From: =?utf-8?b?R3Vhbiw=?= Xin Subject: Squashfs: add LZ4 compression Date: Sat, 6 Sep 2014 20:58:43 +0000 (UTC) Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: linux-fsdevel@vger.kernel.org Return-path: Received: from plane.gmane.org ([80.91.229.3]:60935 "EHLO plane.gmane.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751521AbaIFU65 (ORCPT ); Sat, 6 Sep 2014 16:58:57 -0400 Received: from list by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1XQN4R-0003hy-W8 for linux-fsdevel@vger.kernel.org; Sat, 06 Sep 2014 22:58:56 +0200 Received: from 5mao.bbsindex.com ([74.121.150.153]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sat, 06 Sep 2014 22:58:55 +0200 Received: from guanx.bac by 5mao.bbsindex.com with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sat, 06 Sep 2014 22:58:55 +0200 Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Hi, More than one year ago Phillip Lougher proposed LZ4 compression in his squashfs: http://thread.gmane.org/gmane.linux.file-systems/76409 The discussion above shows that people wished that LZ4 support be added to the Linux kernel after it's officially supported by squahsfs-tools. Also, some benchmarks were requested. Version 4.3 of squahsfs-tools, released on 2014-05-12 includes LZ4 support. I have posted my benchmarks in that thread. The HC mode of LZ4 compression came out to be 4x faster (=5x speed) than LZO. The default mode of LZ4 (i.e. the non-high-compression) is ~50x faster than LZO. Decompression speed is also noticeably faster. Not only that applications start faster, but the occasional lags during normal desktop operations on LZO compressed squashfs completely disappeared. So, can we agree now, to merge the LZ4 support of squashfs into mainline? Thanks and Best regards, Guan