Re: [Xen-devel] [PATCH v2 0/7] x86: Meltdown band-aid overhead reduction

2018-02-14 Thread Jan Beulich
>>> On 13.02.18 at 19:39,  wrote:
> On Feb 7, 2018, at 11:05, Jan Beulich  wrote:
>> 
>> 1: slightly reduce Meltdown band-aid overhead
>> 2: remove CR reads from exit-to-guest path
>> 3: introduce altinstruction_nop assembler macro
>> 4: NOP out most XPTI entry/exit code when it's not in use
>> 5: avoid double CR3 reload when switching to guest user mode
>> 6: disable XPTI when RDCL_NO
>> 7: x86: log XPTI enabled status
> 
> Since work on XPTI is ongoing, will these improvements to XPTI-stage-1 be 
> published via http://xenbits.xen.org/xsa/xsa254/README.pti? 

They first of all need to go in (or whatever sub-portion of it is
deemed acceptable). And then I'm not sure we should further
clutter the XSA with performance improvements - I'd be
inclined to add further commits to the list only if we find issues
with the current approach.

Jan


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

Re: [Xen-devel] [PATCH v2 0/7] x86: Meltdown band-aid overhead reduction

2018-02-13 Thread Rich Persaud
On Feb 7, 2018, at 11:05, Jan Beulich  wrote:
> 
> 1: slightly reduce Meltdown band-aid overhead
> 2: remove CR reads from exit-to-guest path
> 3: introduce altinstruction_nop assembler macro
> 4: NOP out most XPTI entry/exit code when it's not in use
> 5: avoid double CR3 reload when switching to guest user mode
> 6: disable XPTI when RDCL_NO
> 7: x86: log XPTI enabled status

Since work on XPTI is ongoing, will these improvements to XPTI-stage-1 be 
published via http://xenbits.xen.org/xsa/xsa254/README.pti?

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