From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:42186) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cZcFI-0000sU-Qg for qemu-devel@nongnu.org; Fri, 03 Feb 2017 06:41:41 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cZcFF-0000dk-Kr for qemu-devel@nongnu.org; Fri, 03 Feb 2017 06:41:40 -0500 Received: from mx1.redhat.com ([209.132.183.28]:43156) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1cZcFF-0000cj-Bm for qemu-devel@nongnu.org; Fri, 03 Feb 2017 06:41:37 -0500 Received: from int-mx14.intmail.prod.int.phx2.redhat.com (int-mx14.intmail.prod.int.phx2.redhat.com [10.5.11.27]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 47E7C3D945 for ; Fri, 3 Feb 2017 11:41:37 +0000 (UTC) Date: Fri, 3 Feb 2017 11:41:32 +0000 From: "Dr. David Alan Gilbert" Message-ID: <20170203114132.GH3208@work-vm> References: <1485207141-1941-1-git-send-email-quintela@redhat.com> <1485207141-1941-17-git-send-email-quintela@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1485207141-1941-17-git-send-email-quintela@redhat.com> Subject: Re: [Qemu-devel] [PATCH 16/17] migration: [HACK]Transfer pages over new channels List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Juan Quintela Cc: qemu-devel@nongnu.org, amit.shah@redhat.com * Juan Quintela (quintela@redhat.com) wrote: > We switch for sending the page number to send real pages. > > [HACK] > How we calculate the bandwidth is beyond repair, there is a hack there > that would work for x86 and archs that have 4kb pages. > > If you are having a nice day just go to migration/ram.c and look at > acct_update_position(). Now you are depressed, right? > > Signed-off-by: Juan Quintela > --- > migration/migration.c | 15 +++++++++++---- > migration/ram.c | 25 +++++++++---------------- > 2 files changed, 20 insertions(+), 20 deletions(-) > > diff --git a/migration/migration.c b/migration/migration.c > index 1d62b91..cbbf2a3 100644 > --- a/migration/migration.c > +++ b/migration/migration.c > @@ -1839,7 +1839,8 @@ static void *migration_thread(void *opaque) > /* Used by the bandwidth calcs, updated later */ > int64_t initial_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME); > int64_t setup_start = qemu_clock_get_ms(QEMU_CLOCK_HOST); > - int64_t initial_bytes = 0; > + int64_t qemu_file_bytes = 0; > + int64_t multifd_pages = 0; > int64_t max_size = 0; > int64_t start_time = initial_time; > int64_t end_time; > @@ -1923,9 +1924,14 @@ static void *migration_thread(void *opaque) > } > current_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME); > if (current_time >= initial_time + BUFFER_DELAY) { > - uint64_t transferred_bytes = qemu_ftell(s->to_dst_file) - > - initial_bytes; > uint64_t time_spent = current_time - initial_time; > + uint64_t qemu_file_bytes_now = qemu_ftell(s->to_dst_file); > + uint64_t multifd_pages_now = multifd_mig_pages_transferred(); > + /* Hack ahead. Why the hell we don't have a function to now the > + target_page_size. Hard coding it to 4096 */ > + uint64_t transferred_bytes = > + (qemu_file_bytes_now - qemu_file_bytes) + > + (multifd_pages_now - multifd_pages) * 4096; The problem was that it was target specific so the macros aren't allowed in migration.c that's target independent; but you'll see I added the qemu_target_page_bits() function, so you can do 1ul << qemu_target_page_bits() Still, I don't quite understand why we need the transferred_bytes hack; can't you just accumulate them as you pass them to multifd_send_page ? > double bandwidth = (double)transferred_bytes / time_spent; > max_size = bandwidth * s->parameters.downtime_limit; > > @@ -1942,7 +1948,8 @@ static void *migration_thread(void *opaque) > > qemu_file_reset_rate_limit(s->to_dst_file); > initial_time = current_time; > - initial_bytes = qemu_ftell(s->to_dst_file); > + qemu_file_bytes = qemu_file_bytes_now; > + multifd_pages = multifd_pages_now; > } > if (qemu_file_rate_limit(s->to_dst_file)) { > /* usleep expects microseconds */ > diff --git a/migration/ram.c b/migration/ram.c > index 95af694..28d099f 100644 > --- a/migration/ram.c > +++ b/migration/ram.c > @@ -441,9 +441,9 @@ static void *multifd_send_thread(void *opaque) > > for(i=0; i < num; i++) { > if (qio_channel_write(params->c, > - (const char *)¶ms->pages.address[i], > - sizeof(uint8_t *), &error_abort) > - != sizeof(uint8_t*)) { > + (const char *)params->pages.address[i], > + TARGET_PAGE_SIZE, &error_abort) > + != TARGET_PAGE_SIZE) { > /* Shuoudn't ever happen */ > exit(-1); > } > @@ -608,7 +608,6 @@ QemuCond multifd_recv_cond; > static void *multifd_recv_thread(void *opaque) > { > MultiFDRecvParams *params = opaque; > - uint8_t *recv_address; > char start; > > qio_channel_read(params->c, &start, 1, &error_abort); > @@ -629,20 +628,13 @@ static void *multifd_recv_thread(void *opaque) > > for(i = 0; i < num; i++) { > if (qio_channel_read(params->c, > - (char *)&recv_address, > - sizeof(uint8_t*), &error_abort) > - != sizeof(uint8_t *)) { > + (char *)params->pages.address[i], > + TARGET_PAGE_SIZE, &error_abort) > + != TARGET_PAGE_SIZE) { > /* shouldn't ever happen */ > exit(-1); > } > - if (recv_address != params->pages.address[i]) { > - printf("We received %p what we were expecting %p (%d)\n", > - recv_address, > - params->pages.address[i], i); > - exit(-1); > - } > } > - > qemu_mutex_lock(&multifd_recv_mutex); > params->done = true; > qemu_cond_signal(&multifd_recv_cond); > @@ -1195,8 +1187,10 @@ static int ram_multifd_page(QEMUFile *f, PageSearchStatus *pss, > save_page_header(f, block, offset | RAM_SAVE_FLAG_MULTIFD_PAGE); > fd_num = multifd_send_page(p, migration_dirty_pages == 1); > qemu_put_be16(f, fd_num); > + if (fd_num != UINT16_MAX) { > + qemu_fflush(f); > + } > *bytes_transferred += 2; /* size of fd_num */ > - qemu_put_buffer(f, p, TARGET_PAGE_SIZE); > *bytes_transferred += TARGET_PAGE_SIZE; > pages = 1; > acct_info.norm_pages++; > @@ -3017,7 +3011,6 @@ static int ram_load(QEMUFile *f, void *opaque, int version_id) > case RAM_SAVE_FLAG_MULTIFD_PAGE: > fd_num = qemu_get_be16(f); > multifd_recv_page(host, fd_num); I think also at some point you need to validate that fd_num that's come off the wire. Dave > - qemu_get_buffer(f, host, TARGET_PAGE_SIZE); > break; > > case RAM_SAVE_FLAG_EOS: > -- > 2.9.3 > -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK