[PATCH v3] rtc: armada38x: Follow the new recommendation for errata implementation

[alexandre.belloni@free-electrons.com (Alexandre Belloni)]

On 21/12/2016 at 11:28:16 +0100, Gregory CLEMENT wrote :
> According to RES-3124064:
> 
> The device supports CPU write and read access to the RTC time register.
> However, due to this restriction, read and write from/to internal RTC
> register may fail.
> 
> Workaround:
> General setup:
> 1. Configure the RTC Mbus Bridge Timing Control register (offset 0x184A0)
>    to value 0xFD4D4FFF
>    Write RTC WRCLK Period to its maximum value (0x3FF)
>    Write RTC WRCLK setup to 0x29
>    Write RTC WRCLK High Time to 0x53 (default value)
>    Write RTC Read Output Delay to its maximum value (0x1F)
>    Mbus - Read All Byte Enable to 0x1 (default value)
> 2. Configure the RTC Test Configuration Register (offset 0xA381C) bit3
>    to '1' (Reserved, Marvell internal)
> 
> For any RTC register read operation:
> 1. Read the requested register 100 times.
> 2. Find the result that appears most frequently and use this result
>    as the correct value.
> 
> For any RTC register write operation:
> 1. Issue two dummy writes of 0x0 to the RTC Status register (offset
>    0xA3800).
> 2. Write the time to the RTC Time register (offset 0xA380C).
> 
> This patch is based on the work of Shaker Daibes
> 
> Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
> ---
> Hi,
> 
> this patch followed the patch series sent here:
> http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473232.html
> 
> For now I kept the patch 2 converting to time64_t apart after Russell
> King feedback.
> 
> Changelog:
> 
> v2 -> v3:
> - Removed unrelated change
> - Remove Unnecessary blank line
> - Used u32 instead of uint32_t
> - use the new errata name: RES-3124064 instead of FE-3124064
> 
> v1 -> v2:
> - merged the patch 1 and 3
> - substantially modified the 1st patch with I think a better
> implementation:
>   - First I do not put anymore more a big array onto the stack as
>     suggested by Andrew Lunn.
>   - Then I optimize the way to find the correct value.
> 
> Gregory
> 
>  drivers/rtc/rtc-armada38x.c | 118 +++++++++++++++++++++++++++++++++++---------
>  1 file changed, 95 insertions(+), 23 deletions(-)
> 
Applied, thanks.

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com