3.5-rc6 fb hw inteldrmfb

As said, with 3.5-rc6 with vga=791 , on someone else's computer with intel grafics , it sticks/crashs within 5 (or even 3) seconds after start booting, long time before trying to switch into grafic mode. That happened today again, like yesterday. In the logfiles yesterday I found nothing.

Now I tried also the same, 3.5-rc6 but with vga=0 . Then the booting process goes forward, but later, at the time it tries to switch into the grafic mode, it sticks in the same way. Reboot with CTRL+ALT+C is not possible, one have to press the reboot button.

With 3.5-rc1 and vga=791 (what on my own computer with nv grafic card runs normally), on that said computer with intel grafic card, it don't stop in early stage, but continues, but at the time of switching in grafic mode, the screen stays black. However, now it's possible to reboot wich CTRL+ALT+C , so I suppose that different to 3.5-rc6 the computer didn't really stick/crash (but the grafic stays just invisible)

With 3.4-rc5 that computer runs normal.

I forget to mention, that on my own computer, what works almost normally with 3.5-rc6 , rsync don't return (i.e. after type rsync in text mode, it continues eternally, neither can be stopped with CTRL-C). That didn't happen with a next laptop of someone else where I installed 3.5-rc6 yesterday (which has ati grafics). However, I suppose that has nothing to do with the current problem, but with some other hardware.

Enclosed are the new logfiles, of today, of the trying out 's explained above, of that computer which sticks/crashs with 3.5-rc6. I send you that quickly but I don't have tiome now to inspect them; I suppose there is again nothing from the 3.5-rc6 vga=791 run which crashs in early stage of booting, but perhaps there is something from 3.5-rc6 vga=0 which crashs only later (at the time switching to grafics) as explained above. Added is also, for comparison, dmesg of the run with 3.4-rc5 what's without that grafic problem.

wl

I've looked a bit at the logfiles and dmesg contains tons of stuff that I didn't expect to be there and the other's contain tons of unrelated stuff (e.g. they contain a few days worth of data). So I couldn't figure out anything with that.
The difference betwen rc5 and rc6 is simply due to the firmware kicking, but since rc5 dies in a similar fashion (but simply later) I think the bug has been introduced earlier.

I guess we're down to a bisect, since 3.4-rc5 seems to work for you. Can you please to that?

Thanks, Daniel

Daniel Vetter
[email protected] - +41 (0) 79 365 57 48 - http://blog.ffwll.ch