Re: [RFC PATCH] x86 NMI-safe INT3 and Page Fault (v7)

[Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>]

* Jeremy Fitzhardinge (jeremy@goop.org) wrote:
> Mathieu Desnoyers wrote:
>> * Jeremy Fitzhardinge (jeremy@goop.org) wrote:
>>   
>>> Mathieu Desnoyers wrote:
>>>     
>>>> +/*
>>>> + * Only returns from a trap or exception to a NMI context 
>>>> (intra-privilege
>>>> + * level near return) to the same SS and CS segments. Should be used
>>>> + * upon trap or exception return when nested over a NMI context so no 
>>>> iret is
>>>> + * issued. It takes care of modifying the eflags, rsp and returning to 
>>>> the
>>>> + * previous function.
>>>> + *
>>>> + * The stack, at that point, looks like :
>>>> + *
>>>> + * 0(rsp)  RIP
>>>> + * 8(rsp)  CS
>>>> + * 16(rsp) EFLAGS
>>>> + * 24(rsp) RSP
>>>> + * 32(rsp) SS
>>>> + *
>>>> + * Upon execution :
>>>> + * Copy EIP to the top of the return stack
>>>> + * Update top of return stack address
>>>> + * Pop eflags into the eflags register
>>>> + * Make the return stack current
>>>> + * Near return (popping the return address from the return stack)
>>>> + */
>>>> +#define INTERRUPT_RETURN_NMI_SAFE	pushq %rax;		\
>>>> +					mov %rsp, %rax;		\
>>>> +					mov 24+8(%rax), %rsp;	\
>>>> +					pushq 0+8(%rax);	\
>>>> +					pushq 16+8(%rax);	\
>>>> +					movq (%rax), %rax;	\
>>>> +					popfq;			\
>>>> +					ret;
>>>>         
>>> I got this right first go?  I must be getting good at this...
>>>
>>>     
>>
>> Yes, it looked good to me at least :)
>>
>>   
>>> Anyway, trailing ';', and perhaps use a consistent form for mov (either 
>>> movq or mov in all three instances).
>>>
>>>     
>>
>> Ok, fixed. Here is the update.
>>
>> The last issue standing would be the paravirt code. Any ideas about how
>> it's best to do it ? It would be good to be able to get the nmi-safe
>> version on bare metal, patched with a standard iret emulation in
>> paravirt code.
>>   
>
> #ifdef CONFIG_PARAVIRT
> #define NMI_RETURN   paravirt_nmi_return
>
> /* elsewhere */
>    .nmi_return = native_iret_nmi_return;
> #else
> #define NMI_RETURN   NMI_SAFE_NMI_RETURN
> #endif
>
> ?
>

Then I guess we would have to start dealing with faults caused by popf
if it happens to be run before the code patching is done ?

Mathieu

>    J

-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68