[PATCH] code cleanup

linas at austin.ibm.com linas at austin.ibm.com
Tue Apr 13 02:43:15 EST 2004

On Fri, Apr 09, 2004 at 02:18:20PM -0500, jschopp at austin.ibm.com wrote:
> I was looking at rtas serialization for reasons I won't go into here.

I'd like to know why ... the LTE and others have a stack of bugs
that have MSR=9000... and there is some suspicion that bad args sent
sent to rtas might be one of the causes of this.

> While wandering through the code I found that two functions were not
> properly serialized.  phys_call_rtas and phys_call_rtas_display_status are

While wandering through the same code, I noticed that some of the
locking fails to lock sections of the paca as well, which might (??)
cause races in the future (??) if CONFIG_PREEMPT is enabled (maybe).

If you're in a cleanup mood, I'd suggest moving those locks to
protect the reading/writing of the paca data.

> below removes them, unless I hear some objections I will push it
> early next week.

I don't have any objections.


** Sent via the linuxppc64-dev mail list. See http://lists.linuxppc.org/

More information about the Linuxppc64-dev mailing list