x86: don't special-case pmd allocations as much

In x86 PAE mode, stop treating pmds as a special case.  Previously
they were always allocated and freed with the pgd.  The modifies the
code to be the same as 64-bit mode, where they are allocated on
demand.

This is a step on the way to unifying 32/64-bit pagetable allocation
as much as possible.

There is a complicating wart, however.  When you install a new
reference to a pmd in the pgd, the processor isn't guaranteed to see
it unless you reload cr3.  Since reloading cr3 also has the
side-effect of flushing the tlb, this is an expense that we want to
avoid whereever possible.

This patch simply avoids reloading cr3 unless the update is to the
current pagetable.  Later patches will optimise this further.

Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: William Irwin <wli@holomorphy.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

2 files changed, 0 insertions, 81 deletions

diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 02d269c07b9..da524fb2242 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -704,19 +704,6 @@ int arch_add_memory(int nid, u64 start, u64 size)
 }
 #endif
 
-struct kmem_cache *pmd_cache;
-
-void __init pgtable_cache_init(void)
-{
-        if (PTRS_PER_PMD > 1) {
-                pmd_cache = kmem_cache_create("pmd",
-                                              PTRS_PER_PMD*sizeof(pmd_t),
-                                              PTRS_PER_PMD*sizeof(pmd_t),
-                                              SLAB_PANIC,
-                                              pmd_ctor);
-        }
-}
-
 /*
  * This function cannot be __init, since exceptions don't work in that
  * section.  Put this after the callers, so that it cannot be inlined.
diff --git a/arch/x86/mm/pgtable_32.c b/arch/x86/mm/pgtable_32.c
index 3a6c9200058..5ca3552474a 100644
--- a/arch/x86/mm/pgtable_32.c
+++ b/arch/x86/mm/pgtable_32.c
@@ -195,11 +195,6 @@ struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address)
         return pte;
 }
 
-void pmd_ctor(struct kmem_cache *cache, void *pmd)
-{
-        memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
-}
-
 /*
  * List of all pgd's needed for non-PAE so it can invalidate entries
  * in both cached and uncached pgd's; not needed for PAE since the
@@ -285,7 +280,6 @@ static void pgd_dtor(void *pgd)
         if (SHARED_KERNEL_PMD)
                 return;
 
-        paravirt_release_pd(__pa(pgd) >> PAGE_SHIFT);
         spin_lock_irqsave(&pgd_lock, flags);
         pgd_list_del(pgd);
         spin_unlock_irqrestore(&pgd_lock, flags);
@@ -367,84 +361,22 @@ static void pgd_mop_up_pmds(pgd_t *pgd)
 }
 #endif  /* CONFIG_X86_PAE */
 
-/* If we allocate a pmd for part of the kernel address space, then
-   make sure its initialized with the appropriate kernel mappings.
-   Otherwise use a cached zeroed pmd.  */
-static pmd_t *pmd_cache_alloc(int idx)
-{
-        pmd_t *pmd;
-
-        if (idx >= USER_PTRS_PER_PGD) {
-                pmd = (pmd_t *)__get_free_page(GFP_KERNEL);
-
-                if (pmd)
-                        memcpy(pmd,
-                               (void *)pgd_page_vaddr(swapper_pg_dir[idx]),
-                               sizeof(pmd_t) * PTRS_PER_PMD);
-        } else
-                pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
-
-        return pmd;
-}
-
-static void pmd_cache_free(pmd_t *pmd, int idx)
-{
-        if (idx >= USER_PTRS_PER_PGD)
-                free_page((unsigned long)pmd);
-        else
-                kmem_cache_free(pmd_cache, pmd);
-}
-
 pgd_t *pgd_alloc(struct mm_struct *mm)
 {
-        int i;
         pgd_t *pgd = quicklist_alloc(0, GFP_KERNEL, pgd_ctor);
 
-        if (PTRS_PER_PMD == 1 || !pgd)
-                return pgd;
-
         mm->pgd = pgd;          /* so that alloc_pd can use it */
 
-        for (i = 0; i < UNSHARED_PTRS_PER_PGD; ++i) {
-                pmd_t *pmd = pmd_cache_alloc(i);
-
-                if (!pmd)
-                        goto out_oom;
-
-                paravirt_alloc_pd(mm, __pa(pmd) >> PAGE_SHIFT);
-                set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
-        }
         if (pgd && !pgd_prepopulate_pmd(mm, pgd)) {
                 quicklist_free(0, pgd_dtor, pgd);
                 pgd = NULL;
         }
 
         return pgd;
-
-out_oom:
-        for (i--; i >= 0; i--) {
-                pgd_t pgdent = pgd[i];
-                void* pmd = (void *)__va(pgd_val(pgdent)-1);
-                paravirt_release_pd(__pa(pmd) >> PAGE_SHIFT);
-                pmd_cache_free(pmd, i);
-        }
-        quicklist_free(0, pgd_dtor, pgd);
-        return NULL;
 }
 
 void pgd_free(pgd_t *pgd)
 {
-        int i;
-
-        /* in the PAE case user pgd entries are overwritten before usage */
-        if (PTRS_PER_PMD > 1)
-                for (i = 0; i < UNSHARED_PTRS_PER_PGD; ++i) {
-                        pgd_t pgdent = pgd[i];
-                        void* pmd = (void *)__va(pgd_val(pgdent)-1);
-                        paravirt_release_pd(__pa(pmd) >> PAGE_SHIFT);
-                        pmd_cache_free(pmd, i);
-                }
-        /* in the non-PAE case, free_pgtables() clears user pgd entries */
         pgd_mop_up_pmds(pgd);
         quicklist_free(0, pgd_dtor, pgd);
 }