From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Holden Karau" Subject: Re: [PATCH 1/1] fat: improve sync performance by grouping writes in fat_mirror_bhs [really unmangled] Date: Fri, 27 Oct 2006 14:56:16 -0400 Message-ID: References: <4540A32E.5050602@pigscanfly.ca> <20061026153037.GB12596@wohnheim.fh-wedel.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "Josef Sipek" , hirofumi@mail.parknet.co.jp, linux-kernel@vger.kernel.org, holdenk@xandros.com, "akpm@osdl.org" , linux-fsdevel@vger.kernel.org, holden.karau@gmail.com Return-path: To: "=?ISO-8859-1?Q?J=F6rn_Engel?=" In-Reply-To: <20061026153037.GB12596@wohnheim.fh-wedel.de> Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Hi J=F6rn, Thanks for your time, I'll make those changes [along with a few other things I noticed while benchmarking it]. Before I put together a patch, does anyone else see any obvious stuff I should clean up? Cheers, Holden :-) On 10/26/06, J=F6rn Engel wrote: > I didn't pay too much attention, but found some low hanging fruits. > > On Thu, 26 October 2006 07:59:42 -0400, Holden Karau wrote: > > > > -/* FIXME: We can write the blocks as more big chunk. */ > > static int fat_mirror_bhs(struct super_block *sb, struct buffer_he= ad **bhs, > > - int nr_bhs) > > + int nr_bhs ) { > > + return fat_mirror_bhs_optw(sb , bhs , nr_bhs, 0); > > +} > > + > > +static int fat_mirror_bhs_optw(struct super_block *sb, struct buff= er_head **bhs, > > + int nr_bhs , int wait) > > Does this compile without warnings? Looks as if you should reverse > the order of the two functions. > =46or some reason it compiles without warnings for me, but I'll switch = the order. > > { > > struct msdos_sb_info *sbi =3D MSDOS_SB(sb); > > - struct buffer_head *c_bh; > > + struct buffer_head *c_bh[nr_bhs]; > > int err, n, copy; > > > > + /* Always wait if mounted -o sync */ > > + if (sb->s_flags & MS_SYNCHRONOUS ) { > > + wait =3D 1; > > + } > > Coding style. Use a tab for indentation and don't use braces for > single-line conditional statements. > Sorry about that. A lot of the places where I used braces are because I had some debugging output in there while I was hacking on it. I'll change it. > > + > > err =3D 0; > > + err =3D fat_sync_bhs_optw( bhs , nr_bhs , wait); > > The err=3D0; is superfluous now, isn't it? > =2E... no comment :-) > > + if (err) > > + goto error; > > Indentation. > oops :-) I'll fix that. > J=F6rn > > -- > Fantasy is more important than knowledge. Knowledge is limited, > while fantasy embraces the whole world. > -- Albert Einstein > --=20 Cell: 613-276-1645