I was tring to get a bootm command to work and had a problem until I called the fluch_cache routine to force the results of the data move into ram. This way the execution of the image used the code I had moved into ram rather then whatever was sitting there.
 
 
I see a reference to another user who had this problem but didn't see any resolution.. It seems like a rather serious thing and should be fixed.
 
 

Email Archive: u-boot-users (read-only)

Re: do_bootm() and flush_cache()
From: Scott McNutt <smcnutt@ps...> - 2005-03-23 18:56 
 >>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