[PATCH -fixes 3/3] riscv: Optimize kernel virtual address conversion macro

[Palmer Dabbelt]

On Sat, 07 Aug 2021 12:31:24 PDT (-0700), alex at ghiti.fr wrote:
> Le 7/08/2021 à 18:36, Palmer Dabbelt a écrit :
>> On Wed, 21 Jul 2021 00:59:37 PDT (-0700), alex at ghiti.fr wrote:
>>> The current test in kernel_mapping_va_to_pa only applies when
>>> CONFIG_XIP_KERNEL is set, so use IS_ENABLED to optimize this macro at
>>> compile-time in standard kernels that do not require this test.
>>>
>>> Signed-off-by: Alexandre Ghiti <alex at ghiti.fr>
>>> ---
>>>  arch/riscv/include/asm/page.h | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/arch/riscv/include/asm/page.h
>>> b/arch/riscv/include/asm/page.h
>>> index b0ca5058e7ae..10dc063868f6 100644
>>> --- a/arch/riscv/include/asm/page.h
>>> +++ b/arch/riscv/include/asm/page.h
>>> @@ -123,7 +123,7 @@ extern phys_addr_t phys_ram_base;
>>>  #define linear_mapping_va_to_pa(x)    ((unsigned long)(x) -
>>> kernel_map.va_pa_offset)
>>>  #define kernel_mapping_va_to_pa(y) ({                        \
>>>      unsigned long _y = y;                            \
>>> -    (_y < kernel_map.virt_addr + XIP_OFFSET) ?                    \
>>> +    (IS_ENABLED(CONFIG_XIP_KERNEL) && _y < kernel_map.virt_addr +
>>> XIP_OFFSET) ?    \
>>>          ((unsigned long)(_y) - kernel_map.va_kernel_xip_pa_offset)
>>> :        \
>>>          ((unsigned long)(_y) - kernel_map.va_kernel_pa_offset -
>>> XIP_OFFSET);    \
>>>      })
>>
>> IIUC this isn't actually a fix?  The other two are, though, so they're
>> on fixes.
>
> Indeed this is not a fix, this is symmetrical to what I have done in
> patch 1 so that seemed natural to land in the same patchset. Let me know
> if you want me to resend it on its own.

That's fine, it was just explicitly called out as a fix in the subject 
line so I wanted to make sure I wasn't missing something.  This is on 
for-next.

Thanks!