[Letux-kernel] 2016.11 uboot considered harmful

[H. Nikolaus Schaller]

> Am 03.02.2017 um 07:53 schrieb Andreas Kemnade <andreas at kemnade.info>:
> 
> On Fri, 3 Feb 2017 07:31:14 +0100
> Andreas Kemnade <andreas at kemnade.info> wrote:
> 
>> Hi,
>> 
>> On Thu, 2 Feb 2017 19:27:03 +0100
>> "Dr. H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>> 
>>> Amorpher interesting question is if we still have to Enable Backup
>>> charging at all. It might have been missing in 2011 uboot but fixed
>>> now.
>>> 
>>> On the Road --- hns
>>> 
>> 
>> I compared with old uboot.
>> The parameters to twl4030_i2c_write_u8 seem to have been reversed.
>> twl4030-additions.c was converted, gta04.c not.
>> 
>> So the backup battery was charging in earlier uboots.
>> 
> commit 0208aaf6c2e0f346e8410c0e5adb0ea784f00829
> Author: Nishanth Menon <nm at ti.com>
> Date:   Tue Mar 26 05:20:49 2013 +0000
> 
>    twl4030: make twl4030_i2c_write_u8 prototype consistent
> 
>    u-boot standard i2c register write prototype is
>    i2c_reg_write(u8 addr, u8 reg, u8 val)
> 
>    twl4030_i2c_write_u8(u8 addr, u8 val, u8 reg)
>    does not provide consistency, so switch the prototype to be
>    consistent with rest of u-boot i2c operations:
>    twl4030_i2c_write_u8(u8 addr, u8 reg, u8 val)
> 
>    Signed-off-by: Nishanth Menon <nm at ti.com>
> 

Ok, I will add this link to the commit message.

And I have tested the patch. Now, RTC works again :)
And even ./measure-suspend. Now I understand why:
measure-suspend calculates from sw-clock a time position
when the wakeup should occur.

Since the hw-clock is wrong, this will not happen
in sync with the sw-clock...

BR and thanks for debugging this issue!

Nikolaus

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.goldelico.com/pipermail/letux-kernel/attachments/20170203/443021bc/attachment.asc>