ping 127.0.0.1 takes too much

ping 127.0.0.1 takes too much

[DI BACCO ANTONIO - technolabs]

[-- Attachment #1: Type: text/plain, Size: 717 bytes --]

I have two different boards: one with MPC875 running at 66 MHz cpu/66 MHz bus, one with MPC880 running at 132 MHz cpu /66MHz bus, I load the same kernel and root file system on both the boards that have only different u-boot(s).

If I do a ping 127.0.0.1 on the MPC875 I get a round trip of 0.300 ms, while on the other one I get 0.900 ms, I expected the opposite.

Could be an issue with memory configuration? In the formula to calculate the PTx on the MPC885RM manual I used 66 MHz as "system clock" for the first board and 132 MHz for the second board. Probably this is a mistake because "system clock" should be the bus clock but how this can so adversely affect the ping round trip?

Thank you,
Antonio.

[-- Attachment #2: Type: text/html, Size: 1134 bytes --]

Re: ping 127.0.0.1 takes too much

[Scott Wood]

DI BACCO ANTONIO - technolabs wrote:
> I have two different boards: one with MPC875 running at 66 MHz cpu/66 
> MHz bus, one with MPC880 running at 132 MHz cpu /66MHz bus, I load the 
> same kernel and root file system on both the boards that have only 
> different u-boot(s).
> 
> If I do a ping 127.0.0.1 on the MPC875 I get a round trip of 0.300 ms, 
> while on the other one I get 0.900 ms, I expected the opposite.
> 
> Could be an issue with memory configuration? In the formula to calculate 
> the PTx on the MPC885RM manual I used 66 MHz as "system clock" for the 
> first board and 132 MHz for the second board. Probably this is a mistake 
> because "system clock" should be the bus clock but how this can so 
> adversely affect the ping round trip?

Are you sure the timebase clock is what the kernel thinks it is?  Does 
the time-of-day clock progress at the correct rate?

-Scott

RE: ping 127.0.0.1 takes too much

[DI BACCO ANTONIO - technolabs]

[-- Attachment #1: Type: text/plain, Size: 267 bytes --]


>Are you sure the timebase clock is what the kernel thinks it is?  Does 
>the time-of-day clock progress at the correct rate?

The kernel receives the CPU clock from the bdinfo structure and thus I think the kernel knows how fast the CPU is. 

Bye,
Antonio.

[-- Attachment #2: Type: text/html, Size: 683 bytes --]

Re: ping 127.0.0.1 takes too much

[Scott Wood]

DI BACCO ANTONIO - technolabs wrote:
> 
>  >Are you sure the timebase clock is what the kernel thinks it is?  Does
>  >the time-of-day clock progress at the correct rate?
> 
> The kernel receives the CPU clock from the bdinfo structure and thus I 
> think the kernel knows how fast the CPU is.

That wasn't what I asked.  I asked whether you checked whether it was 
actually correct or not.  I've seen it be wrong quite often.

-Scott