From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53332) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1eOKa2-0003G4-43 for qemu-devel@nongnu.org; Mon, 11 Dec 2017 04:40:59 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1eOKZz-0007BH-1i for qemu-devel@nongnu.org; Mon, 11 Dec 2017 04:40:58 -0500 Received: from mx1.redhat.com ([209.132.183.28]:50650) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1eOKZy-0007AR-OO for qemu-devel@nongnu.org; Mon, 11 Dec 2017 04:40:54 -0500 Received: from smtp.corp.redhat.com (int-mx03.intmail.prod.int.phx2.redhat.com [10.5.11.13]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id DE0E3C0587E1 for ; Mon, 11 Dec 2017 09:40:53 +0000 (UTC) From: Juan Quintela In-Reply-To: <20171017145151.GD2326@work-vm> (David Alan Gilbert's message of "Tue, 17 Oct 2017 15:51:52 +0100") References: <20171004104636.7963-1-quintela@redhat.com> <20171004104636.7963-12-quintela@redhat.com> <20171017145151.GD2326@work-vm> Reply-To: quintela@redhat.com Date: Mon, 11 Dec 2017 10:40:49 +0100 Message-ID: <87609dr49q.fsf@secure.laptop> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Qemu-devel] [PATCH v9 11/12] migration: Flush receive queue List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Dr. David Alan Gilbert" Cc: qemu-devel@nongnu.org, lvivier@redhat.com, peterx@redhat.com "Dr. David Alan Gilbert" wrote: > * Juan Quintela (quintela@redhat.com) wrote: >> +/* We are getting low on pages flags, so we start using combinations >> + When we need to flush a page, we sent it as >> + RAM_SAVE_FLAG_MULTIFD_PAGE | RAM_SAVE_FLAG_COMPRESS_PAGE >> + We don't allow that combination >> +*/ >> +#define RAM_SAVE_FLAG_MULTIFD_SYNC \ >> + (RAM_SAVE_FLAG_MULTIFD_PAGE | RAM_SAVE_FLAG_ZERO) > > Good that's better than last time; note you're using FLAG_ZERO where > the comment says COMPRESS_PAGE and the commit message says COMPRESSED. Fixed. > >> + >> static inline bool is_zero_range(uint8_t *p, uint64_t size) >> { >> return buffer_is_zero(p, size); >> @@ -194,6 +202,9 @@ struct RAMState { >> uint64_t iterations_prev; >> /* Iterations since start */ >> uint64_t iterations; >> + /* Indicates if we have synced the bitmap and we need to assure that >> + target has processeed all previous pages */ >> + bool multifd_needs_flush; >> /* number of dirty bits in the bitmap */ >> uint64_t migration_dirty_pages; >> /* protects modification of the bitmap */ >> @@ -614,9 +625,11 @@ struct MultiFDRecvParams { >> QIOChannel *c; >> QemuSemaphore ready; >> QemuSemaphore sem; >> + QemuCond cond_sync; >> QemuMutex mutex; >> /* proteced by param mutex */ >> bool quit; >> + bool sync; >> multifd_pages_t pages; >> bool done; >> }; >> @@ -669,6 +682,7 @@ int multifd_load_cleanup(Error **errp) >> qemu_thread_join(&p->thread); >> qemu_mutex_destroy(&p->mutex); >> qemu_sem_destroy(&p->sem); >> + qemu_cond_destroy(&p->cond_sync); >> socket_recv_channel_destroy(p->c); >> g_free(p->name); >> p->name = NULL; >> @@ -707,6 +721,10 @@ static void *multifd_recv_thread(void *opaque) >> return NULL; >> } >> p->done = true; >> + if (p->sync) { >> + qemu_cond_signal(&p->cond_sync); >> + p->sync = false; >> + } >> qemu_mutex_unlock(&p->mutex); >> qemu_sem_post(&p->ready); >> continue; >> @@ -752,9 +770,11 @@ void multifd_new_channel(QIOChannel *ioc) >> qemu_mutex_init(&p->mutex); >> qemu_sem_init(&p->sem, 0); >> qemu_sem_init(&p->ready, 0); >> + qemu_cond_init(&p->cond_sync); >> p->quit = false; >> p->id = msg.id; >> p->done = false; >> + p->sync = false; >> multifd_init_pages(&p->pages); >> p->c = ioc; >> multifd_recv_state->count++; >> @@ -819,6 +839,27 @@ static void multifd_recv_page(uint8_t *address, uint16_t fd_num) >> qemu_sem_post(&p->sem); >> } >> >> +static int multifd_flush(void) >> +{ >> + int i, thread_count; >> + >> + if (!migrate_use_multifd()) { >> + return 0; >> + } >> + thread_count = migrate_multifd_channels(); >> + for (i = 0; i < thread_count; i++) { >> + MultiFDRecvParams *p = &multifd_recv_state->params[i]; >> + >> + qemu_mutex_lock(&p->mutex); >> + while (!p->done) { >> + p->sync = true; >> + qemu_cond_wait(&p->cond_sync, &p->mutex); >> + } >> + qemu_mutex_unlock(&p->mutex); >> + } >> + return 0; >> +} pD> > I wonder if we need some way of terminating this on error > (e.g. if terminate_multifd_recev_threads is called for an error > case). It could be, I have to think about this. Later, Juan.