[RTEMS 5 PATCH] bsp/motorola_powerp: Print RTEMS_VERSION from the bootloader

Gedare Bloom gedare at rtems.org
Mon Feb 8 21:50:59 UTC 2021


On Mon, Feb 8, 2021 at 2:45 PM Gedare Bloom <gedare at rtems.org> wrote:

>
> On Mon, Feb 8, 2021 at 2:04 PM Chris Johns <chrisj at rtems.org> wrote:
>
>> On 9/2/21 5:18 am, Gedare Bloom wrote:
>> > If you put this on master/6, you can switch that align to use
>> > RTEMS_ALIGN_UP(addr, PAGE_SIZE) -- not PAGE_MASK
>>
>> Sorry, I had already pushed the patch.
>>
>> A quick grep shows there are a few of these ...
>>
>>  $ grep -r PAGE_MASK *
>> 0001-bsp-motorola_powerp-Print-RTEMS_VERSION-from-the-boo.patch: #define
>> PAGE_ALIGN(addr)       (((addr) + PAGE_MASK) & ~PAGE_MASK)
>> 0002-bsp-motorola_powerp-Print-RTEMS_VERSION-from-the-boo.patch: #define
>> PAGE_ALIGN(addr)       (((addr) + PAGE_MASK) & ~PAGE_MASK)
>> bsps/powerpc/motorola_powerpc/bootloader/mm.c:#define PAGE_ALIGN(addr)
>> (((addr)
>> + PAGE_MASK) & ~PAGE_MASK)
>> bsps/powerpc/motorola_powerpc/bootloader/mm.c:                  rpn =
>> (vaddr&PAGE_MASK)-area->base+area->firstpte;
>> bsps/powerpc/motorola_powerpc/bootloader/mm.c:  if (mask&PAGE_MASK) {
>> bsps/powerpc/motorola_powerpc/bootloader/misc.c:#define PAGE_ALIGN(addr)
>> (((addr) + PAGE_MASK) & ~PAGE_MASK)
>> bsps/powerpc/include/libcpu/pgtable.h:#define _PAGE_CHG_MASK
>> (PAGE_MASK |
>> _PAGE_ACCESSED | _PAGE_DIRTY)
>>
> This is suspect. PAGE_MASK coming through newlib is 0xFFF. the ORs are
redundant.



> bsps/powerpc/shared/start/pgtbl_setup.c:#define PAGE_ALIGN(addr)
>> (((addr)
>> + PAGE_MASK) & ~PAGE_MASK)
>> cpukit/libnetworking/vm/vm_param.h:
>>  ((vm_offset_t)((((vm_offset_t)(x)) +
>> PAGE_MASK) >> PAGE_SHIFT))
>>
> ignore networking


> cpukit/posix/src/mmap.c:    if ((uintptr_t)addr & PAGE_MASK) {
>
>
>> Do all these needs to be looked at as well?
>>
>> Probably should. This is only for 6+ version, to help with consistency.
> Some of those appear to be actual PAGE_MASK uses, but the alignment macros
> should use the shared RTEMS_ALIGN* for easier validation.
>
>

FWIW, we also have round_page() defined in newlib.


> Chris
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20210208/a12b63e9/attachment-0001.html>


More information about the devel mailing list