From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33411) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XyeiS-00036u-4M for qemu-devel@nongnu.org; Wed, 10 Dec 2014 05:42:01 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XyeiM-0003EH-PM for qemu-devel@nongnu.org; Wed, 10 Dec 2014 05:41:56 -0500 Received: from mx1.redhat.com ([209.132.183.28]:42009) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XyeiM-0003E2-Iv for qemu-devel@nongnu.org; Wed, 10 Dec 2014 05:41:50 -0500 Date: Wed, 10 Dec 2014 10:41:40 +0000 From: "Dr. David Alan Gilbert" Message-ID: <20141210104139.GB4662@work-vm> References: <1416830152-524-1-git-send-email-arei.gonglei@huawei.com> <1416830152-524-5-git-send-email-arei.gonglei@huawei.com> <87r3w7akgy.fsf@elfo.elfo> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <87r3w7akgy.fsf@elfo.elfo> Subject: Re: [Qemu-devel] [PATCH RESEND for 2.3 4/6] xbzrle: check 8 bytes at a time after an concurrency scene List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Juan Quintela Cc: ChenLiang , weidong.huang@huawei.com, qemu-devel@nongnu.org, dgilbert@redhat.com, arei.gonglei@huawei.com, amit.shah@redhat.com, pbonzini@redhat.com, peter.huangpeng@huawei.com * Juan Quintela (quintela@redhat.com) wrote: > wrote: > > From: ChenLiang > > > > The logic of old code is correct. But Checking byte by byte will > > consume time after an concurrency scene. > > > > Signed-off-by: ChenLiang > > Signed-off-by: Gonglei > > --- > > xbzrle.c | 28 ++++++++++++++++++---------- > > 1 file changed, 18 insertions(+), 10 deletions(-) > > > > diff --git a/xbzrle.c b/xbzrle.c > > index d27a140..0477367 100644 > > --- a/xbzrle.c > > +++ b/xbzrle.c > > @@ -50,16 +50,24 @@ int xbzrle_encode_buffer(uint8_t *old_buf, uint8_t *new_buf, int slen, > > > > /* word at a time for speed */ > > if (!res) { > > - while (i < slen && > > - (*(long *)(old_buf + i)) == (*(long *)(new_buf + i))) { > > - i += sizeof(long); > > - zrun_len += sizeof(long); > > - } > > - > > - /* go over the rest */ > > - while (i < slen && old_buf[i] == new_buf[i]) { > > - zrun_len++; > > - i++; > > + while (i < slen) { > > + if ((*(long *)(old_buf + i)) == (*(long *)(new_buf + i))) { > > + i += sizeof(long); > > + zrun_len += sizeof(long); > > + } else { > > + /* go over the rest */ > > + for (j = 0; j < sizeof(long); j++) { > > + if (old_buf[i] == new_buf[i]) { > > + i++; > > + zrun_len++; > > + } else { > > + break; > > + } > > + } > > + if (j != sizeof(long)) { > > + break; > > + } > > + } > > } > > } > > This still does misaligned reads. If we want to do aligned stuff, > something like that looks much better, no? Notice that where I put > "break", I mean we have finished, but you get the idea. Or I am missing something? You're missing the loop just above this code that's changed that guarantees alginment by the start of this code. Dave > > while(i % sizeof(long) != 0) { > if (old_buf[i] == new_buf[i]) { > i++; > zrun_len++; > } else { > break; > } > } > > while (i < slen) { > if ((*(long *)(old_buf + i)) == (*(long *)(new_buf + i))) { > i += sizeof(long); > zrun_len += sizeof(long); > } else { > break; > } > } > > for (j = 0; j < sizeof(long); j++) { > if (old_buf[i] == new_buf[i]) { > i++; > zrun_len++; > } else { > break; > } > } -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK