[PATCH] ARM: BCMRING: Fix UART declaration and clk dev name mismatch

Jiandong Zheng jdzheng at broadcom.com
Mon May 7 19:14:06 EDT 2012


On 07/05/12 16:00, Paul Gortmaker wrote:
> [Re: [PATCH] ARM: BCMRING: Fix UART declaration and clk dev name mismatch] On 07/05/2012 (Mon 14:50) Jiandong Zheng wrote:
>
>> On 5/7/2012 2:39 PM, Paul Gortmaker wrote:
>>> Do things work if you simply do a partial revert of 888073d41 ?
>>>
>>> Meaning:
>>>
>>> -static AMBA_APB_DEVICE(uartA, "uartA", 0, MM_ADDR_IO_UARTA, {IRQ_UARTA}, NULL);
>>> -static AMBA_APB_DEVICE(uartB, "uartB", 0, MM_ADDR_IO_UARTB, {IRQ_UARTB}, NULL);
>>> +static AMBA_APB_DEVICE(uartA, "uarta", 0, MM_ADDR_IO_UARTA, {IRQ_UARTA}, NULL);
>>> +static AMBA_APB_DEVICE(uartB, "uartb", 0, MM_ADDR_IO_UARTB, {IRQ_UARTB}, NULL);
>>>
>>> The original error messages led me to believe it was two errors;
>>> i.e. (1) the wrong # of args error, and (2) the "a" vs. "A" error.  But
>>> I think Olof was right on his 1st instinct, that it was just the
>>> single error, and the a<--->   A thing was a side effect.
>>>
>>> In which case the above should just work, and it will also keep
>>> the naming consistent with what it was before any of these
>>> commits.
>> Yes.  Either way works. Just in original code, the name uartA and
>> the name string "uarta" looks confusing and easy to get it wrong.
> Agreed, as it confused me as well.  But I don't want to break any
> existing userspace.  So if people expect the name "uarta" and the
> name "uartb" then we really should go with that.  But I have no
> insight into the use case or userspace.  So I leave that part of
> the input up to you.
I don't see user space has direct usage of these strings and the console 
runs fine so that it should be safe.

Thanks,
JD

>
> Thanks,
> Paul.
>
>> Thanks,
>> JD
>>> Paul.
>>> --
>>>
>>





More information about the linux-arm-kernel mailing list