From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [PATCH] mm, page_alloc: re-enable softirq use of per-cpu page allocator Date: Mon, 10 Apr 2017 14:26:16 -0700 Message-ID: <20170410142616.6d37a11904dd153298cf7f3b@linux-foundation.org> References: <20170410150821.vcjlz7ntabtfsumm@techsingularity.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: brouer@redhat.com, willy@infradead.org, peterz@infradead.org, pagupta@redhat.com, ttoukan.linux@gmail.com, tariqt@mellanox.com, netdev@vger.kernel.org, saeedm@mellanox.com, linux-kernel@vger.kernel.org, linux-mm@kvack.org To: Mel Gorman Return-path: In-Reply-To: <20170410150821.vcjlz7ntabtfsumm@techsingularity.net> Sender: owner-linux-mm@kvack.org List-Id: netdev.vger.kernel.org On Mon, 10 Apr 2017 16:08:21 +0100 Mel Gorman wrote: > IRQ context were excluded from using the Per-Cpu-Pages (PCP) lists caching > of order-0 pages in commit 374ad05ab64d ("mm, page_alloc: only use per-cpu > allocator for irq-safe requests"). > > This unfortunately also included excluded SoftIRQ. This hurt the performance > for the use-case of refilling DMA RX rings in softirq context. Out of curiosity: by how much did it "hurt"? Tariq found: : I disabled the page-cache (recycle) mechanism to stress the page : allocator, and see a drastic degradation in BW, from 47.5 G in v4.10 to : 31.4 G in v4.11-rc1 (34% drop). then with this patch he found : It looks very good! I get line-rate (94Gbits/sec) with 8 streams, in : comparison to less than 55Gbits/sec before. Can I take this to mean that the page allocator's per-cpu-pages feature ended up doubling the performance of this driver? Better than the driver's private page recycling? I'd like to believe that, but am having trouble doing so ;) > This patch re-allow softirq context, which should be safe by disabling > BH/softirq, while accessing the list. PCP-lists access from both hard-IRQ > and NMI context must not be allowed. Peter Zijlstra says in_nmi() code > never access the page allocator, thus it should be sufficient to only test > for !in_irq(). > > One concern with this change is adding a BH (enable) scheduling point at > both PCP alloc and free. If further concerns are highlighted by this patch, > the result wiill be to revert 374ad05ab64d and try again at a later date > to offset the irq enable/disable overhead. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org