Hmm,
So i looked into the code and it would appear to be a bug.
ocd_mem2array is implemented via jim_mem2array which eventually calls
target_mem2array. When this happens, I am pretty confident if
ocd_mem2array works, that argc will be 5.
BUT
if I use omap3.cpu mem2array, this is implemented through a call to
tcl_target_func, which ends up calling target_mem2array as well. But in
this case argc will be 4. So there is a mismatch between what
jim_mem2array and tcl_target_func expect with regard to target_mem2array
I just proved this by changing my call to:
omap3.cpu mem2array junk dataval 32 [expr "0x54011000 + $reg_num * 4"] 1
"junk" is just crap to make the argc=5 check work, and to align the
variables to the right places in argv[]
What would the most appropriate fix be?
Strontium
Strontium wrote:
> Howdy,
>
> I have a script that goes:
>
> proc omap3_ReadDebugRegister { reg_num } {
> # read the value of the debug register reg_num at address reg_num << 2
> omap3.cpu mem2array dataval 32 [expr "0x54011000 + $reg_num * 4"] 1
> }
>
> It errors and crashes openocd with the following message:
>
> Runtime error, file "script//omap3.tcl", line 11:
> wrong # args: should be "dataval varname width addr nelems"
>
> I cant see anything wrong with it, but obviously I have stuffed up
> somehow. Any suggestions on fixing it would be mightily appreciated.
>
> I will add processing to dataval after this, but until i can read
> dataval its a bit pointless.
>
> Strontium
>
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development
