All times are UTC-06:00




Post new topic  Reply to topic  [ 7 posts ] 
Author Message
 Post subject: Efika's IPB clock
PostPosted: Sun Dec 17, 2006 4:14 pm 
Offline

Joined: Mon Jan 30, 2006 1:30 am
Posts: 43
Hello,
from the SFW bootup output, it seems core clock is 396 and at least XPB clock seems to be 132 MHz (based on MPC5200B user manual) The question is: is IPB clock 132 or 66 MHz?
Thanks,
Karel


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Mon Dec 18, 2006 4:38 am 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
Hello,
from the SFW bootup output, it seems core clock is 396 and at least XPB clock seems to be 132 MHz (based on MPC5200B user manual) The question is: is IPB clock 132 or 66 MHz?
Thanks,
Karel
Why do you need to know that? :)

_________________
Matt Sealey


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Mon Dec 18, 2006 4:45 am 
Offline

Joined: Mon Jan 30, 2006 1:30 am
Posts: 43
Quote:
Quote:
Hello,
from the SFW bootup output, it seems core clock is 396 and at least XPB clock seems to be 132 MHz (based on MPC5200B user manual) The question is: is IPB clock 132 or 66 MHz?
Thanks,
Karel
Why do you need to know that? :)
Because in current generic 5200 BSP, it's needed to configure those three clocks. For example following comments in code, IPB clock is the same as in Bestcomm and Bestcomm clock is what you need to know to configure serial port bauds correctly.

Thanks,
Karel


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Tue Dec 19, 2006 10:05 am 
Offline

Joined: Mon Jan 30, 2006 1:30 am
Posts: 43
Quote:
Quote:
Quote:
Hello,
from the SFW bootup output, it seems core clock is 396 and at least XPB clock seems to be 132 MHz (based on MPC5200B user manual) The question is: is IPB clock 132 or 66 MHz?
Thanks,
Karel
Why do you need to know that? :)
Because in current generic 5200 BSP, it's needed to configure those three clocks. For example following comments in code, IPB clock is the same as in Bestcomm and Bestcomm clock is what you need to know to configure serial port bauds correctly.

Thanks,
Karel
So it seems it's 66 MHz if anyone also needs this.
Karel


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Fri May 18, 2007 6:22 am 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
So it seems it's 66 MHz if anyone also needs this.
If you're coding in Linux using the platform support provided by 2.6.20 onwards, you can use the function;
Code:
if ((ipb_freq = mpc52xx_find_ipb_freq(np)) == 0) {
pr_debug("Could not find IPB bus frequency!\n");
return -EINVAL;
}
I got that from the serial port driver. You should pass in the node of the device you're using, which should probably be a child node of /builtin for it to work.

Essentially, the value is the 'bus-frequency' property of the /builtin node.

_________________
Matt Sealey


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Thu May 24, 2007 12:44 am 
Offline

Joined: Mon Jan 30, 2006 1:30 am
Posts: 43
Quote:
Quote:
So it seems it's 66 MHz if anyone also needs this.
If you're coding in Linux using the platform support provided by 2.6.20 onwards, you can use the function;
Code:
if ((ipb_freq = mpc52xx_find_ipb_freq(np)) == 0) {
pr_debug("Could not find IPB bus frequency!\n");
return -EINVAL;
}
I got that from the serial port driver. You should pass in the node of the device you're using, which should probably be a child node of /builtin for it to work.

Essentially, the value is the 'bus-frequency' property of the /builtin node.
Hello,
that's interesting observation. Anyway, I'm not sure if it's correct, since returned 'bus-frequency' is 132MHz. The problem is that the driver author a few lines below divide it twice and use it as an 'uart' frequency. See:
Code:
port->uartclk = ipb_freq / 2;
It's interesting since if I understand mpc5200b user's manual well, then uart clock is exactly the same like bestcomm clock which is again the same like IPB clock. (Paragraph 5.3.5 IPB Clock Domain). So If I'm right Linux folks just get XLB clock frequency and assume that IPB = XLB /2, which don't need to be true (Tables 5-4 and 5-5).

Please correct me if I'm wrong.

Thanks!
Karel


Top
   
 Post subject: Re: Efika's IPB clock
PostPosted: Thu May 24, 2007 2:56 am 
Offline
Site Admin

Joined: Fri Sep 24, 2004 1:39 am
Posts: 1589
Location: Austin, TX
Quote:
that's interesting observation. Anyway, I'm not sure if it's correct, since returned 'bus-frequency' is 132MHz. The problem is that the driver author a few lines below divide it twice and use it as an 'uart' frequency. See:
Code:
port->uartclk = ipb_freq / 2;
It's interesting since if I understand mpc5200b user's manual well, then uart clock is exactly the same like bestcomm clock which is again the same like IPB clock. (Paragraph 5.3.5 IPB Clock Domain). So If I'm right Linux folks just get XLB clock frequency and assume that IPB = XLB /2, which don't need to be true (Tables 5-4 and 5-5).

Please correct me if I'm wrong.
The way I understand it, the XLB runs at a certain clock, and the IPB runs at a derivative and divided clock (same or half) of the XLB, and the PCI bus runs at yet another derivative clock of the IPB (same or 2 or 4).

If the XLB runs at 132MHz, the IPB can run at 132MHz, 66MHz or 33MHz. The PCI bus runs at the same or half or a quarter of the IPB but has to be 66MHz or 33MHz (to fit spec) and no more than the IPB.

As far as I understand, Efika's XLB is 132MHz, the IPB is 132MHz (no divider), and the PCI bus is 66MHz (2x divider).

I'm not sure why the UART driver divides the IPB by 2 - but it is storing 'uartclk' and not 'ipbclock' in that structure. I have to read a chapter of the manual and check the code to work it out (15.3.1.2) to see what prescalers and dividers it's using in the clock generation.

If the division by 2 was merely to get the correct IPB frequency from a misreporting device tree, then the UART driver would be broken on Lite5200, so I don't think you're correct :)

It may be just simpler to deal with the value by dividing it there. If you are really concerned of the real values, read the CDM configuration register (mbar + 0x20c) and check your values.

_________________
Matt Sealey


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 7 posts ] 

All times are UTC-06:00


Who is online

Users browsing this forum: No registered users and 2 guests


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

Search for:
Jump to:  
PowerDeveloper.org: Copyright © 2004-2012, Genesi USA, Inc. The Power Architecture and Power.org wordmarks and the Power and Power.org logos and related marks are trademarks and service marks licensed by Power.org.
All other names and trademarks used are property of their respective owners. Privacy Policy
Powered by phpBB® Forum Software © phpBB Group