[U-Boot-Users] do_bootm() and flush_cache()

[U-Boot-Users] do_bootm() and flush_cache()

[Scott McNutt]

I'm looking for some guidance:

On some processors (e.g. Nios-II), the instruction and
data caches can't be disabled. The code for loadb, loads,
etc. call flush_cache() ... so life is good.

But after reviewing do_bootm(), I don't see where the
caches are flushed. A call to flush_cache() after
moving/decompressing the image seems appropriate, but
I'm not sure if this would break other architectures ...

... do I need to wrap an ifdef around a call to
flush_cache() ?

Regards,
--Scott

[U-Boot-Users] do_bootm() and flush_cache()

[Wolfgang Denk]

In message <4240C916.30901@psyent.com> you wrote:
> 
> But after reviewing do_bootm(), I don't see where the
> caches are flushed. A call to flush_cache() after
> moving/decompressing the image seems appropriate, but

Why?


Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"Open the pod bay doors, HAL."                    - Dave Bowman, 2001

[U-Boot-Users] do_bootm() and flush_cache()

[Scott McNutt]

>>But after reviewing do_bootm(), I don't see where the
>>caches are flushed. A call to flush_cache() after
>>moving/decompressing the image seems appropriate, but
> 
> 
> Why?

After the move/decompression, the data cache must be
flushed to main memory ... otherwise an instruction
cache miss (after jumping into the kernel) may read
in a (potentially) stale line. (Neither cache can be
disabled.)

Regards,
--Scott