fio/bdev: Increase available queue depth for message passing

For malloc backends without IOAT a message gets queued up
for every completion. This can quickly run the message
passing ring out of entries. Increase it's size
substantially.

Change-Id: Ib09a0887d3d843a768156c9221a4f43c6cb43a4d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/374353
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This commit is contained in:
Ben Walker 2017-08-15 14:14:54 -07:00 committed by Jim Harris
parent 264a611c49
commit 0d3a55b79e

View File

@ -189,7 +189,7 @@ spdk_fio_init_thread(struct thread_data *td)
fio_thread->td = td;
td->io_ops_data = fio_thread;
fio_thread->ring = spdk_ring_create(SPDK_RING_TYPE_MP_SC, 128, SPDK_ENV_SOCKET_ID_ANY);
fio_thread->ring = spdk_ring_create(SPDK_RING_TYPE_MP_SC, 4096, SPDK_ENV_SOCKET_ID_ANY);
if (!fio_thread->ring) {
SPDK_ERRLOG("failed to allocate ring\n");
free(fio_thread);