From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751832AbcFNHb1 (ORCPT ); Tue, 14 Jun 2016 03:31:27 -0400 Received: from mail-it0-f66.google.com ([209.85.214.66]:35001 "EHLO mail-it0-f66.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751791AbcFNHbZ (ORCPT ); Tue, 14 Jun 2016 03:31:25 -0400 MIME-Version: 1.0 In-Reply-To: <20160614062456.GB13753@js1304-P5Q-DELUXE> References: <20160614062456.GB13753@js1304-P5Q-DELUXE> From: Geert Uytterhoeven Date: Tue, 14 Jun 2016 09:31:23 +0200 X-Google-Sender-Auth: x9XbkJTcFElNOXpLyDlwu_ZbuUc Message-ID: Subject: Re: Boot failure on emev2/kzm9d (was: Re: [PATCH v2 11/11] mm/slab: lockless decision to grow cache) To: Joonsoo Kim Cc: Andrew Morton , Christoph Lameter , Pekka Enberg , David Rientjes , Jesper Dangaard Brouer , Linux MM , "linux-kernel@vger.kernel.org" , linux-renesas-soc@vger.kernel.org, "linux-arm-kernel@lists.infradead.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Joonsoo, On Tue, Jun 14, 2016 at 8:24 AM, Joonsoo Kim wrote: > On Mon, Jun 13, 2016 at 09:43:13PM +0200, Geert Uytterhoeven wrote: >> On Tue, Apr 12, 2016 at 6:51 AM, wrote: >> > From: Joonsoo Kim >> > To check whther free objects exist or not precisely, we need to grab a >> > lock. But, accuracy isn't that important because race window would be >> > even small and if there is too much free object, cache reaper would reap >> > it. So, this patch makes the check for free object exisistence not to >> > hold a lock. This will reduce lock contention in heavily allocation case. >> > >> > Note that until now, n->shared can be freed during the processing by >> > writing slabinfo, but, with some trick in this patch, we can access it >> > freely within interrupt disabled period. >> > >> > Below is the result of concurrent allocation/free in slab allocation >> > benchmark made by Christoph a long time ago. I make the output simpler. >> > The number shows cycle count during alloc/free respectively so less is >> > better. >> > >> > * Before >> > Kmalloc N*alloc N*free(32): Average=248/966 >> > Kmalloc N*alloc N*free(64): Average=261/949 >> > Kmalloc N*alloc N*free(128): Average=314/1016 >> > Kmalloc N*alloc N*free(256): Average=741/1061 >> > Kmalloc N*alloc N*free(512): Average=1246/1152 >> > Kmalloc N*alloc N*free(1024): Average=2437/1259 >> > Kmalloc N*alloc N*free(2048): Average=4980/1800 >> > Kmalloc N*alloc N*free(4096): Average=9000/2078 >> > >> > * After >> > Kmalloc N*alloc N*free(32): Average=344/792 >> > Kmalloc N*alloc N*free(64): Average=347/882 >> > Kmalloc N*alloc N*free(128): Average=390/959 >> > Kmalloc N*alloc N*free(256): Average=393/1067 >> > Kmalloc N*alloc N*free(512): Average=683/1229 >> > Kmalloc N*alloc N*free(1024): Average=1295/1325 >> > Kmalloc N*alloc N*free(2048): Average=2513/1664 >> > Kmalloc N*alloc N*free(4096): Average=4742/2172 >> > >> > It shows that allocation performance decreases for the object size up to >> > 128 and it may be due to extra checks in cache_alloc_refill(). But, with >> > considering improvement of free performance, net result looks the same. >> > Result for other size class looks very promising, roughly, 50% performance >> > improvement. >> > >> > v2: replace kick_all_cpus_sync() with synchronize_sched(). >> > >> > Signed-off-by: Joonsoo Kim >> >> I've bisected a boot failure (no output at all) in v4.7-rc2 on emev2/kzm9d >> (Renesas dual Cortex A9) to this patch, which is upstream commit >> 801faf0db8947e01877920e848a4d338dd7a99e7. >> >> I've attached my .config. I don't know if it also happens with >> shmobile_defconfig, as something went wrong with my remote access to the board, >> preventing further testing. I also couldn't verify if the issue persists in >> v4.7-rc3. In the mean time, I've verified it also happens with shmobile_defconfig. >> >> Do you have a clue? > > I don't have yet. Could you help me to narrow down the problem? > Following diff is half-revert change to check that synchronize_sched() > has no problem. Thanks! Unfortunately the half revert is not sufficient. The full revert is. > ----->8----- > diff --git a/mm/slab.c b/mm/slab.c > index 763096a..257a0eb 100644 > --- a/mm/slab.c > +++ b/mm/slab.c > @@ -3016,9 +3016,6 @@ static void *cache_alloc_refill(struct kmem_cache *cachep, gfp_t flags) > n = get_node(cachep, node); > > BUG_ON(ac->avail > 0 || !n); > - shared = READ_ONCE(n->shared); > - if (!n->free_objects && (!shared || !shared->avail)) > - goto direct_grow; > > spin_lock(&n->list_lock); > shared = READ_ONCE(n->shared); > @@ -3047,7 +3044,6 @@ alloc_done: > spin_unlock(&n->list_lock); > fixup_objfreelist_debug(cachep, &list); > > -direct_grow: > if (unlikely(!ac->avail)) { > /* Check if we can use obj in pfmemalloc slab */ > if (sk_memalloc_socks()) { Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds