~stoth/cx23885-dev Changelog

cx25840: Improvements to the cx23885/7/8 chip detection mechanism.

Sun, 27 Sep 2009 14:54:47 -0400

cx25840: Improvements to the cx23885/7/8 chip detection mechanism.

From: Steven Toth

The prior mechanism wasn't reliable with the cx23887. This new mechanism
tests correctly against cx23885, 7 and 8 based products.

Priority: normal

Signed-off-by: Steven Toth

cx23885: Enable HVR-1850 CX23888 A/V core to get VID_CLK running for IR

Sat, 26 Sep 2009 23:14:33 -0400

cx23885: Enable HVR-1850 CX23888 A/V core to get VID_CLK running for IR

From: Andy Walls

For the IR controller on the CX23888 to work for the HVR-1850 and in
general, the A/V core must be initialized to get the VID_CLK going.
The VID_CLK is the timing reference for the CX23888's integrated
IR controller.

Priority: normal

Signed-off-by: Andy Walls

cx25840: Init PLLs properly for CX2388[578] A/V cores

Sat, 26 Sep 2009 23:00:48 -0400

cx25840: Init PLLs properly for CX2388[578] A/V cores

From: Andy Walls

The SYS and AUX PLLs need to be initialized to different values based on
the chip: CX23885, CX23887, CX23888, as each uses a different crystal
frequency: 28.6363 MHz, 25.0 MHz, 50.0 MHz.

Priority: normal

Signed-off-by: Andy Walls

cx25840: Separate set_audclk_freq functionality for the different chips

Sat, 26 Sep 2009 22:54:20 -0400

cx25840: Separate set_audclk_freq functionality for the different chips

From: Andy Walls

Separate out the set_audclk_freq() function into separate functions for the
four families of cores. These cores all use slightly different sample clock
schemes and may be assuming slightly (+/- 3 Hz) different reference frequencies.
The code resuse was not worth the maintenance and testing headache of have all
chips use the same function peppered with conditional logic.

Added comments on how PLL and SRC parameters values are computed. Fixed a few
bugs related to the shared code having a large number of conditional statements.
Noted inconsistencies with FIXME in the comments.

This is done in preparation for getting the CX2388[578] PLL/clock setting logic
cleaned up for CX23888 analog video and IR (which need the VID PLL set right).

Priority: normal

Signed-off-by: Andy Walls

cx25840: Convert chip/core family checks to static inline functions

Sat, 26 Sep 2009 22:47:21 -0400

cx25840: Convert chip/core family checks to static inline functions

From: Andy Walls

Change logic to check for various chip or core families to inline functions.
Checks for specific chips should be made against the state->id field now. This
is in preparation for chip/core specific code for setting up PLLs for the
CX2388[578] family of cores, that all run with different crystal frequencies.

Priority: normal

Signed-off-by: Andy Walls

cx25840: Improve detection of CX2388[578] A/V cores

Sat, 26 Sep 2009 22:32:54 -0400

cx25840: Improve detection of CX2388[578] A/V cores

From: Andy Walls

Add improved logic to detect the exact CX2388[578] A/V core that is being
probed. Also cleaned up detection and logging for CX2310[12], CX2583[67],
and CX2584[0123] cores and chips.

Also added code to identify a CX2388[578] A/V decoder core that is not
responding properly. Typical symptoms include registers 0x00-0xff
responding properly but all other registers returning the same value
(0x13 and 0x5 have been observed). This state will cause proper detection of
'885 vs. '887 vs. '888 to fail and the chip won't respond to get configured
properly anyway. I have no method of reseting the core to a working state at
this time; but I didn't try too hard to work one out either. The problem
likely only occurs in development. I suspect configuring the SYS PLL VCO to
oscillate too slowly (286.3 MHz?) before post divide may be the root cause,
when encountered.

Priority: normal

Signed-off-by: Andy Walls

cx23885: Add skeleton v4l2_subdev for the CX23888 integrated IR controller

Sat, 26 Sep 2009 22:17:30 -0400

cx23885: Add skeleton v4l2_subdev for the CX23888 integrated IR controller

From: Andy Walls

This change adds a skeletal implementation of a v4l2_subdevice to provide
encapsulation and abstraction of the CX23888's integrated consumer infrared
controller. This change also instantiates the cx23888_ir subdev for the
HVR-1850 which has IR hardware physically wired up to a CX23888.

The cx23888_ir subdev code is being written with long-term objectives to:
1. port it to the cx25840 module for the CX2584x, CX2583x, CX23885, & CX231xx
IR controllers
2. possibly port it to the cx18 module for the CX23418 IR controller
3. have the IR subdevice accessed abstractly in the cx23885 module, so the
driver can ignore the difference between the CX23885 and CX23888.

Priority: normal

Signed-off-by: Andy Walls

cx23885: Fix support for v4l2-dbg access to CX2388[578] and CX23417 regs

Sat, 26 Sep 2009 21:50:44 -0400

cx23885: Fix support for v4l2-dbg access to CX2388[578] and CX23417 regs

From: Andy Walls

This changes corrects the ioctl() operations for both the CX2388[578] analog
video and MPEG video device nodes to properly and consistently support
VIDIOC_G_CHIP_IDENT, VIDIOC_DBG_G_REGISTER and VIDIOC_DBG_S_REGISTER ioctl()s.

This caused some ioctl() support routines to be broken out into a separate
source file.

Now v4l2-dbg can be used to manipulate CX2388[578] and CX23417 registers
including the CX2388[57] functions handled by the cx25840 module.

This was done in anticipation of developing a new v4l2_subdev for the
integrated IR controller of the CX23888.

Priority: normal

Signed-off-by: Andy Walls

v4l2-chip-ident: Add ID's needed for the cx23885 and cx25840 modules

Sat, 26 Sep 2009 21:07:10 -0400

v4l2-chip-ident: Add ID's needed for the cx23885 and cx25840 modules

From: Andy Walls

Add identifiers for CX2388[578] chips, CX2310[012] chips, integrated
A/V decoders cores, integrated IR controller core, and the CX23417
MPEG encoder. The cx23885 module and cx25840 module will use these
identifiers in upcoming changes.

Priority: normal

Signed-off-by: Andy Walls

dib0700: Be less verbose when an IR key is pressed

Sat, 26 Sep 2009 13:45:03 -0300

dib0700: Be less verbose when an IR key is pressed

From: Mauro Carvalho Chehab

Priority: normal

Signed-off-by: Mauro Carvalho Chehab