Re: [linux-sunxi] Mainline work status

2015-01-07 Thread wens Tsai
Hi,

On Thu, Jan 8, 2015 at 6:21 AM, Iain Paton ipat...@gmail.com wrote:
 On 24/12/14 15:29, wens Tsai wrote:

 - AXP20x (PEK and docs) series, originally by Carlo, posted v8
 - AXP20x regulator driver cleanup, finished and tested
 - AXP221 support, originally by Boris, finished and tested
   * Also enables WiFi on the Hummingbird A31
 - sunxi (sun[457]i) cpufreq support, finished and testing (CB/CB2 tested)
   * Only added support for the boards I own. It's just a matter of adding
 the regulator nodes with the proper constraints

 Hi,

 I've been testing your patches with a20 olinuxino lime2  a10 lime

 I notice your cpu opp data reduces the voltage down to 0.9v for the lowest
 frequency, but all of the regulator nodes you added for cubieboard cubietruck
 etc, have the lower limit set to 1v.  Did you test 0.9v on anything?

Not before. The assumption was that we provide a generic set of OPPs.
Board files then decide which ones are actually usable via regulator
constraints.

 The lime2 dts currently has 0.7v as the lower limit for dcdc2 and when it's
 dropped to an opp with cpu core voltage set to 0.9v the board starts crashing
 randomly very soon after.

I just tested my cubieboard2, and it crashes as soon as the voltage
is lowered to 0.9v by cpufrequtils.

0.9v is below the recommended voltage range for VDD-CPU, according to
the datasheet. Running my board at 144 MHz / 1.0v seems fine though.

 I'll post a patch shortly to raise the lime2 dcdc2 lower limit to 1v, but
 I think it's probably worth collecting some more data on how many boards
 will really be able to use that 0.9v setting.

I can increase the voltage setting for 144 MHz to 1.0v in the dtsi.

This brings out additional problems though. The clock driver doesn't
support re-calculating the dividers ATM. So when the cpu clock is
at 144 MHz, the ahb clocks run at 24 MHz, which IIRC is too slow for
USB.

Maybe I just remove it from the OPP set for the time being.

 With the exception of altering the dcdc2 limits, I've not encountered any
 other problems.

Thanks for the feedback!

ChenYu

-- 
You received this message because you are subscribed to the Google Groups 
linux-sunxi group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to linux-sunxi+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


[linux-sunxi] Mainline work status

2014-12-24 Thread wens Tsai
Hi everyone,

As some of you might have noticed, I've sent out patches for A80 MMC and AXP20x.
In addition, I've picked up Boris' AXP221 patches to resend after the
AXP patches
are merged. The following is a list of things I'm working on or
waiting to submit:

- AXP20x (PEK and docs) series, originally by Carlo, posted v8
- AXP20x regulator driver cleanup, finished and tested
- AXP221 support, originally by Boris, finished and tested
  * Also enables WiFi on the Hummingbird A31
- sunxi (sun[457]i) cpufreq support, finished and testing (CB/CB2 tested)
  * Only added support for the boards I own. It's just a matter of adding
the regulator nodes with the proper constraints
- sun6i cpufreq support, WiP
- sun8i cpufreq support, minimal work only
- sun9i mmc support, posted v2
- sun9i usb support, v3 WiP
- RSB driver, not working yet

All the above, excluding RSB, can be found in my repository:
https://github.com/wens/linux
Each topic is a separate branch, except for the AXP branches, which
have dependencies.

Testing and feedback is more than welcome.

Regards
ChenYu

-- 
You received this message because you are subscribed to the Google Groups 
linux-sunxi group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to linux-sunxi+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.