On 02/14/2017 08:16, James Hogan wrote:
> Hi Joshua,
>
> On Tue, Feb 07, 2017 at 12:57:50AM -0500, Joshua Kinard wrote:
>> From: Joshua Kinard <kumba@xxxxxxxxxx>
>>
>> Clean-up several macros in arch/mips/include/asm/xtalk/xtalk.h:
>> - Hex addresses are lowercased.
>> - Added whitespace around several operators.
>> - Removed bridge_probe declaration.
>>
>> Signed-off-by: Joshua Kinard <kumba@xxxxxxxxxx>
>> ---
>> arch/mips/include/asm/xtalk/xtalk.h | 23 +++++++----------------
>> 1 file changed, 7 insertions(+), 16 deletions(-)
>>
>> diff --git a/arch/mips/include/asm/xtalk/xtalk.h b/arch/mips/include/asm/xtalk/xtalk.h
>> index 9125bd85514d..627ed91b2880 100644
>> --- a/arch/mips/include/asm/xtalk/xtalk.h
>> +++ b/arch/mips/include/asm/xtalk/xtalk.h
>> @@ -21,24 +21,15 @@
>> #define XWIDGET_MFG_NUM_NONE -1
>>
>> /* It is often convenient to fold the XIO target port */
>> -#define XIO_NOWHERE (0xFFFFFFFFFFFFFFFFull)
>> -#define XIO_ADDR_BITS (0x0000FFFFFFFFFFFFull)
>> -#define XIO_PORT_BITS (0xF000000000000000ull)
>> +#define XIO_NOWHERE (0xffffffffffffffffULL)
>> +#define XIO_ADDR_BITS (0x0000ffffffffffffULL)
>> +#define XIO_PORT_BITS (0xf000000000000000ULL)
>> #define XIO_PORT_SHIFT (60)
>>
>> -#define XIO_PACKED(x) (((x)&XIO_PORT_BITS) != 0)
>> -#define XIO_ADDR(x) ((x)&XIO_ADDR_BITS)
>> -#define XIO_PORT(x) ((s8)(((x)&XIO_PORT_BITS) >> XIO_PORT_SHIFT))
>> -#define XIO_PACK(p, o) ((((uint64_t)(p))<<XIO_PORT_SHIFT) | ((o)&XIO_ADDR_BITS))
>> -
>> -#ifdef CONFIG_PCI
>> -extern int bridge_probe(nasid_t nasid, int widget, int masterwid);
>> -#else
>> -static inline int bridge_probe(nasid_t nasid, int widget, int masterwid)
>> -{
>> - return 0;
>> -}
>> -#endif
>
> Won't this break the build when CONFIG_PCI=n?
>
> Cheers
> James
It might, though I believe my BRIDGE updates addressed that. Though, these
machines are pretty unusable without PCI anyways.
Hold on these patches for now (same for the BRIDGE series). I need to re-spin
them once I tackle fixing the PCI Bridge window bits this weekend. I think
I'll make the window selection a platform_data parameter and have a built-in
default for small windows, since those are usually fixed by hardware, with each
BRIDGE-capable platform passing in its own big window range.
Then, IP27 is going to require porting additional code from Linux-2.5.70/IA64
to get its big windows to work...
--J
