Re: [patch] Real-Time Preemption, -RT-2.6.10-rc1-mm2-V0.7.7

[Ingo Molnar <mingo@elte.hu>]

* Michael J. Cohen <mjc@unre.st> wrote:

> Turned off the debugging stuff to fix this one :/
> 
> might_sleep issue at swap_on and firefox causes oopsen.
> 
> dmesg is 120k+ so here:
> 
> http://325i.org/software/2.6.10-rc1-mm2-RT-V0.7.8.dmesg

does the patch below fix the fbcon problem? (if any new oops happens or
old one triggers again then please re-post the syslog or serial console
capture)

	Ingo

--- linux/drivers/video/console/fbcon.c.orig
+++ linux/drivers/video/console/fbcon.c
@@ -1051,7 +1051,14 @@ static void fbcon_cursor(struct vc_data 
 	struct display *p = &fb_display[vc->vc_num];
 	int y = real_y(p, vc->vc_y);
  	int c = scr_readw((u16 *) vc->vc_pos);
+#ifdef CONFIG_PREEMPT_REALTIME
+	unsigned long flags;
+#endif
 
+#ifdef CONFIG_PREEMPT_REALTIME
+	local_save_flags(flags);
+	local_irq_enable();
+#endif
 	ops->cursor_flash = 1;
 	if (mode & CM_SOFTBACK) {
 		mode &= ~CM_SOFTBACK;
@@ -1069,6 +1076,9 @@ static void fbcon_cursor(struct vc_data 
 	ops->cursor(vc, info, p, mode, get_color(vc, info, c, 1),
 		    get_color(vc, info, c, 0));
 	vbl_cursor_cnt = CURSOR_DRAW_DELAY;
+#ifdef CONFIG_PREEMPT_REALTIME
+	local_irq_restore(flags);
+#endif
 }
 
 static int scrollback_phys_max = 0;