Cannot enable the caches on MPC8260A

cancel
Showing results for 
Search instead for 
Did you mean: 

Cannot enable the caches on MPC8260A

Jump to solution
199 Views
Contributor III

Hello everybody,

I work with a specific board at the following frequencies:

MPC8260 Clock Configuration

- Bus-to-Core Mult 3x, VCO Div 2, 60x Bus Freq  33-100, Core Freq 100-300

- dfbrg 0, corecnf 0x08, busdf 3, cpmdf 1, plldf 0, pllmf 1, pcidf 3

- vco_out  133332000, scc_clk   33333000, brg_clk   33333000

- cpu_clk   99999000, cpm_clk   66666000, bus_clk   33333000

CPU:   MPC8260 (HiP4 Rev 14, Mask B.1 4K25A) at 99.999 MHz

Quote from the MPC8260AEC:

"The user should choose the input clock frequency and the multiplication factors such that the frequency of the CPU is equal to or greater than 150 MHz and the CPM ranges between 66–233 MHz."

Is it possible that my CPU's low frequency has somehow caused problems with caches?

Thanks

Tags (3)
1 Solution
21 Views
NXP TechSupport
NXP TechSupport

Setting core PLL frequency below specified causes excessive PLL output jitter.

This jitter may be tolerated by core on internal core operations, but affects core to bus operations, because core and bus clock sources are different.

Core to bus operations depend on cache state - when cache is enabled, cache controller uses burst transactions for cache to bus operatons, so bus

throughput increases, and the probability of bus error due to big frequecy jitter increases also

View solution in original post

3 Replies
22 Views
NXP TechSupport
NXP TechSupport

Setting core PLL frequency below specified causes excessive PLL output jitter.

This jitter may be tolerated by core on internal core operations, but affects core to bus operations, because core and bus clock sources are different.

Core to bus operations depend on cache state - when cache is enabled, cache controller uses burst transactions for cache to bus operatons, so bus

throughput increases, and the probability of bus error due to big frequecy jitter increases also

View solution in original post

21 Views
Contributor III

Thank you very much!

Can you explain the following?

U-boot is run from flash at 0xfff00100, then moved to the end of the SDRAM (0x0 - 0x04000000). Without the cache everything works. If I turn on the instruction cache u-boot executes exactly one instruction at the address of SDRAM, and then flies to 0xfff00000 and falls. Why is this happening?

0 Kudos
21 Views
NXP TechSupport
NXP TechSupport

There are several possible reasons for incorrect behavior after enabling instruction/data cache:

1. MMU error or disabled MMU. There are several memory-mapped device registers in IMMR space, this space should be configured as cache-inhibited. This can be done only in MMU settings. If you enable cache without MMU enabled, all memory space, including internal memory space, will become cacheable and this will result to erratic behavior.

If you need an example for MMU initialization, please look CodeWarrior installation folder, \Templates\PowerPC_EABI\Sources\G2_core\

2. SDRAM configuration error. Cache controller, when enabled, starts using burst transactions to access external SDRAM memory. If your SDRAM memory is not properly configured and can support only single transactions, than there will be an instruction/data corruption when fetching instructions from this SDRAM with instruction cache enabled.

To check this, please test your SDRAM memory with data cache enaled.

This community was created to support customers worldwide, to let customers to communicate together and to re-use information, previously posted by/for another customers.

Because of above reasons, this community language is English-only. Therefore, I would ask you to use English for your further communications, just for information re-use reasons.

Thank you in advance.