Re: [PATCH v2 1/7] cpu: use ROUND_UP() to define xxx_PAGE_ALIGN

[Paolo Bonzini]

On 24/10/19 16:06, Richard Henderson wrote:
> On 10/24/19 8:04 AM, Paolo Bonzini wrote:
>> On 24/10/19 13:52, Philippe Mathieu-Daudé wrote:
>>>> diff --git a/include/exec/cpu-all.h b/include/exec/cpu-all.h
>>>> index ad9ab85eb3..255bb186ac 100644
>>>> --- a/include/exec/cpu-all.h
>>>> +++ b/include/exec/cpu-all.h
>>>> @@ -220,7 +220,7 @@ extern int target_page_bits;
>>>>     #define TARGET_PAGE_SIZE (1 << TARGET_PAGE_BITS)
>>>>   #define TARGET_PAGE_MASK ~(TARGET_PAGE_SIZE - 1)
>>>> -#define TARGET_PAGE_ALIGN(addr) (((addr) + TARGET_PAGE_SIZE - 1) &
>>>> TARGET_PAGE_MASK)
>>>> +#define TARGET_PAGE_ALIGN(addr) ROUND_UP((addr), TARGET_PAGE_SIZE)
>>>>     /* Using intptr_t ensures that qemu_*_page_mask is sign-extended even
>>>>    * when intptr_t is 32-bit and we are aligning a long long.
>>>> @@ -228,9 +228,8 @@ extern int target_page_bits;
>>>>   extern uintptr_t qemu_host_page_size;
>>>>   extern intptr_t qemu_host_page_mask;
>>>>   -#define HOST_PAGE_ALIGN(addr) (((addr) + qemu_host_page_size - 1) &
>>>> qemu_host_page_mask)
>>>> -#define REAL_HOST_PAGE_ALIGN(addr) (((addr) +
>>>> qemu_real_host_page_size - 1) & \
>>>> -                                    qemu_real_host_page_mask)
>>>> +#define HOST_PAGE_ALIGN(addr) ROUND_UP((addr), qemu_host_page_size)
>>>> +#define REAL_HOST_PAGE_ALIGN(addr) ROUND_UP((addr),
>>>> qemu_real_host_page_size)
>>>>     /* same as PROT_xxx */
>>>>   #define PAGE_READ      0x0001
>>>>
>>
>> Isn't this the patch where Richard pointed out that the compiler
>> generates worse code?
> 
> Richard confused ROUND_UP with QEMU_ALIGN_UP.

Uh, those are both really badly named.  Especially considering that
DIV_ROUND_UP divides the result of QEMU_ALIGN_UP.  Should we consider
swapping them?!?

Paolo