[feisty]Screen corruption when usplash is enable

[Florent Mertens]

Binary package hint: usplash

When booting with the splash option, i get a corrupted screen (see screenshots).
Restarting X solve the issue, but it's begin to be quite anoying to restarting X each time i boot.
This also seems to happens only when DefaultDepth is set to 24 in xorg.conf.

System :
laptop inspiron 510m
Graphical card : intel 855GM
Feisty up to date

[didier (did447-deactivatedaccount)]

Hi,
Can you post your xorg.conf?
And does switching VT solve the issue? CTRL+ALT+F1 ALT+F7

[Florent Mertens (givre)]

switching VT do also the trick (that's much better than restarting X ;) ).

I attach you my xorg.conf.

One precision : i force vga=773 to get a good usplash resolution (1024*768). Without that, screen is not corrupt, but boot splash is awful.

[didier (did447-deactivatedaccount)]

I'm sure it's a duplicate but I can't find it. Maybe it should be in Xorg, it's not an i810 driver issue, I have the same thing with an ATI on ppc.

[Florent Mertens (givre)]

It seams that this bug is due to :

usplash (0.4-35) feisty; urgency=low

  * Add the new console-setup initramfs hook as a prerequisite rather than
    running setupcon after shutting down usplash.

 -- Colin Watson <email address hidden> Mon, 22 Jan 2007 14:27:46 +0000

In the usplash init script, the only change since edgy is that setupcon is not run after usplash. Adding it like in the edgy init script solve my problem.

Could you try that ?

[Colin Watson (cjwatson)]

Thanks for the useful lead, Florent. While that does fix it, it's not the real problem. The real problem is that unless you set CONSOLE_SCREEN to something then the usplash init script doesn't shut down usplash at all! I've fixed this in bzr now, and am just chasing up some related problems in gdm's init script uncovered by that change.

[Colin Watson (cjwatson)]

usplash (0.4-43) feisty; urgency=low

  [ Ben Collins ]
  * Fix two serious bugs in handle_input:
    - It was free()ing a stack allocated char array.
    - Initialize fifo_outfd to -1 so we can conditionally close() it only
      when it was valid. In the INPUTENTER path, it would close() on an
      uninitialized value in fifo_outfd, causing all sorts of unpredictable
      problems.
  * Do not skip over writing to fifo_outfd in handle_input for INPUTENTER.
    Because of this, users of INPUTENTER could not block on the outfd to see
    when enter was actually pressed by the user.

  [ Colin Watson ]
  * Fix the init script to shut down usplash even if CONSOLE_SCREEN is empty
    (LP: #86606).

 -- Colin Watson <email address hidden> Fri, 23 Feb 2007 12:31:39 +0000