Will this patch to windows for Intel chips impact speed for nvda?


Brian's Mail list account <bglists@...>
 

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm not saying this sounds like paranoia, as I do not know enough to make such a comment, but if the security hole has been in the chips for a decade, I'm not so sure its that much of an issue. at any rate I guess all those out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel open to action to give everyone new chips!
Brian
bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.


Antony Stone
 

A large number of security flaws (in both software and hardware) do not get
discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not so
sure its that much of an issue" is a dangerous assumption, and not borne out
by Microsoft's sudden activity to patch Windows in such a way that it will
have a definite performance penalty. If the 10-year old security flaw were not
much of a problem, Microsoft wouldn't react in a way that unavoidably impacts
the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm
not saying this sounds like paranoia, as I do not know enough to make such
a comment, but if the security hole has been in the chips for a decade,
I'm not so sure its that much of an issue. at any rate I guess all those
out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!
Brian
--
Someone has stolen all the toilets from New Scotland Yard. Police say they
have absolutely nothing to go on.

Please reply to the list;
please *don't* CC me.


Brian's Mail list account <bglists@...>
 

Oh really, well, I'm not so sure. It feeds into the mindset that says Microsofts eventual aim is to lock people in to getting software from its store and nowhere else in the 'interests' of security, when in reality they want to do what Apple is doing, create an ecosystem and make it hard for third parties to just allow their software to run on a system.
Whether you suspect ulterior motives or not is your choice.

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


A large number of security flaws (in both software and hardware) do not get
discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not so
sure its that much of an issue" is a dangerous assumption, and not borne out
by Microsoft's sudden activity to patch Windows in such a way that it will
have a definite performance penalty. If the 10-year old security flaw were not
much of a problem, Microsoft wouldn't react in a way that unavoidably impacts
the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm
not saying this sounds like paranoia, as I do not know enough to make such
a comment, but if the security hole has been in the chips for a decade,
I'm not so sure its that much of an issue. at any rate I guess all those
out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!
Brian
--
Someone has stolen all the toilets from New Scotland Yard. Police say they
have absolutely nothing to go on.

Please reply to the list;
please *don't* CC me.


Antony Stone
 

I do not suspect ulterior motives on Microsoft's part in this case, because
the Linux kernel is being patched in just the same way for just the same
reason.


Antony.

On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
Groups.Io wrote:

Oh really, well, I'm not so sure. It feeds into the mindset that says
Microsofts eventual aim is to lock people in to getting software from its
store and nowhere else in the 'interests' of security, when in reality they
want to do what Apple is doing, create an ecosystem and make it hard for
third parties to just allow their software to run on a system.
Whether you suspect ulterior motives or not is your choice.

Brian

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?

A large number of security flaws (in both software and hardware) do not
get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not
so sure its that much of an issue" is a dangerous assumption, and not
borne out by Microsoft's sudden activity to patch Windows in such a way
that it will have a definite performance penalty. If the 10-year old
security flaw were not much of a problem, Microsoft wouldn't react in a way
that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via

Groups.Io wrote:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case.
I'm not saying this sounds like paranoia, as I do not know enough to make
such a comment, but if the security hole has been in the chips for a
decade, I'm not so sure its that much of an issue. at any rate I guess
all those out there with AMD chips will be chortling loudly.

If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!

Brian
--
When you find yourself arguing with an idiot,
you should first of all make sure that the other person isn't doing the same
thing.

Please reply to the list;
please *don't* CC me.


Brian's Mail list account <bglists@...>
 

It just seems odd that we have if you believe all this lived with this for a decade and I'd have thought there was now a bit of the shutting the stable door when all the horses have gone about trying to retro patch it I understand the biggest losers could be the big server companies who might be using Intel chipped hardware, and hence need things sorting out.
So Maybe I should get my claim for a new processor in tomorrow to be in time....:-)

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


I do not suspect ulterior motives on Microsoft's part in this case, because
the Linux kernel is being patched in just the same way for just the same
reason.


Antony.

On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
Groups.Io wrote:

Oh really, well, I'm not so sure. It feeds into the mindset that says
Microsofts eventual aim is to lock people in to getting software from its
store and nowhere else in the 'interests' of security, when in reality they
want to do what Apple is doing, create an ecosystem and make it hard for
third parties to just allow their software to run on a system.
Whether you suspect ulterior motives or not is your choice.

Brian

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?

A large number of security flaws (in both software and hardware) do not
get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not
so sure its that much of an issue" is a dangerous assumption, and not
borne out by Microsoft's sudden activity to patch Windows in such a way
that it will have a definite performance penalty. If the 10-year old
security flaw were not much of a problem, Microsoft wouldn't react in a way
that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via

Groups.Io wrote:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case.
I'm not saying this sounds like paranoia, as I do not know enough to make
such a comment, but if the security hole has been in the chips for a
decade, I'm not so sure its that much of an issue. at any rate I guess
all those out there with AMD chips will be chortling loudly.

If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!

Brian
--
When you find yourself arguing with an idiot,
you should first of all make sure that the other person isn't doing the same
thing.

Please reply to the list;
please *don't* CC me.


Alex Kenny
 

While this has existed for the past decade, it has only been discovered recently. We have been "living with it", because, as far as we know no one had discover and exploit it. These patches, while slowing performance slightly and only in some use cases, will fix the problem.


Also, jumping immediately to some conspiracy by Microsoft doesn't make sense. First, as Anthony previously stated, Linux is being patched. Mac OS will also be updated.

Second, this issue affects the kernel, which has nothing to do with whether traditional Win32 or Windows Store apps are running on top of it. Deliberately slowing the kernel to induce people to use Windows Store apps is completely nonsensical.

This patch should only significantly affect use cases doing a lot of work in kernel space. NVDA will likely not be affected by this.  

On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:
It just seems odd that we have if you believe all this lived with this for a
decade and I'd have thought there was now a bit of the shutting the stable
door when all the horses have gone about trying to retro patch it I
understand the biggest losers could be the big server companies who might be
using Intel chipped hardware, and hence need things sorting out.
 So Maybe I should get my claim for a new processor in tomorrow to be in
time....:-)

Brian

bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
----- Original Message -----
From: "Antony Stone" <antony.stone@...>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?


>I do not suspect ulterior motives on Microsoft's part in this case, because
> the Linux kernel is being patched in just the same way for just the same
> reason.
>
>
> Antony.
>
> On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
> Groups.Io wrote:
>
>> Oh really, well, I'm not so sure. It feeds into the mindset that says
>> Microsofts eventual aim is to lock people in to getting software from its
>> store and nowhere else in the 'interests' of security, when in reality
>> they
>> want to do what Apple is doing, create an ecosystem and make it hard for
>> third parties to just allow their software to run on a system.
>>  Whether you suspect ulterior motives or not is your  choice.
>>
>> Brian
>>
>> ----- Original Message -----
>> From: "Antony Stone" <antony.stone@...>
>> To: <nvda@nvda.groups.io>
>> Sent: Wednesday, January 03, 2018 10:03 AM
>> Subject: Re: [nvda] Will this patch to windows for Intel chips impact
>> speed
>> for nvda?
>>
>> >A large number of security flaws (in both software and hardware) do not
>> >get discovered for a long time after they crept into the design.
>> >
>> > I think "if the security hole has been in the chips for a decade, I'm
>> > not
>> > so sure its that much of an issue" is a dangerous assumption, and not
>> > borne out by Microsoft's sudden activity to patch Windows in such a way
>> > that it will have a definite performance penalty.  If the 10-year old
>> > security flaw were not much of a problem, Microsoft wouldn't react in a
>> > way
>> > that unavoidably impacts the majority of Windows users.
>> >
>> >
>> > Antony.
>> >
>> > On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
>> >
>> > Groups.Io wrote:
>> >> https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
>> >>
>> >> I hope that link works, but its easy to find on that site in any case.
>> >> I'm not saying this sounds like paranoia, as I do not know enough to
>> >> make
>> >> such a comment, but if the security hole has been in the chips for a
>> >> decade, I'm not so sure its that much of an issue. at any rate I guess
>> >> all those out there with AMD chips will be chortling loudly.
>> >>
>> >>  If its a fault in the chip this seems that it could leave Intel open
>> >> to
>> >>  action to give everyone new chips!
>> >>
>> >>  Brian
>
> --
> When you find yourself arguing with an idiot,
> you should first of all make sure that the other person isn't doing the
> same
> thing.
>
>                                                   Please reply to the
> list;
>                                                         please *don't* CC
> me.
>
>
>





Gene
 

The conspiracy theory makes me wonder if you, Brian, read the article before you expressed the conspiracy hypothesis.  the facts in the article, such as that this affects Linux as well, and that the problem wasn't discovered by Microsoft and that Linux developers are working on a software fix as well, all demolish the conspiracy idea. 
 
Gene
----- Original Message -----

From: Alex Kenny
Sent: Wednesday, January 03, 2018 9:26 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

While this has existed for the past decade, it has only been discovered recently. We have been "living with it", because, as far as we know no one had discover and exploit it. These patches, while slowing performance slightly and only in some use cases, will fix the problem.


Also, jumping immediately to some conspiracy by Microsoft doesn't make sense. First, as Anthony previously stated, Linux is being patched. Mac OS will also be updated.

Second, this issue affects the kernel, which has nothing to do with whether traditional Win32 or Windows Store apps are running on top of it. Deliberately slowing the kernel to induce people to use Windows Store apps is completely nonsensical.

This patch should only significantly affect use cases doing a lot of work in kernel space. NVDA will likely not be affected by this.  

On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:
It just seems odd that we have if you believe all this lived with this for a
decade and I'd have thought there was now a bit of the shutting the stable
door when all the horses have gone about trying to retro patch it I
understand the biggest losers could be the big server companies who might be
using Intel chipped hardware, and hence need things sorting out.
 So Maybe I should get my claim for a new processor in tomorrow to be in
time....:-)

Brian

bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
----- Original Message -----
From: "Antony Stone" <antony.stone@...>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?


>I do not suspect ulterior motives on Microsoft's part in this case, because
> the Linux kernel is being patched in just the same way for just the same
> reason.
>
>
> Antony.
>
> On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
> Groups.Io wrote:
>
>> Oh really, well, I'm not so sure. It feeds into the mindset that says
>> Microsofts eventual aim is to lock people in to getting software from its
>> store and nowhere else in the 'interests' of security, when in reality
>> they
>> want to do what Apple is doing, create an ecosystem and make it hard for
>> third parties to just allow their software to run on a system.
>>  Whether you suspect ulterior motives or not is your  choice.
>>
>> Brian
>>
>> ----- Original Message -----
>> From: "Antony Stone" <antony.stone@...>
>> To: <nvda@nvda.groups.io>
>> Sent: Wednesday, January 03, 2018 10:03 AM
>> Subject: Re: [nvda] Will this patch to windows for Intel chips impact
>> speed
>> for nvda?
>>
>> >A large number of security flaws (in both software and hardware) do not
>> >get discovered for a long time after they crept into the design.
>> >
>> > I think "if the security hole has been in the chips for a decade, I'm
>> > not
>> > so sure its that much of an issue" is a dangerous assumption, and not
>> > borne out by Microsoft's sudden activity to patch Windows in such a way
>> > that it will have a definite performance penalty.  If the 10-year old
>> > security flaw were not much of a problem, Microsoft wouldn't react in a
>> > way
>> > that unavoidably impacts the majority of Windows users.
>> >
>> >
>> > Antony.
>> >
>> > On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
>> >
>> > Groups.Io wrote:
>> >> https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
>> >>
>> >> I hope that link works, but its easy to find on that site in any case.
>> >> I'm not saying this sounds like paranoia, as I do not know enough to
>> >> make
>> >> such a comment, but if the security hole has been in the chips for a
>> >> decade, I'm not so sure its that much of an issue. at any rate I guess
>> >> all those out there with AMD chips will be chortling loudly.
>> >>
>> >>  If its a fault in the chip this seems that it could leave Intel open
>> >> to
>> >>  action to give everyone new chips!
>> >>
>> >>  Brian
>
> --
> When you find yourself arguing with an idiot,
> you should first of all make sure that the other person isn't doing the
> same
> thing.
>
>                                                   Please reply to the
> list;
>                                                         please *don't* CC
> me.
>
>
>





Gene
 

I just looked at the original message.  it appears you did read the article before commenting on it.  In that case, a conspiracy theory is ruled out by the contents of the article. 
 
Gene

----- Original Message -----
Sent: Wednesday, January 03, 2018 3:52 AM
Subject: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm
not saying this sounds like paranoia, as I do not know enough to make such a
comment, but if the security hole has been in the chips for a decade, I'm
not so sure its that much of an issue. at any rate I guess all those out
there with AMD chips will be chortling loudly.
 If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!
 Brian
bglists@...
Sent via blueyonder.
Please address personal email to:-
briang1@..., putting 'Brian Gaff'
in the display name field.




 

Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.

On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do not get
discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not so
sure its that much of an issue" is a dangerous assumption, and not borne out
by Microsoft's sudden activity to patch Windows in such a way that it will
have a definite performance penalty. If the 10-year old security flaw were not
much of a problem, Microsoft wouldn't react in a way that unavoidably impacts
the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm
not saying this sounds like paranoia, as I do not know enough to make such
a comment, but if the security hole has been in the chips for a decade,
I'm not so sure its that much of an issue. at any rate I guess all those
out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!
Brian


 

Who knows, its even hitting the i8 series, they do say anything 4th gen and over may be less effected which means me and a few in my family with 3rd generations will be out of luck.

It does make me wander though they descover this now even on mac, maybe have been incubating a fix, just like y2k for ages before and I have used my system for well everything.

Ofcause the issue is once its out, people will try to exploit it.

On 4/01/2018 4:00 a.m., Brian's Mail list account via Groups.Io wrote:
It just seems odd that we have if you believe all this lived with this for a decade and I'd have thought there was now a bit of the shutting the stable door when all the horses have gone about trying to retro patch it I understand the biggest losers could be the big server companies who might be using Intel chipped hardware, and hence need things sorting out.
So Maybe I should get my claim for a new processor in tomorrow to be in time....:-)

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message ----- From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


I do not suspect ulterior motives on Microsoft's part in this case, because
the Linux kernel is being patched in just the same way for just the same
reason.


Antony.

On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
Groups.Io wrote:

Oh really, well, I'm not so sure. It feeds into the mindset that says
Microsofts eventual aim is to lock people in to getting software from its
store and nowhere else in the 'interests' of security, when in reality they
want to do what Apple is doing, create an ecosystem and make it hard for
third parties to just allow their software to run on a system.
 Whether you suspect ulterior motives or not is your  choice.

Brian

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?

A large number of security flaws (in both software and hardware) do
not
get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade,
I'm > not
so sure its that much of an issue" is a dangerous assumption, and not
borne out by Microsoft's sudden activity to patch Windows in such
a way
that it will have a definite performance penalty.  If the 10-year old
security flaw were not much of a problem, Microsoft wouldn't react
in a > way
that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list
account via

Groups.Io wrote:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case.
I'm not saying this sounds like paranoia, as I do not know enough
to >> make
such a comment, but if the security hole has been in the chips for a
decade, I'm not so sure its that much of an issue. at any rate I
guess
all those out there with AMD chips will be chortling loudly.

  If its a fault in the chip this seems that it could leave Intel
open >> to
  action to give everyone new chips!

  Brian
--
When you find yourself arguing with an idiot,
you should first of all make sure that the other person isn't doing the same
thing.

                                                  Please reply to the list;
                                                        please *don't* CC me.




.


 

Hi,
The slow down has to do with kernel mitigations to work around this problem.
Technical: as I noted on another forum (can get very geeky in the end but please bare with me):
One way to speed up programs is doing things beforehand. On hardware, this is called "speculative execution". Basically, computer programs can perform tasks based on a condition (or a set of them). When a program comes across a fork, it can walk down a path based on the condition it is looking for (usually prefixed with the word "if"). Processors can take advantage of this and say, "oh, let me walk down a random path that can lead me to a jackpot so I don't have to come here again." If the processor gambled right, it leads to huge performance boost, because the program can traverse this path without giving a second thought. But what if the chosen path leads to thorn bushes? The process will then say, "oops, I was wrong, so let me retrace my steps and wait for my guest (program) to show me where to go next."
Under some circumstances, when the processor executes the "wrong" path (although it'll think it was right then), the processor will gather the things needed to travel this path. They include next set of instructions and data needed to carry out what it thinks is the "best" match. The thing is, the only thing the processor can understand is 1's and 0's, so it won't care what it will meet down the road, or the fact that it can leak sensitive information (that's the high-level overview of the vulnerability disclosed by Google). In a nutshell, now the processor will be told by its boss (operating system kernel) that it needs to carry some protective gear when traversing conditional paths in order to make sure the processor won't leak sensitive information again (if possible, but it won't work 100 percent of the time), hence the potential performance degradation in some situations.
As for this affecting NVDA's performance, it could, but we won't know the full extent until next week at the earliest once everyone gets the patch.
Cheers,
Joseph

-----Original Message-----
From: nvda@nvda.groups.io [mailto:nvda@nvda.groups.io] On Behalf Of Shaun Everiss
Sent: Wednesday, January 3, 2018 10:44 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.




On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do
not get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm
not so sure its that much of an issue" is a dangerous assumption, and
not borne out by Microsoft's sudden activity to patch Windows in such
a way that it will have a definite performance penalty. If the
10-year old security flaw were not much of a problem, Microsoft
wouldn't react in a way that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account
via Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case. I'm not saying this sounds like paranoia, as I do not know
enough to make such a comment, but if the security hole has been in
the chips for a decade, I'm not so sure its that much of an issue. at
any rate I guess all those out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel
open to action to give everyone new chips!
Brian


 

Well You can apply the patch manually if you know how.

Any gen4 processer and higher should in theory be impacted less but still.

For me, with things, say a game on first start, there will be some lag, a couple seconds or so, nvda is not effected, obviously as the patch gets adjusted things will continue to evolve.

Servers and low grade processers may be the ones with issues, I have an genuine x86 first gen i3 here and that I admit could have issues.

On one of my win10 systems I have I have noticed that things like transfering files, and other things are more jerky and pronounced, I have not tried with skype but you may just have to wait just that tincy bit longer for things to start, there is a noticable lag when things start.

Ofcause it goes without saying that for the newer systems especially those under support, that bios updates to chip firmwars will probably mitigate this even more.

Aparently they knew as early as december, and one of the hps I have has a system upgrade.

For those on win7 and up search for windows 7 or 8 update history, for those  on win10 hit learn more from windows update then the catalog to get that update after you click on the top article.

If you feel comfortable with this then I would strongly recomend you get this if you can.

Why?

The net is going to struggle with a world wide release I am patching all my systems because I can.

Once you have the patches for everything you need put them on a flash drive and update the systems.

Interestingly, win10 has a full cu, win7 and so on have a security only update.

This does suggest though that this rushed patch will be improved as things come on stream.

On 4/01/2018 8:18 p.m., Joseph Lee wrote:
Hi,
The slow down has to do with kernel mitigations to work around this problem.
Technical: as I noted on another forum (can get very geeky in the end but please bare with me):
One way to speed up programs is doing things beforehand. On hardware, this is called "speculative execution". Basically, computer programs can perform tasks based on a condition (or a set of them). When a program comes across a fork, it can walk down a path based on the condition it is looking for (usually prefixed with the word "if"). Processors can take advantage of this and say, "oh, let me walk down a random path that can lead me to a jackpot so I don't have to come here again." If the processor gambled right, it leads to huge performance boost, because the program can traverse this path without giving a second thought. But what if the chosen path leads to thorn bushes? The process will then say, "oops, I was wrong, so let me retrace my steps and wait for my guest (program) to show me where to go next."
Under some circumstances, when the processor executes the "wrong" path (although it'll think it was right then), the processor will gather the things needed to travel this path. They include next set of instructions and data needed to carry out what it thinks is the "best" match. The thing is, the only thing the processor can understand is 1's and 0's, so it won't care what it will meet down the road, or the fact that it can leak sensitive information (that's the high-level overview of the vulnerability disclosed by Google). In a nutshell, now the processor will be told by its boss (operating system kernel) that it needs to carry some protective gear when traversing conditional paths in order to make sure the processor won't leak sensitive information again (if possible, but it won't work 100 percent of the time), hence the potential performance degradation in some situations.
As for this affecting NVDA's performance, it could, but we won't know the full extent until next week at the earliest once everyone gets the patch.
Cheers,
Joseph

-----Original Message-----
From: nvda@nvda.groups.io [mailto:nvda@nvda.groups.io] On Behalf Of Shaun Everiss
Sent: Wednesday, January 3, 2018 10:44 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.




On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do
not get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm
not so sure its that much of an issue" is a dangerous assumption, and
not borne out by Microsoft's sudden activity to patch Windows in such
a way that it will have a definite performance penalty. If the
10-year old security flaw were not much of a problem, Microsoft
wouldn't react in a way that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account
via Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case. I'm not saying this sounds like paranoia, as I do not know
enough to make such a comment, but if the security hole has been in
the chips for a decade, I'm not so sure its that much of an issue. at
any rate I guess all those out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel
open to action to give everyone new chips!
Brian





Brian's Mail list account <bglists@...>
 

That is not what I meant. Its more likely that when a new chip is out people will use it as a selling point to drive people into buying a new system... again...
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Alex Kenny" <alexkenny08@gmail.com>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 3:26 PM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


While this has existed for the past decade, it has only been discovered
recently. We have been "living with it", because, as far as we know no one
had discover and exploit it. These patches, while slowing performance
slightly and only in some use cases, will fix the problem.


Also, jumping immediately to some conspiracy by Microsoft doesn't make
sense. First, as Anthony previously stated, Linux is being patched. Mac OS
will also be updated.

Second, this issue affects the kernel, which has nothing to do with whether
traditional Win32 or Windows Store apps are running on top of it.
Deliberately slowing the kernel to induce people to use Windows Store apps
is completely nonsensical.

This patch should only significantly affect use cases doing a lot of work
in kernel space. NVDA will likely not be affected by this.

On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io
<bglists=blueyonder.co.uk@groups.io> wrote:

It just seems odd that we have if you believe all this lived with this for
a
decade and I'd have thought there was now a bit of the shutting the stable
door when all the horses have gone about trying to retro patch it I
understand the biggest losers could be the big server companies who might
be
using Intel chipped hardware, and hence need things sorting out.
So Maybe I should get my claim for a new processor in tomorrow to be in
time....:-)

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?


I do not suspect ulterior motives on Microsoft's part in this case,
because
the Linux kernel is being patched in just the same way for just the same
reason.


Antony.

On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
Groups.Io wrote:

Oh really, well, I'm not so sure. It feeds into the mindset that says
Microsofts eventual aim is to lock people in to getting software from
its
store and nowhere else in the 'interests' of security, when in reality
they
want to do what Apple is doing, create an ecosystem and make it hard for
third parties to just allow their software to run on a system.
Whether you suspect ulterior motives or not is your choice.

Brian

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact
speed
for nvda?

A large number of security flaws (in both software and hardware) do not
get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm
not
so sure its that much of an issue" is a dangerous assumption, and not
borne out by Microsoft's sudden activity to patch Windows in such a
way
that it will have a definite performance penalty. If the 10-year old
security flaw were not much of a problem, Microsoft wouldn't react in
a
way
that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account
via

Groups.Io wrote:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case.
I'm not saying this sounds like paranoia, as I do not know enough to
make
such a comment, but if the security hole has been in the chips for a
decade, I'm not so sure its that much of an issue. at any rate I
guess
all those out there with AMD chips will be chortling loudly.

If its a fault in the chip this seems that it could leave Intel open
to
action to give everyone new chips!

Brian
--
When you find yourself arguing with an idiot,
you should first of all make sure that the other person isn't doing the
same
thing.

Please reply to the
list;
please *don't* CC
me.






Brian's Mail list account <bglists@...>
 

I did read the article, yes. However just as with that network bug a few months ago, it had been known about in security circles and was only patched when it became common knowledge.
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Gene" <gsasner@ripco.com>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 5:03 PM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


The conspiracy theory makes me wonder if you, Brian, read the article before you expressed the conspiracy hypothesis. the facts in the article, such as that this affects Linux as well, and that the problem wasn't discovered by Microsoft and that Linux developers are working on a software fix as well, all demolish the conspiracy idea.

Gene
----- Original Message -----


From: Alex Kenny
Sent: Wednesday, January 03, 2018 9:26 AM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


While this has existed for the past decade, it has only been discovered recently. We have been "living with it", because, as far as we know no one had discover and exploit it. These patches, while slowing performance slightly and only in some use cases, will fix the problem.




Also, jumping immediately to some conspiracy by Microsoft doesn't make sense. First, as Anthony previously stated, Linux is being patched. Mac OS will also be updated.


Second, this issue affects the kernel, which has nothing to do with whether traditional Win32 or Windows Store apps are running on top of it. Deliberately slowing the kernel to induce people to use Windows Store apps is completely nonsensical.


This patch should only significantly affect use cases doing a lot of work in kernel space. NVDA will likely not be affected by this.


On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:

It just seems odd that we have if you believe all this lived with this for a
decade and I'd have thought there was now a bit of the shutting the stable
door when all the horses have gone about trying to retro patch it I
understand the biggest losers could be the big server companies who might be
using Intel chipped hardware, and hence need things sorting out.
So Maybe I should get my claim for a new processor in tomorrow to be in
time....:-)

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?


>I do not suspect ulterior motives on Microsoft's part in this case, because
> the Linux kernel is being patched in just the same way for just the same
> reason.
>
>
> Antony.
>
> On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
> Groups.Io wrote:
>
>> Oh really, well, I'm not so sure. It feeds into the mindset that says
>> Microsofts eventual aim is to lock people in to getting software from its
>> store and nowhere else in the 'interests' of security, when in reality
>> they
>> want to do what Apple is doing, create an ecosystem and make it hard for
>> third parties to just allow their software to run on a system.
>> Whether you suspect ulterior motives or not is your choice.
>>
>> Brian
>>
>> ----- Original Message -----
>> From: "Antony Stone" <antony.stone@nvda.open.source.it>
>> To: <nvda@nvda.groups.io>
>> Sent: Wednesday, January 03, 2018 10:03 AM
>> Subject: Re: [nvda] Will this patch to windows for Intel chips impact
>> speed
>> for nvda?
>>
>> >A large number of security flaws (in both software and hardware) do not
>> >get discovered for a long time after they crept into the design.
>> >
>> > I think "if the security hole has been in the chips for a decade, I'm
>> > not
>> > so sure its that much of an issue" is a dangerous assumption, and not
>> > borne out by Microsoft's sudden activity to patch Windows in such a way
>> > that it will have a definite performance penalty. If the 10-year old
>> > security flaw were not much of a problem, Microsoft wouldn't react in a
>> > way
>> > that unavoidably impacts the majority of Windows users.
>> >
>> >
>> > Antony.
>> >
>> > On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
>> >
>> > Groups.Io wrote:
>> >> https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
>> >>
>> >> I hope that link works, but its easy to find on that site in any case.
>> >> I'm not saying this sounds like paranoia, as I do not know enough to
>> >> make
>> >> such a comment, but if the security hole has been in the chips for a
>> >> decade, I'm not so sure its that much of an issue. at any rate I guess
>> >> all those out there with AMD chips will be chortling loudly.
>> >>
>> >> If its a fault in the chip this seems that it could leave Intel open
>> >> to
>> >> action to give everyone new chips!
>> >>
>> >> Brian
>
> --
> When you find yourself arguing with an idiot,
> you should first of all make sure that the other person isn't doing the
> same
> thing.
>
> Please reply to the
> list;
> please *don't* CC
> me.
>
>
>


Brian's Mail list account <bglists@...>
 

Yes that was my whole point really. It all a bit late. Who knows how long the security authorities have been using these vulnerabilities. The point I made some time ago is that there is never ever going to be anything that is totally secure, no matter what you do, and its realising this and not getting paranoid that is at issue a lot of the time
The main vulnerability these days is the internet of things, which give entry points to systems that are going to be hard to plug.
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Shaun Everiss" <sm.everiss@gmail.com>
To: <nvda@nvda.groups.io>
Sent: Thursday, January 04, 2018 6:44 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.




On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do not get
discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm not so
sure its that much of an issue" is a dangerous assumption, and not borne out
by Microsoft's sudden activity to patch Windows in such a way that it will
have a definite performance penalty. If the 10-year old security flaw were not
much of a problem, Microsoft wouldn't react in a way that unavoidably impacts
the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any case. I'm
not saying this sounds like paranoia, as I do not know enough to make such
a comment, but if the security hole has been in the chips for a decade,
I'm not so sure its that much of an issue. at any rate I guess all those
out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel open to
action to give everyone new chips!
Brian


Brian's Mail list account <bglists@...>
 

I'd imagine this is the very confusing differences between multithreading, Multi cores and multi tasking.
My core is not multithreaded, so if that is the case, and its a 32 bit app, one supposes it uses a core and stays there with its own memory map etc, but 64 bit processes can share so this might be an issue as well, since in your example it would need to read the command its going to execute at the same time to make sure no parameters are needed from not yet done bits of calculations running elsewhere, if there are these provisos then its not going to be able to do that command obviously.
However my reading of the bug was that despite it pulling back and not doing the calc, it does store lots of data to make that operation faster even it does need to do it and as it may or may not be that core then that has to be able to be seen by the overall system, breaking the 'secret' hidden parts of a process in the way it allows this.
I'd suggest UIA will be the loser, so more sluggish explorer windows and that sort of thing.
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Joseph Lee" <joseph.lee22590@gmail.com>
To: <nvda@nvda.groups.io>
Sent: Thursday, January 04, 2018 7:18 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


Hi,
The slow down has to do with kernel mitigations to work around this problem.
Technical: as I noted on another forum (can get very geeky in the end but please bare with me):
One way to speed up programs is doing things beforehand. On hardware, this is called "speculative execution". Basically, computer programs can perform tasks based on a condition (or a set of them). When a program comes across a fork, it can walk down a path based on the condition it is looking for (usually prefixed with the word "if"). Processors can take advantage of this and say, "oh, let me walk down a random path that can lead me to a jackpot so I don't have to come here again." If the processor gambled right, it leads to huge performance boost, because the program can traverse this path without giving a second thought. But what if the chosen path leads to thorn bushes? The process will then say, "oops, I was wrong, so let me retrace my steps and wait for my guest (program) to show me where to go next."
Under some circumstances, when the processor executes the "wrong" path (although it'll think it was right then), the processor will gather the things needed to travel this path. They include next set of instructions and data needed to carry out what it thinks is the "best" match. The thing is, the only thing the processor can understand is 1's and 0's, so it won't care what it will meet down the road, or the fact that it can leak sensitive information (that's the high-level overview of the vulnerability disclosed by Google). In a nutshell, now the processor will be told by its boss (operating system kernel) that it needs to carry some protective gear when traversing conditional paths in order to make sure the processor won't leak sensitive information again (if possible, but it won't work 100 percent of the time), hence the potential performance degradation in some situations.
As for this affecting NVDA's performance, it could, but we won't know the full extent until next week at the earliest once everyone gets the patch.
Cheers,
Joseph

-----Original Message-----
From: nvda@nvda.groups.io [mailto:nvda@nvda.groups.io] On Behalf Of Shaun Everiss
Sent: Wednesday, January 3, 2018 10:44 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.




On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do
not get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm
not so sure its that much of an issue" is a dangerous assumption, and
not borne out by Microsoft's sudden activity to patch Windows in such
a way that it will have a definite performance penalty. If the
10-year old security flaw were not much of a problem, Microsoft
wouldn't react in a way that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account
via Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case. I'm not saying this sounds like paranoia, as I do not know
enough to make such a comment, but if the security hole has been in
the chips for a decade, I'm not so sure its that much of an issue. at
any rate I guess all those out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel
open to action to give everyone new chips!
Brian


Brian's Mail list account <bglists@...>
 

How would you know what level processor you have in a given machine? Obviously it tells you you have four cores etc in my I5, but there are different issues to these chips.
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Shaun Everiss" <sm.everiss@gmail.com>
To: <nvda@nvda.groups.io>
Sent: Thursday, January 04, 2018 8:40 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


Well You can apply the patch manually if you know how.

Any gen4 processer and higher should in theory be impacted less but still.

For me, with things, say a game on first start, there will be some lag, a couple seconds or so, nvda is not effected, obviously as the patch gets adjusted things will continue to evolve.

Servers and low grade processers may be the ones with issues, I have an genuine x86 first gen i3 here and that I admit could have issues.

On one of my win10 systems I have I have noticed that things like transfering files, and other things are more jerky and pronounced, I have not tried with skype but you may just have to wait just that tincy bit longer for things to start, there is a noticable lag when things start.

Ofcause it goes without saying that for the newer systems especially those under support, that bios updates to chip firmwars will probably mitigate this even more.

Aparently they knew as early as december, and one of the hps I have has a system upgrade.

For those on win7 and up search for windows 7 or 8 update history, for those on win10 hit learn more from windows update then the catalog to get that update after you click on the top article.

If you feel comfortable with this then I would strongly recomend you get this if you can.

Why?

The net is going to struggle with a world wide release I am patching all my systems because I can.

Once you have the patches for everything you need put them on a flash drive and update the systems.

Interestingly, win10 has a full cu, win7 and so on have a security only update.

This does suggest though that this rushed patch will be improved as things come on stream.




On 4/01/2018 8:18 p.m., Joseph Lee wrote:
Hi,
The slow down has to do with kernel mitigations to work around this problem.
Technical: as I noted on another forum (can get very geeky in the end but please bare with me):
One way to speed up programs is doing things beforehand. On hardware, this is called "speculative execution". Basically, computer programs can perform tasks based on a condition (or a set of them). When a program comes across a fork, it can walk down a path based on the condition it is looking for (usually prefixed with the word "if"). Processors can take advantage of this and say, "oh, let me walk down a random path that can lead me to a jackpot so I don't have to come here again." If the processor gambled right, it leads to huge performance boost, because the program can traverse this path without giving a second thought. But what if the chosen path leads to thorn bushes? The process will then say, "oops, I was wrong, so let me retrace my steps and wait for my guest (program) to show me where to go next."
Under some circumstances, when the processor executes the "wrong" path (although it'll think it was right then), the processor will gather the things needed to travel this path. They include next set of instructions and data needed to carry out what it thinks is the "best" match. The thing is, the only thing the processor can understand is 1's and 0's, so it won't care what it will meet down the road, or the fact that it can leak sensitive information (that's the high-level overview of the vulnerability disclosed by Google). In a nutshell, now the processor will be told by its boss (operating system kernel) that it needs to carry some protective gear when traversing conditional paths in order to make sure the processor won't leak sensitive information again (if possible, but it won't work 100 percent of the time), hence the potential performance degradation in some situations.
As for this affecting NVDA's performance, it could, but we won't know the full extent until next week at the earliest once everyone gets the patch.
Cheers,
Joseph

-----Original Message-----
From: nvda@nvda.groups.io [mailto:nvda@nvda.groups.io] On Behalf Of Shaun Everiss
Sent: Wednesday, January 3, 2018 10:44 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?

Well its for all oses, linux macs maybe even amd cpus, its huge.

We say a 30% slowdown but then no one knows what that means.

Suggestions with games on linux and a few other things running in it suggest that we won't see a problem.

Cloud systems will probably slow down a bit on initial patch.

And every manufacturer is different.

Windows will have a patch next week, mac allready has one about to go, everyone will follow suit from phones to whatnot.

And its not all of a sudden, we are hearing it from the manufacturers now because they are almost ready for this thing to go.

Its complex as hell, but basically the kernal chip has a memmory leak that means people could read data.

Memmory leaks will be fixed going forward but there needs to be a patch reguardless even if the firmwares and such on various chips are patched, most will be if you have a modern pc, I don't though and so for most of us it will be the os alone.

As for slowing nvda down, we can run on slow boxes, I just pitty everything from games to jaws, but then no one knows information is being gleened from articles and such, And we can't do much about it, releases will be done next week and they will be on top of it, theuy have to.

As for it taking 10 years, you do wander, but then let me remind you of the issues with routers, attached storage boxes, and ssl.

Maybe someone saw that coming but maybe they didn't.




On 3/01/2018 11:03 p.m., Antony Stone wrote:
A large number of security flaws (in both software and hardware) do
not get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a decade, I'm
not so sure its that much of an issue" is a dangerous assumption, and
not borne out by Microsoft's sudden activity to patch Windows in such
a way that it will have a definite performance penalty. If the
10-year old security flaw were not much of a problem, Microsoft
wouldn't react in a way that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account
via Groups.Io wrote:

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case. I'm not saying this sounds like paranoia, as I do not know
enough to make such a comment, but if the security hole has been in
the chips for a decade, I'm not so sure its that much of an issue. at
any rate I guess all those out there with AMD chips will be chortling loudly.
If its a fault in the chip this seems that it could leave Intel
open to action to give everyone new chips!
Brian






 

Hmmph, maybe they will, maybe they won't.

They will anyway, so far I havn'

't had issues in standard mode though on a couple systems as well as the update, where the machine didn't need a lot of network related services, I have removed a lot of extra packages including the multipoint servers, all server including iis webcore, all vertual machine support and a few other things like message cue.

Not sure if in the long run I will get bitten in the ass for removing it but the system had already got slow, including with office, I have even removed the google drive plugin, for office as excell seemed to be crashing.

Both units are low usage but to be honest, you almost have to take things you had off just to have things work now.

I pitty server users.

I can probably get buy with my usb mainly.

As for new chips, every system from first gen to 8th gen has issues, a lot of the older chips most 4th gen up could get patched certainly 5th-8th will have firmware patches if not allready.

On 4/01/2018 10:32 p.m., Brian's Mail list account via Groups.Io wrote:
That is not what I meant. Its  more likely that when a new chip is out people will use it as a selling point to drive people into buying a new system... again...
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message ----- From: "Alex Kenny" <alexkenny08@gmail.com>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 3:26 PM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


While this has existed for the past decade, it has only been discovered
recently. We have been "living with it", because, as far as we know no one
had discover and exploit it. These patches, while slowing performance
slightly and only in some use cases, will fix the problem.


Also, jumping immediately to some conspiracy by Microsoft doesn't make
sense. First, as Anthony previously stated, Linux is being patched. Mac OS
will also be updated.

Second, this issue affects the kernel, which has nothing to do with whether
traditional Win32 or Windows Store apps are running on top of it.
Deliberately slowing the kernel to induce people to use Windows Store apps
is completely nonsensical.

This patch should only significantly affect use cases doing a lot of work
in kernel space. NVDA will likely not be affected by this.

On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io
<bglists=blueyonder.co.uk@groups.io> wrote:

It just seems odd that we have if you believe all this lived with this for
a
decade and I'd have thought there was now a bit of the shutting the stable
door when all the horses have gone about trying to retro patch it I
understand the biggest losers could be the big server companies who might
be
using Intel chipped hardware, and hence need things sorting out.
 So Maybe I should get my claim for a new processor in tomorrow to be in
time....:-)

Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:23 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
for nvda?


I do not suspect ulterior motives on Microsoft's part in this case,
because
the Linux kernel is being patched in just the same way for just
the > same
reason.


Antony.

On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list
account via
Groups.Io wrote:

Oh really, well, I'm not so sure. It feeds into the mindset that
says
Microsofts eventual aim is to lock people in to getting software
from
its
store and nowhere else in the 'interests' of security, when in
reality
they
want to do what Apple is doing, create an ecosystem and make it
hard >> for
third parties to just allow their software to run on a system.
  Whether you suspect ulterior motives or not is your choice.

Brian

----- Original Message -----
From: "Antony Stone" <antony.stone@nvda.open.source.it>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 10:03 AM
Subject: Re: [nvda] Will this patch to windows for Intel chips
impact
speed
for nvda?

A large number of security flaws (in both software and hardware)
do >> >not
get discovered for a long time after they crept into the design.

I think "if the security hole has been in the chips for a
decade, >> > I'm
not
so sure its that much of an issue" is a dangerous assumption,
and >> > not
borne out by Microsoft's sudden activity to patch Windows in
such a
way
that it will have a definite performance penalty.  If the
10-year >> > old
security flaw were not much of a problem, Microsoft wouldn't
react >> > in
a
way
that unavoidably impacts the majority of Windows users.


Antony.

On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list
account
via

Groups.Io wrote:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

I hope that link works, but its easy to find on that site in any
case.
I'm not saying this sounds like paranoia, as I do not know
enough >> >> to
make
such a comment, but if the security hole has been in the chips
for >> >> a
decade, I'm not so sure its that much of an issue. at any rate I
guess
all those out there with AMD chips will be chortling loudly.

  If its a fault in the chip this seems that it could leave
Intel >> >> open
to
  action to give everyone new chips!

  Brian
--
When you find yourself arguing with an idiot,
you should first of all make sure that the other person isn't
doing the
same
thing.

                                                   Please reply to the
list;
please *don't* > CC
me.







.


 

This bug has been known by those that needed to in december, but my suspician was that they only patched it now because it took them time to think what to do about it.

I mean, you don'

't chuck a problem then say but it may take a few months to figure out things, meanwhile you can't do anything its not really good for business.

On 4/01/2018 10:34 p.m., Brian's Mail list account via Groups.Io wrote:
I did read the article, yes. However just as with that network bug a few months ago, it had been known about in security circles and was only patched when it became common knowledge.
Brian

bglists@blueyonder.co.uk
Sent via blueyonder.
Please address personal email to:-
briang1@blueyonder.co.uk, putting 'Brian Gaff'
in the display name field.
----- Original Message ----- From: "Gene" <gsasner@ripco.com>
To: <nvda@nvda.groups.io>
Sent: Wednesday, January 03, 2018 5:03 PM
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


The conspiracy theory makes me wonder if you, Brian, read the article before you expressed the conspiracy hypothesis.  the facts in the article, such as that this affects Linux as well, and that the problem wasn't discovered by Microsoft and that Linux developers are working on a software fix as well, all demolish the conspiracy idea.

Gene
----- Original Message -----


From: Alex Kenny
Sent: Wednesday, January 03, 2018 9:26 AM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed for nvda?


While this has existed for the past decade, it has only been discovered recently. We have been "living with it", because, as far as we know no one had discover and exploit it. These patches, while slowing performance slightly and only in some use cases, will fix the problem.




Also, jumping immediately to some conspiracy by Microsoft doesn't make sense. First, as Anthony previously stated, Linux is being patched. Mac OS will also be updated.


Second, this issue affects the kernel, which has nothing to do with whether traditional Win32 or Windows Store apps are running on top of it. Deliberately slowing the kernel to induce people to use Windows Store apps is completely nonsensical.


This patch should only significantly affect use cases doing a lot of work in kernel space. NVDA will likely not be affected by this.


On Wed, Jan 3, 2018 at 8:00 AM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:

 It just seems odd that we have if you believe all this lived with this for a
 decade and I'd have thought there was now a bit of the shutting the stable
 door when all the horses have gone about trying to retro patch it I
 understand the biggest losers could be the big server companies who might be
 using Intel chipped hardware, and hence need things sorting out.
  So Maybe I should get my claim for a new processor in tomorrow to be in
 time....:-)

 Brian

 bglists@blueyonder.co.uk
 Sent via blueyonder.
 Please address personal email to:-
 briang1@blueyonder.co.uk, putting 'Brian Gaff'
 in the display name field.
 ----- Original Message -----
 From: "Antony Stone" <antony.stone@nvda.open.source.it>
 To: <nvda@nvda.groups.io>
 Sent: Wednesday, January 03, 2018 10:23 AM
 Subject: Re: [nvda] Will this patch to windows for Intel chips impact speed
 for nvda?


 >I do not suspect ulterior motives on Microsoft's part in this case, because
 > the Linux kernel is being patched in just the same way for just the same
 > reason.
 >
 >
 > Antony.
 >
 > On Wednesday 03 January 2018 at 11:17:37, Brian's Mail list account via
 > Groups.Io wrote:
 >
 >> Oh really, well, I'm not so sure. It feeds into the mindset that says
 >> Microsofts eventual aim is to lock people in to getting software from its
 >> store and nowhere else in the 'interests' of security, when in reality
 >> they
 >> want to do what Apple is doing, create an ecosystem and make it hard for
 >> third parties to just allow their software to run on a system.
 >>  Whether you suspect ulterior motives or not is your choice.
 >>
 >> Brian
 >>
 >> ----- Original Message -----
 >> From: "Antony Stone" <antony.stone@nvda.open.source.it>
 >> To: <nvda@nvda.groups.io>
 >> Sent: Wednesday, January 03, 2018 10:03 AM
 >> Subject: Re: [nvda] Will this patch to windows for Intel chips impact
 >> speed
 >> for nvda?
 >>
 >> >A large number of security flaws (in both software and hardware) do not
 >> >get discovered for a long time after they crept into the design.
 >> >
 >> > I think "if the security hole has been in the chips for a decade, I'm
 >> > not
 >> > so sure its that much of an issue" is a dangerous assumption, and not
 >> > borne out by Microsoft's sudden activity to patch Windows in such a way
 >> > that it will have a definite performance penalty. If the 10-year old
 >> > security flaw were not much of a problem, Microsoft wouldn't react in a
 >> > way
 >> > that unavoidably impacts the majority of Windows users.
 >> >
 >> >
 >> > Antony.
 >> >
 >> > On Wednesday 03 January 2018 at 10:52:15, Brian's Mail list account via
 >> >
 >> > Groups.Io wrote:
 >> >> https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
 >> >>
 >> >> I hope that link works, but its easy to find on that site in any case.
 >> >> I'm not saying this sounds like paranoia, as I do not know enough to
 >> >> make
 >> >> such a comment, but if the security hole has been in the chips for a
 >> >> decade, I'm not so sure its that much of an issue. at any rate I guess
 >> >> all those out there with AMD chips will be chortling loudly.
 >> >>
 >> >>  If its a fault in the chip this seems that it could leave Intel open
 >> >> to
 >> >>  action to give everyone new chips!
 >> >>
 >> >>  Brian
 >
 > --
 > When you find yourself arguing with an idiot,
 > you should first of all make sure that the other person isn't doing the
 > same
 > thing.
 >
 >                                                   Please reply to the
 > list;
 > please *don't* CC
 > me.
 >
 >
 >








.