W dniu 2013-01-27 15:20, Andreas Fritiofson pisze:
> On Sun, Jan 27, 2013 at 9:17 AM, Freddie Chopin <[email protected]
>     I guess it's official now - OpenOCD is NEVER* going to have SWD
>     support as:
>     a. no one is working on it
>
>
> Because of c. And because several people have already done a before c.

All of these are either hardware specific or a fork, only Tomek's 
solution tries to be generic...

> He's not dismissed, he's gotten some negative feedback and some
> complaints that his patches are hard to review (which I agree with).
> He's even got help with fixing up the patches so we can see more clearly
> what parts are good and what needs more work.

To be perfectly honest I also don't understand most of the comments he 
got, because these are either editorial trivia which is easily fixed 
(and this was/is/will be done) or just something like "it's wrong" - 
your last comment on gerrit being the best example. I don't know what's 
a step backward here? SWD support in general? Extending jtag_interface? 
Doing anything in ft2232.c? Using LibSWD?

>     c. OpenOCDs design does not fit anything other than JTAG, and no one's
>     working on changing that
>
>
> But at least I am, and I guess several others have brilliant ideas and
> opinions, it's just not going very fast. I'll never complete it on my
> own. And it wouldn't be great if I did. It's not a one man's job. But if
> we work together towards a common goal we should be able to do it. That
> means we have to be more organized. We need to define that common goal
> first. Anything we merge before that which is a hack, "proof-of-concept"
> or otherwise in the wrong direction will limit our freedom when we
> define that goal, because people tend to be nostalgic about code that is
> already merged. But as you have already noticed, OpenOCD may not have
> the momentum to maintain these kinds of discussions.

Hmm... It's the first time I hear that you would like some help with 
such redesign and the first time I hear about any redesign going on 
(note - yes, all devs are working on something, but "something" != 
"redesign or architecture"). Is there any post on the ML that outlines 
your ideas or calls for help (I'd be glad to help in the aspects that 
I'm capable of)?

Yes, the architecture of OpenOCD is broken, JTAG-centric or whatever you 
call it. There are people that would be willing to improve that - just 
need some guidance and a plan, which could be used to partition the job 
among many people.

>     * - in next 10 years, until a "perfect" architecture redesign will be
>     decided upon
>
>
> Yes, it'll probably take 10 years if no-one's willing to present and
> discuss ideas instead of using their time to blindly code towards their
> own goal.

Just try to be objective - there's the topic by Tomek, there are 
absolutely NO constructive messages there, right? Or did I miss some 
messages? I don't fully agree with Tomek's way of doing things (working 
in his own branch, then throwing that all into gerrit), but the "right 
approach" (discuss first, ...) doesn't seem to be working either...

>     BTW - sorry to ask - does "ST-link interface as a target" with it's
>     duplicated config files for targets is OK for this "nothing but perfect"
>     approach? I guess it's a hack too, with lot's of code duplication (like
>     RTOS does need to support "st-link" target, not only "cortex-m3"), yet
>     hundreds of OpenOCD users are happily using it (me included)... Adding
>     SWD would no doubt make that people miserable and furious.
>
>
> No-one is stopping anyone from using Tomek's patches, or anyone else's.
> That's the beauty of free software. If you just want SWD up and running
> so you can debug or program your SWD-only target, there are solutions
> out there to do it. I've no doubt that Tomek's work does the job
> sufficiently well. If you want to improve OpenOCD in the long run,
> there's no benefit to merging work in the wrong direction.

On the other hand it would be perfectly sound expectation to think that 
adding SWD would attract more users, more developers, more momentum, 
more anything... while NOT adding SWD would probably do nothing at best, 
or - even worse - push more ppl to other software options that DO 
support SWD, perfectly or not, and forget about OpenOCD for good.

> My
> problem with Tomek's patches is that while being presented as
> non-invasive patches to add SWD, they actually introduce significant
> architectural changes that seem to be part of a bigger plan (his coming
> "redesign" work) that we've had absolutely zero (chance of) discussion
> about.

I don't see any "invasion" there - just the interface-features, which is 
just one thing...

On the other hand - this series is too-invasive, so how can one expect 
to do any REAL architecture change?

Just as Tomek says - we can cooperate, there are people willing to help, 
but we need a plan (a plan is not "don't do this, don't do that, this is 
stupid, that is wrong") - what to change - and we need a roadmap - order 
of changes. And good communication (information other than "don't" and 
something constructive).

As I said - I could help with some tasks (that I'd be capable of) but 
this should be a group's effort to create a plan and partition it among 
volunteers.

If you (in general, not only Andreas) think that current "attitude" is 
right and works as a charm, please take a look at this site:
https://www.ohloh.net/p/openocd
and comment on just one fact:

> 12 Month Summary Jan 26 2012 — Jan 26 2013
> 413 Commits
> Down -205 (33%) from previous 12 months
> 59 Contributors
> Down -32 (35%) from previous 12 months

Does a 33% drop mean that current policies are good or bad for the 
project in long term?

Isn't it the case where we should do one step back to let some fresh air 
to the project, even at the price of going in "not-perfect-direction" 
for a moment? I'm not saying that SWD changes should be merged as-is, 
today, without review! I'm just saying that in case of something as 
important (essential!) for OpenOCD as SWD we should all cooperate to 
have it as soon as possible, even without proper architecture redesign, 
as it's past reasonable time for that to appear anyway... It's not a 
feature that will be used by 0.1% of users with some exotic tools and 
exotic chips - it's the new standard interface for the most popular ARM 
microcontrollers (Cortex-Mx) that should be supported by the most 
popular debug adapter for OpenOCD.

Cortex-M3 chips started to be popular when? Was it around 2008? 2009? So 
it's been at least 4 years and OpenOCD is starting to be no-good for 
today's chips... I also think that SWD has priority over any 
architecture stuff that can be worked around (cleanly), as users DON'T 
care for architecture, they only care whether it works or not. So - is 
OpenOCD being developed for users or for software perfection?

4\/3!!

------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnnow-d2d
_______________________________________________
OpenOCD-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openocd-devel

Reply via email to