[media] lirc_zilog: Update TODO list based on work completed and revised plans

Update the TODO.lirc_zilog based on what has been completed.  Also revised
the development plan for lirc_zilog to not try and split Tx/Rx for one IR
transceiver unit between lirc_zilog and ir-kbd-i2c, since that would be a
ref-counting nightmare.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
This commit is contained in:
Andy Walls 2011-02-17 21:32:06 -03:00 committed by Mauro Carvalho Chehab
parent 4933fc9d72
commit aa3f4e446e

View file

@ -1,34 +1,33 @@
1. Both ir-kbd-i2c and lirc_zilog provide support for RX events.
The 'tx_only' lirc_zilog module parameter will allow ir-kbd-i2c
and lirc_zilog to coexist in the kernel, if the user requires such a set-up.
However the IR unit will not work well without coordination between the
two modules. A shared mutex, for transceiver access locking, needs to be
supplied by bridge drivers, in struct IR_i2_init_data, to both ir-kbd-i2c
and lirc_zilog, before they will coexist usefully. This should be fixed
before moving out of staging.
1. Both ir-kbd-i2c and lirc_zilog provide support for RX events for
the chips supported by lirc_zilog. Before moving lirc_zilog out of staging:
2. References and locking need careful examination. For cx18 and ivtv PCI
cards, which are not easily "hot unplugged", the imperfect state of reference
counting and locking is acceptable if not correct. For USB connected units
like HD PVR, PVR USB2, HVR-1900, and HVR1950, the likelyhood of an Ooops on
unplug is probably great. Proper reference counting and locking needs to be
implemented before this module is moved out of staging.
a. ir-kbd-i2c needs a module parameter added to allow the user to tell
ir-kbd-i2c to ignore Z8 IR units.
3. The binding between hdpvr and lirc_zilog is currently disabled,
due to an OOPS reported a few years ago when both the hdpvr and cx18
drivers were loaded in his system. More details can be seen at:
http://www.mail-archive.com/linux-media@vger.kernel.org/msg09163.html
More tests need to be done, in order to fix the reported issue.
b. lirc_zilog should provide Rx key presses to the rc core like ir-kbd-i2c
does.
4. In addition to providing a shared mutex for transceiver access
locking, bridge drivers, if able, should provide a chip reset() callback
2. lirc_zilog module ref-counting need examination. It has not been
verified that cdev and lirc_dev will take the proper module references on
lirc_zilog to prevent removal of lirc_zilog when the /dev/lircN device node
is open.
(The good news is ref-counting of lirc_zilog internal structures appears to be
complete. Testing has shown the cx18 module can be unloaded out from under
irw + lircd + lirc_dev, with the /dev/lirc0 device node open, with no adverse
effects. The cx18 module could then be reloaded and irw properly began
receiving button presses again and ir_send worked without error.)
3. Bridge drivers, if able, should provide a chip reset() callback
to lirc_zilog via struct IR_i2c_init_data. cx18 and ivtv already have routines
to perform Z8 chip resets via GPIO manipulations. This will allow lirc_zilog
to perform Z8 chip resets via GPIO manipulations. This would allow lirc_zilog
to bring the chip back to normal when it hangs, in the same places the
original lirc_pvr150 driver code does. This is not strictly needed, so it
is not required to move lirc_zilog out of staging.
5. Both lirc_zilog and ir-kbd-i2c support the Zilog Z8 for IR, as programmed
Note: Both lirc_zilog and ir-kbd-i2c support the Zilog Z8 for IR, as programmed
and installed on Hauppauge products. When working on either module, developers
must consider at least the following bridge drivers which mention an IR Rx unit
at address 0x71 (indicative of a Z8):