From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:36770) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YJhoU-0006ka-S0 for qemu-devel@nongnu.org; Fri, 06 Feb 2015 07:15:11 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YJhoR-00071P-I7 for qemu-devel@nongnu.org; Fri, 06 Feb 2015 07:15:10 -0500 Received: from mx1.redhat.com ([209.132.183.28]:34420) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YJhoR-00070o-Ah for qemu-devel@nongnu.org; Fri, 06 Feb 2015 07:15:07 -0500 Date: Fri, 6 Feb 2015 12:15:00 +0000 From: "Dr. David Alan Gilbert" Message-ID: <20150206121500.GI2364@work-vm> References: <1422875149-13198-1-git-send-email-liang.z.li@intel.com> <1422875149-13198-10-git-send-email-liang.z.li@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1422875149-13198-10-git-send-email-liang.z.li@intel.com> Subject: Re: [Qemu-devel] [v4 09/13] migration: Make compression co-work with xbzrle List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Liang Li Cc: quintela@redhat.com, armbru@redhat.com, qemu-devel@nongnu.org, lcapitulino@redhat.com, Yang Zhang , amit.shah@redhat.com * Liang Li (liang.z.li@intel.com) wrote: > Now, multiple thread compression can co-work with xbzrle. when > xbzrle is on, multiple thread compression will only work at the > first round of RAM data sync. > > Signed-off-by: Liang Li > Signed-off-by: Yang Zhang > --- > arch_init.c | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/arch_init.c b/arch_init.c > index b8bdb16..8ef0315 100644 > --- a/arch_init.c > +++ b/arch_init.c > @@ -364,6 +364,7 @@ static QemuCond *comp_done_cond; > /* The empty QEMUFileOps will be used by file in CompressParam */ > static const QEMUFileOps empty_ops = { }; > static bool quit_thread; > +static bool compression_switch_on; Yes, OK; a bit of an odd name, but I can see what it's trying to do. Reviewed-by: Dr. David Alan Gilbert > static int one_byte_count; > static DecompressParam *decomp_param; > static QemuThread *decompress_threads; > @@ -440,6 +441,7 @@ void migrate_compress_threads_create(MigrationState *s) > return; > } > quit_thread = false; > + compression_switch_on = true; > thread_count = migrate_compress_threads(); > s->compress_thread = g_new0(QemuThread, thread_count); > comp_param = g_new0(CompressParam, thread_count); > @@ -957,9 +959,16 @@ static int ram_find_and_save_block(QEMUFile *f, bool last_stage) > block = QTAILQ_FIRST(&ram_list.blocks); > complete_round = true; > ram_bulk_stage = false; > + if (migrate_use_xbzrle()) { > + /* If xbzrle is on, stop using the data compression at this > + * point. In theory, xbzrle can do better than compression. > + */ > + flush_compressed_data(f); > + compression_switch_on = false; > + } > } > } else { > - if (migrate_use_compression()) { > + if (compression_switch_on && migrate_use_compression()) { > bytes_sent = ram_save_compressed_page(f, block, offset, > last_stage); > } else { > -- > 1.9.1 > -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK