[U-Boot] [PATCH v2 3/5] dm: ns16550: Don't map_physmem for I/O ports

[Simon Glass]

Hi Daniel,

On 18 May 2016 at 04:04, Daniel Schwierzeck
<daniel.schwierzeck at gmail.com> wrote:
>
>
> Am 17.05.2016 um 18:00 schrieb Simon Glass:
>> Hi Paul,
>>
>> On 17 May 2016 at 09:58, Paul Burton <paul.burton at imgtec.com> wrote:
>>> On Tue, May 17, 2016 at 09:54:21AM -0600, Simon Glass wrote:
>>>>>>>> diff --git a/drivers/serial/ns16550.c b/drivers/serial/ns16550.c
>>>>>>>> index 28da9dd..e58e6aa 100644
>>>>>>>> --- a/drivers/serial/ns16550.c
>>>>>>>> +++ b/drivers/serial/ns16550.c
>>>>>>>> @@ -100,7 +100,11 @@ static void ns16550_writeb(NS16550_t port, int offset, int value)
>>>>>>>>         unsigned char *addr;
>>>>>>>>
>>>>>>>>         offset *= 1 << plat->reg_shift;
>>>>>>>> +#ifdef CONFIG_SYS_NS16550_PORT_MAPPED
>>>>>>>> +       addr = (unsigned char *)plat->base + offset;
>>>>>>>> +#else
>>>>>>>>         addr = map_physmem(plat->base, 0, MAP_NOCACHE) + offset;
>>>>>>>> +#endif
>>>>>>>
>>>>>>> Please don't add CONFIG #ifdefs in these functions. Perhaps it needs
>>>>>>> to be another parameter? Possibly a flag. But with driver-model we
>>>>>>> need to be able to support both options in the core code.
>>>>>>
>>>>>> Hi Simon,
>>>>>>
>>>>>> Are you sure systems rely on using I/O ports with map_physmem? The only
>>>>>> other systems that define CONFIG_SYS_NS16550_PORT_MAPPED are x86 ones,
>>>>>> in include/configs/x86-common.h, and so far as I can tell they don't use
>>>>>> device model which suggests this code has simply been untested before. I
>>>>>> don't see why you would use map_physmem on an I/O port address that is
>>>>>> then going to be passed to inb/outb & I think the code here is simply
>>>>>> wrong to do so.
>>>>>
>>>>> the current code looks wrong. serial_in_shift() is expanded to inb()
>>>>> in case of CONFIG_SYS_NS16550_PORT_MAPPED and to
>>>>> in_le32()/in_be32()/readl()/readb() otherwise. Only in the latter case
>>>>> a map_physmem() is required and should be done in serial_in_shift()
>>>>> itself or preferrably only once in
>>>>> ns16550_serial_ofdata_to_platdata().
>>>>>
>>>>> I think the correct approach would be the following:
>>>>
>>>> This is better I think. But how about adding a device tree binding to
>>>> select I/O access? In principle each device might have its own
>>>> settings.
>>>
>>> Note that's what I worked towards last time I had a crack at this, but
>>> it just expanded into an attempt to tackle the mess that is ns16550.c &
>>> rather lost sight of the original goal of making Malta work.
>>>
>>> https://patchwork.ozlabs.org/patch/575643/
>>> https://patchwork.ozlabs.org/patch/577194/
>>
>> Yes it is tricky. What do you think about the suggestions above?
>
> for now we should only fix the broken port-based I/O. Additional DT
> bindings could be added later. I'd like to merge the DM support on MIPS
> Malta in this merge window ;)

In that case your patch of moving it to ofdata_to_platdata() looks OK to me.

Regards,
Simon