(); lockup_detector_init(); diff -puN mm/vmstat.c~mm-move-mm_percpu_wq-initialization-earlier mm/vmstat.c --- a/mm/vmstat.c~mm-move-mm_percpu_wq-initialization-earlier +++ a/mm/vmstat.c @@ -1764,7 +1764,7 @@ static int vmstat_cpu_dead(unsigned int #endif -static int __init setup_vmstat(void) +void __init init_mm_internals(void) { #ifdef CONFIG_SMP int ret; @@ -1792,9 +1792,7 @@ static int __init setup_vmstat(void) proc_create("vmstat", S_IRUGO, NULL, &proc_vmstat_file_operations); proc_create("zoneinfo", S_IRUGO, NULL, &proc_zoneinfo_file_operations); #endif - return 0; } -module_init(setup_vmstat) #if defined(CONFIG_DEBUG_FS) && defined(CONFIG_COMPACTION) _ -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
