[Xen-ia64-devel] [PATCH] skip MCA setup on PV domU

[Alex Williamson <alex.williamson@xxxxxx>]

   Is there any reason to setup MCA support on PV domUs?  I can't think
of any, and it wastes 3 dynamic IRQs per vCPU to support it.  That
currently limits us to 42 vCPUs on PV guests.  The patch below
automatically disables MCA support on domUs, allowing us to
theoretically hit 85 vCPUs before hitting that limit again.  We probably
need to think about per CPU event channels to remove the limit entirely.
I've tested up to 64 vCPUs, at which point the xm tools appear to
prevent us from going further.  Is anyone thinking of doing something
with MCA support in PV domains, or should I check this in?  Thanks,

        Alex


Signed-off-by: Alex Williamson <alex.williamson@xxxxxx>
---

diff -r 1bde28f762a6 linux-2.6-xen-sparse/arch/ia64/kernel/setup.c
--- a/linux-2.6-xen-sparse/arch/ia64/kernel/setup.c     Mon Apr 09 13:40:25 
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/ia64/kernel/setup.c     Tue Apr 10 11:42:56 
2007 -0600
@@ -594,6 +594,10 @@ setup_arch (char **cmdline_p)
 
 
        /* enable IA-64 Machine Check Abort Handling unless disabled */
+#ifdef CONFIG_XEN
+       if (is_running_on_xen() && !is_initial_xendomain())
+               nomca = 1;
+#endif
        if (!nomca)
                ia64_mca_init();
 



_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel