On Sun, Mar 4, 2012 at 8:12 PM, Rusty Russell <rusty@xxxxxxxxxxxxxxx> wrote:
> On Thu, 23 Feb 2012 02:32:39 -0500, Christoffer Dall <c.dall@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>> From: Christoffer Dall <c.dall@xxxxxxxxxxxxxxxxxxxxxx>
>>
>> +static void cpu_set_vector(void *vector)
>> +{
>> + unsigned long vector_ptr;
>> + unsigned long smc_hyp_nr;
>> +
>> + vector_ptr = (unsigned long)vector;
>> + smc_hyp_nr = SMCHYP_HVBAR_W;
>> +
>> + /*
>> + * Set the HVBAR
>> + */
>> + asm volatile (
>> + "mov r0, %[vector_ptr]\n\t"
>> + "mov r7, %[smc_hyp_nr]\n\t"
>> + "smc #0\n\t" : :
>> + [vector_ptr] "r" (vector_ptr),
>> + [smc_hyp_nr] "r" (smc_hyp_nr) :
>> + "r0", "r1");
>
> Testing this recently, trying to get a kvm guest to boot. No success
> yet, but I did get another issue when I compiled the host non-SMP: you
> need to mark r7 as clobbered here (on SMP we didn't clobber anything
> important, but on UP it hit init_end_phys_addr and changed it to
> 0xffffffff and we died horribly in hyp_idmap_del().
>
you're absolutely right. Nice catch. Thanks.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
