On Tue, May 19, 2015 at 12:42:03PM +0200, Takashi Iwai wrote:
At Tue, 19 May 2015 15:00:40 +0530, Vinod Koul wrote:
From: "Subhransu S. Prusty" subhransu.s.prusty@intel.com
Signed-off-by: Subhransu S. Prusty subhransu.s.prusty@intel.com Signed-off-by: Vinod Koul vinod.koul@intel.com
sound/soc/intel/common/sst-ipc.c | 29 ++++++++++++++++++++++++++++- sound/soc/intel/common/sst-ipc.h | 4 ++-- 2 files changed, 30 insertions(+), 3 deletions(-)
diff --git a/sound/soc/intel/common/sst-ipc.c b/sound/soc/intel/common/sst-ipc.c index a7699f35a8d2..a12c7bb08d3b 100644 --- a/sound/soc/intel/common/sst-ipc.c +++ b/sound/soc/intel/common/sst-ipc.c @@ -129,11 +129,31 @@ static int msg_empty_list_init(struct sst_generic_ipc *ipc) return -ENOMEM;
for (i = 0; i < IPC_EMPTY_LIST_SIZE; i++) {
ipc->msg[i].tx_data = kzalloc(ipc->tx_data_max_size, GFP_KERNEL);if (ipc->msg[i].tx_data == NULL)goto free_mem;ipc->msg[i].rx_data = kzalloc(ipc->rx_data_max_size, GFP_KERNEL);if (ipc->msg[i].rx_data == NULL) {kfree(ipc->msg[i].tx_data);goto free_mem;}- init_waitqueue_head(&ipc->msg[i].waitq); list_add(&ipc->msg[i].list, &ipc->empty_list);
Wouldn't this leave stale linked list entries at error?
Yes it will, but this is happening at device probe, so the probe will fail and device will no longer be usable. We do free the memory so shouldn't leak.