Page 6 of 6 FirstFirst ... 456
Results 51 to 56 of 56

Thread: Canonical's Mir Project Retracts Wayland Criticism

  1. #51
    Join Date
    May 2007
    Location
    Nurnberg.
    Posts
    313

    Default

    Quote Originally Posted by YoungManKlaus View Post
    so, classical NIH at its finest
    The worst thing is, some of those shouting loudest right now are the same people who sided with ATI, against AMD and SuSE, 5 years ago. They did a _huge_ NIH on the RadeonHD driver, with reduced open source goals and they were much less technically advanced. The fglrx driver is still going very strong today because of what they pulled. Yet still people listen to them when they now cry foul few years later, and shout along. It is massively hypocritical.

  2. #52
    Join Date
    Jul 2009
    Posts
    261

    Default

    Quote Originally Posted by libv View Post
    The worst thing is, some of those shouting loudest right now are the same people who sided with ATI, against AMD and SuSE, 5 years ago. They did a _huge_ NIH on the RadeonHD driver, with reduced open source goals and they were much less technically advanced. The fglrx driver is still going very strong today because of what they pulled. Yet still people listen to them when they now cry foul few years later, and shout along. It is massively hypocritical.
    radeonhd itself being a massive NIH of avivo, driven 100% by corporate/business reasons.

  3. #53
    Join Date
    May 2007
    Location
    Nurnberg.
    Posts
    313

    Default

    Quote Originally Posted by daniels View Post
    radeonhd itself being a massive NIH of avivo, driven 100% by corporate/business reasons.
    Really?

    And this by the guy who said that he had an impossible time to set up basic display stuff as it would hardlock at every corner. My view: the avivo hardware was really quite nice and somewhat resilient. I believe i would've managed quite a lot more out of that display engine (with the massive amount of display insight and knowledge that i had at the time), even without the _extremely_ low level information ATI provided (which ATI hoped would stop us in our tracks).

    But honestly, partly due to ATI, partly due to AMD, we were just not allowed to talk to others, and were not allowed to contribute to anything else before the big announcement by AMD. We at SuSE were not even told when that announcement was, we kind of guessed and aimed for september, but we were never told until it happened. That was 2-3 months of possibly disjointed development, for very little gain. As the stuff that Avivo did provide was actually pretty minimal.

    I am not saying that avivo was not important. Avivo was massively important in showing the world that a free driver for ATI was needed. But from an actual driver point of view it only achieved some limited things that could easily be gleaned by reading the 32bit registers directly. There was little that would warrant continued code re-use, and a continuation of the avivo driver would in effect only retain the name.

    Then... You yourself chose to use the GPL, which is not too acceptable for an open source graphics driver. We at SuSE were not allowed to try and talk to you guys to go and change it. If you had chosen to use MIT however...

    All in all, the corporate issues with respect to the avivo driver were definitely not on the SuSE side. And they definitely do not provide any justification to side with ATI on a "divide and conquer" tour which reduced the open source story to a figleaf. And _you_ actively contributed to that my friend, all the way down to taking part in hacking a repo of a long dead project, abusing fd.o root, and then not owning up for 3 days with the security of fd.o in limbo.

    Edit:

    You and some others NIHed, for no other reason than trumping us at SuSE, and being able to make the most noise. There is simply no other sustainable explanation.
    Last edited by libv; 03-06-2013 at 09:03 PM.

  4. #54
    Join Date
    Jul 2009
    Posts
    261

    Default

    Quote Originally Posted by libv View Post
    Really?

    And this by the guy who said that he had an impossible time to set up basic display stuff as it would hardlock at every corner.
    Your employer bought you the specs. I reverse-engineered it in my spare time. Bit of a difference there. And I never said impossible, not least because I actually did it. Sometimes remotely, as I couldn't afford to buy a MacBook out of my own pocket - no-one was buying me hardware. Incidentally, until you switched to ATOM, Avivo worked on a bunch of MacBooks whereas RadeonHD didn't at all. I guess that's just dumb register bashing that any idiot could do, though.

    (Also, yes, it did deadlock whilst busmastering if you set some parameters up incorrectly. Which is pretty easy to do when you have to guess what the register's actually for. It's much easier with specs and already-working code, but you already know that.)

    Quote Originally Posted by libv View Post
    Then... You yourself chose to use the GPL, which is not too acceptable for an open source graphics driver. We at SuSE were not allowed to try and talk to you guys to go and change it. If you had chosen to use MIT however...
    We said we'd happily relicense it if people wanted it to be relicensed. Whatever your reasons may be, you never asked, so you don't get to complain about it. I'm also not really sure what your specific problem was with the GPL, but that's neither here nor there: it could easily have been MIT.

    Quote Originally Posted by libv View Post
    You and some others NIHed, for no other reason than trumping us at SuSE, and being able to make the most noise. There is simply no other sustainable explanation.
    We developed in the open, with our code public, while you worked in secret and didn't tell anyone what you were doing. And we started first. Even by your standards of bending reality to match your wounded ego, this is nonsensical.

  5. #55
    Join Date
    May 2007
    Location
    Nurnberg.
    Posts
    313

    Default

    Quote Originally Posted by daniels View Post
    Your employer bought you the specs. I reverse-engineered it in my spare time. Bit of a difference there. And I never said impossible, not least because I actually did it. Sometimes remotely, as I couldn't afford to buy a MacBook out of my own pocket - no-one was buying me hardware. Incidentally, until you switched to ATOM, Avivo worked on a bunch of MacBooks whereas RadeonHD didn't at all. I guess that's just dumb register bashing that any idiot could do, though.

    (Also, yes, it did deadlock whilst busmastering if you set some parameters up incorrectly. Which is pretty easy to do when you have to guess what the register's actually for. It's much easier with specs and already-working code, but you already know that.)
    Did you read the proposal we at SuSE sent to AMD? We at SuSE made the open specs happen, in as far as ATI wanted to play along (not much, and they stopped when radeonhd died, and only the AMD GPGPU people continued). We also only took those specs when they were bound to become free, we very strictly made sure that we never got tainted. Sadly though, ATI did not keep its side of the bargain, and some later rv6xx and r7xx low level register stuff never got freed.

    Quote Originally Posted by daniels View Post
    We said we'd happily relicense it if people wanted it to be relicensed. Whatever your reasons may be, you never asked, so you don't get to complain about it. I'm also not really sure what your specific problem was with the GPL, but that's neither here nor there: it could easily have been MIT.


    We developed in the open, with our code public, while you worked in secret and didn't tell anyone what you were doing. And we started first. Even by your standards of bending reality to match your wounded ego, this is nonsensical.
    That one is on ATI/AMD.

    We actually got badly rapped when a KDE guy, 2 doors down from us, posted some rumours about the something upcoming at AMD/ATI. Turns out, that this guy just saw a huge pile of graphics cards in the room with the X geeks and thought nothing of it. The "rumour" was instead spread by someone from AMD at Akademy 2007. This was how this thing worked, and we at SuSE already had achieved loads with respect to openness, we just had to keep things under wraps until AMD could make their big noise (and we were never even told when to expect this).

    All of the above explains why we couldn't use the avivo shell.

    It does not however warrant a full and very aggressive fork and NIH, or all the shitthrowing contests afterwards. This NIH was about as damaging to the free ATI/AMD driver as AMDs lack of control over ATI and the financial crisis was.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •