Re: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Andrew S. Baker]

They may not.   It all depends on how they handle context switching for
kernel mode access.

This is going to be very interesting at the CPU maker level, too...

The GPU vendors may not be out of the woods, either.

Regards,

 *ASB*



On Thu, Jan 4, 2018 at 10:41 AM, Jack Kramer <j...@smalltype.net> wrote:

> It wouldn’t surprise me if Apple actually ends up with a speed advantage
> after all this shakes out given their kernel architecture. The Mach guys
> are finally going to have a win of sorts.
>
> 
> Jack Kramer
>
>
> On Jan 4, 2018, at 10:09 AM, Michael B. Smith <mich...@smithcons.com>
> wrote:
>
> None.
>
>
>
> Microsoft says it’s “minimal but dependent on customer workload”. This is
> confirmed by what I’ve seen posted on twitter so far. A couple of
> percentage points.
>
>
>
> Linux is seeing a minimum of 5% performance reduction and as high as 23%
> (so far). It may, for once, finally be true to say: Windows is faster than
> Linux. J
>
>
>
> (Paul Thurrott is gloating about this on twitter.)
>
>
>
> *From:* listsad...@lists.myitforum.com [mailto:listsadmin@lists.
> myitforum.com <listsad...@lists.myitforum.com>] *On Behalf Of *Joseph L.
> Casale
> *Sent:* Thursday, January 4, 2018 9:54 AM
> *To:* ntsysadm@lists.myitforum.com
> *Subject:* RE: [NTSysADM] New blog post: Windows Speculative Execution
> Client/Server Patches/Mitigations/Detection Summary
>
>
>
> Which article quantifies the performance degradation?
>
>
>
> *From:* listsad...@lists.myitforum.com [mailto:listsadmin@lists.
> myitforum.com <listsad...@lists.myitforum.com>] *On Behalf Of *Ed Ziots
> *Sent:* Thursday, January 4, 2018 9:34 AM
> *To:* ntsysadm@lists.myitforum.com
> *Subject:* Re: [NTSysADM] New blog post: Windows Speculative Execution
> Client/Server Patches/Mitigations/Detection Summary
>
>
>
> Yeah this one is going to be a nasty one but looks like cloud service
> providers are patching for it now. But saw there might be a 30 percebt
> performance hit with windows.. is pretty ugly..
>
>
>
> Nice job intel.. way to mess it up again.
>
>
>
> Ed
>
>
>
> On Jan 4, 2018 8:54 AM, "Michael B. Smith" <mich...@smithcons.com> wrote:
>
> New blog post: #Windows Speculative Execution Client/Server
> Patches/Mitigations/Detection Summary
>
> http://www.essential.exchange/2018/01/04/windows-
> speculative-execution-client-server-patches-mitigations-detection-summary/
>
>
>
> Not intended to be comprehensive, but as of the morning of 4-January, all
> the worthwhile information I can find published by Microsoft on the
> Speculative Execution issue…
>
>

RE: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Michael B. Smith]

Yeah, DragonFlyBSD also re-implemented the BSD kernel as a message passing 
architecture. They’ve declared a win as well. ☺

From: listsad...@lists.myitforum.com [mailto:listsad...@lists.myitforum.com] On 
Behalf Of Jack Kramer
Sent: Thursday, January 4, 2018 10:41 AM
To: ntsysadm@lists.myitforum.com
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

It wouldn’t surprise me if Apple actually ends up with a speed advantage after 
all this shakes out given their kernel architecture. The Mach guys are finally 
going to have a win of sorts.

Jack Kramer


On Jan 4, 2018, at 10:09 AM, Michael B. Smith 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
None.

Microsoft says it’s “minimal but dependent on customer workload”. This is 
confirmed by what I’ve seen posted on twitter so far. A couple of percentage 
points.

Linux is seeing a minimum of 5% performance reduction and as high as 23% (so 
far). It may, for once, finally be true to say: Windows is faster than Linux. ☺

(Paul Thurrott is gloating about this on twitter.)

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Joseph L. Casale
Sent: Thursday, January 4, 2018 9:54 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: RE: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Which article quantifies the performance degradation?

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Ed Ziots
Sent: Thursday, January 4, 2018 9:34 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Yeah this one is going to be a nasty one but looks like cloud service providers 
are patching for it now. But saw there might be a 30 percebt performance hit 
with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith" 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
New blog post: #Windows Speculative Execution Client/Server 
Patches/Mitigations/Detection Summary
http://www.essential.exchange/2018/01/04/windows-speculative-execution-client-server-patches-mitigations-detection-summary/

Not intended to be comprehensive, but as of the morning of 4-January, all the 
worthwhile information I can find published by Microsoft on the Speculative 
Execution issue…

Re: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Jack Kramer]

It wouldn’t surprise me if Apple actually ends up with a speed advantage after 
all this shakes out given their kernel architecture. The Mach guys are finally 
going to have a win of sorts.


Jack Kramer


On Jan 4, 2018, at 10:09 AM, Michael B. Smith 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:

None.

Microsoft says it’s “minimal but dependent on customer workload”. This is 
confirmed by what I’ve seen posted on twitter so far. A couple of percentage 
points.

Linux is seeing a minimum of 5% performance reduction and as high as 23% (so 
far). It may, for once, finally be true to say: Windows is faster than Linux. ☺

(Paul Thurrott is gloating about this on twitter.)

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Joseph L. Casale
Sent: Thursday, January 4, 2018 9:54 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: RE: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Which article quantifies the performance degradation?

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Ed Ziots
Sent: Thursday, January 4, 2018 9:34 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Yeah this one is going to be a nasty one but looks like cloud service providers 
are patching for it now. But saw there might be a 30 percebt performance hit 
with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith" 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
New blog post: #Windows Speculative Execution Client/Server 
Patches/Mitigations/Detection Summary
http://www.essential.exchange/2018/01/04/windows-speculative-execution-client-server-patches-mitigations-detection-summary/

Not intended to be comprehensive, but as of the morning of 4-January, all the 
worthwhile information I can find published by Microsoft on the Speculative 
Execution issue…

RE: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Michael B. Smith]

So I looked at the at patches posted on the Linux Kernel mailing list, and 
their “quick solution” was to convert everything to thunks. Which effectively 
discards any benefit offered by the kernel page table cache (which is the cause 
of this problem). They have plans to improve those patches and reduce the 
performance hit. Microsoft probably had a better solution. I’m sure people are 
out there reverse-engineering their little fingers off to figure it out.

From: Michael B. Smith
Sent: Thursday, January 4, 2018 10:01 AM
To: ntsysadm@lists.myitforum.com
Subject: RE: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

None.

Microsoft says it’s “minimal but dependent on customer workload”. This is 
confirmed by what I’ve seen posted on twitter so far. A couple of percentage 
points.

Linux is seeing a minimum of 5% performance reduction and as high as 23% (so 
far). It may, for once, finally be true to say: Windows is faster than Linux. ☺

(Paul Thurrott is gloating about this on twitter.)

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Joseph L. Casale
Sent: Thursday, January 4, 2018 9:54 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: RE: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Which article quantifies the performance degradation?

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Ed Ziots
Sent: Thursday, January 4, 2018 9:34 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Yeah this one is going to be a nasty one but looks like cloud service providers 
are patching for it now. But saw there might be a 30 percebt performance hit 
with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith" 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
New blog post: #Windows Speculative Execution Client/Server 
Patches/Mitigations/Detection Summary
http://www.essential.exchange/2018/01/04/windows-speculative-execution-client-server-patches-mitigations-detection-summary/

Not intended to be comprehensive, but as of the morning of 4-January, all the 
worthwhile information I can find published by Microsoft on the Speculative 
Execution issue…

RE: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Michael B. Smith]

None.

Microsoft says it’s “minimal but dependent on customer workload”. This is 
confirmed by what I’ve seen posted on twitter so far. A couple of percentage 
points.

Linux is seeing a minimum of 5% performance reduction and as high as 23% (so 
far). It may, for once, finally be true to say: Windows is faster than Linux. ☺

(Paul Thurrott is gloating about this on twitter.)

From: listsad...@lists.myitforum.com [mailto:listsad...@lists.myitforum.com] On 
Behalf Of Joseph L. Casale
Sent: Thursday, January 4, 2018 9:54 AM
To: ntsysadm@lists.myitforum.com
Subject: RE: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Which article quantifies the performance degradation?

From: listsad...@lists.myitforum.com<mailto:listsad...@lists.myitforum.com> 
[mailto:listsad...@lists.myitforum.com] On Behalf Of Ed Ziots
Sent: Thursday, January 4, 2018 9:34 AM
To: ntsysadm@lists.myitforum.com<mailto:ntsysadm@lists.myitforum.com>
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Yeah this one is going to be a nasty one but looks like cloud service providers 
are patching for it now. But saw there might be a 30 percebt performance hit 
with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith" 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
New blog post: #Windows Speculative Execution Client/Server 
Patches/Mitigations/Detection Summary
http://www.essential.exchange/2018/01/04/windows-speculative-execution-client-server-patches-mitigations-detection-summary/

Not intended to be comprehensive, but as of the morning of 4-January, all the 
worthwhile information I can find published by Microsoft on the Speculative 
Execution issue…

RE: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Joseph L. Casale]

Which article quantifies the performance degradation?

From: listsad...@lists.myitforum.com [mailto:listsad...@lists.myitforum.com] On 
Behalf Of Ed Ziots
Sent: Thursday, January 4, 2018 9:34 AM
To: ntsysadm@lists.myitforum.com
Subject: Re: [NTSysADM] New blog post: Windows Speculative Execution 
Client/Server Patches/Mitigations/Detection Summary

Yeah this one is going to be a nasty one but looks like cloud service providers 
are patching for it now. But saw there might be a 30 percebt performance hit 
with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith" 
<mich...@smithcons.com<mailto:mich...@smithcons.com>> wrote:
New blog post: #Windows Speculative Execution Client/Server 
Patches/Mitigations/Detection Summary
http://www.essential.exchange/2018/01/04/windows-speculative-execution-client-server-patches-mitigations-detection-summary/

Not intended to be comprehensive, but as of the morning of 4-January, all the 
worthwhile information I can find published by Microsoft on the Speculative 
Execution issue…

Re: [NTSysADM] New blog post: Windows Speculative Execution Client/Server Patches/Mitigations/Detection Summary

[Ed Ziots]

Yeah this one is going to be a nasty one but looks like cloud service
providers are patching for it now. But saw there might be a 30 percebt
performance hit with windows.. is pretty ugly..

Nice job intel.. way to mess it up again.

Ed

On Jan 4, 2018 8:54 AM, "Michael B. Smith"  wrote:

> New blog post: #Windows Speculative Execution Client/Server
> Patches/Mitigations/Detection Summary
>
> http://www.essential.exchange/2018/01/04/windows-
> speculative-execution-client-server-patches-mitigations-detection-summary/
>
>
>
> Not intended to be comprehensive, but as of the morning of 4-January, all
> the worthwhile information I can find published by Microsoft on the
> Speculative Execution issue…
>