From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([209.51.188.92]:43727) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1h1UxI-0006Wz-N4 for qemu-devel@nongnu.org; Wed, 06 Mar 2019 06:43:25 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1h1UxH-0005K0-Or for qemu-devel@nongnu.org; Wed, 06 Mar 2019 06:43:24 -0500 Received: from mx1.redhat.com ([209.132.183.28]:50048) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1h1UxH-0005JV-EP for qemu-devel@nongnu.org; Wed, 06 Mar 2019 06:43:23 -0500 From: "Dr. David Alan Gilbert (git)" Date: Wed, 6 Mar 2019 11:42:21 +0000 Message-Id: <20190306114227.9125-17-dgilbert@redhat.com> In-Reply-To: <20190306114227.9125-1-dgilbert@redhat.com> References: <20190306114227.9125-1-dgilbert@redhat.com> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Subject: [Qemu-devel] [PULL 16/22] migration: API to clear bits of guest free pages from the dirty bitmap List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org, quintela@redhat.com, peterx@redhat.com, marcel.apfelbaum@gmail.com, wei.w.wang@intel.com, yury-kotov@yandex-team.ru, chen.zhang@intel.com Cc: eblake@redhat.com, armbru@redhat.com From: Wei Wang This patch adds an API to clear bits corresponding to guest free pages from the dirty bitmap. Spilt the free page block if it crosses the QEMU RAMBlock boundary. Signed-off-by: Wei Wang CC: Dr. David Alan Gilbert CC: Juan Quintela CC: Michael S. Tsirkin CC: Peter Xu Reviewed-by: Peter Xu Message-Id: <1544516693-5395-5-git-send-email-wei.w.wang@intel.com> Signed-off-by: Dr. David Alan Gilbert --- include/migration/misc.h | 2 ++ migration/ram.c | 47 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 49 insertions(+) diff --git a/include/migration/misc.h b/include/migration/misc.h index 6f9df74436..81ee347e35 100644 --- a/include/migration/misc.h +++ b/include/migration/misc.h @@ -14,12 +14,14 @@ #ifndef MIGRATION_MISC_H #define MIGRATION_MISC_H =20 +#include "exec/cpu-common.h" #include "qemu/notify.h" #include "qapi/qapi-types-net.h" =20 /* migration/ram.c */ =20 void ram_mig_init(void); +void qemu_guest_free_page_hint(void *addr, size_t len); =20 /* migration/block.c */ =20 diff --git a/migration/ram.c b/migration/ram.c index 0747873ca9..32c0dbb98a 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -3181,6 +3181,53 @@ static void ram_state_resume_prepare(RAMState *rs,= QEMUFile *out) trace_ram_state_resume_prepare(pages); } =20 +/* + * This function clears bits of the free pages reported by the caller fr= om the + * migration dirty bitmap. @addr is the host address corresponding to th= e + * start of the continuous guest free pages, and @len is the total bytes= of + * those pages. + */ +void qemu_guest_free_page_hint(void *addr, size_t len) +{ + RAMBlock *block; + ram_addr_t offset; + size_t used_len, start, npages; + MigrationState *s =3D migrate_get_current(); + + /* This function is currently expected to be used during live migrat= ion */ + if (!migration_is_setup_or_active(s->state)) { + return; + } + + for (; len > 0; len -=3D used_len, addr +=3D used_len) { + block =3D qemu_ram_block_from_host(addr, false, &offset); + if (unlikely(!block || offset >=3D block->used_length)) { + /* + * The implementation might not support RAMBlock resize duri= ng + * live migration, but it could happen in theory with future + * updates. So we add a check here to capture that case. + */ + error_report_once("%s unexpected error", __func__); + return; + } + + if (len <=3D block->used_length - offset) { + used_len =3D len; + } else { + used_len =3D block->used_length - offset; + } + + start =3D offset >> TARGET_PAGE_BITS; + npages =3D used_len >> TARGET_PAGE_BITS; + + qemu_mutex_lock(&ram_state->bitmap_mutex); + ram_state->migration_dirty_pages -=3D + bitmap_count_one_with_offset(block->bmap, start, n= pages); + bitmap_clear(block->bmap, start, npages); + qemu_mutex_unlock(&ram_state->bitmap_mutex); + } +} + /* * Each of ram_save_setup, ram_save_iterate and ram_save_complete has * long-running RCU critical section. When rcu-reclaims in the code --=20 2.20.1