| Message ID | 146239057139.11271.9011797645454781543.stgit@bahia.huguette.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, 04 May 2016 21:44:19 +0200 Greg Kurz <gkurz@linux.vnet.ibm.com> wrote: > QEMU has currently two ways to prevent migration to occur: > - migration blocker when it depends on runtime state > - VMStateDescription.unmigratable when migration is not supported at all > > This patch gathers all the logic into a single function to be called from > both the savevm and the migrate paths. > > This fixes a bug with 9p, at least, where savevm would succeed and the > following would happen in the guest after loadvm: > > $ ls /host > ls: cannot access /host: Protocol error > > With this patch: > > (qemu) savevm foo > Migration is disabled when VirtFS export path '/' is mounted in the guest > using mount_tag 'host' > > Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> > --- > Ping ? > Since there has been some activity on 9p again since a few weeks, maybe it > is not too late for 2.6, if everybody agrees the fix is ok ? > > include/migration/migration.h | 1 + > migration/migration.c | 21 +++++++++++++++------ > migration/savevm.c | 2 +- > 3 files changed, 17 insertions(+), 7 deletions(-) > > diff --git a/include/migration/migration.h b/include/migration/migration.h > index ac2c12c2a518..9e36a97fc5e5 100644 > --- a/include/migration/migration.h > +++ b/include/migration/migration.h > @@ -210,6 +210,7 @@ int migrate_fd_close(MigrationState *s); > void add_migration_state_change_notifier(Notifier *notify); > void remove_migration_state_change_notifier(Notifier *notify); > MigrationState *migrate_init(const MigrationParams *params); > +bool migration_is_blocked(Error **errp); > bool migration_in_setup(MigrationState *); > bool migration_has_finished(MigrationState *); > bool migration_has_failed(MigrationState *); > diff --git a/migration/migration.c b/migration/migration.c > index 991313a8629a..afd883b78179 100644 > --- a/migration/migration.c > +++ b/migration/migration.c > @@ -992,6 +992,20 @@ void qmp_migrate_incoming(const char *uri, Error **errp) > once = false; > } > > +bool migration_is_blocked(Error **errp) > +{ > + if (qemu_savevm_state_blocked(errp)) { > + return true; > + } > + > + if (migration_blockers) { > + *errp = error_copy(migration_blockers->data); > + return true; > + } > + > + return false; > +} > + > void qmp_migrate(const char *uri, bool has_blk, bool blk, > bool has_inc, bool inc, bool has_detach, bool detach, > Error **errp) > @@ -1014,12 +1028,7 @@ void qmp_migrate(const char *uri, bool has_blk, bool blk, > return; > } > > - if (qemu_savevm_state_blocked(errp)) { > - return; > - } > - > - if (migration_blockers) { > - *errp = error_copy(migration_blockers->data); > + if (migration_is_blocked(errp)) { > return; > } > > diff --git a/migration/savevm.c b/migration/savevm.c > index 16ba44379870..834664936dfc 100644 > --- a/migration/savevm.c > +++ b/migration/savevm.c > @@ -1169,7 +1169,7 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp) > MigrationState *ms = migrate_init(¶ms); > ms->to_dst_file = f; > > - if (qemu_savevm_state_blocked(errp)) { > + if (migration_is_blocked(errp)) { > return -EINVAL; > } > > >
On 04/05/2016 21:44, Greg Kurz wrote: > QEMU has currently two ways to prevent migration to occur: > - migration blocker when it depends on runtime state > - VMStateDescription.unmigratable when migration is not supported at all > > This patch gathers all the logic into a single function to be called from > both the savevm and the migrate paths. > > This fixes a bug with 9p, at least, where savevm would succeed and the > following would happen in the guest after loadvm: > > $ ls /host > ls: cannot access /host: Protocol error > > With this patch: > > (qemu) savevm foo > Migration is disabled when VirtFS export path '/' is mounted in the guest > using mount_tag 'host' > > Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> > --- > > Since there has been some activity on 9p again since a few weeks, maybe it > is not too late for 2.6, if everybody agrees the fix is ok ? The patch is good, but I think the subject should be savevm: fail if migration blockers are present The subject you used would be good for the include/migration and migration/migration.c parts alone (where you introduce the function without any semantic change). Apart from this, Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Thanks, Paolo > include/migration/migration.h | 1 + > migration/migration.c | 21 +++++++++++++++------ > migration/savevm.c | 2 +- > 3 files changed, 17 insertions(+), 7 deletions(-) > > diff --git a/include/migration/migration.h b/include/migration/migration.h > index ac2c12c2a518..9e36a97fc5e5 100644 > --- a/include/migration/migration.h > +++ b/include/migration/migration.h > @@ -210,6 +210,7 @@ int migrate_fd_close(MigrationState *s); > void add_migration_state_change_notifier(Notifier *notify); > void remove_migration_state_change_notifier(Notifier *notify); > MigrationState *migrate_init(const MigrationParams *params); > +bool migration_is_blocked(Error **errp); > bool migration_in_setup(MigrationState *); > bool migration_has_finished(MigrationState *); > bool migration_has_failed(MigrationState *); > diff --git a/migration/migration.c b/migration/migration.c > index 991313a8629a..afd883b78179 100644 > --- a/migration/migration.c > +++ b/migration/migration.c > @@ -992,6 +992,20 @@ void qmp_migrate_incoming(const char *uri, Error **errp) > once = false; > } > > +bool migration_is_blocked(Error **errp) > +{ > + if (qemu_savevm_state_blocked(errp)) { > + return true; > + } > + > + if (migration_blockers) { > + *errp = error_copy(migration_blockers->data); > + return true; > + } > + > + return false; > +} > + > void qmp_migrate(const char *uri, bool has_blk, bool blk, > bool has_inc, bool inc, bool has_detach, bool detach, > Error **errp) > @@ -1014,12 +1028,7 @@ void qmp_migrate(const char *uri, bool has_blk, bool blk, > return; > } > > - if (qemu_savevm_state_blocked(errp)) { > - return; > - } > - > - if (migration_blockers) { > - *errp = error_copy(migration_blockers->data); > + if (migration_is_blocked(errp)) { > return; > } > > diff --git a/migration/savevm.c b/migration/savevm.c > index 16ba44379870..834664936dfc 100644 > --- a/migration/savevm.c > +++ b/migration/savevm.c > @@ -1169,7 +1169,7 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp) > MigrationState *ms = migrate_init(¶ms); > ms->to_dst_file = f; > > - if (qemu_savevm_state_blocked(errp)) { > + if (migration_is_blocked(errp)) { > return -EINVAL; > } > > > >
On Wed, 11 May 2016 13:59:19 +0200 Paolo Bonzini <pbonzini@redhat.com> wrote: > On 04/05/2016 21:44, Greg Kurz wrote: > > QEMU has currently two ways to prevent migration to occur: > > - migration blocker when it depends on runtime state > > - VMStateDescription.unmigratable when migration is not supported at all > > > > This patch gathers all the logic into a single function to be called from > > both the savevm and the migrate paths. > > > > This fixes a bug with 9p, at least, where savevm would succeed and the > > following would happen in the guest after loadvm: > > > > $ ls /host > > ls: cannot access /host: Protocol error > > > > With this patch: > > > > (qemu) savevm foo > > Migration is disabled when VirtFS export path '/' is mounted in the guest > > using mount_tag 'host' > > > > Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> > > --- > > > > Since there has been some activity on 9p again since a few weeks, maybe it > > is not too late for 2.6, if everybody agrees the fix is ok ? > > The patch is good, but I think the subject should be > > savevm: fail if migration blockers are present > > The subject you used would be good for the include/migration and > migration/migration.c parts alone (where you introduce the function > without any semantic change). > Yeah, I agree the subject you are suggesting is far more informative :) > Apart from this, > > Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> > Thanks for the review ! > Thanks, > > Paolo > Cheers. -- Greg > > include/migration/migration.h | 1 + > > migration/migration.c | 21 +++++++++++++++------ > > migration/savevm.c | 2 +- > > 3 files changed, 17 insertions(+), 7 deletions(-) > > > > diff --git a/include/migration/migration.h b/include/migration/migration.h > > index ac2c12c2a518..9e36a97fc5e5 100644 > > --- a/include/migration/migration.h > > +++ b/include/migration/migration.h > > @@ -210,6 +210,7 @@ int migrate_fd_close(MigrationState *s); > > void add_migration_state_change_notifier(Notifier *notify); > > void remove_migration_state_change_notifier(Notifier *notify); > > MigrationState *migrate_init(const MigrationParams *params); > > +bool migration_is_blocked(Error **errp); > > bool migration_in_setup(MigrationState *); > > bool migration_has_finished(MigrationState *); > > bool migration_has_failed(MigrationState *); > > diff --git a/migration/migration.c b/migration/migration.c > > index 991313a8629a..afd883b78179 100644 > > --- a/migration/migration.c > > +++ b/migration/migration.c > > @@ -992,6 +992,20 @@ void qmp_migrate_incoming(const char *uri, Error **errp) > > once = false; > > } > > > > +bool migration_is_blocked(Error **errp) > > +{ > > + if (qemu_savevm_state_blocked(errp)) { > > + return true; > > + } > > + > > + if (migration_blockers) { > > + *errp = error_copy(migration_blockers->data); > > + return true; > > + } > > + > > + return false; > > +} > > + > > void qmp_migrate(const char *uri, bool has_blk, bool blk, > > bool has_inc, bool inc, bool has_detach, bool detach, > > Error **errp) > > @@ -1014,12 +1028,7 @@ void qmp_migrate(const char *uri, bool has_blk, bool blk, > > return; > > } > > > > - if (qemu_savevm_state_blocked(errp)) { > > - return; > > - } > > - > > - if (migration_blockers) { > > - *errp = error_copy(migration_blockers->data); > > + if (migration_is_blocked(errp)) { > > return; > > } > > > > diff --git a/migration/savevm.c b/migration/savevm.c > > index 16ba44379870..834664936dfc 100644 > > --- a/migration/savevm.c > > +++ b/migration/savevm.c > > @@ -1169,7 +1169,7 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp) > > MigrationState *ms = migrate_init(¶ms); > > ms->to_dst_file = f; > > > > - if (qemu_savevm_state_blocked(errp)) { > > + if (migration_is_blocked(errp)) { > > return -EINVAL; > > } > > > > > > > > >
On (Wed) 11 May 2016 [14:10:34], Greg Kurz wrote: > On Wed, 11 May 2016 13:59:19 +0200 > Paolo Bonzini <pbonzini@redhat.com> wrote: > > > On 04/05/2016 21:44, Greg Kurz wrote: > > > QEMU has currently two ways to prevent migration to occur: > > > - migration blocker when it depends on runtime state > > > - VMStateDescription.unmigratable when migration is not supported at all > > > > > > This patch gathers all the logic into a single function to be called from > > > both the savevm and the migrate paths. > > > > > > This fixes a bug with 9p, at least, where savevm would succeed and the > > > following would happen in the guest after loadvm: > > > > > > $ ls /host > > > ls: cannot access /host: Protocol error > > > > > > With this patch: > > > > > > (qemu) savevm foo > > > Migration is disabled when VirtFS export path '/' is mounted in the guest > > > using mount_tag 'host' > > > > > > Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> > > > --- > > > > > > Since there has been some activity on 9p again since a few weeks, maybe it > > > is not too late for 2.6, if everybody agrees the fix is ok ? > > > > The patch is good, but I think the subject should be > > > > savevm: fail if migration blockers are present > > > > The subject you used would be good for the include/migration and > > migration/migration.c parts alone (where you introduce the function > > without any semantic change). > > > > Yeah, I agree the subject you are suggesting is far more informative :) I'll apply this with the suggested subject. Please CC the migration maintainers for faster processing next time around. Thanks, Amit
On Mon, 23 May 2016 21:42:23 +0530 Amit Shah <amit.shah@redhat.com> wrote: > On (Wed) 11 May 2016 [14:10:34], Greg Kurz wrote: > > On Wed, 11 May 2016 13:59:19 +0200 > > Paolo Bonzini <pbonzini@redhat.com> wrote: > > > > > On 04/05/2016 21:44, Greg Kurz wrote: > > > > QEMU has currently two ways to prevent migration to occur: > > > > - migration blocker when it depends on runtime state > > > > - VMStateDescription.unmigratable when migration is not supported at all > > > > > > > > This patch gathers all the logic into a single function to be called from > > > > both the savevm and the migrate paths. > > > > > > > > This fixes a bug with 9p, at least, where savevm would succeed and the > > > > following would happen in the guest after loadvm: > > > > > > > > $ ls /host > > > > ls: cannot access /host: Protocol error > > > > > > > > With this patch: > > > > > > > > (qemu) savevm foo > > > > Migration is disabled when VirtFS export path '/' is mounted in the guest > > > > using mount_tag 'host' > > > > > > > > Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> > > > > --- > > > > > > > > Since there has been some activity on 9p again since a few weeks, maybe it > > > > is not too late for 2.6, if everybody agrees the fix is ok ? > > > > > > The patch is good, but I think the subject should be > > > > > > savevm: fail if migration blockers are present > > > > > > The subject you used would be good for the include/migration and > > > migration/migration.c parts alone (where you introduce the function > > > without any semantic change). > > > > > > > Yeah, I agree the subject you are suggesting is far more informative :) > > I'll apply this with the suggested subject. > > Please CC the migration maintainers for faster processing next time > around. > I had Cc'd Juan (who happens to be away) but I forgot your address... my bad :\ > Thanks, > > Amit > Cheers ! -- Greg
diff --git a/include/migration/migration.h b/include/migration/migration.h index ac2c12c2a518..9e36a97fc5e5 100644 --- a/include/migration/migration.h +++ b/include/migration/migration.h @@ -210,6 +210,7 @@ int migrate_fd_close(MigrationState *s); void add_migration_state_change_notifier(Notifier *notify); void remove_migration_state_change_notifier(Notifier *notify); MigrationState *migrate_init(const MigrationParams *params); +bool migration_is_blocked(Error **errp); bool migration_in_setup(MigrationState *); bool migration_has_finished(MigrationState *); bool migration_has_failed(MigrationState *); diff --git a/migration/migration.c b/migration/migration.c index 991313a8629a..afd883b78179 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -992,6 +992,20 @@ void qmp_migrate_incoming(const char *uri, Error **errp) once = false; } +bool migration_is_blocked(Error **errp) +{ + if (qemu_savevm_state_blocked(errp)) { + return true; + } + + if (migration_blockers) { + *errp = error_copy(migration_blockers->data); + return true; + } + + return false; +} + void qmp_migrate(const char *uri, bool has_blk, bool blk, bool has_inc, bool inc, bool has_detach, bool detach, Error **errp) @@ -1014,12 +1028,7 @@ void qmp_migrate(const char *uri, bool has_blk, bool blk, return; } - if (qemu_savevm_state_blocked(errp)) { - return; - } - - if (migration_blockers) { - *errp = error_copy(migration_blockers->data); + if (migration_is_blocked(errp)) { return; } diff --git a/migration/savevm.c b/migration/savevm.c index 16ba44379870..834664936dfc 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -1169,7 +1169,7 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp) MigrationState *ms = migrate_init(¶ms); ms->to_dst_file = f; - if (qemu_savevm_state_blocked(errp)) { + if (migration_is_blocked(errp)) { return -EINVAL; }
QEMU has currently two ways to prevent migration to occur: - migration blocker when it depends on runtime state - VMStateDescription.unmigratable when migration is not supported at all This patch gathers all the logic into a single function to be called from both the savevm and the migrate paths. This fixes a bug with 9p, at least, where savevm would succeed and the following would happen in the guest after loadvm: $ ls /host ls: cannot access /host: Protocol error With this patch: (qemu) savevm foo Migration is disabled when VirtFS export path '/' is mounted in the guest using mount_tag 'host' Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> --- Since there has been some activity on 9p again since a few weeks, maybe it is not too late for 2.6, if everybody agrees the fix is ok ? include/migration/migration.h | 1 + migration/migration.c | 21 +++++++++++++++------ migration/savevm.c | 2 +- 3 files changed, 17 insertions(+), 7 deletions(-)