[IA64] relax per-cpu TLB requirement to DTC

Instead of pinning per-cpu TLB into a DTR, use DTC.  This will free up
one TLB entry for application, or even kernel if access pattern to
per-cpu data area has high temporal locality.

Since per-cpu is mapped at the top of region 7 address, we just need to
add special case in alt_dtlb_miss.  The physical address of per-cpu data
is already conveniently stored in IA64_KR(PER_CPU_DATA).  Latency for
alt_dtlb_miss is not affected as we can hide all the latency.  It was
measured that alt_dtlb_miss handler has 23 cycles latency before and
after the patch.

The performance effect is massive for applications that put lots of tlb
pressure on CPU.  Workload environment like database online transaction
processing or application uses tera-byte of memory would benefit the most.
Measurement with industry standard database benchmark shown an upward
of 1.6% gain.  While smaller workloads like cpu, java also showing small
improvement.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

1 files changed, 0 insertions, 24 deletions

diff --git a/arch/ia64/kernel/mca_asm.S b/arch/ia64/kernel/mca_asm.S
index c6b607c00dee..8c9c26aa6ae0 100644
--- a/arch/ia64/kernel/mca_asm.S
+++ b/arch/ia64/kernel/mca_asm.S
@@ -101,14 +101,6 @@ ia64_do_tlb_purge:
         ;;
         srlz.d
         ;;
-        // 2. Purge DTR for PERCPU data.
-        movl r16=PERCPU_ADDR
-        mov r18=PERCPU_PAGE_SHIFT<<2
-        ;;
-        ptr.d r16,r18
-        ;;
-        srlz.d
-        ;;
         // 3. Purge ITR for PAL code.
         GET_THIS_PADDR(r2, ia64_mca_pal_base)
         ;;
@@ -196,22 +188,6 @@ ia64_reload_tr:
         srlz.i
         srlz.d
         ;;
-        // 2. Reload DTR register for PERCPU data.
-        GET_THIS_PADDR(r2, ia64_mca_per_cpu_pte)
-        ;;
-        movl r16=PERCPU_ADDR            // vaddr
-        movl r18=PERCPU_PAGE_SHIFT<<2
-        ;;
-        mov cr.itir=r18
-        mov cr.ifa=r16
-        ;;
-        ld8 r18=[r2]                    // load per-CPU PTE
-        mov r16=IA64_TR_PERCPU_DATA;
-        ;;
-        itr.d dtr[r16]=r18
-        ;;
-        srlz.d
-        ;;
         // 3. Reload ITR for PAL code.
         GET_THIS_PADDR(r2, ia64_mca_pal_pte)
         ;;