Re: [microblaze-uclinux] "Value too large to represent" error duringFPGA 'libs' build

[John Williams <jwilliams@xxxxxxxxxxxxxx>]

Hi Kevin,

Kevin Somervill wrote:
> Hello David,
> |
> |OK, I've traced this error:
> |
> |to this parameter definition:
> |
> |>C_TRI_DEFAULT = 11111111111111111111111111111111
> |
> |which is a binary definition, but is missing the '0b' prefix 
> |that the Tcl script expects to be in front of binary constants.
> |
> |However, I can't find this parameter's definition anywhere! Is 
> |this a default? If so, can anyone point me to the file that 
> |defines it, so that I can add the '0b' prefix?
> 
> The file is in
> $EDK\hw\XilinxProcessorIPLib\pcores\opb_gpio_v3_01_a\data\opb_gpio_v2_1_
> 0.mpd, but it won't have a 0b in front.  This is probably a bug in the
> TCL script in that it should be able to translate the value since it's
> defined as DT = std_logic_vector.  You could try to put the 0b in the
> file, but this is not a long term solution since a lot of files use this
> (the default for hdl std_logic_vector is bit fields).

Hmm, I'll have a look into it.  I found another limitation of the script 
yesterday, when specifying base addresses (or hex values generally) the 
EDK allows you to insert '_' characters to space out the digits, like 
0x80000_0000.

However, the TCL functions (and I think the EDK's back-end TCL 
libraries), fail to parse any hex constants with '_' in them.  For now, 
the easiest fix is just to remove the underscores from hex constants, we 
can fix it properly some time down the track.

I appreciate people's patience with this new tool, despite the expected 
teething problems I still think it's supererior to the old hack'n'slash 
method!

Regards,

John
___________________________
microblaze-uclinux mailing list
microblaze-uclinux@itee.uq.edu.au
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/