Re: [Xen-devel] [PATCH-for-4.11 v2] xpti: fix bug in double fault handling

[Juergen Gross <jgross@xxxxxxxx>]

On 23/04/18 15:06, Andrew Cooper wrote:
> On 23/04/18 13:38, Juergen Gross wrote:
>> When entering the hypervisor via the double fault handler resetting
>> xen_cr3 was missing. This led to switching to pv_cr3 when returning
>> from the next following interrupt. So repair this in order to allow
>> interrupt handling to work even after a double fault.
>>
>> Signed-off-by: Juergen Gross <jgross@xxxxxxxx>
>> ---
>>  xen/arch/x86/x86_64/entry.S | 1 +
>>  1 file changed, 1 insertion(+)
>>
>> diff --git a/xen/arch/x86/x86_64/entry.S b/xen/arch/x86/x86_64/entry.S
>> index 45d9842d09..25427b0cec 100644
>> --- a/xen/arch/x86/x86_64/entry.S
>> +++ b/xen/arch/x86/x86_64/entry.S
>> @@ -773,6 +773,7 @@ ENTRY(double_fault)
>>          neg   %rbx
>>  .Ldblf_cr3_load:
>>          mov   %rbx, %cr3
>> +        movq $0, STACK_CPUINFO_FIELD(xen_cr3)(%r14)
>>  .Ldblf_cr3_okay:
>>  
>>          movq  %rsp,%rdi
> 
> What about the other write into xen_cr3 with a negated value?  Won't
> this still explode if we get an NMI or MCE at the wrong moment?

Hmm, you mean a NMI between the mov to %cr3 and zeroing xen_cr3?
Could be an issue, yes. Okay, V3 then...


Juergen


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel