On (Fri) 26 Feb 2016 [17:58:34], Denis V. Lunev wrote:
> On 02/26/2016 04:15 PM, Amit Shah wrote:
> >On (Fri) 26 Feb 2016 [12:23:01], Peter Maydell wrote:
> >>Hi. I'm afraid this causes make check to fail:
> >Oops, I should've caught that.
> >
> >> /i386/ahci/flush/migrate:
> >> /i386/ahci/migrate/sanity:
> >> /i386/ahci/migrate/dma/simple:
> >> /i386/ahci/migrate/dma/halted:
> >> /i386/ahci/migrate/ncq/simple:
> >> /i386/ahci/migrate/ncq/halted:
> >>Broken pipe
> >>FAIL
> >>GTester: last random seed: R02S34da744548f394684dbb7b6d11e14043
> >>(pid=56820)
> >>
> >>(failure seen on ppc64be, x86-64 linux, aarch64, but not aarch32 or OSX).
> >AHCI tests are failing due to Denis's patches. I'm checking why.
> >
> > Amit
> oops. Interesting :(
Right - use-after-free; this on top of your patch fixes it.
diff --git a/migration/migration.c b/migration/migration.c
index f82fdf6..0129d9f 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -366,9 +366,8 @@ static void process_incoming_migration_bh(void *opaque)
*/
migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
MIGRATION_STATUS_COMPLETED);
- migration_incoming_state_destroy();
-
qemu_bh_delete(mis->bh);
+ migration_incoming_state_destroy();
}
static void process_incoming_migration_co(void *opaque)
Amit
