[SPARC64]: Move away from virtual page tables, part 1.

We now use the TSB hardware assist features of the UltraSPARC
MMUs.

SMP is currently knowingly broken, we need to find another place
to store the per-cpu base pointers.  We hid them away in the TSB
base register, and that obviously will not work any more :-)

Another known broken case is non-8KB base page size.

Also noticed that flush_tlb_all() is not referenced anywhere, only
the internal __flush_tlb_all() (local cpu only) is used by the
sparc64 port, so we can get rid of flush_tlb_all().

The kernel gets it's own 8KB TSB (swapper_tsb) and each address space
gets it's own private 8K TSB.  Later we can add code to dynamically
increase the size of per-process TSB as the RSS grows.  An 8KB TSB is
good enough for up to about a 4MB RSS, after which the TSB starts to
incur many capacity and conflict misses.

We even accumulate OBP translations into the kernel TSB.

Another area for refinement is large page size support.  We could use
a secondary address space TSB to handle those.

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 2 insertions, 59 deletions

diff --git a/arch/sparc64/mm/tlb.c b/arch/sparc64/mm/tlb.c
index 8b104be4662b..78357cc2a0b7 100644
--- a/arch/sparc64/mm/tlb.c
+++ b/arch/sparc64/mm/tlb.c
@@ -25,6 +25,8 @@ void flush_tlb_pending(void)
         struct mmu_gather *mp = &__get_cpu_var(mmu_gathers);
 
         if (mp->tlb_nr) {
+                flush_tsb_user(mp);
+
                 if (CTX_VALID(mp->mm->context)) {
 #ifdef CONFIG_SMP
                         smp_flush_tlb_pending(mp->mm, mp->tlb_nr,
@@ -89,62 +91,3 @@ no_cache_flush:
         if (nr >= TLB_BATCH_NR)
                 flush_tlb_pending();
 }
-
-void flush_tlb_pgtables(struct mm_struct *mm, unsigned long start, unsigned long end)
-{
-        struct mmu_gather *mp = &__get_cpu_var(mmu_gathers);
-        unsigned long nr = mp->tlb_nr;
-        long s = start, e = end, vpte_base;
-
-        if (mp->fullmm)
-                return;
-
-        /* If start is greater than end, that is a real problem.  */
-        BUG_ON(start > end);
-
-        /* However, straddling the VA space hole is quite normal. */
-        s &= PMD_MASK;
-        e = (e + PMD_SIZE - 1) & PMD_MASK;
-
-        vpte_base = (tlb_type == spitfire ?
-                     VPTE_BASE_SPITFIRE :
-                     VPTE_BASE_CHEETAH);
-
-        if (unlikely(nr != 0 && mm != mp->mm)) {
-                flush_tlb_pending();
-                nr = 0;
-        }
-
-        if (nr == 0)
-                mp->mm = mm;
-
-        start = vpte_base + (s >> (PAGE_SHIFT - 3));
-        end = vpte_base + (e >> (PAGE_SHIFT - 3));
-
-        /* If the request straddles the VA space hole, we
-         * need to swap start and end.  The reason this
-         * occurs is that "vpte_base" is the center of
-         * the linear page table mapping area.  Thus,
-         * high addresses with the sign bit set map to
-         * addresses below vpte_base and non-sign bit
-         * addresses map to addresses above vpte_base.
-         */
-        if (end < start) {
-                unsigned long tmp = start;
-
-                start = end;
-                end = tmp;
-        }
-
-        while (start < end) {
-                mp->vaddrs[nr] = start;
-                mp->tlb_nr = ++nr;
-                if (nr >= TLB_BATCH_NR) {
-                        flush_tlb_pending();
-                        nr = 0;
-                }
-                start += PAGE_SIZE;
-        }
-        if (nr)
-                flush_tlb_pending();
-}