From patchwork Tue Dec 3 10:43:53 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Taejoon Song X-Patchwork-Id: 11271001 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 62DED138D for ; Tue, 3 Dec 2019 11:15:26 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4B3352073B for ; Tue, 3 Dec 2019 11:15:26 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726251AbfLCLPZ (ORCPT ); Tue, 3 Dec 2019 06:15:25 -0500 Received: from lgeamrelo11.lge.com ([156.147.23.51]:44364 "EHLO lgeamrelo11.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726139AbfLCLPZ (ORCPT ); Tue, 3 Dec 2019 06:15:25 -0500 Received: from unknown (HELO lgemrelse6q.lge.com) (156.147.1.121) by 156.147.23.51 with ESMTP; 3 Dec 2019 19:45:22 +0900 X-Original-SENDERIP: 156.147.1.121 X-Original-MAILFROM: taejoon.song@lge.com Received: from unknown (HELO ubuntu0.156.147.1.1) (10.177.220.34) by 156.147.1.121 with ESMTP; 3 Dec 2019 19:45:22 +0900 X-Original-SENDERIP: 10.177.220.34 X-Original-MAILFROM: taejoon.song@lge.com From: Taejoon Song To: minchan@google.com, ngupta@vflare.org, sergey.senozhatsky.work@gmail.com, axboe@kernel.dk Cc: linux-kernel@vger.kernel.org, linux-block@vger.kernel.org, yjay.kim@lge.com, Taejoon Song Subject: [PATCH] zram: try to avoid worst-case scenario on same element pages Date: Tue, 3 Dec 2019 19:43:53 +0900 Message-Id: <1575369833-16424-1-git-send-email-taejoon.song@lge.com> X-Mailer: git-send-email 2.7.4 Sender: linux-block-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-block@vger.kernel.org The worst-case scenario on finding same element pages is that almost all elements are same at the first glance but only last few elements are different. Since the same element tends to be grouped from the beginning of the pages, if we check the first element with the last element before looping through all elements, we might have some chances to quickly detect non-same element pages. 1. Test is done under LG webOS TV (64-bit arch) 2. Dump the swap-out pages (~819200 pages) 3. Analyze the pages with simple test script which counts the iteration number and measures the speed at off-line Under 64-bit arch, the worst iteration count is PAGE_SIZE / 8 bytes = 512. The speed is based on the time to consume page_same_filled() function only. The result, on average, is listed as below: Num of Iter Speed(MB/s) Looping-Forward (Orig) 38 99265 Looping-Backward 36 102725 Last-element-check (This Patch) 33 125072 The result shows that the average iteration count decreases by 13% and the speed increases by 25% with this patch. This patch does not increase the overall time complexity, though. I also ran simpler version which uses backward loop. Just looping backward also makes some improvement, but less than this patch. Signed-off-by: Taejoon Song --- drivers/block/zram/zram_drv.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c index 4285e75..afadd7f 100644 --- a/drivers/block/zram/zram_drv.c +++ b/drivers/block/zram/zram_drv.c @@ -207,14 +207,17 @@ static inline void zram_fill_page(void *ptr, unsigned long len, static bool page_same_filled(void *ptr, unsigned long *element) { - unsigned int pos; unsigned long *page; unsigned long val; + unsigned int pos, last_pos = PAGE_SIZE / sizeof(*page) - 1; page = (unsigned long *)ptr; val = page[0]; - for (pos = 1; pos < PAGE_SIZE / sizeof(*page); pos++) { + if (val != page[last_pos]) + return false; + + for (pos = 1; pos < last_pos - 1; pos++) { if (val != page[pos]) return false; }