greybus: operation: fix response-cancellation race

Make sure the request handler has submitted the response before
cancelling it during operation cancellation.

This prevents cancelling not-yet-submitted messages. It currently also
avoids us ending up with an active message on a stalled connection (e.g.
due to E2EFC).

Note that the call to gb_operation_result_set() is now redundant but is
kept as a precaution to guarantee that a response has indeed been
allocated as part of response submission.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
This commit is contained in:
Johan Hovold 2015-07-14 15:43:34 +02:00 committed by Greg Kroah-Hartman
parent 3325a4ad71
commit 0eb8c11598
1 changed files with 8 additions and 4 deletions

View File

@ -928,10 +928,14 @@ void gb_connection_recv(struct gb_connection *connection,
void gb_operation_cancel(struct gb_operation *operation, int errno)
{
if (gb_operation_is_incoming(operation)) {
/* Cancel response if it has been allocated */
if (!gb_operation_result_set(operation, errno) &&
!gb_operation_is_unidirectional(operation)) {
gb_message_cancel(operation->response);
if (!gb_operation_is_unidirectional(operation)) {
/*
* Make sure the request handler has submitted the
* response before cancelling it.
*/
flush_work(&operation->work);
if (!gb_operation_result_set(operation, errno))
gb_message_cancel(operation->response);
}
} else {
if (gb_operation_result_set(operation, errno)) {