Real time clock on ia64 platforms

Real time clock on ia64 platforms

[Christian Hildner]

Hi,

someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
+ i2000) use. I want to access it directly, without using EFI calls.

Christian

Re: Real time clock on ia64 platforms

[Matthew Wilcox]

On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
> someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
> + i2000) use. I want to access it directly, without using EFI calls.

It's probably different on each of them.  Why do you want to access it
without going through firmware?

-- 
"Next the statesmen will invent cheap lies, putting the blame upon 
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince 
himself that the war is just, and will thank God for the better sleep 
he enjoys after this process of grotesque self-deception." -- Mark Twain

Re: Real time clock on ia64 platforms

[Alex Williamson]

On Tue, 2004-06-01 at 01:50, Christian Hildner wrote:
> Hi,
> 
> someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
> + i2000) use. I want to access it directly, without using EFI calls.

   This is strongly not recommended.  On HP chipset boxes, the devices
are only designed to be accessed via the EFI calls (and yes, EFI
firmware can and does work around bugs in the RTC).  There's also no
architected way to find the real RTC in the address map.  Why do you
need to poke it directly?

	Alex

Re: Real time clock on ia64 platforms

[Christoph Lameter]

The portable way to access the RTC is either gettimeofday()
or clock_gettime(CLOCK_REALTIME). The gettimeofday caall is a fast system
call on many ia64 platforms and does not use EFI calls. EFI calls are only
used on bootup to determine the current RTC clock and the clock characteristics.

On Tue, 1 Jun 2004, Alex Williamson wrote:

> On Tue, 2004-06-01 at 01:50, Christian Hildner wrote:
> > Hi,
> >
> > someone may know what kind of RTC device the ia64 platforms (HP ZX 6000
> > + i2000) use. I want to access it directly, without using EFI calls.
>
>    This is strongly not recommended.  On HP chipset boxes, the devices
> are only designed to be accessed via the EFI calls (and yes, EFI
> firmware can and does work around bugs in the RTC).  There's also no
> architected way to find the real RTC in the address map.  Why do you
> need to poke it directly?
>
> 	Alex
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: Real time clock on ia64 platforms

[Christian Hildner]

Matthew Wilcox schrieb:

>On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
>  
>
>>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
>>+ i2000) use. I want to access it directly, without using EFI calls.
>>    
>>
>
>It's probably different on each of them.  Why do you want to access it
>without going through firmware?
>
It's just because I need rtc information in an environment where no EFI 
calls are possible. I am aware of the non-portability of doing so. But 
there is no other way.

Christian

Re: Real time clock on ia64 platforms

[Stephane Eranian]

Christian,

On Wed, Jun 02, 2004 at 07:56:49AM +0200, Christian Hildner wrote:
> Matthew Wilcox schrieb:
> 
> >On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
> > 
> >
> >>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
> >>+ i2000) use. I want to access it directly, without using EFI calls.
> >>   
> >>
> >
> >It's probably different on each of them.  Why do you want to access it
> >without going through firmware?
> >
> It's just because I need rtc information in an environment where no EFI 
> calls are possible. I am aware of the non-portability of doing so. But 
> there is no other way.
> 
What do you mean by "rtc information"? Are you after a very fine granularity
timer?

-- 
-Stephane

Re: Real time clock on ia64 platforms

[Christian Hildner]

Alex Williamson schrieb:

>On Tue, 2004-06-01 at 01:50, Christian Hildner wrote:
>  
>
>>Hi,
>>
>>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
>>+ i2000) use. I want to access it directly, without using EFI calls.
>>    
>>
>
>   This is strongly not recommended.  On HP chipset boxes, the devices
>are only designed to be accessed via the EFI calls (and yes, EFI
>firmware can and does work around bugs in the RTC).  There's also no
>architected way to find the real RTC in the address map.
>
So the only chance to access the rtc directly is to do it the same way 
the firmware does. Since the firmware is that hardware specific it does 
not need to "find" it because the firmware already knows where the rtc 
is and how to access it. If the rtc is not mapped in the conventional 
port address map, how does the firmware access it? Is there a 
firmware-only accessible io memory map or is it done with machine 
specific registers (msr)?

>  Why do you
>need to poke it directly?
>
Because I need to call it from an environment, where no EFI calls are 
possible. I know that the solution is absolutely not portable in any way 
and in common it is bad doing that way but in that special case there is 
no other way to go. And  the portability is given at another layer on 
top of it.

Christian

>	Alex
>

Re: Real time clock on ia64 platforms

[Christian Hildner]

Stephane Eranian schrieb:

>Christian,
>
>On Wed, Jun 02, 2004 at 07:56:49AM +0200, Christian Hildner wrote:
>  
>
>>Matthew Wilcox schrieb:
>>
>>    
>>
>>>On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
>>>
>>>
>>>      
>>>
>>>>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
>>>>+ i2000) use. I want to access it directly, without using EFI calls.
>>>>  
>>>>
>>>>        
>>>>
>>>It's probably different on each of them.  Why do you want to access it
>>>without going through firmware?
>>>
>>>      
>>>
>>It's just because I need rtc information in an environment where no EFI 
>>calls are possible. I am aware of the non-portability of doing so. But 
>>there is no other way.
>>
>>    
>>
>What do you mean by "rtc information"? Are you after a very fine granularity
>timer?
>
I just need access to a rtc device on the HP ZX 6000 comparable to that 
Dallas devices on PC boards.

Christian

Re: Real time clock on ia64 platforms

[Stephane Eranian]

Christian,

On Wed, Jun 02, 2004 at 08:41:53AM +0200, Christian Hildner wrote:
> Stephane Eranian schrieb:
> 
> >Christian,
> >
> >On Wed, Jun 02, 2004 at 07:56:49AM +0200, Christian Hildner wrote:
> > 
> >
> >>Matthew Wilcox schrieb:
> >>
> >>   
> >>
> >>>On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
> >>>
> >>>
> >>>     
> >>>
> >>>>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
> >>>>+ i2000) use. I want to access it directly, without using EFI calls.
> >>>> 
> >>>>
> >>>>       
> >>>>
> >>>It's probably different on each of them.  Why do you want to access it
> >>>without going through firmware?
> >>>
> >>>     
> >>>
> >>It's just because I need rtc information in an environment where no EFI 
> >>calls are possible. I am aware of the non-portability of doing so. But 
> >>there is no other way.
> >>
> >>   
> >>
> >What do you mean by "rtc information"? Are you after a very fine 
> >granularity
> >timer?
> >
> I just need access to a rtc device on the HP ZX 6000 comparable to that 
> Dallas devices on PC boards.
> 
I am sorry but I don't know that device. What functionality of that device
is of interest to you?

-- 
-Stephane

Re: Real time clock on ia64 platforms

[Christian Hildner]

Stephane Eranian schrieb:

>Christian,
>
>On Wed, Jun 02, 2004 at 08:41:53AM +0200, Christian Hildner wrote:
>  
>
>>Stephane Eranian schrieb:
>>
>>    
>>
>>>Christian,
>>>
>>>On Wed, Jun 02, 2004 at 07:56:49AM +0200, Christian Hildner wrote:
>>>
>>>
>>>      
>>>
>>>>Matthew Wilcox schrieb:
>>>>
>>>>  
>>>>
>>>>        
>>>>
>>>>>On Tue, Jun 01, 2004 at 09:50:48AM +0200, Christian Hildner wrote:
>>>>>
>>>>>
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>>>>someone may know what kind of RTC device the ia64 platforms (HP ZX 6000 
>>>>>>+ i2000) use. I want to access it directly, without using EFI calls.
>>>>>>
>>>>>>
>>>>>>      
>>>>>>
>>>>>>            
>>>>>>
>>>>>It's probably different on each of them.  Why do you want to access it
>>>>>without going through firmware?
>>>>>
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>>It's just because I need rtc information in an environment where no EFI 
>>>>calls are possible. I am aware of the non-portability of doing so. But 
>>>>there is no other way.
>>>>
>>>>  
>>>>
>>>>        
>>>>
>>>What do you mean by "rtc information"? Are you after a very fine 
>>>granularity
>>>timer?
>>>
>>>      
>>>
>>I just need access to a rtc device on the HP ZX 6000 comparable to that 
>>Dallas devices on PC boards.
>>
>>    
>>
>I am sorry but I don't know that device. What functionality of that device
>is of interest to you?
>
It is just getting an absolute time without any calls to firmware or linux.

Christian

Re: Real time clock on ia64 platforms

[Stephane Eranian]

Christian,

> >>Dallas devices on PC boards.
> >>
> >>   
> >>
> >I am sorry but I don't know that device. What functionality of that device
> >is of interest to you?
> >
> It is just getting an absolute time without any calls to firmware or linux.
> 
at least on UP, can't you use a combination of gettimeofday() once and then use
ar.itc (equivalent of TSC) to make the appropriate adjustments. Also note
that on 2.6 with the right version of glibc, gettimeofday() is extremely fast.

-- 
-Stephane

Re: Real time clock on ia64 platforms

[Robin Holt]

On Wed, Jun 02, 2004 at 08:26:53AM +0200, Christian Hildner wrote:
> Alex Williamson schrieb:
> 
> >On Tue, 2004-06-01 at 01:50, Christian Hildner wrote:
> > Why do you
> >need to poke it directly?
> >
> Because I need to call it from an environment, where no EFI calls are 
> possible. I know that the solution is absolutely not portable in any way 
> and in common it is bad doing that way but in that special case there is 
> no other way to go. And  the portability is given at another layer on 
> top of it.

What are you doing with it?  Can you just use ITC in the realtime sensitive
portion of your code and then convert ITC ticks back into clock time as
needed?

Robin Holt