[ECOS] problem enabling the caches on powerpc

Gary Thomas gary@mlbassoc.com
Thu Jun 29 11:46:00 GMT 2006

Please keep your replies on the mailing list so that all may benefit.
Private support is available only with a support contract.

saurabh prakash wrote:
> How do i test that burst acces to dram is working properly?
> I have enabled the burst for the device by OR/BR registers.

You'll need to check your settings carefully and may need to enlist the
aid of a hardware expert with appropriate test equipment.

Verify what you have vs other known working designs - there are many
in the eCos source tree.

> The board is working with MMU disabled.

There can be no cache/bursting with the MMU disabled, so this is expected.

> thnx in advance,
> saurabh
> On 6/29/06, Gary Thomas <gary@mlbassoc.com> wrote:
>> saurabh prakash wrote:
>> > hi,
>> > i am working on  mbx type board, with mpc860P, i had successfully 
>> worked on
>> > similar board  in the past. but i am having problems enabling the
>> > caches on the current
>> > board. As soon as the macro HAL_ICACHE_ENABLE() executes, the cpu
>> > gives an exception(SEI). To be precise the second "isync" gives an
>> > exception. I have modified the
>> according to
>> > the variant i am using but without any results. Can anybody suggest
>> > any solution.
>> >
>> >
>> > #define HAL_ICACHE_ENABLE()
>> >    asm volatile ("isync;"
>> >                  "mtspr %0, %1;"
>> >                  "isync"   <----- Executing this gives an exception
>> >                  : : "I" (CYGARC_REG_IC_CST), "r" 
>> The most likely cause is your hardware timings, in particular DRAM
>> setups.  When you enable the cache(s), the CPU will start using burst
>> access to memory.  If you've not set this up properly, exceptions will
>> occur.
>> Check your UPM tables.

Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world

Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

More information about the Ecos-discuss mailing list