Forums

HP "owning" the software for Xilinx-FTDI drivers???

Started by Rick C October 27, 2020
Someone new on the project is talking about how HP owns the driver code for=
 the FTDI JTAG chip used to program Xilinx parts.  Has anyone heard of this=
? =20

Even if that is true, I'm not sure how relevant it is.  Is there some issue=
 with HP owning the software for FTDI devices even if that is true?  Why wo=
uld it matter to a user of Xilinx FPGAs???

I did try getting an answer and failed.  The guy bringing it up is from a b=
ackground of big dollar development and I don't know if there was some sort=
 of IP ownership issue.  I can't see how it would impact us even if there w=
ere.  I really can't imagine anyone other than FTDI owns the drivers for th=
e FTDI chips, but maybe there's some sort of JTAG software involved. =20

Sometimes it is strange working with strangers.=20

--=20

  Rick C.

  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C:
> Someone new on the project is talking about how HP owns the driver code f=
or the FTDI JTAG chip used to program Xilinx parts. Has anyone heard of th= is? =20
>=20 > Even if that is true, I'm not sure how relevant it is. Is there some iss=
ue with HP owning the software for FTDI devices even if that is true? Why = would it matter to a user of Xilinx FPGAs???
>=20 > I did try getting an answer and failed. The guy bringing it up is from a=
background of big dollar development and I don't know if there was some so= rt of IP ownership issue. I can't see how it would impact us even if there= were. I really can't imagine anyone other than FTDI owns the drivers for = the FTDI chips, but maybe there's some sort of JTAG software involved. =20
>=20 > Sometimes it is strange working with strangers.=20
standard ftdi chips can talk jtag with the standard ftdi driver, you just h= ave to make you own code to drive it and it won't work from inside the xilinx t= ools I know digilent has a jtag based on ftdi that works with Xilinx tools=20 and is often part off xilinx boards (redacted from the schematic), it=20 requires something special in the ftdi eeprom to load their driver and=20 software instead of just being an ftdi device, basically to verify that it= =20 is a licensed digilent programmer and not just a generic ftdi bought from d= igikey
On Tuesday, October 27, 2020 at 5:43:10 PM UTC-4, lasselangwad...@gmail.com=
 wrote:
> tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C: > > Someone new on the project is talking about how HP owns the driver code=
for the FTDI JTAG chip used to program Xilinx parts. Has anyone heard of = this? =20
> >=20 > > Even if that is true, I'm not sure how relevant it is. Is there some i=
ssue with HP owning the software for FTDI devices even if that is true? Wh= y would it matter to a user of Xilinx FPGAs???
> >=20 > > I did try getting an answer and failed. The guy bringing it up is from=
a background of big dollar development and I don't know if there was some = sort of IP ownership issue. I can't see how it would impact us even if the= re were. I really can't imagine anyone other than FTDI owns the drivers fo= r the FTDI chips, but maybe there's some sort of JTAG software involved. = =20
> >=20 > > Sometimes it is strange working with strangers.=20 >=20 > standard ftdi chips can talk jtag with the standard ftdi driver, you just=
have
> to make you own code to drive it and it won't work from inside the xilinx=
tools Your own code to drive it? That's pretty rad, but I think I understand wha= t you are saying.=20
> I know digilent has a jtag based on ftdi that works with Xilinx tools=20 > and is often part off xilinx boards (redacted from the schematic), it=20 > requires something special in the ftdi eeprom to load their driver and=20 > software instead of just being an ftdi device, basically to verify that i=
t=20
> is a licensed digilent programmer and not just a generic ftdi bought from=
digikey I think Trenz has some interface board with the FTDI chip and a small FPGA = (not for the user to muck with, but for the module operation) that costs a = few bucks more for Xilinx capability. I guess this is what they are talkin= g about, something special to work with the Xilinx drivers. I guess Xilinx= has some detail they will share under license to add this capability to yo= ur boards.=20 I still have no idea what he is talking about regarding this being "HP" IP.= I'm starting to get tired of the guy. He seems to really want to deep si= x the Gowin part even though the Xilinx part will cost four times as much a= nd take up twice the board space or require more expensive board fabricatio= n. =20 I'm starting to get tired of this guy. I guess this is what to expect from= part time help... --=20 Rick C. + Get 1,000 miles of free Supercharging + Tesla referral code - https://ts.la/richard11209
onsdag den 28. oktober 2020 kl. 00.19.29 UTC+1 skrev Rick C:
> On Tuesday, October 27, 2020 at 5:43:10 PM UTC-4, lasselangwad...@gmail.c=
om wrote:
> > tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C: > > > Someone new on the project is talking about how HP owns the driver co=
de for the FTDI JTAG chip used to program Xilinx parts. Has anyone heard o= f this? =20
> > >=20 > > > Even if that is true, I'm not sure how relevant it is. Is there some=
issue with HP owning the software for FTDI devices even if that is true? = Why would it matter to a user of Xilinx FPGAs???
> > >=20 > > > I did try getting an answer and failed. The guy bringing it up is fr=
om a background of big dollar development and I don't know if there was som= e sort of IP ownership issue. I can't see how it would impact us even if t= here were. I really can't imagine anyone other than FTDI owns the drivers = for the FTDI chips, but maybe there's some sort of JTAG software involved. = =20
> > >=20 > > > Sometimes it is strange working with strangers.=20 > >=20 > > standard ftdi chips can talk jtag with the standard ftdi driver, you ju=
st have
> > to make you own code to drive it and it won't work from inside the xili=
nx tools
>=20 > Your own code to drive it? That's pretty rad, but I think I understand w=
hat you are saying.=20
>=20
if you just want to configure a part it is pretty straight forward, but if= =20 you want to talk to a build in logic analyser, program one time fuses witho= ut ruining boards, and all the other things the tools can do I'm sure it ca= n get pretty hairy
On Tuesday, October 27, 2020 at 8:03:33 PM UTC-4, lasselangwad...@gmail.com=
 wrote:
> onsdag den 28. oktober 2020 kl. 00.19.29 UTC+1 skrev Rick C: > > On Tuesday, October 27, 2020 at 5:43:10 PM UTC-4, lasselangwad...@gmail=
.com wrote:
> > > tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C: > > > > Someone new on the project is talking about how HP owns the driver =
code for the FTDI JTAG chip used to program Xilinx parts. Has anyone heard= of this? =20
> > > >=20 > > > > Even if that is true, I'm not sure how relevant it is. Is there so=
me issue with HP owning the software for FTDI devices even if that is true?= Why would it matter to a user of Xilinx FPGAs???
> > > >=20 > > > > I did try getting an answer and failed. The guy bringing it up is =
from a background of big dollar development and I don't know if there was s= ome sort of IP ownership issue. I can't see how it would impact us even if= there were. I really can't imagine anyone other than FTDI owns the driver= s for the FTDI chips, but maybe there's some sort of JTAG software involved= . =20
> > > >=20 > > > > Sometimes it is strange working with strangers.=20 > > >=20 > > > standard ftdi chips can talk jtag with the standard ftdi driver, you =
just have
> > > to make you own code to drive it and it won't work from inside the xi=
linx tools
> >=20 > > Your own code to drive it? That's pretty rad, but I think I understand=
what you are saying.=20
> >=20 >=20 > if you just want to configure a part it is pretty straight forward, but i=
f=20
> you want to talk to a build in logic analyser, program one time fuses wit=
hout ruining boards, and all the other things the tools can do I'm sure it = can get pretty hairy I'm just trying to figure out what this guy was talking about saying that H= P owned the software. I would expect either FTDI or Xilinx would own it. = He never answered me when I asked why it mattered. I think he is trying to= find some reason to nix the approach we are taking with the FPGA on this p= roject. He's also seems to be digging to find things with the Gowin chip t= hat he can object to. Or maybe I am just being paranoid. I guess I find s= ome of his comments weird and don't know what to make of it. =20 The trouble is we don't have a bunch of funding to pull off this project an= d the mechanical engineers have to spend a fair bit on the stainless steel = cabinet. So we are trying to get the boards made for free by JLCPCB. We'l= l have to assemble the parts they don't carry, but they have offered to mak= e some of our boards with the parts they do carry. That's pretty generous.= I hope they stick to that when they see the updated board. It's 100x200 = mm and will end up pretty full of parts. =20 --=20 Rick C. -- Get 1,000 miles of free Supercharging -- Tesla referral code - https://ts.la/richard11209
On 28/10/2020 03:48, Rick C wrote:
> On Tuesday, October 27, 2020 at 8:03:33 PM UTC-4, lasselangwad...@gmail.com wrote: >> onsdag den 28. oktober 2020 kl. 00.19.29 UTC+1 skrev Rick C: >>> On Tuesday, October 27, 2020 at 5:43:10 PM UTC-4, lasselangwad...@gmail.com wrote: >>>> tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C: >>>>> Someone new on the project is talking about how HP owns the driver code for the FTDI JTAG chip used to program Xilinx parts. Has anyone heard of this? >>>>> >>>>> Even if that is true, I'm not sure how relevant it is. Is there some issue with HP owning the software for FTDI devices even if that is true? Why would it matter to a user of Xilinx FPGAs??? >>>>> >>>>> I did try getting an answer and failed. The guy bringing it up is from a background of big dollar development and I don't know if there was some sort of IP ownership issue. I can't see how it would impact us even if there were. I really can't imagine anyone other than FTDI owns the drivers for the FTDI chips, but maybe there's some sort of JTAG software involved. >>>>> >>>>> Sometimes it is strange working with strangers. >>>> >>>> standard ftdi chips can talk jtag with the standard ftdi driver, you just have >>>> to make you own code to drive it and it won't work from inside the xilinx tools >>> >>> Your own code to drive it? That's pretty rad, but I think I understand what you are saying. >>> >> >> if you just want to configure a part it is pretty straight forward, but if >> you want to talk to a build in logic analyser, program one time fuses without ruining boards, and all the other things the tools can do I'm sure it can get pretty hairy > > I'm just trying to figure out what this guy was talking about saying that HP owned the software. I would expect either FTDI or Xilinx would own it. He never answered me when I asked why it mattered. I think he is trying to find some reason to nix the approach we are taking with the FPGA on this project. He's also seems to be digging to find things with the Gowin chip that he can object to. Or maybe I am just being paranoid. I guess I find some of his comments weird and don't know what to make of it. > > The trouble is we don't have a bunch of funding to pull off this project and the mechanical engineers have to spend a fair bit on the stainless steel cabinet. So we are trying to get the boards made for free by JLCPCB. We'll have to assemble the parts they don't carry, but they have offered to make some of our boards with the parts they do carry. That's pretty generous. I hope they stick to that when they see the updated board. It's 100x200 mm and will end up pretty full of parts. >
@Rick I can now confirm that a blank FTDI FT2232H chip will work with the Gowin tools and will program a GW1N-9 via JTAG, either SRAM or the on chip Flash. I used a bog standard FTDI module (about £18 from Farnell) and removed one resistor to disable the on board eeprom (may not have been necessary). I blogged a bit about the project here: https://www.element14.com/community/groups/fpga-group/blog/2020/10/29/another-cheap-and-simple-fpga-board There is a bit of explanation and some pictures. MK
On Thursday, October 29, 2020 at 1:41:13 PM UTC-4, Michael Kellett wrote:
> On 28/10/2020 03:48, Rick C wrote: > > On Tuesday, October 27, 2020 at 8:03:33 PM UTC-4, lasselangwad...@gmail=
.com wrote:
> >> onsdag den 28. oktober 2020 kl. 00.19.29 UTC+1 skrev Rick C: > >>> On Tuesday, October 27, 2020 at 5:43:10 PM UTC-4, lasselangwad...@gma=
il.com wrote:
> >>>> tirsdag den 27. oktober 2020 kl. 21.31.20 UTC+1 skrev Rick C: > >>>>> Someone new on the project is talking about how HP owns the driver =
code for the FTDI JTAG chip used to program Xilinx parts. Has anyone heard= of this?
> >>>>> > >>>>> Even if that is true, I'm not sure how relevant it is. Is there so=
me issue with HP owning the software for FTDI devices even if that is true?= Why would it matter to a user of Xilinx FPGAs???
> >>>>> > >>>>> I did try getting an answer and failed. The guy bringing it up is =
from a background of big dollar development and I don't know if there was s= ome sort of IP ownership issue. I can't see how it would impact us even if= there were. I really can't imagine anyone other than FTDI owns the driver= s for the FTDI chips, but maybe there's some sort of JTAG software involved= .
> >>>>> > >>>>> Sometimes it is strange working with strangers. > >>>> > >>>> standard ftdi chips can talk jtag with the standard ftdi driver, you=
just have
> >>>> to make you own code to drive it and it won't work from inside the x=
ilinx tools
> >>> > >>> Your own code to drive it? That's pretty rad, but I think I understa=
nd what you are saying.
> >>> > >> > >> if you just want to configure a part it is pretty straight forward, bu=
t if
> >> you want to talk to a build in logic analyser, program one time fuses =
without ruining boards, and all the other things the tools can do I'm sure = it can get pretty hairy
> >=20 > > I'm just trying to figure out what this guy was talking about saying th=
at HP owned the software. I would expect either FTDI or Xilinx would own i= t. He never answered me when I asked why it mattered. I think he is tryin= g to find some reason to nix the approach we are taking with the FPGA on th= is project. He's also seems to be digging to find things with the Gowin ch= ip that he can object to. Or maybe I am just being paranoid. I guess I fi= nd some of his comments weird and don't know what to make of it.
> >=20 > > The trouble is we don't have a bunch of funding to pull off this projec=
t and the mechanical engineers have to spend a fair bit on the stainless st= eel cabinet. So we are trying to get the boards made for free by JLCPCB. = We'll have to assemble the parts they don't carry, but they have offered to= make some of our boards with the parts they do carry. That's pretty gener= ous. I hope they stick to that when they see the updated board. It's 100x= 200 mm and will end up pretty full of parts.
> >=20 > @Rick >=20 > I can now confirm that a blank FTDI FT2232H chip will work with the=20 > Gowin tools and will program a GW1N-9 via JTAG, either SRAM or the on=20 > chip Flash. >=20 > I used a bog standard FTDI module (about =C2=A318 from Farnell) and remov=
ed=20
> one resistor to disable the on board eeprom (may not have been necessary)=
.
>=20 > I blogged a bit about the project here: >=20 > https://www.element14.com/community/groups/fpga-group/blog/2020/10/29/ano=
ther-cheap-and-simple-fpga-board
>=20 > There is a bit of explanation and some pictures. >=20 > MK
Great! I appreciate the heads up. I think most of the resistance to using= the Gowin devices is behind us, but was just asking about "ILA" which I su= ppose means the software logic analyzer. =20 We may have some long logic chains that will have a bit of trouble with 33 = MHz, but the in general we are using enables of 1 ms and slower. So timing= just should not be an issue except on the timing counters and the bits doi= ng some math in the MACs. =20 I do worry a bit about the poor documentation. This guy found a line that = said the dual ported rams don't support true dual port mode in this and a f= ew other parts. The contact said that's only in the parts prior to rev C. = The docs just aren't up to date or even very accurate. The guides say to = instantiate, but there is example code (in Verilog) in one that doesn't mat= ch the circuit diagram or even the port signals in the code provided by the= IP generator tool. =20 I ended up writing something pretty generic that should match the semi-dual= port which is what I need for this part of the design. It synthesized no = problem. I believe to use the -9 parts you need to select -9C. There are = some other chips that you specify the C revision too.=20 So far I'm very pleased that they have someone to answer my stupid question= s. He has to contact the mother ship for some of them, but I get answers..= . that I like. First of the year I would have been learning Verilog, lol.= =20 --=20 Rick C. -+ Get 1,000 miles of free Supercharging -+ Tesla referral code - https://ts.la/richard11209