Petri, please review that change.
queue_destroy_finalize(qe) calls odp_buffer_free(buf) if queue is
QUEUE_STATUS_DESTROYED.
So that that case there will be double free of that buffer.
Thanks,
Maxim.
On 06/22/15 11:19, Nicolas Morey-Chaisemartin wrote:
Ping
On 06/15/2015 03:30 PM, Nicolas Morey-Chaisemartin wrote:
Buffer double free on schedule terminate introduced by
295d894a3e170f9f30261dc272182fb5eebd64b0
Signed-off-by: Nicolas Morey-Chaisemartin <[email protected]>
---
This fixes an issue that can easily be reproduced by running the odp_scheduler
validation test.
Adding an assert allocator != ODP_FREEBUF in the odp_buffer_free function shows
a double free in the odp_schedule_term_global function.
From what I could gather, the issue comes from the qe cmd_ev being freed by
schedule_queue_destroy,
and then freed again by odp_buffer_free (as it is the command pulled from the
prio queue).
Running odp_buffer_free only if schedule_queue_destroy was not called fixes the
issue.
However I am not sure this won't cause mem leaks.
platform/linux-generic/odp_schedule.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/platform/linux-generic/odp_schedule.c
b/platform/linux-generic/odp_schedule.c
index 7fe42d7..087ee67 100644
--- a/platform/linux-generic/odp_schedule.c
+++ b/platform/linux-generic/odp_schedule.c
@@ -202,9 +202,8 @@ int odp_schedule_term_global(void)
if (num > 0)
ODP_ERR("Queue not empty\n");
- }
-
- odp_buffer_free(buf);
+ } else
+ odp_buffer_free(buf);
}
if (odp_queue_destroy(pri_q)) {
_______________________________________________
lng-odp mailing list
[email protected]
https://lists.linaro.org/mailman/listinfo/lng-odp
_______________________________________________
lng-odp mailing list
[email protected]
https://lists.linaro.org/mailman/listinfo/lng-odp
_______________________________________________
lng-odp mailing list
[email protected]
https://lists.linaro.org/mailman/listinfo/lng-odp
