[OpenBIOS] [PATCH v2] ppc: Convert I/O macros to inline functions

[Alexander Graf]

On 21.11.2010, at 14:56, Andreas Färber wrote:

> Am 21.11.2010 um 14:51 schrieb Alexander Graf:
> 
>> On 13.11.2010, at 00:54, Andreas Färber wrote:
>> 
>>> Suggested by Blue.
>>> 
>>> v2:
>>> * Make port uint16_t, suggested by Alex.
>>> * Adapt isa_io_base for ppc64.
>>> 
>>> Cc: Blue Swirl <blauwirbel at gmail.com>
>>> Cc: Alexander Graf <agraf at suse.de>
>>> Signed-off-by: Andreas Färber <andreas.faerber at web.de>
>>> ---
>>> arch/ppc/qemu/init.c  |    2 +-
>>> include/arch/ppc/io.h |   68 ++++++++++++++++++++++++++++++++++++------------
>>> 2 files changed, 52 insertions(+), 18 deletions(-)
>>> 
>>> diff --git a/arch/ppc/qemu/init.c b/arch/ppc/qemu/init.c
>>> index ba215bc..ab1ac69 100644
>>> --- a/arch/ppc/qemu/init.c
>>> +++ b/arch/ppc/qemu/init.c
>>> @@ -152,7 +152,7 @@ static const pci_arch_t known_arch[] = {
>>>       .irqs = { 21, 22, 23, 24 }
>>>   },
>>> };
>>> -uint32_t isa_io_base;
>>> +uintptr_t isa_io_base;
>>> 
>>> void
>>> entry( void )
>>> diff --git a/include/arch/ppc/io.h b/include/arch/ppc/io.h
>>> index e5180f2..d8f9dcc 100644
>>> --- a/include/arch/ppc/io.h
>>> +++ b/include/arch/ppc/io.h
>>> @@ -14,23 +14,7 @@ extern unsigned long virt_offset;
>>> 
>>> #ifndef BOOTSTRAP
>>> 
>>> -extern uint32_t isa_io_base;
>>> -
>>> -/*
>>> - * The insw/outsw/insl/outsl macros don't do byte-swapping.
>>> - * They are only used in practice for transferring buffers which
>>> - * are arrays of bytes, and byte-swapping is not appropriate in
>>> - * that case.  - paulus
>>> - */
>>> -#define insw(port, buf, ns)	_insw((uint16_t *)(uintptr_t)((port)+isa_io_base), (buf), (ns))
>>> -#define outsw(port, buf, ns)	_outsw((uint16_t *)(uintptr_t)((port)+isa_io_base), (buf), (ns))
>>> -
>>> -#define inb(port)		in_8((uint8_t *)(uintptr_t)((port)+isa_io_base))
>>> -#define outb(val, port)		out_8((uint8_t *)(uintptr_t)((port)+isa_io_base), (val))
>>> -#define inw(port)		in_le16((uint16_t *)(uintptr_t)((port)+isa_io_base))
>>> -#define outw(val, port)		out_le16((uint16_t *)(uintptr_t)((port)+isa_io_base), (val))
>>> -#define inl(port)		in_le32((uint32_t *)(uintptr_t)((port)+isa_io_base))
>>> -#define outl(val, port)		out_le32((uint32_t *)(uintptr_t)((port)+isa_io_base), (val))
>>> +extern uintptr_t isa_io_base;
>>> 
>>> /*
>>> * 8, 16 and 32 bit, big and little endian I/O operations, with barrier.
>>> @@ -145,6 +129,56 @@ static inline void _outsw(volatile uint16_t * port, const void *buf,
>>> 		ns--;
>>> 	}
>>> }
>>> +
>>> +
>>> +/*
>>> + * The insw/outsw/insl/outsl functions don't do byte-swapping.
>>> + * They are only used in practice for transferring buffers which
>>> + * are arrays of bytes, and byte-swapping is not appropriate in
>>> + * that case.  - paulus
>>> + */
>>> +
>>> +static inline void insw(uint16_t port, void *buf, int ns)
>>> +{
>>> +	_insw((uint16_t *)(port + isa_io_base), buf, ns);
>>> +}
>>> +
>>> +static inline void outsw(uint16_t port, void *buf, int ns)
>>> +{
>>> +	_outsw((uint16_t *)(port + isa_io_base), buf, ns);
>>> +}
>>> +
>>> +
>>> +static inline int inb(uint16_t port)
>> 
>> Shouldn't inb return an uint8_t?
>> 
>>> +{
>>> +	return in_8((uint8_t *)(port + isa_io_base));
>>> +}
>>> +
>>> +static inline void outb(int val, uint16_t port)
>>> +{
>>> +	out_8((uint8_t *)(port + isa_io_base), val);
>>> +}
>>> +
>>> +static inline int inw(uint16_t port)
>> 
>> I'd expect uint16_t to be returned here.
>> 
>>> +{
>>> +	return in_le16((uint16_t *)(port + isa_io_base));
>>> +}
>>> +
>>> +static inline void outw(int val, uint16_t port)
>>> +{
>>> +	out_le16((uint16_t *)(port + isa_io_base), val);
>>> +}
>>> +
>>> +static inline unsigned inl(uint16_t port)
>> 
>> 
>> and uint32_t here.
>> 
>>> +{
>>> +	return in_le32((uint32_t *)(port + isa_io_base));
>>> +}
>>> +
>>> +static inline void outl(int val, uint16_t port)
>> 
>> The same goes for the out values. They should only be as big as they can really be and preferably unsigned :).
> 
> I used the types used by the functions my functions wrap. I noticed that myself but wanted to keep the patch as small as possible. I'll change the existing functions then.

Ah, makes sense :). Yeah, just change them while at it :). The closer we are to reality, the better, right? :)


Alex