[Top] [All Lists]

Re: [RFC] SMP initialization order fixes.

Subject: Re: [RFC] SMP initialization order fixes.
From: Stuart Anderson <>
Date: Wed, 22 Feb 2006 16:41:43 -0500 (EST)
In-reply-to: <>
Original-recipient: rfc822;
References: <>
On Wed, 22 Feb 2006, Ralf Baechle wrote:

This one should hopefully fix the SMP problems of the resent times.  It
works on Malta with 34K, it seems to work on IP27 (the kernel is
presumably failing due to other issues), so now I'd ask especially
RM9000 & BCM1250 users for testing.  This really needs to be fixed for

I'm not sure if this is the specific fix or not, but I can report that git
as of today (approx 2pm est) is working better than is has since 2.6.14 for
me on a bcm1480. I had tried git a couple of weeks ago, and it still hung
when I stressed it.

I use NFS root, and the stress test that would hang the system is simply
"make -j 4" of the kernel. Previously this would hang the syste before

Now that things seem to be better, I'll leave this looping for a while,
and then run some other tests. It's time to run the ltp on the 3 different
ABIs again anyway.


Stuart R. Anderson                     
Network & Software Engineering         
1024D/37A79149:                                  0791 D3B8 9A4C 2CDC A31F
                                                  BD03 0A62 E534 37A7 9149

<Prev in Thread] Current Thread [Next in Thread>