02:24 airlied: dliviu: added you to it
07:05 tzimmermann: jfalempe, hi. do you follow the discussion about buggy vblank irqs on the matrox g200se?
07:43 tzimmermann: vsyrjala, i'm looking again at patch 4/8 of your crtc->mode series. you mention that the code holds crtc->mutex. i can't seem to locate where we take and release that lock. do you have a pointer?
09:50 jfalempe: tzimmermann: Yes, I followed the discussion about buggy vblank. I had the same timeout issue when testing DMA on different servers. That's one of the reasons I've abandonned the series.
10:28 tzimmermann: jfalempe, yeah. i saw your reply. thanks a lot. i guess i simply drop the vblank patch.
10:31 jfalempe: tzimmermann: yes, unfortunately this hw is a bit too broken :/
10:33 tzimmermann: jfalempe, in its heydeys in the 90s, matrox cards had a high reputation. now that i've come to maintain their driver, my opinion has changed somewhat :D
10:35 tzimmermann: i'll wait of tony sends out another response to my updated patch and i'll probably send out a revert next week
10:35 tzimmermann: s/of/for
10:35 tzimmermann: s/of/if
10:37 jfalempe: ok that's fine.
10:59 cousin_luigi: tzimmermann: What is it they do these days?
11:02 jfalempe: cousin_luigi: the matrox G200 cards are still used in servers, for remote control. That's why we're still maintaining the driver.
14:46 gfxstrand: dliviu: Done
14:48 gfxstrand: anarsoul: I just plugged a 1650 into my desktop and I'm CTSing it. It's a pretty new one so it may not match older 1650s. IDK if they've changed or not. In either case, we'll see what it does.
16:43 lyndonthe3: so as to what happens to the most of the combinations (i actually told you that), has mathematical proof that i am not interested to riddle out (it is too long proof), but it's just collisions very highly likely, cause 1+4 is same as 2+3, so there is only some number of results which internal adder can yield into which are not canceled as duplicates. as long as the decoder and intrinsics
16:43 lyndonthe3: can understand and manipulate the real IO value it's all fine.
17:50 ElementW: I'm using a FreeSync display at 96Hz (since I don't notice the difference much above that) and I've noticed switching from Linux 6.6 to 6.9 removes the 96Hz from the modelist, as visible in KDE's system settings and using `drm_info`, is this normal?
20:08 ccr:looks around.
22:43 dliviu: gfxstrand: thanks
22:45 dliviu: airlied: thanks
23:54 kode54: why is that project private?
23:54 kode54: does it contain NDA'd documentation?