The quick fix for this is:

diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 490dac6..06a2dac 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -497,6 +497,9 @@ static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c)
               if (!cpu_devs[i])
                       break;

+               if (!cpu_devs[i]->c_ident[0])
+                       continue;
+
               if (!strcmp(v, cpu_devs[i]->c_ident[0]) ||
                   (cpu_devs[i]->c_ident[1] &&
                    !strcmp(v, cpu_devs[i]->c_ident[1]))) {


however, Peter's requirement to remove c_ident renders my patch 
somewhat less useful back on RDC.  /proc/cpuinfo then lists the 
vendor as 'RDC' (an improvement over not applying the patch) and 
the model name as '486' (also an improvement), but for some 
reason the variants in my case statement then no longer make it 
through to /proc/info.  I don't think this is a big deal, and I
still think the patch is worth applying with this change.

regards,
Mark




      
--

[ message continues ]

Well, as I said, the generic code probably needs to change somewhat to
print that stuff.

	-hpa
--

[ message continues ]