#include <asm/cputable.h>
#include <asm/abs_addr.h>
#include <asm/sections.h>
+#include <asm/spu.h>
#ifdef DEBUG
#define DBG(fmt...) udbg_printf(fmt)
#ifdef CONFIG_PPC_64K_PAGES
int mmu_ci_restrictions;
#endif
+#ifdef CONFIG_DEBUG_PAGEALLOC
+static u8 *linear_map_hash_slots;
+static unsigned long linear_map_hash_count;
+static DEFINE_SPINLOCK(linear_map_hash_lock);
+#endif /* CONFIG_DEBUG_PAGEALLOC */
/* There are definitions of page sizes arrays to be used when none
* is provided by the firmware.
for (vaddr = vstart, paddr = pstart; vaddr < vend;
vaddr += step, paddr += step) {
- unsigned long vpn, hash, hpteg;
+ unsigned long hash, hpteg;
unsigned long vsid = get_kernel_vsid(vaddr);
unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
- vpn = va >> shift;
tmp_mode = mode;
/* Make non-kernel text non-executable */
if (ret < 0)
break;
+#ifdef CONFIG_DEBUG_PAGEALLOC
+ if ((paddr >> PAGE_SHIFT) < linear_map_hash_count)
+ linear_map_hash_slots[paddr >> PAGE_SHIFT] = ret | 0x80;
+#endif /* CONFIG_DEBUG_PAGEALLOC */
}
return ret < 0 ? ret : 0;
}
* Not in the device-tree, let's fallback on known size
* list for 16M capable GP & GR
*/
- if (cpu_has_feature(CPU_FTR_16M_PAGE) && !machine_is(iseries))
+ if (cpu_has_feature(CPU_FTR_16M_PAGE))
memcpy(mmu_psize_defs, mmu_psize_defaults_gp,
sizeof(mmu_psize_defaults_gp));
found:
+#ifndef CONFIG_DEBUG_PAGEALLOC
/*
* Pick a size for the linear mapping. Currently, we only support
* 16M, 1M and 4K which is the default
mmu_linear_psize = MMU_PAGE_16M;
else if (mmu_psize_defs[MMU_PAGE_1M].shift)
mmu_linear_psize = MMU_PAGE_1M;
+#endif /* CONFIG_DEBUG_PAGEALLOC */
#ifdef CONFIG_PPC_64K_PAGES
/*
if (mmu_psize_defs[MMU_PAGE_64K].shift) {
mmu_virtual_psize = MMU_PAGE_64K;
mmu_vmalloc_psize = MMU_PAGE_64K;
+ if (mmu_linear_psize == MMU_PAGE_4K)
+ mmu_linear_psize = MMU_PAGE_64K;
if (cpu_has_feature(CPU_FTR_CI_LARGE_PAGE))
mmu_io_psize = MMU_PAGE_64K;
else
mmu_ci_restrictions = 1;
}
-#endif
+#endif /* CONFIG_PPC_64K_PAGES */
printk(KERN_DEBUG "Page orders: linear mapping = %d, "
"virtual = %d, io = %d\n",
extern unsigned int *htab_call_hpte_remove;
extern unsigned int *htab_call_hpte_updatepp;
-#ifdef CONFIG_PPC_64K_PAGES
+#ifdef CONFIG_PPC_HAS_HASH_64K
extern unsigned int *ht64_call_hpte_insert1;
extern unsigned int *ht64_call_hpte_insert2;
extern unsigned int *ht64_call_hpte_remove;
mode_rw = _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_COHERENT | PP_RWXX;
+#ifdef CONFIG_DEBUG_PAGEALLOC
+ linear_map_hash_count = lmb_end_of_DRAM() >> PAGE_SHIFT;
+ linear_map_hash_slots = __va(lmb_alloc_base(linear_map_hash_count,
+ 1, lmb.rmo_size));
+ memset(linear_map_hash_slots, 0, linear_map_hash_count);
+#endif /* CONFIG_DEBUG_PAGEALLOC */
+
/* On U3 based machines, we need to reserve the DART area and
* _NOT_ map it to avoid cache paradoxes as it's remapped non
* cacheable later on
return pp;
}
+/*
+ * Demote a segment to using 4k pages.
+ * For now this makes the whole process use 4k pages.
+ */
+#ifdef CONFIG_PPC_64K_PAGES
+static void demote_segment_4k(struct mm_struct *mm, unsigned long addr)
+{
+ if (mm->context.user_psize == MMU_PAGE_4K)
+ return;
+#ifdef CONFIG_PPC_MM_SLICES
+ slice_set_user_psize(mm, MMU_PAGE_4K);
+#else /* CONFIG_PPC_MM_SLICES */
+ mm->context.user_psize = MMU_PAGE_4K;
+ mm->context.sllp = SLB_VSID_USER | mmu_psize_defs[MMU_PAGE_4K].sllp;
+#endif /* CONFIG_PPC_MM_SLICES */
+
+#ifdef CONFIG_SPE_BASE
+ spu_flush_all_slbs(mm);
+#endif
+}
+#endif /* CONFIG_PPC_64K_PAGES */
+
/* Result code is:
* 0 - handled
* 1 - normal page fault
return 1;
}
vsid = get_vsid(mm->context.id, ea);
+#ifdef CONFIG_PPC_MM_SLICES
+ psize = get_slice_psize(mm, ea);
+#else
psize = mm->context.user_psize;
+#endif
break;
case VMALLOC_REGION_ID:
mm = &init_mm;
if (user_region && cpus_equal(mm->cpu_vm_mask, tmp))
local = 1;
+#ifdef CONFIG_HUGETLB_PAGE
/* Handle hugepage regions */
- if (unlikely(in_hugepage_area(mm->context, ea))) {
+ if (HPAGE_SHIFT && psize == mmu_huge_psize) {
DBG_LOW(" -> huge page !\n");
return hash_huge_page(mm, access, ea, vsid, local, trap);
}
+#endif /* CONFIG_HUGETLB_PAGE */
+
+#ifndef CONFIG_PPC_64K_PAGES
+ /* If we use 4K pages and our psize is not 4K, then we are hitting
+ * a special driver mapping, we need to align the address before
+ * we fetch the PTE
+ */
+ if (psize != MMU_PAGE_4K)
+ ea &= ~((1ul << mmu_psize_defs[psize].shift) - 1);
+#endif /* CONFIG_PPC_64K_PAGES */
/* Get PTE and page size from page tables */
ptep = find_linux_pte(pgdir, ea);
}
/* Do actual hashing */
-#ifndef CONFIG_PPC_64K_PAGES
- rc = __hash_page_4K(ea, access, vsid, ptep, trap, local);
-#else
- if (mmu_ci_restrictions) {
- /* If this PTE is non-cacheable, switch to 4k */
- if (psize == MMU_PAGE_64K &&
- (pte_val(*ptep) & _PAGE_NO_CACHE)) {
- if (user_region) {
- psize = MMU_PAGE_4K;
- mm->context.user_psize = MMU_PAGE_4K;
- mm->context.sllp = SLB_VSID_USER |
- mmu_psize_defs[MMU_PAGE_4K].sllp;
- } else if (ea < VMALLOC_END) {
- /*
- * some driver did a non-cacheable mapping
- * in vmalloc space, so switch vmalloc
- * to 4k pages
- */
- printk(KERN_ALERT "Reducing vmalloc segment "
- "to 4kB pages because of "
- "non-cacheable mapping\n");
- psize = mmu_vmalloc_psize = MMU_PAGE_4K;
- }
- }
+#ifdef CONFIG_PPC_64K_PAGES
+ /* If _PAGE_4K_PFN is set, make sure this is a 4k segment */
+ if (pte_val(*ptep) & _PAGE_4K_PFN) {
+ demote_segment_4k(mm, ea);
+ psize = MMU_PAGE_4K;
+ }
+
+ /* If this PTE is non-cacheable and we have restrictions on
+ * using non cacheable large pages, then we switch to 4k
+ */
+ if (mmu_ci_restrictions && psize == MMU_PAGE_64K &&
+ (pte_val(*ptep) & _PAGE_NO_CACHE)) {
if (user_region) {
- if (psize != get_paca()->context.user_psize) {
- get_paca()->context = mm->context;
- slb_flush_and_rebolt();
- }
- } else if (get_paca()->vmalloc_sllp !=
- mmu_psize_defs[mmu_vmalloc_psize].sllp) {
- get_paca()->vmalloc_sllp =
- mmu_psize_defs[mmu_vmalloc_psize].sllp;
+ demote_segment_4k(mm, ea);
+ psize = MMU_PAGE_4K;
+ } else if (ea < VMALLOC_END) {
+ /*
+ * some driver did a non-cacheable mapping
+ * in vmalloc space, so switch vmalloc
+ * to 4k pages
+ */
+ printk(KERN_ALERT "Reducing vmalloc segment "
+ "to 4kB pages because of "
+ "non-cacheable mapping\n");
+ psize = mmu_vmalloc_psize = MMU_PAGE_4K;
+#ifdef CONFIG_SPE_BASE
+ spu_flush_all_slbs(mm);
+#endif
+ }
+ }
+ if (user_region) {
+ if (psize != get_paca()->context.user_psize) {
+ get_paca()->context.user_psize =
+ mm->context.user_psize;
slb_flush_and_rebolt();
}
+ } else if (get_paca()->vmalloc_sllp !=
+ mmu_psize_defs[mmu_vmalloc_psize].sllp) {
+ get_paca()->vmalloc_sllp =
+ mmu_psize_defs[mmu_vmalloc_psize].sllp;
+ slb_flush_and_rebolt();
}
+#endif /* CONFIG_PPC_64K_PAGES */
+
+#ifdef CONFIG_PPC_HAS_HASH_64K
if (psize == MMU_PAGE_64K)
rc = __hash_page_64K(ea, access, vsid, ptep, trap, local);
else
+#endif /* CONFIG_PPC_HAS_HASH_64K */
rc = __hash_page_4K(ea, access, vsid, ptep, trap, local);
-#endif /* CONFIG_PPC_64K_PAGES */
#ifndef CONFIG_PPC_64K_PAGES
DBG_LOW(" o-pte: %016lx\n", pte_val(*ptep));
unsigned long flags;
int local = 0;
- /* We don't want huge pages prefaulted for now
- */
- if (unlikely(in_hugepage_area(mm->context, ea)))
+ BUG_ON(REGION_ID(ea) != USER_REGION_ID);
+
+#ifdef CONFIG_PPC_MM_SLICES
+ /* We only prefault standard pages for now */
+ if (unlikely(get_slice_psize(mm, ea) != mm->context.user_psize));
return;
+#endif
DBG_LOW("hash_preload(mm=%p, mm->pgdir=%p, ea=%016lx, access=%lx,"
" trap=%lx\n", mm, mm->pgd, ea, access, trap);
- /* Get PTE, VSID, access mask */
+ /* Get Linux PTE if available */
pgdir = mm->pgd;
if (pgdir == NULL)
return;
ptep = find_linux_pte(pgdir, ea);
if (!ptep)
return;
+
+#ifdef CONFIG_PPC_64K_PAGES
+ /* If either _PAGE_4K_PFN or _PAGE_NO_CACHE is set (and we are on
+ * a 64K kernel), then we don't preload, hash_page() will take
+ * care of it once we actually try to access the page.
+ * That way we don't have to duplicate all of the logic for segment
+ * page size demotion here
+ */
+ if (pte_val(*ptep) & (_PAGE_4K_PFN | _PAGE_NO_CACHE))
+ return;
+#endif /* CONFIG_PPC_64K_PAGES */
+
+ /* Get VSID */
vsid = get_vsid(mm->context.id, ea);
- /* Hash it in */
+ /* Hash doesn't like irqs */
local_irq_save(flags);
+
+ /* Is that local to this CPU ? */
mask = cpumask_of_cpu(smp_processor_id());
if (cpus_equal(mm->cpu_vm_mask, mask))
local = 1;
-#ifndef CONFIG_PPC_64K_PAGES
- __hash_page_4K(ea, access, vsid, ptep, trap, local);
-#else
- if (mmu_ci_restrictions) {
- /* If this PTE is non-cacheable, switch to 4k */
- if (mm->context.user_psize == MMU_PAGE_64K &&
- (pte_val(*ptep) & _PAGE_NO_CACHE)) {
- mm->context.user_psize = MMU_PAGE_4K;
- mm->context.sllp = SLB_VSID_USER |
- mmu_psize_defs[MMU_PAGE_4K].sllp;
- get_paca()->context = mm->context;
- slb_flush_and_rebolt();
- }
- }
+
+ /* Hash it in */
+#ifdef CONFIG_PPC_HAS_HASH_64K
if (mm->context.user_psize == MMU_PAGE_64K)
__hash_page_64K(ea, access, vsid, ptep, trap, local);
else
- __hash_page_4K(ea, access, vsid, ptep, trap, local);
#endif /* CONFIG_PPC_64K_PAGES */
+ __hash_page_4K(ea, access, vsid, ptep, trap, local);
+
local_irq_restore(flags);
}
}
bad_page_fault(regs, address, SIGBUS);
}
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+static void kernel_map_linear_page(unsigned long vaddr, unsigned long lmi)
+{
+ unsigned long hash, hpteg, vsid = get_kernel_vsid(vaddr);
+ unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
+ unsigned long mode = _PAGE_ACCESSED | _PAGE_DIRTY |
+ _PAGE_COHERENT | PP_RWXX | HPTE_R_N;
+ int ret;
+
+ hash = hpt_hash(va, PAGE_SHIFT);
+ hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP);
+
+ ret = ppc_md.hpte_insert(hpteg, va, __pa(vaddr),
+ mode, HPTE_V_BOLTED, mmu_linear_psize);
+ BUG_ON (ret < 0);
+ spin_lock(&linear_map_hash_lock);
+ BUG_ON(linear_map_hash_slots[lmi] & 0x80);
+ linear_map_hash_slots[lmi] = ret | 0x80;
+ spin_unlock(&linear_map_hash_lock);
+}
+
+static void kernel_unmap_linear_page(unsigned long vaddr, unsigned long lmi)
+{
+ unsigned long hash, hidx, slot, vsid = get_kernel_vsid(vaddr);
+ unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
+
+ hash = hpt_hash(va, PAGE_SHIFT);
+ spin_lock(&linear_map_hash_lock);
+ BUG_ON(!(linear_map_hash_slots[lmi] & 0x80));
+ hidx = linear_map_hash_slots[lmi] & 0x7f;
+ linear_map_hash_slots[lmi] = 0;
+ spin_unlock(&linear_map_hash_lock);
+ if (hidx & _PTEIDX_SECONDARY)
+ hash = ~hash;
+ slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
+ slot += hidx & _PTEIDX_GROUP_IX;
+ ppc_md.hpte_invalidate(slot, va, mmu_linear_psize, 0);
+}
+
+void kernel_map_pages(struct page *page, int numpages, int enable)
+{
+ unsigned long flags, vaddr, lmi;
+ int i;
+
+ local_irq_save(flags);
+ for (i = 0; i < numpages; i++, page++) {
+ vaddr = (unsigned long)page_address(page);
+ lmi = __pa(vaddr) >> PAGE_SHIFT;
+ if (lmi >= linear_map_hash_count)
+ continue;
+ if (enable)
+ kernel_map_linear_page(vaddr, lmi);
+ else
+ kernel_unmap_linear_page(vaddr, lmi);
+ }
+ local_irq_restore(flags);
+}
+#endif /* CONFIG_DEBUG_PAGEALLOC */