Re: [Tails-testers] Nvidia Pascal no longer loads gdm

If I may toss some unsolicited thoughts regarding what to do into the mix... I'm not sure how deep the EGL issues creep into the mix, but I also was unable to load Tails with a GTX 610 without modesetting (didn't get to try it on the 710 since I recently got rid of it). With that being such an old card (and should in theory have decent support), it does give me some pause. EGL seems fairly baked into the drivers these days (since it's apparently integrated directly into newer Xorg versions now), so it seems like it may hurt more devices than it helps by having "updated" drivers. Ultimately, I'm of the mindset that until Xorg gets updated, the Nouveau drivers should probably stay as they are (unless there is a security issue, of course). You'd have a better gauge at how many people are having issues with nvidia drivers, since you're receiving the reports. Now that I know modesetting works, ultimately I have no self-interest in choosing a side, since that'll be what I use anyways, but from a standpoint of "how many issues does it fix vs how many issues it causes", I err on the side of "I want Tails easily accessible to as many people as possible". I don't maintain my own distro. I did used to bake my own custom kernels with custom patchsets, have built my own custom "hybrid" distros for personal use, and know that making a decision for a mass-user product is much harder than a solo endeavor, so my thoughts do have limited value. But I gave it anyways, and I hope that you come to a decision you're comfortable with. Regardless, again, I thank you for your time and was glad I could help in identifying the issue.30.08.2018, 00:32, "Soul Null" <soulnull@???>:You're very welcome. Anything you think I may be of assistance with in the future, please don't hesitate to get in touch.30.08.2018, 00:30, "intrigeri" <intrigeri@???>:Soul Null: Here are the results: * No changes to kernel command line - Tails starts as 3.8 did. Seems to fall back to modesetting driver, ESID detected, no vsync, etc.Thanks! This confirms that the regression you're experiencing with3.9~rc1 is solely caused by the nouveau driver upgrade.I'll take a step back before I decide what seems best for 3.9 final.