[SRC] cvs commit: src/sys/i386/acpica acpi_wakeup.c src/sys/i386/i386 pmap.c

[Nate Lawson]

Julian Elischer wrote:
> Nate Lawson wrote:
>> njl         2007-03-14 22:30:02 UTC
>>
>>   FreeBSD src repository
>>
>>   Modified files:
>>     sys/i386/acpica      acpi_wakeup.c     sys/i386/i386        pmap.c
>>   Log:
>>   Create an identity mapping (V=P) super page for the low memory
>> region on
>>   boot.  Then, just switch to the kernel pmap when suspending instead of
>>   allocating/freeing our own mapping every time.  This should solve a
>> panic
>>   of pmap_remove() being called with interrupts disabled.  Thanks to Alan
>>   Cox for developing this patch.
>>     Note: this means that ACPI requires super page (PG_PS) support in
>> the CPU.
>>   This has been present since the Pentium and first documented in the
>>   Pentium Pro.  However, it may need to be revisited later.
> 
> I thought we were still supporting 486 class machines.
> if so, shouldn't this be somehow ifdef'd?

Yes, that part needs to be covered although an ifdef is not necessary
(PG_PS is defined for even 486's so this will compile).

>> Index: src/sys/i386/i386/pmap.c
>> diff -u src/sys/i386/i386/pmap.c:1.582 src/sys/i386/i386/pmap.c:1.583
>> --- src/sys/i386/i386/pmap.c:1.582    Mon Mar  5 21:40:10 2007
>> +++ src/sys/i386/i386/pmap.c    Wed Mar 14 22:30:02 2007
>> @@ -422,6 +422,13 @@
>>  
>>      /* Turn on PG_G on kernel page(s) */
>>      pmap_set_pg();
>> +
>> +    /*
>> +     * Create an identity mapping (virt == phys) for the low 1 MB
>> +     * physical memory region that is used by the ACPI wakeup code.
>> +     * This mapping must not have PG_G set. +     */
>> +    kernel_pmap->pm_pdir[0] = PG_PS | PG_RW | PG_V;
>>  }
>>  
>>  /*

I propose conditionalizing this code on "if (pseflag)".  Of course, the
acpi suspend code will fail on 486's but we disable acpi entirely if the
bios date < 1999/1/1 and acpi isn't supported on the 486.

-- 
Nate