Re: [Xen-ia64-devel] alt_itlb_miss?

[Masaki Kanno <kanno.masaki@xxxxxxxxxxxxxx>]

Hi,

Because I was a business trip yesterday, I was behind with reading an 
email of all of you.

I'm working on FPSWA support now.
I inserted FPSWA codes in ITC with the below patch, and I passed tests 
of a floating point by LTP. Please give me comment.

diff -r 7ed6c203efe9 xen/arch/ia64/xen/ivt.S
--- a/xen/arch/ia64/xen/ivt.S   Wed Apr 19 10:39:15 2006 -0600
+++ b/xen/arch/ia64/xen/ivt.S   Thu Apr 20 17:17:23 2006 +0900
@@ -465,12 +465,21 @@ late_alt_itlb_miss:
        or r19=r19,r18          // set bit 4 (uncached) if the access was to 
region 6
 (p8)   br.cond.spnt page_fault
 #ifdef XEN
-       FORCE_CRASH
+       ;;
+       // Test for Xen address, if not handle via page_fault
+       // note that 0xf000 (cached) addresses should be OK.
+       extr.u r22=r16,59,5;;
+       cmp.ne p8,p0=0x1e,r22
+(p8)   br.cond.spnt 1f
 #endif 
        ;;
        itc.i r19               // insert the TLB entry
        mov pr=r31,-1
        rfi
+#ifdef XEN
+1:
+       FORCE_CRASH
+#endif
 END(alt_itlb_miss)
 
        .org ia64_ivt+0x1000

Best regars,
 Kan

Alex Williamson wrote:
>On Wed, 2006-04-19 at 11:50 +0900, Isaku Yamahata wrote:
>> On Wed, Apr 19, 2006 at 09:05:49AM +0800, Tian, Kevin wrote:
>> > 
>> > So the logic to support identity mapped ITC insertion should be 
>> > added to itlb_miss handler, however the interesting thing is why 
>> > you encounter it in alt_itlb_miss handler. As above debug output, 
>> > the rr7 is also configured with vhpt table enabled...
>> 
>> PSR.dt = 0 or PSR.ic = 0?
>> If so, the assumption that alt_itlb_miss handler dosn't occur is incorrect.
>
>   I traced through one more time.  Looks like I missed the initial
>entry into the IVT.  I'm taking an itlb_miss, then branching to the
>alt_itlb_miss in VHPT_CCHAIN_LOOKUP.  PSR.dt = PSR.ic = 1.  Hopefully
>that clarifies how we're getting into alt_itlb_miss in a vhpt covered
>region.  AFAICT, the logic exists in alt_itlb_miss to fill the itc for
>this, but it's disable by the FORCE_CRASH.  Thanks,
>
>       Alex
>
>-- 
>Alex Williamson                             HP Linux & Open Source Lab
>
>
>_______________________________________________
>Xen-ia64-devel mailing list
>Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>http://lists.xensource.com/xen-ia64-devel
>


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