Description of problem: During startup/shutdown, the screen flickers like crazy. I managed to track down what was causing it: Init scripts -> /etc/init.d/functions -> /etc/profile.d/lang.sh -> /bin/unicode_start -> /bin/setfont. Every time /bin/setfont is called, the screen "resets", flickering. This may be because of my graphics hardware, but the problem should be avoidable -- if the font hasn't been changed, don't ask the kernel to change it. If it's not possible to tell what the previous font was from userspace, this bug will need to be fixed in the kernel (please change the bug component in that case). Version-Release number of selected component (if applicable): kbd-1.12-21 kernel-2.6.20-1.3104.fc7
Fixed summary -- sorry for the bugspam.
Hi, there's no problem to read current font to user space. But your proposed solution will not help, because the screen resets when you are reading font to user space too, not only when you are loading new one into kernes space. Try: setfont -o foo.bar Also it looks like very hardware depending problem, on some machines no flickering, on my machine time to time... On your machine screen resets everytime you call setfont?
You're right that there's not much that can be done about the problem in setfont if reading the font causes the flicker. There are two other solutions to the problem though: 1) Only call setfont once on boot, and then don't call it again. This would fix the described problem of constant flickering on boot. There should be no reason to need to repeatedly set the font -- if there is, that's a bug in something else. 2) Implement a solution in the kernel The vgacon_do_font_op function makes a lot of VGA calls regardless of whether the font is being 'get' or 'set'. http://lxr.linux.no/source/drivers/video/console/vgacon.c?v=2.6.18#L1078 It's very likely that some of those commands are not necessary when just reading the font memory. I'm not sure who would be most qualified to make some changes here.
*** Bug 245478 has been marked as a duplicate of this bug. ***
The bug is filled against Fedora rawhide, but I have this bug now in RHEL 5.1. I had previously F7 and F8. F7 had this bug for sure. But I remember somewhere between F7 and F8 this bug was solved. ¿Perhaps initscripts-8.54.1-1.i386.rpm? I can do some testing in RHEL 5.1 if required.
Some research here. Look at bug #237839 Comment #14 From Bill Nottingham (notting) on 2008-02-27 11:32 EST [reply] Fixed in 8.64-1.
*** This bug has been marked as a duplicate of 237839 ***