As far as JTAG interfaces on development boards, we need to develop
on our target product, a complicated piece.
Our other engineers also have problems from time to time with the software
drivers.
Luckily we haven't broken any hardware lately. It's always drivers
and on windoze XP or 7 of some sort.
We tend to use these entire parts... Most of the peripherals and a lot of memory
and pretty much all out speed wise.
This may be one reason why I was evidently the first (and only ?) one in north
America to find a silicon bug in the old ATmega 32 Atmel chip which almost put
my last company out of business.
I tend to bring out the worst in products I guess.
boB
--- In l..., FreeRTOS Info wrote: >
> >
> > It amazes me that this stuff works at all. No wonder it is buggy as heck
> > and crashes so much for unknown reasons.
> >
> > I would like to turn these unknown reasons into known reasons so they
> > can be dealt with.
> >
> > I don't mind having bugs as long as I know WHY and/or what is
happening.
> >
> > I would rather not to have to re-install and re-boot my computer once
> > per week if I can have some other way of fixing the problem. To do that,
> > I have to have a better understanding of the problem, usually.
> > Right now, I would settle for a work around or a fix but nobody seems to
> > understand the WHY of the problem.
> >
> > I fear that technology is taking a step backwards these days.
> I have to say my experience with Segger interfaces and drivers differs
> from yours. I 'officially' support 18 different tools chains, and
who
> knows how many different development boards. Many many, different USB
> interface with their drivers installed, etc. etc. You get the picture,
> there is a *lot* of mixed up stuff on my computer.
>
> There seems to be a trend towards having mini J-Links like interfaces
> build directly onto development boards now too - so the variety of board
> I connect to one way or another through Segger software is large - and I
> can honestly say I have never had a problem, that I can recall anyway.
> I would go further and say I am pleased when I see a Segger interface on
> a board because I have confidence that there is one less thing that is
> gong to give me hassle.
>
> I do find, as I switch from board to board, the emulator firmware
> version gets re-programmed sometimes, but it just does it all itself. I
> even have some old KickStart J-links from probably 8 years ago that
> still work, and used beta drivers on chips like the RX, and they worked
too.
>
> Compare that to my experience with things like the numerous FTDI
> interfaces versions, which show no DLL mercy, well....
>
> Regards,
> Richard.
>
> It amazes me that this stuff works at all. No wonder it is buggy as heck
> and crashes so much for unknown reasons.
>
> I would like to turn these unknown reasons into known reasons so they
> can be dealt with.
>
> I don't mind having bugs as long as I know WHY and/or what is
happening.
>
> I would rather not to have to re-install and re-boot my computer once
> per week if I can have some other way of fixing the problem. To do that,
> I have to have a better understanding of the problem, usually.
> Right now, I would settle for a work around or a fix but nobody seems to
> understand the WHY of the problem.
>
> I fear that technology is taking a step backwards these days. I have to say my experience with Segger interfaces and drivers differs
from yours. I 'officially' support 18 different tools chains, and
who
knows how many different development boards. Many many, different USB
interface with their drivers installed, etc. etc. You get the picture,
there is a *lot* of mixed up stuff on my computer.
There seems to be a trend towards having mini J-Links like interfaces
build directly onto development boards now too - so the variety of board
I connect to one way or another through Segger software is large - and I
can honestly say I have never had a problem, that I can recall anyway.
I would go further and say I am pleased when I see a Segger interface on
a board because I have confidence that there is one less thing that is
gong to give me hassle.
I do find, as I switch from board to board, the emulator firmware
version gets re-programmed sometimes, but it just does it all itself. I
even have some old KickStart J-links from probably 8 years ago that
still work, and used beta drivers on chips like the RX, and they worked too.
Compare that to my experience with things like the numerous FTDI
interfaces versions, which show no DLL mercy, well....
Regards,
Richard.
+ http://www.FreeRTOS.org
Designed for microcontrollers. More than 7000 downloads per month.
>> Rolf and his team put a lot of effort into J-Links
and their other
> software and I assume they are equally proud and protective of their
> creation as anybody else. The correct way to deal with this is for both
> sides to understand why things are the way they are, and that is really
> hard in many cases: broken USB chipsets, dicky USB hardware and hubs,
> recalcitrant target systems.
>>
>> We've found some fixes are as simple as replacing a USB cable. And,
of
> course, it helps if customers turns the target board on! :-) Power
> supplies are the number one cause of much frustration. I'll say it
> again: when we work through an issue with a customer only for them to
> say "sorry, I had a bad power supply, replaced it, and it's all OK
now",
> it makes us fume. Interestingly, if you buy a debug interface from Microchip (RealICE,
ICD3, etc.) they apparently warrant it for life. If it breaks, you just
send it back and they send another one - even if it broke because you
did something daft with it. I have myself sent stuff back, and received
replacements (albeit referbs, rather than brand new), so there is some
substance to the promise.
Regards,
Richard.
+ http://www.FreeRTOS.org
Designed for microcontrollers. More than 7000 downloads per month.
--- In l..., Paul Curtis wrote: > On 10 May 2012, at 23:47, boB G wrote:
>
> >
> > Rolf should also have known that non students can NOT purchase their less
expensive educational version of the programmer. I gather that Rolf is not an
owner of the Segger company.
> > You know, I think he is not *an* owner, he is *the* owner. :-)
>
> -- Paul.
>
Yes, I just now found that out !!
I know what the problem is I think... He is just as frustrated as a lot of
other companies out there, and so are we somewhat, that knowledgeable engineers
and techs are almost impossible to find.
What has probably happened is that whoever did write the drivers and interface
to the IAR C-Spy and debugger has gone somewhere else and there may not be
anybody there that knows how it all works working at the company anymore....
Unless Rolf wrote this himself which I find hard to believe with such a hugely
involved product. If so, he ain't thinking about this particular problem.
Very possibly too many little problems and not enough engineers to fix them.
I know that we certainly can't find anybody that knows anything about
windows drivers around here in the pacific northwest. If they worked for
Microsoft, they've probably retired.
It amazes me that this stuff works at all. No wonder it is buggy as heck and
crashes so much for unknown reasons.
I would like to turn these unknown reasons into known reasons so they can be
dealt with.
I don't mind having bugs as long as I know WHY and/or what is happening.
I would rather not to have to re-install and re-boot my computer once per week
if I can have some other way of fixing the problem. To do that, I have to have
a better understanding of the problem, usually.
Right now, I would settle for a work around or a fix but nobody seems to
understand the WHY of the problem.
I fear that technology is taking a step backwards these days.
boB
Reply by Paul Curtis●May 10, 20122012-05-10
On 10 May 2012, at 23:47, boB G wrote:
>
> Rolf should also have known that non students can NOT purchase their less
expensive educational version of the programmer. I gather that Rolf is not an
owner of the Segger company.
>
You know, I think he is not *an* owner, he is *the* owner. :-)
-- Paul.
Reply by boB G●May 10, 20122012-05-10
--- In l..., Paul Curtis wrote: >
>
> On 10 May 2012, at 23:04, boB G wrote:
>
> >
> > Thought I would mention that Segger has banned me from logging onto their
web site and forum for asking about alternate J-Link solutions.
> >
>
> Defendable if you had stepped beyond the AUP of using their resource. ;-)
>
> > I also received an email from them (Rolf) directly asking why I would want
to purchase a cheap solution when I make so much money programming for a
living.
> >
>
> This doesn't seem an unreasonable question; at least Rolf is engaging
with you now, right from the top.
>
> > I had to tell them about how I like open source solutions and companies with
support, like great forums for products like NXP LPC processors and how I just
started receiving a small salary after 4 years of not receiving any salary and
putting ALL of my money into our company, etc.
> >
> > They just don't understand I guess.
> >
>
> Rolf and his team put a lot of effort into J-Links and their other software
and I assume they are equally proud and protective of their creation as anybody
else. The correct way to deal with this is for both sides to understand why
things are the way they are, and that is really hard in many cases: broken USB
chipsets, dicky USB hardware and hubs, recalcitrant target systems.
>
> We've found some fixes are as simple as replacing a USB cable. And, of
course, it helps if customers turns the target board on! :-) Power supplies
are the number one cause of much frustration. I'll say it again: when we
work through an issue with a customer only for them to say "sorry, I had a bad
power supply, replaced it, and it's all OK now", it makes us fume.
>
> -- Paul.
>
Oh, I KNOW they've put a lot of work into those !
In this case, nobody could make an attempt to help me with
my problem of obviously corrupted drivers. It wasn't a problem with the
hardware in this case... Just drivers gone awry about once per week.
I sent a nice reply to them telling them exactly the problem I had with their
"support" forum. Or lack of support in this case.
I also gave them an idea on a product improvement... To add a
USB isolation circuit to their programmer. We use several of their
$800 (USD) black programmers (same thing as a J-Link with different software
evidently), and they would break some times until we started
using an isolator from PC to the product being programmed.
And Rolf made the assumption that our company was big and making lots of money
and I had to tell him of course that I don't make much if any money (yet)
and have given the last 4 years of my own finances into our company. Maybe some
day it will pay off.
Oh, I also mentioned that competition is good, which it is.
I'm tired of being grabbed by the B*11s by large companies, IAR included.
Xilinx too ! We've had problems with their CPLD JTAG programming cables
and they do not warranty their units so we JUST this week took 5 of our
programmers and finally fixed them. At least they are somewhat fixable.
They're $250 a pop when they break.
Rolf should also have known that non students can NOT purchase their less
expensive educational version of the programmer. I gather that Rolf is not an
owner of the Segger company.
Thanks,
boB
Reply by Paul Curtis●May 10, 20122012-05-10
On 10 May 2012, at 23:04, boB G wrote:
>
> Thought I would mention that Segger has banned me from logging onto their web
site and forum for asking about alternate J-Link solutions.
>
Defendable if you had stepped beyond the AUP of using their resource. ;-)
> I also received an email from them (Rolf) directly
asking why I would want to purchase a cheap solution when I make so much money
programming for a living.
>
This doesn't seem an unreasonable question; at least Rolf is engaging with
you now, right from the top.
> I had to tell them about how I like open source
solutions and companies with support, like great forums for products like NXP
LPC processors and how I just started receiving a small salary after 4 years of
not receiving any salary and putting ALL of my money into our company, etc.
>
> They just don't understand I guess.
>
Rolf and his team put a lot of effort into J-Links and their other software and
I assume they are equally proud and protective of their creation as anybody
else. The correct way to deal with this is for both sides to understand why
things are the way they are, and that is really hard in many cases: broken USB
chipsets, dicky USB hardware and hubs, recalcitrant target systems.
We've found some fixes are as simple as replacing a USB cable. And, of
course, it helps if customers turns the target board on! :-) Power supplies
are the number one cause of much frustration. I'll say it again: when we
work through an issue with a customer only for them to say "sorry, I had a bad
power supply, replaced it, and it's all OK now", it makes us fume.
-- Paul.
Reply by boB G●May 10, 20122012-05-10
Thought I would mention that Segger has banned me from logging onto their web
site and forum for asking about alternate J-Link solutions.
I also received an email from them (Rolf) directly asking why I would want to
purchase a cheap solution when I make so much money programming for a
living....
I had to tell them about how I like open source solutions and companies with
support, like great forums for products like NXP LPC processors and how I just
started receiving a small salary after 4 years of not receiving any salary and
putting ALL of my money into our company, etc.
They just don't understand I guess.
boB
--- In l..., Olivier Gautherot wrote: >
> Hi Xiaofan,
>
> Thanks for the good news - I'm also using the version 0.6.0 (or a beta
> of it).
>
> I'm still struggling to get a reliable download on the LPC1754 (I'm
> still looking for the correct sequence of commands) so I can't give
> precise results yet.
>
> Cheers
> Olivier
> Xiaofan Chen wrote:
> >
> > On Thu, May 10, 2012 at 9:14 AM, Olivier Gautherot
> > > wrote:
> > > Bob,
> > >
> > > note that the download speed of a professional JTAG probe is better
that
> > > OpenOCD: people report a stable use of OpenOCD at up to 500kHz (and
> > > others as low as 10kHz but I seem to be on the lucky side). I assume
> > > professional tools support up to several MHz and enhance the debugging
> > > experience, depending on the type of app that you're debugging. As
the
> > > probe costs only $79 at Mouser, you won't loose much if you give it
> > a try.
> >
> > Please also take note that OpenOCD has made quite some progress
> > over the past years and now it is near to 0.6.0 release.
> >
> > One result I got from recent git version of OpenOCD with
> > TI/Luminary EK-LM3S1968.
> >
> > mymacmini:lm3s1968 xiaofanc$ openocd -f board/ek-lm3s1968.cfg
> > Open On-Chip Debugger 0.6.0-dev-00543-g908ee4d (2012-05-05-20:28)
> > Licensed under GNU GPL v2
> > For bug reports, read
> > http://openocd.sourceforge.net/doc/doxygen/bugs.html
> > Info : only one transport option; autoselect 'jtag'
> > 500 kHz
> > Info : clock speed 500 kHz
> > Info : lm3s1968.cpu: hardware has 6 breakpoints, 4 watchpoints
> > Info : accepting 'telnet' connection from 4444
> > target state: halted
> > target halted due to debug-request, current mode: Thread
> > xPSR: 0x01000000 pc: 0x000011ae msp: 0x200000f0
> > flash 'stellaris' found at 0x00000000
> > 500 kHz
> > auto erase enabled
> > wrote 262144 bytes from file demo4.bin in 8.341610s (30.690 KiB/s)
> > 1000 kHz
> > auto erase enabled
> > wrote 262144 bytes from file demo4.bin in 5.227277s (48.974 KiB/s)
> > 2000 kHz
> > auto erase enabled
> > wrote 262144 bytes from file demo4.bin in 4.353256s (58.807 KiB/s)
> > 3000 kHz
> > auto erase enabled
> > wrote 262144 bytes from file demo4.bin in 4.346641s (58.896 KiB/s)
> >
> > You can see it is much better than last time's result: the flash
write
> > was never above 18KB/sec for LM3S1968 for my test done about
> > a year ago.
> > http://comments.gmane.org/gmane.comp.debugging.openocd.devel/17877
> >
> > --
> > Xiaofan
> >
> >
> --
>
> Olivier Gautherot
> *Email:* olivier@...
> *Cel:* +56 98 730 9361
> *Web:* www.gautherot.net
> *LinkedIn:* http://www.linkedin.com/in/ogautherot
>
>
>
Reply by Olivier Gautherot●May 10, 20122012-05-10
Hi Xiaofan,
Thanks for the good news - I'm also using the version 0.6.0 (or a beta
of it).
I'm still struggling to get a reliable download on the LPC1754 (I'm
still looking for the correct sequence of commands) so I can't give
precise results yet.
Cheers
Olivier
Xiaofan Chen wrote: >
> On Thu, May 10, 2012 at 9:14 AM, Olivier Gautherot
> > wrote:
> > Bob,
> >
> > note that the download speed of a professional JTAG probe is better that
> > OpenOCD: people report a stable use of OpenOCD at up to 500kHz (and
> > others as low as 10kHz but I seem to be on the lucky side). I assume
> > professional tools support up to several MHz and enhance the debugging
> > experience, depending on the type of app that you're debugging. As
the
> > probe costs only $79 at Mouser, you won't loose much if you give it
> a try.
>
> Please also take note that OpenOCD has made quite some progress
> over the past years and now it is near to 0.6.0 release.
>
> One result I got from recent git version of OpenOCD with
> TI/Luminary EK-LM3S1968.
>
> mymacmini:lm3s1968 xiaofanc$ openocd -f board/ek-lm3s1968.cfg
> Open On-Chip Debugger 0.6.0-dev-00543-g908ee4d (2012-05-05-20:28)
> Licensed under GNU GPL v2
> For bug reports, read
> http://openocd.sourceforge.net/doc/doxygen/bugs.html
> Info : only one transport option; autoselect 'jtag'
> 500 kHz
> Info : clock speed 500 kHz
> Info : lm3s1968.cpu: hardware has 6 breakpoints, 4 watchpoints
> Info : accepting 'telnet' connection from 4444
> target state: halted
> target halted due to debug-request, current mode: Thread
> xPSR: 0x01000000 pc: 0x000011ae msp: 0x200000f0
> flash 'stellaris' found at 0x00000000
> 500 kHz
> auto erase enabled
> wrote 262144 bytes from file demo4.bin in 8.341610s (30.690 KiB/s)
> 1000 kHz
> auto erase enabled
> wrote 262144 bytes from file demo4.bin in 5.227277s (48.974 KiB/s)
> 2000 kHz
> auto erase enabled
> wrote 262144 bytes from file demo4.bin in 4.353256s (58.807 KiB/s)
> 3000 kHz
> auto erase enabled
> wrote 262144 bytes from file demo4.bin in 4.346641s (58.896 KiB/s)
>
> You can see it is much better than last time's result: the flash write
> was never above 18KB/sec for LM3S1968 for my test done about
> a year ago.
> http://comments.gmane.org/gmane.comp.debugging.openocd.devel/17877
>
> --
> Xiaofan --
On Thu, May 10, 2012 at 9:14 AM, Olivier Gautherot
wrote: > Bob,
>
> note that the download speed of a professional JTAG probe is better that
> OpenOCD: people report a stable use of OpenOCD at up to 500kHz (and
> others as low as 10kHz but I seem to be on the lucky side). I assume
> professional tools support up to several MHz and enhance the debugging
> experience, depending on the type of app that you're debugging. As the
> probe costs only $79 at Mouser, you won't loose much if you give it a
try.
Please also take note that OpenOCD has made quite some progress
over the past years and now it is near to 0.6.0 release.
One result I got from recent git version of OpenOCD with
TI/Luminary EK-LM3S1968.
mymacmini:lm3s1968 xiaofanc$ openocd -f board/ek-lm3s1968.cfg
Open On-Chip Debugger 0.6.0-dev-00543-g908ee4d (2012-05-05-20:28)
Licensed under GNU GPL v2
For bug reports, read http://openocd.sourceforge.net/doc/doxygen/bugs.html
Info : only one transport option; autoselect 'jtag'
500 kHz
Info : clock speed 500 kHz
Info : lm3s1968.cpu: hardware has 6 breakpoints, 4 watchpoints
Info : accepting 'telnet' connection from 4444
target state: halted
target halted due to debug-request, current mode: Thread
xPSR: 0x01000000 pc: 0x000011ae msp: 0x200000f0
flash 'stellaris' found at 0x00000000
500 kHz
auto erase enabled
wrote 262144 bytes from file demo4.bin in 8.341610s (30.690 KiB/s)
1000 kHz
auto erase enabled
wrote 262144 bytes from file demo4.bin in 5.227277s (48.974 KiB/s)
2000 kHz
auto erase enabled
wrote 262144 bytes from file demo4.bin in 4.353256s (58.807 KiB/s)
3000 kHz
auto erase enabled
wrote 262144 bytes from file demo4.bin in 4.346641s (58.896 KiB/s)