Re: Safely reading Intel MSRs?
Re: Safely reading Intel MSRs?
- Subject: Re: Safely reading Intel MSRs?
- From: "mm w" <email@hidden>
- Date: Fri, 21 Nov 2008 14:12:36 -0800
Hi 2 , I forgot, if you could choose a proper design and not to read
these values from your kext
it should be great for the safety of your customers
Cheers!
On Fri, Nov 21, 2008 at 2:03 PM, mm w <email@hidden> wrote:
> Hi nop (and yes include what you need),
>
> osfmk/i386/proc_reg.h
> so can use the defined rdmsr, I guess , instead of doing it by hand,
> if I did "well" understand,
>
> unsigned hi, lo, msr;
>
> __asm(
> movl %íx, %
> ...
> );
>
> Cheers!
>
> On Fri, Nov 21, 2008 at 8:58 AM, Brent Burton <email@hidden> wrote:
>> Hi all,
>>
>> I am writing some internal system software including a kernel
>> extension that has the ability to read MSR values.
>> While I have a runtime check for processor family and model, the
>> mapping to valid MSR registers isn't finished (and of course, may have
>> errors when it is finished). If an invalid MSR is read or written, it
>> causes a kernel panic (via GPF), and I'd like to protect against this
>> for all cases.
>>
>> Are there kernel functions, accessible from a kext, that provide safe
>> reading of Intel MSRs?
>>
>> If not, one alternative is to (temporarily) redirect GPF handling. I
>> don't particularly like this path, but I want to get more information
>> on it. Where is documentation or headers that cover this, if any?
>>
>> thanks,
>> -Brent
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> Darwin-kernel mailing list (email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>>
>> This email sent to email@hidden
>>
>
>
>
> --
> -mmw
>
--
-mmw
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden