[U-Boot] Pull request: u-boot-spi/master

[Álvaro Fernández Rojas]

Hi Jagan,

El 24/01/2018 a las 7:57, Jagan Teki escribió:
> On Tue, Jan 23, 2018 at 3:58 PM, Daniel Schwierzeck
> <daniel.schwierzeck at gmail.com> wrote:
>>
>> On 22.01.2018 21:55, Álvaro Fernández Rojas wrote:
>>> Hi Daniel,
>>>
>>>
>>> El 22/01/2018 a las 21:26, Daniel Schwierzeck escribió:
>>>> On 22.01.2018 18:14, Tom Rini wrote:
>>>>> On Mon, Jan 22, 2018 at 05:49:39PM +0100, Daniel Schwierzeck wrote:
>>>>>> On 22.01.2018 13:58, Tom Rini wrote:
>>>>>>> On Mon, Jan 22, 2018 at 11:20:56AM +0530, Jagan Teki wrote:
>>>>>>>
>>>>>>>> Hi Tom,
>>>>>>>>
>>>>>>>> Please pull this PR.
>>>>>>>>
>>>>>>>> thanks!
>>>>>>>> Jagan.
>>>>>>>>
>>>>>>>> The following changes since commit
>>>>>>>> 98691a60abffb44303d7dae6e9e699d0daded930:
>>>>>>>>
>>>>>>>>     Merge git://git.denx.de/u-boot-rockchip (2018-01-09 13:28:51
>>>>>>>> -0500)
>>>>>>>>
>>>>>>>> are available in the git repository at:
>>>>>>>>
>>>>>>>>     git://git.denx.de/u-boot-spi.git master
>>>>>>>>
>>>>>>>> for you to fetch changes up to
>>>>>>>> b23c685c6f295da3c01dd487f0e003b70299af91:
>>>>>>>>
>>>>>>>>     mips: bmips: enable the SPI flash on the Comtrend AR-5387un
>>>>>>>> (2018-01-22 10:39:13 +0530)
>>>>>>>>
>>>>>>> NAK:
>>>>>>>
>>>>>>> commit 19e3a4856c1cba751a9ecb3931ff0d96a7f169be
>>>>>>> Author: Álvaro Fernández Rojas <noltari at gmail.com>
>>>>>>> Date:   Sat Jan 20 02:11:34 2018 +0100
>>>>>>>
>>>>>>>       wait_bit: add 8/16/32 BE/LE versions of wait_for_bit
>>>>>>>
>>>>>>>       Add 8/16/32 bits and BE/LE versions of wait_for_bit.
>>>>>>>       This is needed for reading registers that are not aligned to
>>>>>>> 32 bits, and for
>>>>>>>       Big Endian platforms.
>>>>>>>
>>>>>>>       Signed-off-by: Álvaro Fernández Rojas <noltari at gmail.com>
>>>>>>>       Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck at gmail.com>
>>>>>>>       Reviewed-by: Jagan Teki <jagan at openedev.com>
>>>>>>>
>>>>>>> Adds warnings on almost all platforms:
>>>>>>> w+(ls1088ardb_qspi_SECURE_BOOT) ../include/wait_bit.h: In function
>>>>>>> ?wait_for_bit_be16?:
>>>>>>> w+(ls1088ardb_qspi_SECURE_BOOT) ../include/wait_bit.h:76:31:
>>>>>>> warning: implicit declaration of function ?readw_be?
>>>>>>> [-Wimplicit-function-declaration]
>>>>>>> w+(ls1088ardb_qspi_SECURE_BOOT) ../include/wait_bit.h: In function
>>>>>>> ?wait_for_bit_be32?:         w+(ls1088ardb_qspi_SECURE_BOOT)
>>>>>>> ../include/wait_bit.h:78:31: warning: implicit declaration of
>>>>>>> function ?readl_be? [-Wimplicit-function-declaration]
>>>>>>>
>>>>>>>
>>>>>> Tom, would this change to the patch be acceptable?
>>>>>>
>>>>>> --- a/include/wait_bit.h
>>>>>> +++ b/include/wait_bit.h
>>>>>> @@ -73,8 +73,12 @@ static inline int wait_for_bit_##sfx(const void
>>>>>> *reg,            \
>>>>>>
>>>>>>    BUILD_WAIT_FOR_BIT(8, u8, readb)
>>>>>>    BUILD_WAIT_FOR_BIT(le16, u16, readw)
>>>>>> +#ifdef readw_be
>>>>>>    BUILD_WAIT_FOR_BIT(be16, u16, readw_be)
>>>>>> +#endif
>>>>>>    BUILD_WAIT_FOR_BIT(le32, u32, readl)
>>>>>> +#ifdef readl_be
>>>>>>    BUILD_WAIT_FOR_BIT(be32, u32, readl_be)
>>>>>> +#endif
>>>>>>
>>>>>>    #endif
>>>>>>
>>>>>> This wouldn't define wait_bit_be*() on archs which doesn't implement
>>>>>> readw_be or readl_be.
>>>>>>
>>>>>> A build with the updated patch is scheduled at
>>>>>> https://travis-ci.org/danielschwierzeck/u-boot/builds/331899381
>>>>> That seems reasonable, thanks!
>>>>>
>>>> Álvaro, could you send a v10 where the patch "wait_bit: add 8/16/32
>>>> BE/LE versions of wait_for_bit" is fixed like above? Thanks.
>>> Sure, but I think this alternative would be much cleaner:
>>> https://gist.github.com/Noltari/3e6ed4648b87484c73ca22e2f533f9b0
>>>
>>> What do you think?
>>>
>> I had a similar idea at first but a #ifdef within a #define is not possible and AFAIK not implemented in C standards.
>>
>> Have you tried it? With nested #ifdef I'm getting:
>>
>> $ ./tools/buildman/buildman -o /tmp/bm -b master..wait_for_bit aarch64
>> Building 1 commit for 111 boards (8 threads, 1 job per thread)
>>     50   11   50 /111    lion-rk3368
>>
>> vs.
>>
>> $ ./tools/buildman/buildman -o /tmp/bm -b master..wait_for_bit_2 aarch64
>> Building 1 commit for 111 boards (8 threads, 1 job per thread)
>>     98   13    0 /111    0:00:06  : ls1046ardb_emm
>>
>> Also a Travis CI build already shows several broken builds:
>> https://travis-ci.org/danielschwierzeck/u-boot/builds/332205310
>>
>> Maybe a yet better solution would be if each arch implements the same common set of I/O primitives like Linux is doing.
> Recent patch from Alvaro, seems fine to built all [1]
>
> [1] https://travis-ci.org/openedev/u-boot-spi/builds/332390626
I ended up adding the changes proposed by Daniel to avoid nested ifdefs.
So I guess it can be merged now :).