From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:54858) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1R21lW-0000Zw-6Y for qemu-devel@nongnu.org; Fri, 09 Sep 2011 10:09:15 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1R21lV-0006cL-2F for qemu-devel@nongnu.org; Fri, 09 Sep 2011 10:09:10 -0400 Received: from mail-wy0-f173.google.com ([74.125.82.173]:35314) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1R21lU-0006cG-QJ for qemu-devel@nongnu.org; Fri, 09 Sep 2011 10:09:09 -0400 Received: by wyf22 with SMTP id 22so1682188wyf.4 for ; Fri, 09 Sep 2011 07:09:08 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <20110909094436.GB23929@f15.cn.ibm.com> <20110909103801.GA26148@stefanha-thinkpad.localdomain> Date: Fri, 9 Sep 2011 22:09:07 +0800 Message-ID: From: Zhi Yong Wu Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] Why qemu write/rw speed is so low? List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Stefan Hajnoczi Cc: kwolf@redhat.com, aliguro@us.ibm.com, Stefan Hajnoczi , Zhi Yong Wu , qemu-devel@nongnu.org, ryanh@us.ibm.com On Fri, Sep 9, 2011 at 9:54 PM, Stefan Hajnoczi wrote: > On Fri, Sep 9, 2011 at 2:48 PM, Zhi Yong Wu wrote: >> On Fri, Sep 9, 2011 at 6:38 PM, Stefan Hajnoczi >> wrote: >>> On Fri, Sep 09, 2011 at 05:44:36PM +0800, Zhi Yong Wu wrote: >>>> Today, i did some basical I/O testing, and suddenly found that qemu wr= ite and rw speed is so low now, my qemu binary is built on commit 344eecf69= 95f4a0ad1d887cec922f6806f91a3f8. >>>> >>>> Do qemu have regression? >>>> >>>> The testing data is shown as below: >>>> >>>> 1.) write >>>> >>>> test: (g=3D0): rw=3Dwrite, bs=3D512-512/512-512, ioengine=3Dlibaio, io= depth=3D1 >>> >>> Please post your QEMU command-line. =A0If your -drive is using >>> cache=3Dwritethrough then small writes are slow because they require th= e >>> physical disk to write and then synchronize its write cache. =A0Typical= ly >>> cache=3Dnone is a good setting to use for local disks. >> Now i can not access my workstation in the office. >> -drive if=3Dvirtio,cache=3Dnone,file=3Dxxxx >> >>> >>> The block size of 512 bytes is too small. =A0Ext4 uses a 4 KB block siz= e, >>> so I think a 512 byte write from the guest could cause a 4 KB >>> read-modify-write operation on the host filesystem. >> You mean RCU? What is its work procedure? Can you explain in more >> details if you are available? > > If the host file system manages space in 4 KB blocks, then a 512 byte > to an unallocated part of the file causes the file system to find 4 KB > of free space for this data. =A0Since the write is only 512 bytes and > does not cover the entire 4 KB region, the file system initializes the > remaining 3.5 KB with zeros and writes out the full 4 KB block. > > Now if a 512 byte write comes in for an allocated 4 KB block, then we > need to read in the existing 4 KB, modify the 512 bytes in place, and > write out the 4 KB block again. =A0This is read-modify-write. =A0In this > worst-case scenario a 512 byte write turns into a 4 KB read followed > by a 4 KB write. A 512B write will lead to a 4KB read + 512B modify + a 4KB write. got it. thanks. > > Stefan > --=20 Regards, Zhi Yong Wu