Turn off NMI ???

Index > OS Construction > Turn off NMI ???

Goto page 1, 2 Next

Author

Thread

milind

Joined: 15 Feb 2004
Posts: 33

milind 04 Mar 2004, 12:51

I found this code from one of the posts, it disables the NMI. Would this really work??

in AL,070h ;
or AL,080h ; Turn NMI off
out 070h,AL ;

04 Mar 2004, 12:51

Ralph

Joined: 04 Oct 2003
Posts: 86

Ralph 04 Mar 2004, 17:37

Yes.

04 Mar 2004, 17:37

Adam Kachwalla

Joined: 01 Apr 2006
Posts: 150

Adam Kachwalla 26 May 2007, 11:40

Hang on... Isn't NMI Non-Maskable Interrupt? I think that is a CPU bug, which has most likely been fixed.

26 May 2007, 11:40

Japheth

Joined: 26 Oct 2004
Posts: 151

Japheth 26 May 2007, 17:18

The code above should work on most machines, but it is not 100% safe. A write to port 70h should be followed by an access to port 71h, for example:

Code:

pushf
cli
mov al, 0Fh+80h
out 70h, al
in al,71h
popf

26 May 2007, 17:18

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4623
Location: Argentina

LocoDelAssembly 26 May 2007, 18:10

NMI means that CLI instruction doesn't prevent the processor to handle it. However, if some other chip prevents the processor's NMI pin to be asserted then you have the NMI interrupt masked Very Happy

26 May 2007, 18:10

Japheth

Joined: 26 Oct 2004
Posts: 151

Japheth 26 May 2007, 19:10

> NMI means that CLI instruction doesn't prevent the processor to handle it.

That was absolutely new to me. Sad

(the "CLI" was to mask IRQs, not NMIs, in case the RTC timer /alarm has been enabled.)

26 May 2007, 19:10

mikegonta

Joined: 20 Nov 2005
Posts: 99

mikegonta 31 May 2007, 20:37

[ Post removed by author. ]

Last edited by mikegonta on 28 Jan 2009, 09:15; edited 2 times in total

31 May 2007, 20:37

Japheth

Joined: 26 Oct 2004
Posts: 151

Japheth 03 Jun 2007, 12:17

> Keep in mind that this caveat was designed for the original chip and like
> many other "urban" legends is not applicable to a modern PC.

This might be true, but the OP did not indicate that the code must run on a new PC only.

> I ran the following code in Protected Mode from a boot sector on a Pentium
> for 5 minutes and it had no noticeable effect on the time.

Hm, what did you expect from this code? It loops. What effect should have been seen?

03 Jun 2007, 12:17

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4623
Location: Argentina

LocoDelAssembly 24 Jan 2009, 17:51

Sorry for bumping this thread. The code shown on this thread is the correct to disable NMI even on modern computers or not?

I would like to write my switch-to-protected-mode code, and Intel says NMIs must be disabled during the transition.

24 Jan 2009, 17:51

DJ Mauretto

Joined: 14 Mar 2007
Posts: 464
Location: Rome,Italy

DJ Mauretto 24 Jan 2009, 18:35

Unfortunately this is Chipset specific, on Intel Chipset
you cannot read anymore Port 70h , but you can set bit 7 and
Disable All NMI sources, that is the Port 70h is Write Only, unless
you enable Alt Access Mode or use alias Port 74h ( Intel chipset specific)
anyway that bit is still valid for write operation and disable NMI Wink

Code:

; Disable NMI

mov   al,80h
out   70h,al

; Enable NMI

xor   al,al
out  70h,al

_________________
Nil Volentibus Arduum Razz

24 Jan 2009, 18:35

mikegonta

Joined: 20 Nov 2005
Posts: 99

mikegonta 24 Jan 2009, 18:57

[ Post removed by author. ]

Last edited by mikegonta on 28 Jan 2009, 09:15; edited 2 times in total

24 Jan 2009, 18:57

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20700
Location: In your JS exploiting you and your system

revolution 24 Jan 2009, 19:04

mikegonta wrote:

The chances of an NMI triggered event are slim and by definition they are non recoverable. So the simple thing is to ignore it.

Haven't you ever heard of Murphy's Law?

24 Jan 2009, 19:04

mikegonta

Joined: 20 Nov 2005
Posts: 99

mikegonta 24 Jan 2009, 19:13

[ Post removed by author. ]

Last edited by mikegonta on 28 Jan 2009, 09:15; edited 1 time in total

24 Jan 2009, 19:13

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20700
Location: In your JS exploiting you and your system

revolution 24 Jan 2009, 19:21

mikegonta wrote:

revolution wrote:
Haven't you ever heard of Murphy's Law?
Only happens when you are Out to lunch!

But I am always Out to Lunch, so does that mean it always happens. Sounds quite bad in that case.

BTW, I am using the second to last meaning of Out to lunch in my location.

24 Jan 2009, 19:21

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4623
Location: Argentina

LocoDelAssembly 24 Jan 2009, 19:24

In general, operating systems (both, hobby and professional) do care about NMI while switching?

Thanks for the answers!

24 Jan 2009, 19:24

mikegonta

Joined: 20 Nov 2005
Posts: 99

mikegonta 24 Jan 2009, 21:48

[ Post removed by author. ]

Last edited by mikegonta on 28 Jan 2009, 09:16; edited 2 times in total

24 Jan 2009, 21:48

Coddy41

Joined: 18 Jan 2009
Posts: 384
Location: Ohio, USA

Coddy41 24 Jan 2009, 22:01

So jmp $ wont pause the system on my new computer?

24 Jan 2009, 22:01

mikegonta

Joined: 20 Nov 2005
Posts: 99

mikegonta 24 Jan 2009, 22:06

[ Post removed by author. ]

Last edited by mikegonta on 28 Jan 2009, 09:16; edited 3 times in total

24 Jan 2009, 22:06

baldr

Joined: 19 Mar 2008
Posts: 1651

baldr 24 Jan 2009, 22:13

Coddy41,

jmp $ makes CPU busy, not uninterruptable.

24 Jan 2009, 22:13

Coddy41

Joined: 18 Jan 2009
Posts: 384
Location: Ohio, USA

Coddy41 24 Jan 2009, 22:24

Busy, hmm... --takes notes-- busy doing what?
Mike "hlt" stands for stall in
+==========+
|@@:
| sti
| hlt
| jmp @B
+==========+
And then it jumps back to @, Right????

24 Jan 2009, 22:24

Goto page 1, 2 Next

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum