xc3sprog-users — Generic user and developer list

Re: [Xc3sprog-users] Report about parallel cable in FreeBSD

[Uwe B. <bo...@el...>]

>>>>> "Wojciech" == Wojciech A Koszek <wkoszek@FreeBSD.org> writes:

    Wojciech> Hi, Several days ago I've sent quite long mail to this list.

    Wojciech> It described experiences with xc3sprog + FreeBSD with parallel
    Wojciech> cable.

    Wojciech> I think report was long enough to exceed 40k limit for this
    Wojciech> mailing list. I got a warning about a need for an approval
    Wojciech> from the list's moderator.

    Wojciech> Uwe, any chance to get it accepted? I wouldn't like to loose
    Wojciech> it.

I have no influence on the sourceforge settings, an it seems the mail didn't
get through.

So first send the mail in private email.

Bye
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Report about parallel cable in FreeBSD

[Wojciech A. K. <wkoszek@FreeBSD.org>]

Hi,

Several days ago I've sent quite long mail to this list.

It described experiences with xc3sprog + FreeBSD with parallel cable.

I think report was long enough to exceed 40k limit for this mailing
list. I got a warning about a need for an approval from the list's moderator.

Uwe, any chance to get it accepted? I wouldn't like to loose it.

Thanks,
-- 
Wojciech A. Koszek
wkoszek@FreeBSD.org
http://FreeBSD.czest.pl/~wkoszek/

Re: [Xc3sprog-users] Bringing FreeBSD support to xc3sprog.

[Uwe B. <bo...@el...>]

>>>>> "Wojciech" == Wojciech A Koszek <wkoszek@FreeBSD.org> writes:

...
    Wojciech> Before I prepare a port of xc3sprog for FreeBSD, I'll probably
    Wojciech> send you some more stuff. I want your xc3sprog to be
    Wojciech> identified well (Author+project URL), so that there is no
    Wojciech> confusion with Andrew Rogers' stuff and XUP code.

    Wojciech> Do they have something which your code doesn't have?

I started from Rogers sources, so everything should be included.

XUP uses a different firmware for the DLC9/10 dongles. No need for the a
different firmware with kawk's reverse engineering the Xilinx protocoll. I
didn't check if they added anythging else.

    >>  I added the errno.h in the FreeBSD conditional path. (Rev 428)
    >> 
    >> A CMAKE Toolchain File is only needed for crosscompiling. Did you
    >> crosscompile? Or does CMAKE not work appropriate on FreeBSD?  It's
    >> added as Rev 429.

    Wojciech> I just assumed you pick one CMAKE file for one family of
    Wojciech> systems. Can you use this string (system name) in the xc3sprog
    Wojciech> output somehow?

Normally you start "cmake" and it should configure for the running
system. Toolchainfiles only needed for crosscompiling. As I run on Suse
64-bit, I needed the Toolchain-SUSE32.cmake to compile the 32-bit binary.


...
    Wojciech> At home I'll just comment it out and perform some tests with
    Wojciech> small/big bitfiles and see if everything is more broken.

Look at the -T0 option to keep JTAG busy.

Bye
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Re: [Xc3sprog-users] Bringing FreeBSD support to xc3sprog.

[Uwe B. <bo...@el...>]

(setting the xc3sprog mailing list on the CC)
>>>>> "Wojciech" == Wojciech A Koszek <wkoszek@FreeBSD.org> writes:

    Wojciech> On Sat, Jan 16, 2010 at 12:12:03AM +0100, Uwe Bonnes wrote:
    >> >>>>> "Wojciech" == Wojciech A Koszek <wkoszek@FreeBSD.org> writes:
    >> 
    Wojciech> Hi, I want to have FreeBSD support in xc3sprog.  This is first
    Wojciech> patch that makes you code compile cleanly under FreeBSD:
    >>
    Wojciech> http://freebsd.czest.pl/~wkoszek/fpga/xc3sprog.0.patch

Thanks for reminding! I didn't notice in the first mail.

I added the errno.h in the FreeBSD conditional path. (Rev 428)

A CMAKE Toolchain File is only needed for crosscompiling. Did you
crosscompile? Or does CMAKE not work appropriate on FreeBSD? 
It's added as Rev 429.
    >>
    Wojciech> 0x03fd:0x0008 is for default iMPACT firmware for Cypress's FX2
    Wojciech> chip?  I run Spartan 3E-500 Starter Kit with chip reprogrammed
    Wojciech> with xc3sprog and there it's 0x03fd:0x000d that is detected..
    >>  0x03fd:0x000d is the Xilinx FX2 CHIP-ID. You must load the
    >> appropriate firmware. When you install the impact cable drivers ( in
    >> linux, can you report what happens in BSD when installing the XILINX
    >> driver?) the firmware and udev rules are copied to the ppropriate
    >> places. You can also load the firmware by hand with fxload or maybe
    >> other loaders. After loading the right firmware, the chip
    >> reenumerated and should get listed as 0x03fd:0x0008.

Can't you compile native? Only libusb/libftdi should be needed and available
for BSD too. Please let me know about problems.

    Wojciech> OK, I didn't know that on Linux fxload stage is necessary as a
    Wojciech> first step.

    Wojciech> We have Linux emulation layer, so that you can run Linux
    Wojciech> binaries on FreeBSD without a lot of problems. I have
    Wojciech> /compat/linux directory which can be populated by binary
    Wojciech> packages.  Each package contains Linux binaries and
    Wojciech> libraries. They can be executed from this directory.
    Wojciech> /compat/linux is basically it's where everything gets
    Wojciech> installed. So my ISE is under /compat/linux/opt/Xilinx/... and
    Wojciech> I have all the .hex files there.

    Wojciech> In terms of drivers...

    Wojciech> Layout of /dev in Linux and FreeBSD differs. But I've compiled
    Wojciech> FreeBSD's libusb(3) implementation on Linux machine and got
    Wojciech> USB to work in emulation.  and I also got Impact working
    Wojciech> natively with it--I can right now use it to program my Spartan
    Wojciech> 3E-Starter Kit.

    Wojciech> Today I commited kernel level patch to make this
    Wojciech> happen. User-space patch is being reviewed.

    Wojciech> Delay in a response is caused by the fact that we seem to have
    Wojciech> no native fxload here, so I had to look around and found
    Wojciech> libusb-enabled Windows port of fxload, which compiled for me
    Wojciech> after minor tweaks.

    Wojciech> Does it answer your question?

    Wojciech> Anyway, this is what I get from your code:

    Wojciech> wkoszek@laptop:/media/Multimedia/fpga/xc3sprog/trunk/build#
    Wojciech> ./xc3sprog -c xpc Release $Rev: 426 $ Free software: If you
    Wojciech> contribute nothing, expect nothing!  Please provide feedback
    Wojciech> on success/failure/enhancement requests!  Check Sourceforge
    Wojciech> SVN for updates!  DEBUG(xpcu_read_hid:196)rc = -1, errno = 5
    Wojciech> usb_control_msg(0x42.1 Unknown error 

Did you add the  DEBUG(..? It's not from the xc3sprog code. Is errno =5
#define EIO              5      /* I/O error */
on FreeBSD too? 

Why does it get translated to "Unknown error"?
> DEBUG(xpcu_read_hid:196)rc = -1, errno = 5
> usb_control_msg(0x42.1 Unknown error

The Call 0x42.1 however is only a verbatim reproduction of a reverse
engineer run on the XPC. Maybe it's no needed, maybe it should be another
sequence with the DLC9/10 integrated in the Spartan 3E-500 Starter Kit.


    Wojciech> firmware version = 0x0404
    Wojciech> (1028) CPLD version = 0x0012 (18) JTAG loc.: 0 IDCODE:
    Wojciech> 0x41c22093 Desc: XC3S500E IR length: 6 JTAG loc.: 1 IDCODE:
    Wojciech> 0xf5046093 Desc: XCF04S IR length: 8 JTAG loc.: 2 IDCODE:
    Wojciech> 0x06e5e093 Desc: XC2C64A-VQ44 IR length: 8

    Wojciech> Startan 3E Starter Kit identification.

    Wojciech> wkoszek@laptop:/media/Multimedia/fpga/xc3sprog/trunk/build#
    Wojciech> ./xc3sprog -c xpc
    Wojciech> /home/wkoszek/fpga/fbsd-prog/xup-0.0.2/w/wkoszek_leds/leds.bit
    Wojciech> Release $Rev: 426 $ Free software: If you contribute nothing,
    Wojciech> expect nothing!  Please provide feedback on
    Wojciech> success/failure/enhancement requests!  Check Sourceforge SVN
    Wojciech> for updates!  DEBUG(xpcu_read_hid:196)rc = -1, errno = 5
    Wojciech> usb_control_msg(0x42.1 Unknown error firmware version = 0x0404
    Wojciech> (1028) CPLD version = 0x0012 (18)

    Wojciech> Even though I'm getting an errors, this seems to work just
    Wojciech> fine as well.  LED ruler implemented in this bitstream appears
    Wojciech> after programming.


-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] SVN access spikes

[Uwe B. <bo...@el...>]

Hello,

it seems few people use xc3sprog SVN, as would be best for a developping
program.  But on Dec. 9, for the second time, there was a spike of over 400 SVN
accesses in one day. Does anybody know what is going on?

Thanks
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Re: [Xc3sprog-users] Error handling in Multi Prom Files case

[Joris v. R. <jo...@sr...>]

Hi,

On 10 nov 2009, at 19:05, Uwe Bonnes wrote:
> V418 tries to bring back an error return value when programming  
> flash, as it
> aborts when one prom fails to erase/program. Do you agree with that
> handling?

Just a bit of positive feedback:
I tested SVN r423 with my board and it still works like a charm.

Greetings, Joris.

[Xc3sprog-users] Rev 423

[Uwe B. <bo...@el...>]

Hello,

thanks to Thomas Preußer rev 423 could make the appending of additional PROM
data usable again. It was probably lost with the command line argument
handling rewrite. This is probably usefull for SOC to load up the programm.

In this context, writing to parallel flash wold probably also be
usefull. Does anybody have HDL code to do so?

Thanks 
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Usleep fixes for Win32

[Uwe B. <bo...@el...>]

Hello,

it seems that ZwDelayExecution didn't work on XP, and so Usleep() and
probably most programming. Strange that I didn't hear
reports about that. Hope the new approach works in more cases.

Bye

-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Error handling in Multi Prom Files case

[Uwe B. <bo...@el...>]

Dear Joris,

V418 tries to bring back an error return value when programming flash, as it
aborts when one prom fails to erase/program. Do you agree with that
handling?

Thanks
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] XC18V handled with REV417

[Uwe B. <bo...@el...>]

The XCF 1532 algorithms are a superset of the XC18V algorithms. progalgxcf
now handles XC18V with the non-optimized algorithms, that means e.g. after
erase a fixed time is waited as the chip doesn't provide a  mean to
query the chip if erase is finished.

There is no blank check done yet, as supposed in the 1532 BSDL files. Let
look where this hurts...
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Re: [Xc3sprog-users] speed...

[Uwe B. <bo...@el...>]

>>>>> "Jaap" == Jaap de Jong <jaa...@ne...> writes:

    Jaap> Hi Uwe Can you explain why it takes so long to read the bitfile?
    Jaap> The programming takes 935 ms while the whole procedure takes 3.9
    Jaap> s.  Well, it is no big deal, but I was just wondering...  Regards,
    Jaap> Jaap


    Jaap> time xc3sprog -c ftdi -v xyzzy.bit Release $Rev: 416 $ Free
    Jaap> software: If you contribute nothing, expect nothing!  Please
    Jaap> provide feedback on success/failure/enhancement requests!  Check
    Jaap> Sourceforge SVN for updates!  Using built-in device list JTAG
    Jaap> chainpos: 0 Device IDCODE = 0x01c2e093 Desc: XC3S1200E Created
    Jaap> from NCD file: xyzzy.ncd;HW_TIMEOUT=FALSE Target device:
    Jaap> 3s1200eft256 Created: 2009/02/16 15:15:04 Bitstream length:
    Jaap> 3841184 bits done. Programming time 935.8 ms USB transactions:
    Jaap> Write 244 read 7 retries 0 real 0m 3.90s user 0m 2.85s sys 0m
    Jaap> 0.06s

(forwarded to xc3sprog-users)

Jaap, what was you setup? I guess it was a FT2232D|L linked directly into
the PC. That way, the Full Spped devices is handled as full speed device in
the PC, and transaction take place in one millisecond steps. If you connect
the FT2232D|L to a USB-2.0 hub, as written in the README, the hub stores and
forwards microframes and things happen in 125 Mikrosecond steps, in the best
case a 8 Times enhancement.

The FT2232H will work that fast without the HUB, as it is a high speed
devices per se.

Bye
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Inconsistant license information

[Uwe B. <bo...@el...>]

Hello,

Joris van Rantwijk just pointed to
"inconsistent licensing information" on the tracker.

I obviously copied the License file from a libray project, so there was a
wrong reference to LGPL and not GPL. If there are other flaws, please let me
know.

Thanks
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] spartan 6 support

[Jaap de J. <jaa...@ne...>]

Is it enough to add the following lines to devlist.txt to get devices
supported by this tool?

0x4004093 6 0x9 XC6SLX25 
0x4024093 6 0x9 XC6SLX25T

Thanks!

[Xc3sprog-users] Recognize some more FT2232 Adapters

[Uwe B. <bo...@el...>]

Hello,

xc3sprog SVN now adds some more subtypes (amontec, oimex)for the FTDI
cable type and tries to use the right default USB vid/pid combination and
activate the JTAG buffers. Probaly more elaboration is needed. Feedback
welcome

Thanks
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Xilinx DLC10 cable support from urjtag/src/tap/cable/xpc.c

[Uwe B. <bo...@el...>]

Hello,

I have adapted Kolja Waschk' reversed engineered urjtag/src/tap/cable/xpc.c
for xc3sprog. Tested with a DLC10 by loading the bscan_spi pattern to the
XC3SD1800 from a ADSP Starter board. After substantially increasing the
buffer size, as opposed to Kolja's observation with a DLC9, speed now is only
about half as slow (6 seconds) as impact(3 seconds), while it crawled before
with the original buffer size.

Cheers
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Major XC3SPROG update

[Uwe B. <bo...@el...>]

Hello,

I updated the sourceforge xc3sprog tree with a dump of my SVN tree, starting
with the sourceforge published rev6. Numbering has pitily changed in that
course of matter.

Requirements:
- CMAKE
- libusb  (linked dynamic on Linux, loaded by LoadLibrary() on Win32
- libftdi (static linked)

Supported adapters:
- Byteblaster/DLC5
- FT2232 in MPSSE connection
- Framework for FX2 ( I extended usrp/xguff I2C path )

- Tested on Win32 and Linux

Part     Programm Verify Read
XC3S         X
XC3S50AN-SPI X      X    X
XC95X        X      X    X
XCF          X      X    X

Planned work:
- Experiment with DLC10
- Support XC2C, but fusemapping will chew large amount of memory. Probably
  will start with P/V/R some kind of bitfile and provide a bitfile-jedecfile
  mapper. 
- Support more parts (XC2X, XC18 are available on my desk)
- Support more SPI Flash vendors (e.g. Intel for the ADSP1800 starter kit
  on my desk)
- Have a look at urjtags fjmem to program byte-parallel Flash (also
  on the ADSP starter set)

Cheers
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] More patches

[Uwe B. <bo...@el...>]

Hello,

I have sent more patches to sync my local tree the sourceforge tree. Please
let me know if something still needs changes.

However programming and XCF02 still needs about 22 sec with the patched
sourceforge tree and only 7 second with my local tree. Have to
investigate...

Cheers
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Patches in SVN

[Dmitry T. <di...@gm...>]

Hello,

I've applied all of the pending patches and committed them to SVN.

-- 
Dmitry Teytelman <di...@gm...>

[Xc3sprog-users] Revision 4 of xc3sprog committed to SVN

[Dmitry T. <di...@gm...>]

Hello,

I've just committed revision 4 of xc3sprog to SVN. Here is a (fairly
extensive) list of changes:

  Added FT2232 USB-to-JTAG driver
  Verbose mode for xc3sprog
  Cable driver selection in xc3sprog and detectchain
  Output device selection in xc3sprog and detectchain
  Virtex II device IDs added to devlist.txt
  DeviceDB::loadDevice() modified to ignore 4 MSBs of the device ID
  Schematics of a Xilinx-compatible parallel cable and FTDI cable
  Makefile changes to support "make install"
  Code cleanup for clean -Wall compile

I've tested the updated code in the following configurations:

  FTDI cable on Opal Kelly XEM3001 (XC3S400)
  Integrated FTDI programmer on a custom board (XC3S200)
  My parallel cable version (see schematics) on Opal Kelly XEM3001
  Xilinx parallel cable IV with XC2V4000

-- 
Dmitry Teytelman <di...@gm...>

Re: [Xc3sprog-users] Hello

[Eric J. <jo...@mw...>]

> Eric = Eric Jonas == The one who created the sourceforge project
  Correct!

> There are also some more cleanup patches in my tree. Theu are needed to run
> gcc with the -WALL option.
> Much more delicate will be the needed encapsulation for
> other transport layers, like the JTAG machine of the FTDI FT2232 ( which I
> have pending on my machine) and perhaps the FX2 transport layer used in
> Antti Lukats  Xilinx USB Blaster clone.

I agree; I think there's some substantial work to be done with regard to
factoring out the JTAG controller, the UI, and the low-level transport
drivers. The code is C++ so it would be nice if we could make good use
of the design features that the language provides us. 

Also, I found Andrew Rogers original style a bit difficult to read, so
if anyone else does, as well, we might try just a basic style clean-up
once the latest patches are applied. 

			...Eric

Re: [Xc3sprog-users] Hello

[Uwe B. <bo...@el...>]

>>>>> "Eric" == Eric Jonas <jo...@mw...> writes:

    Eric> XC3Sprog users,

    Eric> I'm hosed with my thesis right now and haven't had time to touch
    Eric> the code; in the meantime I've given dmitry commit access to
    Eric> 1. apply his patches and 2. perhaps begin working through the
    Eric> submitted list of patches from other users, such as uwe. I worry
    Eric> they might overlap a bit, so this isn't a trivial task. Let me
    Eric> know if you have any questions :)

    Eric> If you're receiving this and you're not on the list feel free to
    Eric> add yourself :) And if you submit a patch and feel like we're
    Eric> ignoring you, please ping us.  ...Eric

I tried to have my patches (nearly) independant from each other.

To get names right:

Eric = Eric Jonas == The one who created the sourceforge project

I didn't see Dimitry involved in any newsnet discussion around XC3SPROG yet.

There are also some more cleanup patches in my tree. Theu are needed to run
gcc with the -WALL option.
Much more delicate will be the needed encapsulation for
other transport layers, like the JTAG machine of the FTDI FT2232 ( which I
have pending on my machine) and perhaps the FX2 transport layer used in
Antti Lukats  Xilinx USB Blaster clone.

Cheers
-- 
Uwe Bonnes                bo...@el...

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

[Xc3sprog-users] Hello

[Eric J. <jo...@mw...>]

XC3Sprog users, 

I'm hosed with my thesis right now and haven't had time to touch the
code; in the meantime I've given dmitry commit access to 1. apply his
patches and 2. perhaps begin working through the submitted list of
patches from other users, such as uwe. I worry they might overlap a bit,
so this isn't a trivial task. Let me know if you have any questions :) 

If you're receiving this and you're not on the list feel free to add
yourself :) And if you submit a patch and feel like we're ignoring you,
please ping us. 
		...Eric