How to disable EDK instantiated IOBs?

[MM]

This is a more generic question following up my previous question on
translation failing because of the IOBs instantiated in one of the EDK
cores. I was wondering if there is a way (short of hacking the core) to
either disable IOB instantiation in EDK altogether or vice-versa to disable
IOB instantiation in ISE, but only for those pins, which already have IOBs
instantiated by EDK. I tried implementing the latter idea by assigning
synthesis iob attribute for these pins and setting it to FALSE, but got an
INTERNAL_ERROR in XST... I can't believe there is no simple solution... In
the past ISE somehow managed to figure this out. I also vaguely remember
that at some point long ago I had a similar problem and did something with
some of the cores, but can't remember what... In any case latest MPMC core
has these IOB instantiations scattered through many files and hacking them
all doesn't appear to be a nice solution.

Any ideas will be highly appreciated!

/Mikhail

 

[Kim Povlsen]

Hi MM,

I haven't been able to find your previous posts.

In your attempt to disable the IOB insertion on the IP Core, did you
try to modify the .mhs file in XPS, by assigning BUFFER_TYPE=NONE on
the port giving you problems?

Regards,
Kim P.

On Thu, 14 Oct 2010 17:41:20 -0400, "MM" <mbmsv@yahoo.com> wrote:

This is a more generic question following up my previous question on
translation failing because of the IOBs instantiated in one of the EDK
cores. I was wondering if there is a way (short of hacking the core) to
either disable IOB instantiation in EDK altogether or vice-versa to disable
IOB instantiation in ISE, but only for those pins, which already have IOBs
instantiated by EDK. I tried implementing the latter idea by assigning
synthesis iob attribute for these pins and setting it to FALSE, but got an
INTERNAL_ERROR in XST... I can't believe there is no simple solution... In
the past ISE somehow managed to figure this out. I also vaguely remember
that at some point long ago I had a similar problem and did something with
some of the cores, but can't remember what... In any case latest MPMC core
has these IOB instantiations scattered through many files and hacking them
all doesn't appear to be a nice solution.

Any ideas will be highly appreciated!

/Mikhail

 

[MM]

"Kim Povlsen" <kim@rtldesign.dk> wrote in message
news:9fcgb6dk3277dks6sahi1f9gggsih4n5ai@4ax.com...

In your attempt to disable the IOB insertion on the IP Core, did you
try to modify the .mhs file in XPS, by assigning BUFFER_TYPE=NONE on
the port giving you problems?

I didn't know you could do this in the .mhs file... Thank you for the
suggestion. But it turns out that I don't need to do this at all. See the
following thread: "ISE/EDK12.2 Translate fails with "NgdBuild:467 - output
pad net 'Ddr_Ck_N_0_OBUF' has an illegal buffer" and similar messages."

Thanks,
/Mikhail