[hardy] esc key during fsck check causes usplash to break into text mode

[KrazyPenguin]

During a pre-determined number of boots, Hardy will do a fsck disk check.
This is now done in graphical mode, and the user is given the option to use the [Esc]
key to skip the check and continue booting.
When the [Esc] key is pressed , Hardy goes into a text mode.
This should NOT happen, as Hardy should continue in graphics mode even
if its' skipping the test.

(Alpha 4 - completely updated Feb 21, 2008)

[Saivann Carignan (oxmosys)]

I confirm with alpha 5

[lsmobrian (brian-j-porter)]

Confirmed. After pressing esc to canel, fsck runs in text mode. Does not show text based status bar
I thought I locked my system up when i hit escape so i pressed ctrl+alt+del and let it run in graphical then tried it again
sudo touch /forcefsck

/var/log/fsck/checkfs and /var/log/fsck/checkroot look normal

[Kees Cook (kees)]

I believe this is due to usplash requiring a FIFO output reader when getting the "INPUTCHAR" command. There is a race condition in the fsck script that the reader starts after the write is sent. usplash will exit when it wins the race.

e.g.:

$ sudo -s
# usplash -c -v &
# usplash_write "INPUTCHAR"
[1]+ Exit 1 usplash -c -v

[Launchpad Janitor (janitor)]

This bug was fixed in the package usplash - 0.5.16

---------------
usplash (0.5.16) hardy; urgency=low

  * libusplash.c: INPUTCHAR did not sleep correctly due to animation
    signals (LP: #194017).
  * libusplash.c: get_timeout_char() was not re-enabling signals.
  * usplash.c, usplash_write.c: removed command data-loss race introduced by
    needless re-opening of input fifo (LP: #200299).
  * usplash.c: "TIMEOUT 0" was causing 100% CPU in select with timeout of 0.

 -- Kees Cook <email address hidden> Sun, 09 Mar 2008 14:16:06 -0700

[Janek (thefoxx)]

I don't think it's fixed because I installed Hardy beta today with all updates and I still have this problem. I think after pressing ESC the boot does not only proceed in text mode but it also continues checking the drives because the drive worked for a long time but afterwards the boot continued.

It's for me like for lsmobrian but I waited for the end of the check.

[Sancho (sancho-sancho)]

I can confirm - I have the newest update and both the problems with the text mode AND the proceeding with the disk check remains.

[Dean Loros (autocrosser)]

I have has similar problems. See Bug: https://bugs.launchpad.net/ubuntu/+source/sysvinit/+bug/197667

[Martin Pitt (pitti)]

Closing this again, since the original reason was fixed (see comment 4). Your recent trouble is most likely bug 203322, so please let's continue that discussion there. Thank you!