diff options
author | Jan Beulich <JBeulich@suse.com> | 2016-05-04 16:02:36 +0300 |
---|---|---|
committer | David Vrabel <david.vrabel@citrix.com> | 2016-05-04 18:37:01 +0300 |
commit | 27e0e6385377c4dc68a4ddaf1a35a2dfa951f3c5 (patch) | |
tree | 940f070cecbb3f335be0c1f976a472c990ebf4ab /drivers/xen | |
parent | dfd74a1edfaba5864276a2859190a8d242d18952 (diff) | |
download | linux-27e0e6385377c4dc68a4ddaf1a35a2dfa951f3c5.tar.xz |
xen/evtchn: fix ring resize when binding new events
The copying of ring data was wrong for two cases: For a full ring
nothing got copied at all (as in that case the canonicalized producer
and consumer indexes are identical). And in case one or both of the
canonicalized (after the resize) indexes would point into the second
half of the buffer, the copied data ended up in the wrong (free) part
of the new buffer. In both cases uninitialized data would get passed
back to the caller.
Fix this by simply copying the old ring contents twice: Once to the
low half of the new buffer, and a second time to the high half.
This addresses the inability to boot a HVM guest with 64 or more
vCPUs. This regression was caused by 8620015499101090 (xen/evtchn:
dynamically grow pending event channel ring).
Reported-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: <stable@vger.kernel.org> # 4.4+
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Diffstat (limited to 'drivers/xen')
-rw-r--r-- | drivers/xen/evtchn.c | 20 |
1 files changed, 8 insertions, 12 deletions
diff --git a/drivers/xen/evtchn.c b/drivers/xen/evtchn.c index 38272ad24551..f4edd6df3df2 100644 --- a/drivers/xen/evtchn.c +++ b/drivers/xen/evtchn.c @@ -316,7 +316,6 @@ static int evtchn_resize_ring(struct per_user_data *u) { unsigned int new_size; evtchn_port_t *new_ring, *old_ring; - unsigned int p, c; /* * Ensure the ring is large enough to capture all possible @@ -346,20 +345,17 @@ static int evtchn_resize_ring(struct per_user_data *u) /* * Copy the old ring contents to the new ring. * - * If the ring contents crosses the end of the current ring, - * it needs to be copied in two chunks. + * To take care of wrapping, a full ring, and the new index + * pointing into the second half, simply copy the old contents + * twice. * * +---------+ +------------------+ - * |34567 12| -> | 1234567 | - * +-----p-c-+ +------------------+ + * |34567 12| -> |34567 1234567 12| + * +-----p-c-+ +-------c------p---+ */ - p = evtchn_ring_offset(u, u->ring_prod); - c = evtchn_ring_offset(u, u->ring_cons); - if (p < c) { - memcpy(new_ring + c, u->ring + c, (u->ring_size - c) * sizeof(*u->ring)); - memcpy(new_ring + u->ring_size, u->ring, p * sizeof(*u->ring)); - } else - memcpy(new_ring + c, u->ring + c, (p - c) * sizeof(*u->ring)); + memcpy(new_ring, old_ring, u->ring_size * sizeof(*u->ring)); + memcpy(new_ring + u->ring_size, old_ring, + u->ring_size * sizeof(*u->ring)); u->ring = new_ring; u->ring_size = new_size; |