[patch] i386: remove IOPL check on task switch

[patch] i386: remove IOPL check on task switch

[Chuck Ebbert]

IOPL is implicitly saved and restored on task switch,
so explicit check is no longer needed.

Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>

--- 2.6.19-rc4-32smp.orig/arch/i386/kernel/process.c
+++ 2.6.19-rc4-32smp/arch/i386/kernel/process.c
@@ -681,12 +681,6 @@ struct task_struct fastcall * __switch_t
 		loadsegment(gs, next->gs);
 
 	/*
-	 * Restore IOPL if needed.
-	 */
-	if (unlikely(prev->iopl != next->iopl))
-		set_iopl_mask(next->iopl);
-
-	/*
 	 * Now maybe handle debug registers and/or IO bitmaps
 	 */
 	if (unlikely((task_thread_info(next_p)->flags & _TIF_WORK_CTXSW)
-- 
Chuck
"Even supernovas have their duller moments."

Re: [patch] i386: remove IOPL check on task switch

[Zachary Amsden]

Chuck Ebbert wrote:
> IOPL is implicitly saved and restored on task switch,
> so explicit check is no longer needed.
>
> Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>
>
> --- 2.6.19-rc4-32smp.orig/arch/i386/kernel/process.c
> +++ 2.6.19-rc4-32smp/arch/i386/kernel/process.c
> @@ -681,12 +681,6 @@ struct task_struct fastcall * __switch_t
>  		loadsegment(gs, next->gs);
>  
>  	/*
> -	 * Restore IOPL if needed.
> -	 */
> -	if (unlikely(prev->iopl != next->iopl))
> -		set_iopl_mask(next->iopl);
> -
> -	/*
>  	 * Now maybe handle debug registers and/or IO bitmaps
>  	 */
>  	if (unlikely((task_thread_info(next_p)->flags & _TIF_WORK_CTXSW)
>   

Nack.  This is used for paravirt-ops kernels that use IOPL'd userspace.  
Fixing it would require a fairly heavy penalty on the iret path, since 
every single instruction there contributes to a critical region which 
must have custom fixup code, or some other technique to provide 
protection against interrupt re-entrancy.

At least, let's discuss other potential solutions first - for now it is 
harmless.

Zach

Re: [patch] i386: remove IOPL check on task switch

[Andi Kleen]

 
> Nack.  This is used for paravirt-ops kernels that use IOPL'd userspace.  
> Fixing it would require a fairly heavy penalty on the iret path, since 
> every single instruction there contributes to a critical region which 
> must have custom fixup code, or some other technique to provide 
> protection against interrupt re-entrancy.
> 
> At least, let's discuss other potential solutions first - for now it is 
> harmless.

Ok I will drop the patch again

-Andi

Re: [patch] i386: remove IOPL check on task switch

[Chuck Ebbert]

In-Reply-To: <454B850C.3050402@vmware.com>

On Fri, 03 Nov 2006 10:06:04 -0800, Zachary Amsden wrote:

> Chuck Ebbert wrote:
> > IOPL is implicitly saved and restored on task switch,
> > so explicit check is no longer needed.
> 
> Nack.  This is used for paravirt-ops kernels that use IOPL'd userspace.  

How does that work?  In the stock kernel, anything done by
the call to set_iopl_mask() (that was removed by the patch)
will be nullified by the 'popfl' at the end of the switch_to()
macro.

-- 
Chuck
"Even supernovas have their duller moments."

Re: [patch] i386: remove IOPL check on task switch

[Zachary Amsden]

Chuck Ebbert wrote:
> In-Reply-To: <454B850C.3050402@vmware.com>
>
> On Fri, 03 Nov 2006 10:06:04 -0800, Zachary Amsden wrote:
>
>   
>> Chuck Ebbert wrote:
>>     
>>> IOPL is implicitly saved and restored on task switch,
>>> so explicit check is no longer needed.
>>>       
>> Nack.  This is used for paravirt-ops kernels that use IOPL'd userspace.  
>>     
>
> How does that work?  In the stock kernel, anything done by
> the call to set_iopl_mask() (that was removed by the patch)
> will be nullified by the 'popfl' at the end of the switch_to()
> macro.
>   

Who put a popfl back in switch_to?  I took it out some time ago.  It 
should not be there.  The only reason for it was to stop IOPL leaking 
from one process to another from a sleep during a sysenter based system 
call.

Zach

Re: [patch] i386: remove IOPL check on task switch

[Zachary Amsden]

Zachary Amsden wrote:
> Chuck Ebbert wrote:
>> In-Reply-To: <454B850C.3050402@vmware.com>
>>
>> On Fri, 03 Nov 2006 10:06:04 -0800, Zachary Amsden wrote:
>>
>>  
>>> Chuck Ebbert wrote:
>>>    
>>>> IOPL is implicitly saved and restored on task switch,
>>>> so explicit check is no longer needed.
>>>>       
>>> Nack.  This is used for paravirt-ops kernels that use IOPL'd 
>>> userspace.      
>>
>> How does that work?  In the stock kernel, anything done by
>> the call to set_iopl_mask() (that was removed by the patch)
>> will be nullified by the 'popfl' at the end of the switch_to()
>> macro.
>>   
>
> Who put a popfl back in switch_to?  I took it out some time ago.  It 
> should not be there.  The only reason for it was to stop IOPL leaking 
> from one process to another from a sleep during a sysenter based 
> system call.

Ok, checking shows Linus put it back to stop NT leakage.  This is 
correct, but unlikely.  Would be nice to avoid it unless absolutely 
necessary.  Perhaps xor eflags old and new and only set_system_eflags() 
if non-ALU bits have changed.

Zach

Re: [patch] i386: remove IOPL check on task switch

[Linus Torvalds]

On Sat, 4 Nov 2006, Zachary Amsden wrote:
>
> Ok, checking shows Linus put it back to stop NT leakage.  This is correct, but
> unlikely.  Would be nice to avoid it unless absolutely necessary.  Perhaps xor
> eflags old and new and only set_system_eflags() if non-ALU bits have changed.

Not just NT. AC also leaked, and caused crashes in other programs (Wine) 
that didn't expect AC to be set and did unaligned accesses.

		Linus

Re: [patch] i386: remove IOPL check on task switch

[Zachary Amsden]

Linus Torvalds wrote:
> On Sat, 4 Nov 2006, Zachary Amsden wrote:
>   
>> Ok, checking shows Linus put it back to stop NT leakage.  This is correct, but
>> unlikely.  Would be nice to avoid it unless absolutely necessary.  Perhaps xor
>> eflags old and new and only set_system_eflags() if non-ALU bits have changed.
>>     
>
> Not just NT. AC also leaked, and caused crashes in other programs (Wine) 
> that didn't expect AC to be set and did unaligned accesses.

Yes, AC, NT, IOPL, ID are bad to leak.  DF / TF / RF are impossible to 
leak by privilege contract.  SF, ZF, PF, OF, CF can be clobbered.

VM / VIF / VIP are dealt with in separate switch paths (although I have 
witnessed a VIF leak once from a userspace process that managed to get 
VIF set).  These can't even be set with popf, and require iret to fix.

But 99% of the time, only SF / ZF / PF / OF / CF will be different, so 
you can avoid the popf.

Zach

Re: [patch] i386: remove IOPL check on task switch

[Andi Kleen]

 
> Ok, checking shows Linus put it back to stop NT leakage. 

It is also needed to stop AC leakage

> This is  
> correct, but unlikely.  Would be nice to avoid it unless absolutely 
> necessary.  Perhaps xor eflags old and new and only set_system_eflags() 
> if non-ALU bits have changed.

If it's worth it...

If it's worth for the context switch i think it would make 
more sense to do it for the normal restore_flags() etc. which are much
more common (and could be actually handled in most cases with test + jump + sti)

-Andi

Re: [patch] i386: remove IOPL check on task switch

[Chuck Ebbert]

In-Reply-To: <454B850C.3050402@vmware.com>

On Fri, 03 Nov 2006 10:06:04 -0800, Zachary Amsden wrote:

> Chuck Ebbert wrote:
> > IOPL is implicitly saved and restored on task switch,
> > so explicit check is no longer needed.
> 
> Nack.  This is used for paravirt-ops kernels that use IOPL'd userspace.  

How does that work?  In the stock kernel, anything done by
the call to set_iopl_mask() (that was removed by the patch)
will be nullified by the 'popfl' at the end of the switch_to()
macro.

-- 
Chuck
"Even supernovas have their duller moments."