Re: [RFC V2 3/6] sh/mm: Stop using pte_ERROR()
"David Hildenbrand (Arm)" <[email protected]>
| Newsgroups | gmane.linux.ports.ppc.embedded |
|---|---|
| Message-ID | <f7bab5a5-76c9-4856-b314-49f19f54e229__2481.4380494476$1786605526$gmane$org@kernel.org> |
>>> +}
>>> +
>>> static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot)
>>> {
>>> pte_t *pte;
>>>
>>> pte = __get_pte_phys(addr);
>>> if (!pte_none(*pte)) {
>>> - pte_ERROR(*pte);
>>> + print_bad_pte(*pte);
>>
>> Maybe similar to my reply to patch #1, actually print something more useful than
>> just "bad pte" ? And possibly also just use ptval_to_str().
>
> Probably #ifdef could be dropped here as ptval_to_str() should be able to
> handle pte with two unsiged long elements in there i.e pte_high, pte_low
> along with standard single unsiged long pte_low format.
Yes, that should just work and clean this up!
--
Cheers,
David