Re: Patch: linux-2.5.8-pre1/kernel/exit.c change caused BUG() at boot time

["David C. Hansen" <haveblue@us.ibm.com>]

Adam J. Richter wrote:

>	When I attempted to boot linux-2.5.8-pre1, I got a kernel
>BUG() for exit.c line 519.
>
That bug is hit when the schedule() returns.  In the do_exit() case, 
schedule is not supposed to return.  After the current task is scheduled 
out, it is destroyed.  I guess that most of the freeing of the task's 
resources is done in do_exit(), but I don't see where its kernel stack 
is freed.  

> The was a small change to to kernel/exit.c
>in 2.5.8-pre1 which deleted a kernel_lock() call.  Restoring that line
>resulted in a kernel that booted fine.
>
I take it you don't have a copy of the BUG().   I was going to ask if 
preemption was enabled, but I see that it was from another message.  I 
was guessing that preemption contributed to this, but now I know.  The 
lock_kernel() has 2 different effects here.  It locks the kernel_flag, 
AND it disables preemption.  The correct fix here will probably be to 
disable preemption, rather than readd the lock_kernel().  

For the preemption gurus:
Is a preempt_disable() in do_exit() going to hurt anything?  Shold we 
selectively skip the preempt_disable() in schedule() if it was 
schedule() called from do_exit()?
--
Dave Hansen
haveblue@us.ibm.com