On Tue, Mar 19, 2013 at 4:36 PM, Yu, Fenghua <fenghua...@intel.com> wrote:
>> From: yhlu.ker...@gmail.com [mailto:yhlu.ker...@gmail.com] On Behalf Of
>> Yinghai Lu
>> On Tue, Mar 19, 2013 at 8:04 AM, Fenghua Yu <fenghua...@intel.com>
>> wrote:
>> > From: Fenghua Yu <fenghua...@intel.com>
>> >
>> > In 32-bit, __pa_symbol() in CONFIG_DEBUG_VIRTUAL accesses kernel data
>> (e.g.
>> > max_low_pfn) that haven't been setup yet in such early boot phase. To
>> fix the
>> > issue, __pa_nodebug() replaces __pa_symbol() to get a global symbol's
>> physical
>> > address.
>>
>> That explanation is not precise.
>>
>> __phys_addr in 32bit when DEBUG_VIRTUAL is set, it does have check
>>         /* max_low_pfn is set early, but not _that_ early */
>>         if (max_low_pfn) {
>>          ....
>>         }
>> the reason should be:
>> in head_32.S, 32bit flat mode, you need to switch to phys addr to
>> access
>> global variable max_low_pfn.
>
> Max_low_pfn is not set up yet at this point, but __pa_symbol() reads it
> if CONFIG_DEBUG_VIRTUAL. There is no need to access max_low_pfn or any
> other data that is not setup yet if CONFIG_DEBUG_VIRTUAL.
>
> So I think the explanation is ok.


Changlog said:

 In 32-bit, __pa_symbol() in CONFIG_DEBUG_VIRTUAL accesses kernel data
(e.g. max_low_pfn) that haven't been setup yet in such early boot phase.

---

So do you mean we have to change all __pa_symbol before
setup.c::setup_arch/find_low_pfn_range ?
because only at that time max_low_pfn get setup.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to