Red Hat Bugzilla – Bug 6401
Login hangs after installation of RedHat 6.1
Last modified: 2008-05-01 11:37:52 EDT
Machine: Toshiba Portege 7010CT
When booting up for the first time after a successful
installation the keyboard is completely locked up. No
regular keys, not ctrl-alt-del, nothing.
Workaround: disable GPM.
Yes, we have seen this on a couple of different laptop models. Thanks for the
I think that I know what might be the cause of this. Attached is the note that
I sent to Alan Cox. He has since temporarily reverted the patch because it was
causing characters to be dropped on the serial port.
Try to resolve it with him. This will probably solve quite a few problems.
After two solid days of frustration, a coworker and I found
and fixed a bug in the the Linux ps2 mouse/keyboard driver. After a
certain amount of mouse use, the keyboard and mouse would hang,
leaving the machine unresponsive to keypresses.
The problem was as follows.
The Linux driver polled the ps2 status register faster than the device
could keep up. The FreeBSD ps2 driver inserted a pause between the
polls. By doing the same in Linux, the problem disappears.
This bug showed up on our Digital HiNote CT475 laptops. Without the
fix, the boxes are quite useless.
This fix shouldn't harm anyone, but may help some. (It may eliminate the
"pc_keyboard jammed" warning that appears right below the patch.)
Attached is a patch for both 2.2.13 and 2.3.28.
diff -ur linux-orig/drivers/char/pc_keyb.c linux/drivers/char/pc_keyb.c
--- linux-orig/drivers/char/pc_keyb.c Mon Aug 9 15:04:39 1999
+++ linux/drivers/char/pc_keyb.c Mon Nov 15 19:27:09 1999
@@ -433,7 +433,7 @@
handle_scancode(scancode, !(scancode & 0x80));
status = kbd_read_status();
Yep - this one is probably a 2.4 thing. The delay cannot be simply inserted ,
instead the entire
PS/AUX handler needs to be reconstructed so that it runs with IRQ's enabled and
to do its interrupting. Non trivial- not for 2.2
I feel bad about this one as its our bug and your suffering...
How about submitting a variant where you can do psaux=slow on the init line -
I'd take that
I'm assuming that this is fixed in the 2.4 kernels and we're not going to
go back and fix it for 2.2 kernels. If I'm wrong about 2.4 kernels, let