Blank Screen in 2.6.0
Quoted 5 lines · Dan
Upgraded to the lastest module-init-tools, and disabled the frame buffer support in the kernel. So the only graphic option enabled is text mode selection. But when I boot I still get a blank screen! My lilo.conf contains a line: vga=773, which works beautifully under RedHat's stock 2.4.20-8.
In my experience the vga= option worked with every 2.4 kernel in every distro that I had used, continued working with 2.6 test* and 0 in Red Hat 7.3, but blanked the screen with 2.6 test* and 0 in SuSE 8.1 and SuSE 8.2. Haven't tried other configurations with 2.6.
But now you're getting the same with a Red Hat distro, so it's looking pretty random.
The decision to release 2.6.0 with the same broken vga= option that was reported many times in 2.6.0-test* makes me think that vga= is not intended to work.
By the way, on one machine I experimented with putting the framebuffer driver back into the kernel after deleting the vga= option, and it automatically expanded to all of the screen and it worked. I don't know if this is random behavior either, for example it probably doesn't work on the machines of people who told you to disable your framebuffer and it probably doesn't work on the machines of people who didn't check whether you obeyed the Post Hallowe'en doc. But if you're brave you might try it.
Quoted 13 lines · Norman Diamond
Quoted 5 lines · Dan
Upgraded to the lastest module-init-tools, and disabled the frame buffer support in the kernel. So the only graphic option enabled is text mode selection. But when I boot I still get a blank screen! My lilo.conf contains a line: vga=773, which works beautifully under RedHat's stock 2.4.20-8.
In my experience the vga= option worked with every 2.4 kernel in every distro that I had used, continued working with 2.6 test* and 0 in Red Hat 7.3, but blanked the screen with 2.6 test* and 0 in SuSE 8.1 and SuSE 8.2. Haven't tried other configurations with 2.6.
Quoted 6 lines
But now you're getting the same with a Red Hat distro, so it's looking pretty random.
The decision to release 2.6.0 with the same broken vga= option that was reported many times in 2.6.0-test* makes me think that vga= is not intended to work.
Maybe it has something to do with RedHat 7.3. I've used RH8, RH9, and Fedora Core 1 and haven't had a problem with vga= in any of them during the 2.5/2.6 series, right up through the current one. I've got framebuffer support as a module.
Quoted 14 lines · Thomas Molina
Quoted 13 lines · Norman Diamond
Quoted 5 lines · Dan
Upgraded to the lastest module-init-tools, and disabled the frame buffer support in the kernel. So the only graphic option enabled is text mode selection. But when I boot I still get a blank screen! My lilo.conf contains a line: vga=773, which works beautifully under RedHat's stock 2.4.20-8.
In my experience the vga= option worked with every 2.4 kernel in every distro that I had used, continued working with 2.6 test* and 0 in Red Hat 7.3, but blanked the screen with 2.6 test* and 0 in SuSE 8.1 and SuSE 8.2. Haven't tried other configurations with 2.6.
Quoted 8 lines
Quoted 6 lines
But now you're getting the same with a Red Hat distro, so it's looking pretty random.
The decision to release 2.6.0 with the same broken vga= option that was reported many times in 2.6.0-test* makes me think that vga= is not intended to work.
Maybe it has something to do with RedHat 7.3.
Are you serious? It really is true that vga= isn't supposed to work in 2.6.0, but there is something to do with RedHat 7.3 which caused vga= to continue to work in 2.6.0 with that distro only? Then why hasn't the vga= parameter been removed entirely?
Quoted 2 lines
I've used RH8, RH9, and Fedora Core 1 and haven't had a problem with vga= in any of them during the 2.5/2.6 series, right up through the current one.
I forgot if Dan is using RH8 or RH9. Whichever, you're getting different results than he did, because his failed with one of these and yours works with both. At least my failures with SuSE are consistent. I'd say the overall inconsistency points to a bug in 2.6.0-test* and .0. Especially when no one seems to be reporting similar failures in the 2.4 series.
Quoted 8 lines · Norman Diamond
Quoted 2 lines
I've used RH8, RH9, and Fedora Core 1 and haven't had a problem with vga= in any of them during the 2.5/2.6 series, right up through the current one.
I forgot if Dan is using RH8 or RH9. Whichever, you're getting different results than he did, because his failed with one of these and yours works with both. At least my failures with SuSE are consistent. I'd say the overall inconsistency points to a bug in 2.6.0-test* and .0. Especially when no one seems to be reporting similar failures in the 2.4 series.
I've had problems at various times during 2.5 with a number of card-specific framebuffer drivers so I've backed off to only using the VESA framebuffer driver. Maybe the connection is which framebuffer driver is used.
Quoted 10 lines · Thomas Molina
Quoted 1 line · Norman Diamond
...
I've had problems at various times during 2.5 with a number of card-specific framebuffer drivers so I've backed off to only using the VESA framebuffer driver. Maybe the connection is which framebuffer driver is used.
I've had the same problem. I've tried using radeonfb and it's never worked yet since 2.5.69. In 2.6.0, no matter what mode I select, I get a horizontal frequency of 140kHz and a ~90Hz vertical sync. My monitor blanks and pops up the on-screen menu with the horizontal frequency high-lighted in red. My monitor supports up to 110kHz horizontal according to my X log.
I compiled just the vesafb in as well, but it does the exact same thing. It doesn't matter what values or modes I put in the vga/video parameter. Only the text mode works now, and after only 2 days uptime I've noticed I no longer have the full vertical screen. When I type, random text or colored/flashing blocks appear in the bottom (roughly) 1/5th of the screen. Actually, anytime I run an ncurse app in my console, it gets smaller. Now my console only occupies 1/2 the screen, but it doesn't seem to get any smaller than that. I'm using the boot parameter "vga=0x132".
My point is maybe your screen isn't blank, but rather the horizontal frequency is out of your monitor's range?
Certainly, something has changed, because with 2.6.0-test11 I was running "vga=0x31B video=vesafb:ywrap" and it worked great, whereas now it doesn't work at all with any fb modes.
-Wes-
Quoted 4 lines
Certainly, something has changed, because with 2.6.0-test11 I was running "vga=0x31B video=vesafb:ywrap" and it worked great, whereas now it doesn't work at all with any fb modes.
To my knowledge the code hasn't been touched since then. I have a new patch for you to try.
Quoted 5 lines
Are you serious? It really is true that vga= isn't supposed to work in 2.6.0, but there is something to do with RedHat 7.3 which caused vga= to continue to work in 2.6.0 with that distro only? Then why hasn't the vga= parameter been removed entirely?
vga= still works with the vesa framebuffer. There is a small bug that is causing all the problems.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Thomas Molina wrote:
| On Mon, 29 Dec 2003, Norman Diamond wrote: | |>The decision to release 2.6.0 with the same broken vga= option that was |>reported many times in 2.6.0-test* makes me think that vga= is not intended |>to work. | | | Maybe it has something to do with RedHat 7.3. I've used RH8, RH9, and | Fedora Core 1 and haven't had a problem with vga= in any of them during | the 2.5/2.6 series, right up through the current one. I've got | framebuffer support as a module. |
Not RedHat 9 issue because I'm using RH 9 and I am the one that started this thread.
- --- Dan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQE/8eg9twT22Jak4/4RAuCfAKC6gOC69UgKzO3VU+QcM/qfBxrOGgCeOqOP z8j0QMQAV+G8xVuyH6audgk=
=HgkW
-----END PGP SIGNATURE-----
Quoted 14 lines
|>The decision to release 2.6.0 with the same broken vga= option that was |>reported many times in 2.6.0-test* makes me think that vga= is not intended
|>to work. | | | Maybe it has something to do with RedHat 7.3. I've used RH8, RH9, and | Fedora Core 1 and haven't had a problem with vga= in any of them during | the 2.5/2.6 series, right up through the current one. I've got | framebuffer support as a module. |
Quoted 2 lines
Not RedHat 9 issue because I'm using RH 9 and I am the one that started this thread.
Its a bug in the kernel. Fixed in latest tree.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
James Simmons wrote:
|>|>The decision to release 2.6.0 with the same broken vga= option that was |>|>reported many times in 2.6.0-test* makes me think that vga= is not |>intended |>|>to work. |>| |>| |>| Maybe it has something to do with RedHat 7.3. I've used RH8, RH9, and |>| Fedora Core 1 and haven't had a problem with vga= in any of them during |>| the 2.5/2.6 series, right up through the current one. I've got |>| framebuffer support as a module. |>| |> |> |>Not RedHat 9 issue because I'm using RH 9 and I am the one that started |>this thread. | | | Its a bug in the kernel. Fixed in latest tree. | | http://phoenix.infradead.org/~jsimmons/fbdev.diff.gz | | |
Cool. When the 2.6.1 comes out (I'm assuming that this means it will be in 2.6.1), I'll d/l it and try it out.
- --- Dan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQE/+fTCtwT22Jak4/4RAroCAJkBU2goo0OYMSgRZAo25/zvdpUp+QCfZlLh H9y/WAT8t/hBONnhJ+U6k2I=
=1H0b
-----END PGP SIGNATURE-----