Hi Deepak,
On Wed, 2010-02-10 at 02:01 +0100, ext Deepak Chitriki wrote:
> This patch fixes possible recursive locking detection.The implementation
> in which the spinlock is acquired and released is rectified in WMD_MSG_Get()
> to avoid locking contention.
> Added SYNC_EnterCS() and SYNC_LeaveCS()in WMD_MSG_Get()function.
>
> Cc: Ameya Palande <ameya.palande@xxxxxxxxx>
> Cc: Omar Ramirez Luna <omar.ramirez@xxxxxx>
> Cc: Nishanth Menon <nm@xxxxxx>
>
> Signed-off-by: Deepak Chitriki <deepak.chitriki@xxxxxx>
> ---
> drivers/dsp/bridge/wmd/msg_sm.c | 4 ++++
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/dsp/bridge/wmd/msg_sm.c b/drivers/dsp/bridge/wmd/msg_sm.c
> index 50201e5..8faf5ad 100644
> --- a/drivers/dsp/bridge/wmd/msg_sm.c
> +++ b/drivers/dsp/bridge/wmd/msg_sm.c
> @@ -300,8 +300,10 @@ DSP_STATUS WMD_MSG_Get(struct MSG_QUEUE *hMsgQueue,
> if (LST_IsEmpty(hMsgQueue->msgUsedList))
> SYNC_ResetEvent(hMsgQueue->hSyncEvent);
> else {
> + (void)SYNC_LeaveCS(hMsgMgr->hSyncCS);
> NTFY_Notify(hMsgQueue->hNtfy,
> DSP_NODEMESSAGEREADY);
> + (void)SYNC_EnterCS(hMsgMgr->hSyncCS);
> SYNC_SetEvent(hMsgQueue->hSyncEvent);
> }
>
> @@ -352,8 +354,10 @@ DSP_STATUS WMD_MSG_Get(struct MSG_QUEUE *hMsgQueue,
> hMsgQueue->refCount--;
> /* Reset the event if there are still queued messages */
> if (!LST_IsEmpty(hMsgQueue->msgUsedList)) {
> + (void)SYNC_LeaveCS(hMsgMgr->hSyncCS);
> NTFY_Notify(hMsgQueue->hNtfy,
> DSP_NODEMESSAGEREADY);
> + (void)SYNC_EnterCS(hMsgMgr->hSyncCS);
> SYNC_SetEvent(hMsgQueue->hSyncEvent);
> }
> /* Exit critical section */
Can you explain the need of calling NTFY_Notify() in WMD_MSG_Get()?
I can see that the InputMsg calls NTFY_Notify() already!
Can we get rid of NTFY_Notify() from WMD_MSG_Get() all together?
Cheers,
Ameya.
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
