[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AW: [microblaze-uclinux] ethernet + EMC



Hi John,

Yes, It's very important to check to you don't have unloced pins since they can be placed anywhere and do anything.

There is always a problem with the memcon since it drives 32 address signals and the external memory doesn't have 32 address lines.
Currently you can't slice a signal in .mhs file but I have created a small utility core which will split a bus into two smaller busses and by using two of
these cores, I can trim down to A(0:31) to A(10:29) which is what exists on the P160 board.

I have attached the core and also a code snippet from the .mhs file on how to use it.
This core will be in the coming EDK release but we have a couple of these utility core and therefore we have added a prefix in the names of util_
So the core will be called util_bus_split in EDK 6.1.

This core is for the EDK3.2 release.

Great work
Göran

....
BEGIN bus_split
 PARAMETER INSTANCE = bus_split_0
 PARAMETER HW_VER = 1.00.a
 PARAMETER C_SIZE_IN = 32
 PARAMETER C_SPLIT = 30
 PORT Sig = sram_addr_i
 PORT Out1 = sram_addr_ii
END

BEGIN bus_split
 PARAMETER INSTANCE = bus_split_1
 PARAMETER HW_VER = 1.00.a
 PARAMETER C_SIZE_IN = 30
 PARAMETER C_SPLIT = 10
 PORT Sig = sram_addr_ii
 PORT Out2 = sram_addr
END

....

John Williams wrote:

> Schunke Jan-Hendrik wrote:
> > Hi John,
> >
> > The only way (I know) to check if you have locked all pins is a look at the place and route report (./implementation/*.par). There you will find something like:
> >
> > ...
> > Device utilization summary:
> >
> >    Number of External IOBs            87 out of 812    10%
> >       Number of LOCed External IOBs   87 out of 87    100%
> > ...
> >
> > Maybe you can write a simple script to parse the file and check these lines ;-)
>
> Hey that's a good idea, I could just add it into the makefile to do a
> quick check and report if there's a problem...
>
> also thanks for the emac ucf stuff, i'll give it a go.  i haven't tried
> actually using the emac yet, just the first step which is that uclinux
> is still happy in a system that has an emac on the bus - got to do
> things one step at a time :-)
>
> John
>
> >
> > Jan
> >
> >
> >>-----Ursprüngliche Nachricht-----
> >>Von: John Williams [mailto:jwilliams@itee.uq.edu.au]
> >>Gesendet: Dienstag, 26. August 2003 14:57
> >>An: microblaze-uclinux@itee.uq.edu.au
> >>Betreff: Re: [microblaze-uclinux] ethernet + EMC
> >>
> >>
> >>Hi Jan,
> >>
> >>
> >>>The D0-D7 signals on the P160 are shared for FLASH/SRAM and LCD
> >>>connector.
> >>
> >>The LCD signals are buffered by the cool runner. Did you
> >>accidently reprogram the coolrunner and it pulls down the
> >>lines now? There are also some signals that control the
> >>coolrunner: PLD_LCD_RWN, PLD_LCD_EN ... I would check these, too.
> >>
> >>you are correct!  i am bitten once again by the "random
> >>allocation of unconstrained pins" behaviour of the synthesis
> >>tools.... grrrr... but I should know better.  :-)
> >>
> >>does anyone one know a synthesis switch so unconstrained pin
> >>LOCs cause an error instead of silent potentially dangerous behaviour?
> >>
> >>I had declared the gpio to be 32 bit wide but only
> >>constrained the lowest 24 bits - I was just lucky in earlier
> >>mbvanilla targets, but this time the new io's of the MAC
> >>caused these unconstrained gpio's to move, in this case onto
> >>the CPLD write enable line...
> >>
> >>
> >>>I have taken a quick look at your mhs and that seems to be ok.
> >>
> >>it should be - the ethernet stuff I lifted directly from a
> >>reference design you sent me a while ago :)
> >>
> >>Thanks for your help - uclinux network access gets one step closer...
> >>
> >>John
> >>
> >>
> >>>Jan
> >>>
> >>>
> >>>
> >>>>-----Ursprüngliche Nachricht-----
> >>>>Von: John Williams [mailto:jwilliams@itee.uq.edu.au]
> >>>>Gesendet: Dienstag, 26. August 2003 09:38
> >>>>An: microblaze-uclinux@itee.uq.edu.au
> >>>>Betreff: [microblaze-uclinux] ethernet + EMC
> >>>>
> >>>>
> >>>>Hi folks,
> >>>>
> >>>>Trying to bring up networking on uClinux, first step I've put a
> >>>>xilinx ethernet MAC core in my project (surprisingly enough called
> >>>>mbvanilla_net :).
> >>>>
> >>>>Anyway I've found that when I include the MAC core, all of
> >>
> >>a sudden
> >>
> >>>>my EMC peripheral (which drives the flash and SRAM) doesn't
> >>>>work.  The low
> >>>>byte of all reads in the EMC address space returns
> >>
> >>0xFF...the high 3
> >>
> >>>>bytes are fine...
> >>>>
> >>>>I can query the ethernet MAC device ID register, and the
> >>
> >>BRAM and DDR
> >>
> >>>>memory are unaffected.
> >>>>
> >>>>So, I can upload an image direct into DDR and execute it, no
> >>>>problems, however my bootloading process of reading from
> >>
> >>flash and so
> >>
> >>>>on can't work - all reads from flash are corrupted by this
> >>
> >>0xFF byte.
> >>
> >>>>Has anybody had any experiences, positive or negative, re:
> >>
> >>using the
> >>
> >>>>EMC and EMAC in the same target?  I've attached the mhs
> >>
> >>and ucf files
> >>
> >>>>for reference.
> >>>>
> >>>>Any ideas of where to look would be greatly appreciated.
> >>>>
> >>>>Thanks,
> >>>>
> >>>>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/
> >
> >
> > ___________________________
> > 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/
> >
> >
> >
> > ___________________________
> > 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/
>
> --
> Dr John Williams, Research Fellow,
> Reconfigurable Computing, School of ITEE
> University of Queensland, Brisbane, Australia
> Ph : (07) 3365 8305
>
> ___________________________
> 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/

Attachment: bus_split_v1_00_a.zip
Description: Zip compressed data