From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([209.51.188.92]:45193) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gyemE-0006Vc-84 for qemu-devel@nongnu.org; Tue, 26 Feb 2019 10:36:15 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gyem8-0002Ef-74 for qemu-devel@nongnu.org; Tue, 26 Feb 2019 10:36:14 -0500 From: Stefan Hajnoczi Date: Tue, 26 Feb 2019 15:35:49 +0000 Message-Id: <20190226153549.17867-1-stefanha@redhat.com> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Subject: [Qemu-devel] [PATCH] file-posix: add drop-cache=on|off option List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: qemu-block@nongnu.org, Max Reitz , Markus Armbruster , Kevin Wolf , Eric Blake , Stefan Hajnoczi , Neil Skrypuch Commit dd577a26ff03b6829721b1ffbbf9e7c411b72378 ("block/file-posix: implement bdrv_co_invalidate_cache() on Linux") introduced page cache invalidation so that cache.direct=3Doff live migration is safe on Linux. The invalidation takes a significant amount of time when the file is large and present in the page cache. Normally this is not the case for cross-host live migration but it can happen when migrating between QEMU processes on the same host. On same-host migration we don't need to invalidate pages for correctness anyway, so an option to skip page cache invalidation is useful. I investigated optimizing invalidation and detecting same-host migration, but both are hard to achieve so a user-visible option will suffice. As a bonus this option means that the cache invalidation feature will now be detectable by libvirt via QMP schema introspection. Suggested-by: Neil Skrypuch Signed-off-by: Stefan Hajnoczi --- qapi/block-core.json | 5 +++++ block/file-posix.c | 14 ++++++++++++++ 2 files changed, 19 insertions(+) diff --git a/qapi/block-core.json b/qapi/block-core.json index 2b8afbb924..d4cc3c4294 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -2807,6 +2807,10 @@ # @locking: whether to enable file locking. If set to 'auto', only e= nable # when Open File Descriptor (OFD) locking API is available # (default: auto, since 2.10) +# @drop-cache: invalidate page cache during live migration. This preve= nts +# stale data on the migration destination with cache.direc= t=3Doff. +# Currently only supported on Linux hosts. +# (default: on, since: 4.0) # @x-check-cache-dropped: whether to check that page cache was dropped o= n live # migration. May cause noticeable delays if the= image # file is large, do not use in production. @@ -2819,6 +2823,7 @@ '*pr-manager': 'str', '*locking': 'OnOffAuto', '*aio': 'BlockdevAioOptions', + '*drop-cache': 'bool', '*x-check-cache-dropped': 'bool' } } =20 ## diff --git a/block/file-posix.c b/block/file-posix.c index ba6ab62a38..7bb2c4762f 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -154,6 +154,7 @@ typedef struct BDRVRawState { bool page_cache_inconsistent:1; bool has_fallocate; bool needs_alignment; + bool drop_cache; bool check_cache_dropped; =20 PRManager *pr_mgr; @@ -162,6 +163,7 @@ typedef struct BDRVRawState { typedef struct BDRVRawReopenState { int fd; int open_flags; + bool drop_cache; bool check_cache_dropped; } BDRVRawReopenState; =20 @@ -422,6 +424,11 @@ static QemuOptsList raw_runtime_opts =3D { .type =3D QEMU_OPT_STRING, .help =3D "id of persistent reservation manager object (defa= ult: none)", }, + { + .name =3D "drop-cache", + .type =3D QEMU_OPT_BOOL, + .help =3D "invalidate page cache during live migration (defa= ult: on)", + }, { .name =3D "x-check-cache-dropped", .type =3D QEMU_OPT_BOOL, @@ -511,6 +518,7 @@ static int raw_open_common(BlockDriverState *bs, QDic= t *options, } } =20 + s->drop_cache =3D qemu_opt_get_bool(opts, "drop-cache", true); s->check_cache_dropped =3D qemu_opt_get_bool(opts, "x-check-cache-dr= opped", false); =20 @@ -869,6 +877,7 @@ static int raw_reopen_prepare(BDRVReopenState *state, goto out; } =20 + rs->drop_cache =3D qemu_opt_get_bool_del(opts, "drop-cache", true); rs->check_cache_dropped =3D qemu_opt_get_bool_del(opts, "x-check-cache-dropped", false); =20 @@ -946,6 +955,7 @@ static void raw_reopen_commit(BDRVReopenState *state) BDRVRawState *s =3D state->bs->opaque; Error *local_err =3D NULL; =20 + s->drop_cache =3D rs->drop_cache; s->check_cache_dropped =3D rs->check_cache_dropped; s->open_flags =3D rs->open_flags; =20 @@ -2531,6 +2541,10 @@ static void coroutine_fn raw_co_invalidate_cache(B= lockDriverState *bs, return; } =20 + if (!s->drop_cache) { + return; + } + if (s->open_flags & O_DIRECT) { return; /* No host kernel page cache */ } --=20 2.20.1