Since making that patch a long time ago I have been informed of a
problem that reportedly disappeared when the patch was reverted.
Apparently SDMA engine got locked up and sdma_run_channel0() logged
"Timeout waiting for CH0 ready". Three channels were used in the test.
The reporter suspected this might have happened because rescheduling is
forced no matter which channel is active.
I didn't investigate the problem but I suggested that a better
alternative to forcing reschedule might be to wait for the channel to be
done (either SDMA is in sleep state or the current channel is not the
one that was just disabled). I don't know if this was tried. It looks
like there have been changes to dmaengine (transfer termination
synchronization support) that might make that easier now.
Hannu
--
To unsubscribe from this list: send the line "unsubscribe dmaengine" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
