On Wednesday, September 10, 2014 4:27 AM, Ian Abbott wrote:
> On 2014-09-10 00:15, H Hartley Sweeten wrote:
>> The cmd->stop_src is used to determine how an async command is terminated.
>> Typically the following sources are used:
>>
>> TRIG_NONE: command must be stopped with a (*cancel) operation
>> cmd->stop_arg is always 0
>> TRIG_COUNT: command terminates after cmd->stop_arg scans
>> TRIG_EXT: command terminates with an external signal (not common,
>> cmd->stop_arg should define the signal, i.e. channel,
>> that stops the command)
>>
>> For TRIG_COUNT sources, some of the drivers currently allow a stop_arg == 0.
>> This causes the (*do_cmd) to immediately terminate the command with a
>> COMEDI_CB_EOA, indicating an "empty acquisition". This is not really useful
>> and some of the drivers actually enable the hardware enough that interrupts
>> are generated but no data is actually returned to the user.
>>
>> Remove all the "empty acquisition" code and make sure the (*do_cmdtest) for
>> all drivers validates the cmd->stop_arg is >= 1 when the cmd->stop_src is
>> TRIG_COUNT.
>>
>> Do a bit of additional cleanup regarding the cmd->stop_{src,arg} usage in the
>> drivers.
>
> I'm not keen on patch 07 and think it should be dropped. That makes the
> description of patch 08 slightly wrong, so maybe that should be dropped
> too (it's only a minor improvement).
>
> For the others PATCHES 01-06,09-34:
> Reviewed-by: Ian Abbott <abbotti@xxxxxxxxx>
Greg,
Please drop patches 07 and 08 per Ian's request. I will address patch 08 after
this series is merged.
I you prefer I can repost this series as v2 with the two patches dropped.
Regards,
HaRtley
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
