Comment 96 for bug 625239

@Brian
This bug was split at about comment 30. We have been working on the case where the X crash happens some time after login because it has been started on tty1. Bilal's case start on tty7 is now being discussed in bug #626723. I feel a little bad that we aren't discussing his situation but that's how others split things up. I hope Bilal's not unhappy.

I don't think the current theory for Jamie and Seth's case of this bug necessarily depends on the nvidia driver being used. It is that the driver, whichever one it is, isn't ready to work with X as soon as the current gdm upstart job's preconditions are met and that breaks the assumptions made in the methods used to prevent tty1 from being fought over. I think a good case could be made for delaying gdm's start until X can be successfully run.

If you have the "start on tty1" situation would you mind attaching the output of ps aux gathered while you are still on tty1 (before the crash)? That would be a good start toward figuring out if you match the pattern we are currently discussing or if there is another case to be addressed.

By the way, we know there are two ways the X crash on Enter or 2 key can happen--X started on tty1 or tty7. There may again be several ways either of those might happen. In particular we don't know here for sure that Jamie and Seth are really the same and don't know the root cause of Dino problems.

------------------------------------------

@Dino
I can't speak for all of your fellow bug fighters, but I'm very happy that your problem is gone. I'm glad you are still with us. Since a race condition is involved, I still really don't know whether the problem you were experiencing was the same thing as Jamie's or not. I can't challenge your assumptions in comment 75. The way this is supposed to work is quite complicated and the config files must be correct to avoid problems.

I hope you will again speak up about your situation if the problem comes back, but I hope it doesn't. I was having the tty1 problem with pre-release lucid last Spring until the Plymouth megafix and it's a real nuisance (I had the tty7 problem sometime during testing, too) my tty1 cases were intermittent, where the problem would happen during one boot and everything would work the next--with no change at all. I spent more than a week researching what was happening until the devs were able to fix it.

Thanks for sticking with us and for continuing to contribute (and for your patient contributions above).
-----------------------------------------------

@Seth
Do you think your xorg.conf file could be involved in the problem?