#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
#include <asm/e820.h>
#include <asm/processor.h>
#include <asm/sections.h>
#include <asm/uaccess.h>
#include <asm/pgalloc.h>
+#include <asm/proto.h>
+#include <asm/pat.h>
/*
* The current flushing context - we pass it instead of 5 arguments:
*/
struct cpa_data {
- unsigned long vaddr;
+ unsigned long *vaddr;
pgprot_t mask_set;
pgprot_t mask_clr;
int numpages;
- int flushtlb;
+ int flags;
+ unsigned long pfn;
+ unsigned force_split : 1;
+ int curpage;
};
+/*
+ * Serialize cpa() (for !DEBUG_PAGEALLOC which uses large identity mappings)
+ * using cpa_lock. So that we don't allow any other cpu, with stale large tlb
+ * entries change the page attribute in parallel to some other cpu
+ * splitting a large page entry along with changing the attribute.
+ */
+static DEFINE_SPINLOCK(cpa_lock);
+
+#define CPA_FLUSHTLB 1
+#define CPA_ARRAY 2
+
+#ifdef CONFIG_PROC_FS
+static unsigned long direct_pages_count[PG_LEVEL_NUM];
+
+void update_page_count(int level, unsigned long pages)
+{
+ unsigned long flags;
+
+ /* Protect against CPA */
+ spin_lock_irqsave(&pgd_lock, flags);
+ direct_pages_count[level] += pages;
+ spin_unlock_irqrestore(&pgd_lock, flags);
+}
+
+static void split_page_count(int level)
+{
+ direct_pages_count[level]--;
+ direct_pages_count[level - 1] += PTRS_PER_PTE;
+}
+
+void arch_report_meminfo(struct seq_file *m)
+{
+ seq_printf(m, "DirectMap4k: %8lu kB\n",
+ direct_pages_count[PG_LEVEL_4K] << 2);
+#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
+ seq_printf(m, "DirectMap2M: %8lu kB\n",
+ direct_pages_count[PG_LEVEL_2M] << 11);
+#else
+ seq_printf(m, "DirectMap4M: %8lu kB\n",
+ direct_pages_count[PG_LEVEL_2M] << 12);
+#endif
+#ifdef CONFIG_X86_64
+ if (direct_gbpages)
+ seq_printf(m, "DirectMap1G: %8lu kB\n",
+ direct_pages_count[PG_LEVEL_1G] << 20);
+#endif
+}
+#else
+static inline void split_page_count(int level) { }
+#endif
+
+#ifdef CONFIG_X86_64
+
+static inline unsigned long highmap_start_pfn(void)
+{
+ return __pa(_text) >> PAGE_SHIFT;
+}
+
+static inline unsigned long highmap_end_pfn(void)
+{
+ return __pa(roundup((unsigned long)_end, PMD_SIZE)) >> PAGE_SHIFT;
+}
+
+#endif
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+# define debug_pagealloc 1
+#else
+# define debug_pagealloc 0
+#endif
+
static inline int
within(unsigned long addr, unsigned long start, unsigned long end)
{
{
BUG_ON(irqs_disabled());
- on_each_cpu(__cpa_flush_all, (void *) cache, 1, 1);
+ on_each_cpu(__cpa_flush_all, (void *) cache, 1);
}
static void __cpa_flush_range(void *arg)
BUG_ON(irqs_disabled());
WARN_ON(PAGE_ALIGN(start) != start);
- on_each_cpu(__cpa_flush_range, NULL, 1, 1);
+ on_each_cpu(__cpa_flush_range, NULL, 1);
if (!cache)
return;
}
}
-#define HIGH_MAP_START __START_KERNEL_map
-#define HIGH_MAP_END (__START_KERNEL_map + KERNEL_TEXT_SIZE)
+static void cpa_flush_array(unsigned long *start, int numpages, int cache)
+{
+ unsigned int i, level;
+ unsigned long *addr;
+
+ BUG_ON(irqs_disabled());
+ on_each_cpu(__cpa_flush_range, NULL, 1);
-/*
- * Converts a virtual address to a X86-64 highmap address
- */
-static unsigned long virt_to_highmap(void *address)
-{
-#ifdef CONFIG_X86_64
- return __pa((unsigned long)address) + HIGH_MAP_START - phys_base;
-#else
- return (unsigned long)address;
-#endif
+ if (!cache)
+ return;
+
+ /* 4M threshold */
+ if (numpages >= 1024) {
+ if (boot_cpu_data.x86_model >= 4)
+ wbinvd();
+ return;
+ }
+ /*
+ * We only need to flush on one CPU,
+ * clflush is a MESI-coherent instruction that
+ * will cause all other CPUs to flush the same
+ * cachelines:
+ */
+ for (i = 0, addr = start; i < numpages; i++, addr++) {
+ pte_t *pte = lookup_address(*addr, &level);
+
+ /*
+ * Only flush present addresses:
+ */
+ if (pte && (pte_val(*pte) & _PAGE_PRESENT))
+ clflush_cache_range((void *) *addr, PAGE_SIZE);
+ }
}
/*
* right (again, ioremap() on BIOS memory is not uncommon) so this function
* checks and fixes these known static required protection bits.
*/
-static inline pgprot_t static_protections(pgprot_t prot, unsigned long address)
+static inline pgprot_t static_protections(pgprot_t prot, unsigned long address,
+ unsigned long pfn)
{
pgprot_t forbidden = __pgprot(0);
* The BIOS area between 640k and 1Mb needs to be executable for
* PCI BIOS based config access (CONFIG_PCI_GOBIOS) support.
*/
- if (within(__pa(address), BIOS_BEGIN, BIOS_END))
+ if (within(pfn, BIOS_BEGIN >> PAGE_SHIFT, BIOS_END >> PAGE_SHIFT))
pgprot_val(forbidden) |= _PAGE_NX;
/*
* The kernel text needs to be executable for obvious reasons
- * Does not cover __inittext since that is gone later on
+ * Does not cover __inittext since that is gone later on. On
+ * 64bit we do not enforce !NX on the low mapping
*/
if (within(address, (unsigned long)_text, (unsigned long)_etext))
pgprot_val(forbidden) |= _PAGE_NX;
- /*
- * Do the same for the x86-64 high kernel mapping
- */
- if (within(address, virt_to_highmap(_text), virt_to_highmap(_etext)))
- pgprot_val(forbidden) |= _PAGE_NX;
- /* The .rodata section needs to be read-only */
- if (within(address, (unsigned long)__start_rodata,
- (unsigned long)__end_rodata))
- pgprot_val(forbidden) |= _PAGE_RW;
/*
- * Do the same for the x86-64 high kernel mapping
+ * The .rodata section needs to be read-only. Using the pfn
+ * catches all aliases.
*/
- if (within(address, virt_to_highmap(__start_rodata),
- virt_to_highmap(__end_rodata)))
+ if (within(pfn, __pa((unsigned long)__start_rodata) >> PAGE_SHIFT,
+ __pa((unsigned long)__end_rodata) >> PAGE_SHIFT))
pgprot_val(forbidden) |= _PAGE_RW;
prot = __pgprot(pgprot_val(prot) & ~pgprot_val(forbidden));
return pte_offset_kernel(pmd, address);
}
+EXPORT_SYMBOL_GPL(lookup_address);
/*
* Set the new pmd in all the pgds we know about:
try_preserve_large_page(pte_t *kpte, unsigned long address,
struct cpa_data *cpa)
{
- unsigned long nextpage_addr, numpages, pmask, psize, flags, addr;
+ unsigned long nextpage_addr, numpages, pmask, psize, flags, addr, pfn;
pte_t new_pte, old_pte, *tmp;
pgprot_t old_prot, new_prot;
int i, do_split = 1;
unsigned int level;
+ if (cpa->force_split)
+ return 1;
+
spin_lock_irqsave(&pgd_lock, flags);
/*
* Check for races, another CPU might have split this page
break;
#ifdef CONFIG_X86_64
case PG_LEVEL_1G:
- psize = PMD_PAGE_SIZE;
- pmask = PMD_PAGE_MASK;
+ psize = PUD_PAGE_SIZE;
+ pmask = PUD_PAGE_MASK;
break;
#endif
default:
pgprot_val(new_prot) &= ~pgprot_val(cpa->mask_clr);
pgprot_val(new_prot) |= pgprot_val(cpa->mask_set);
- new_prot = static_protections(new_prot, address);
+
+ /*
+ * old_pte points to the large page base address. So we need
+ * to add the offset of the virtual address:
+ */
+ pfn = pte_pfn(old_pte) + ((address & (psize - 1)) >> PAGE_SHIFT);
+ cpa->pfn = pfn;
+
+ new_prot = static_protections(new_prot, address, pfn);
/*
* We need to check the full range, whether
* the pages in the range we try to preserve:
*/
addr = address + PAGE_SIZE;
- for (i = 1; i < cpa->numpages; i++, addr += PAGE_SIZE) {
- pgprot_t chk_prot = static_protections(new_prot, addr);
+ pfn++;
+ for (i = 1; i < cpa->numpages; i++, addr += PAGE_SIZE, pfn++) {
+ pgprot_t chk_prot = static_protections(new_prot, addr, pfn);
if (pgprot_val(chk_prot) != pgprot_val(new_prot))
goto out_unlock;
*/
new_pte = pfn_pte(pte_pfn(old_pte), canon_pgprot(new_prot));
__set_pmd_pte(kpte, address, new_pte);
- cpa->flushtlb = 1;
+ cpa->flags |= CPA_FLUSHTLB;
do_split = 0;
}
return do_split;
}
-static LIST_HEAD(page_pool);
-static unsigned long pool_size, pool_pages, pool_low;
-static unsigned long pool_used, pool_failed, pool_refill;
-
-static void cpa_fill_pool(void)
-{
- struct page *p;
- gfp_t gfp = GFP_KERNEL;
-
- /* Do not allocate from interrupt context */
- if (in_irq() || irqs_disabled())
- return;
- /*
- * Check unlocked. I does not matter when we have one more
- * page in the pool. The bit lock avoids recursive pool
- * allocations:
- */
- if (pool_pages >= pool_size || test_and_set_bit_lock(0, &pool_refill))
- return;
-
-#ifdef CONFIG_DEBUG_PAGEALLOC
- /*
- * We could do:
- * gfp = in_atomic() ? GFP_ATOMIC : GFP_KERNEL;
- * but this fails on !PREEMPT kernels
- */
- gfp = GFP_ATOMIC | __GFP_NORETRY | __GFP_NOWARN;
-#endif
-
- while (pool_pages < pool_size) {
- p = alloc_pages(gfp, 0);
- if (!p) {
- pool_failed++;
- break;
- }
- spin_lock_irq(&pgd_lock);
- list_add(&p->lru, &page_pool);
- pool_pages++;
- spin_unlock_irq(&pgd_lock);
- }
- clear_bit_unlock(0, &pool_refill);
-}
-
-#define SHIFT_MB (20 - PAGE_SHIFT)
-#define ROUND_MB_GB ((1 << 10) - 1)
-#define SHIFT_MB_GB 10
-#define POOL_PAGES_PER_GB 16
-
-void __init cpa_init(void)
-{
- struct sysinfo si;
- unsigned long gb;
-
- si_meminfo(&si);
- /*
- * Calculate the number of pool pages:
- *
- * Convert totalram (nr of pages) to MiB and round to the next
- * GiB. Shift MiB to Gib and multiply the result by
- * POOL_PAGES_PER_GB:
- */
- gb = ((si.totalram >> SHIFT_MB) + ROUND_MB_GB) >> SHIFT_MB_GB;
- pool_size = POOL_PAGES_PER_GB * gb;
- pool_low = pool_size;
-
- cpa_fill_pool();
- printk(KERN_DEBUG
- "CPA: page pool initialized %lu of %lu pages preallocated\n",
- pool_pages, pool_size);
-}
-
static int split_large_page(pte_t *kpte, unsigned long address)
{
unsigned long flags, pfn, pfninc = 1;
pgprot_t ref_prot;
struct page *base;
- /*
- * Get a page from the pool. The pool list is protected by the
- * pgd_lock, which we have to take anyway for the split
- * operation:
- */
- spin_lock_irqsave(&pgd_lock, flags);
- if (list_empty(&page_pool)) {
- spin_unlock_irqrestore(&pgd_lock, flags);
+ if (!debug_pagealloc)
+ spin_unlock(&cpa_lock);
+ base = alloc_pages(GFP_KERNEL, 0);
+ if (!debug_pagealloc)
+ spin_lock(&cpa_lock);
+ if (!base)
return -ENOMEM;
- }
-
- base = list_first_entry(&page_pool, struct page, lru);
- list_del(&base->lru);
- pool_pages--;
-
- if (pool_pages < pool_low)
- pool_low = pool_pages;
+ spin_lock_irqsave(&pgd_lock, flags);
/*
* Check for races, another CPU might have split this page
* up for us already:
goto out_unlock;
pbase = (pte_t *)page_address(base);
-#ifdef CONFIG_X86_32
- paravirt_alloc_pt(&init_mm, page_to_pfn(base));
-#endif
+ paravirt_alloc_pte(&init_mm, page_to_pfn(base));
ref_prot = pte_pgprot(pte_clrhuge(*kpte));
+ /*
+ * If we ever want to utilize the PAT bit, we need to
+ * update this function to make sure it's converted from
+ * bit 12 to bit 7 when we cross from the 2MB level to
+ * the 4K level:
+ */
+ WARN_ON_ONCE(pgprot_val(ref_prot) & _PAGE_PAT_LARGE);
#ifdef CONFIG_X86_64
if (level == PG_LEVEL_1G) {
for (i = 0; i < PTRS_PER_PTE; i++, pfn += pfninc)
set_pte(&pbase[i], pfn_pte(pfn, ref_prot));
+ if (address >= (unsigned long)__va(0) &&
+ address < (unsigned long)__va(max_low_pfn_mapped << PAGE_SHIFT))
+ split_page_count(level);
+
+#ifdef CONFIG_X86_64
+ if (address >= (unsigned long)__va(1UL<<32) &&
+ address < (unsigned long)__va(max_pfn_mapped << PAGE_SHIFT))
+ split_page_count(level);
+#endif
+
/*
- * Install the new, split up pagetable. Important details here:
+ * Install the new, split up pagetable.
*
- * On Intel the NX bit of all levels must be cleared to make a
- * page executable. See section 4.13.2 of Intel 64 and IA-32
- * Architectures Software Developer's Manual).
+ * We use the standard kernel pagetable protections for the new
+ * pagetable protections, the actual ptes set above control the
+ * primary protection behavior:
+ */
+ __set_pmd_pte(kpte, address, mk_pte(base, __pgprot(_KERNPG_TABLE)));
+
+ /*
+ * Intel Atom errata AAH41 workaround.
*
- * Mark the entry present. The current mapping might be
- * set to not present, which we preserved above.
+ * The real fix should be in hw or in a microcode update, but
+ * we also probabilistically try to reduce the window of having
+ * a large TLB mixed with 4K TLBs while instruction fetches are
+ * going on.
*/
- ref_prot = pte_pgprot(pte_mkexec(pte_clrhuge(*kpte)));
- pgprot_val(ref_prot) |= _PAGE_PRESENT;
- __set_pmd_pte(kpte, address, mk_pte(base, ref_prot));
+ __flush_tlb_all();
+
base = NULL;
out_unlock:
* If we dropped out via the lookup_address check under
* pgd_lock then stick the page back into the pool:
*/
- if (base) {
- list_add(&base->lru, &page_pool);
- pool_pages++;
- } else
- pool_used++;
+ if (base)
+ __free_page(base);
spin_unlock_irqrestore(&pgd_lock, flags);
return 0;
}
-static int __change_page_attr(unsigned long address, struct cpa_data *cpa)
+static int __cpa_process_fault(struct cpa_data *cpa, unsigned long vaddr,
+ int primary)
+{
+ /*
+ * Ignore all non primary paths.
+ */
+ if (!primary)
+ return 0;
+
+ /*
+ * Ignore the NULL PTE for kernel identity mapping, as it is expected
+ * to have holes.
+ * Also set numpages to '1' indicating that we processed cpa req for
+ * one virtual address page and its pfn. TBD: numpages can be set based
+ * on the initial value and the level returned by lookup_address().
+ */
+ if (within(vaddr, PAGE_OFFSET,
+ PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT))) {
+ cpa->numpages = 1;
+ cpa->pfn = __pa(vaddr) >> PAGE_SHIFT;
+ return 0;
+ } else {
+ WARN(1, KERN_WARNING "CPA: called for zero pte. "
+ "vaddr = %lx cpa->vaddr = %lx\n", vaddr,
+ *cpa->vaddr);
+
+ return -EFAULT;
+ }
+}
+
+static int __change_page_attr(struct cpa_data *cpa, int primary)
{
+ unsigned long address;
int do_split, err;
unsigned int level;
- struct page *kpte_page;
- pte_t *kpte;
+ pte_t *kpte, old_pte;
+ if (cpa->flags & CPA_ARRAY)
+ address = cpa->vaddr[cpa->curpage];
+ else
+ address = *cpa->vaddr;
repeat:
kpte = lookup_address(address, &level);
if (!kpte)
- return -EINVAL;
+ return __cpa_process_fault(cpa, address, primary);
- kpte_page = virt_to_page(kpte);
- BUG_ON(PageLRU(kpte_page));
- BUG_ON(PageCompound(kpte_page));
+ old_pte = *kpte;
+ if (!pte_val(old_pte))
+ return __cpa_process_fault(cpa, address, primary);
if (level == PG_LEVEL_4K) {
- pte_t new_pte, old_pte = *kpte;
+ pte_t new_pte;
pgprot_t new_prot = pte_pgprot(old_pte);
-
- if(!pte_val(old_pte)) {
- printk(KERN_WARNING "CPA: called for zero pte. "
- "vaddr = %lx cpa->vaddr = %lx\n", address,
- cpa->vaddr);
- WARN_ON(1);
- return -EINVAL;
- }
+ unsigned long pfn = pte_pfn(old_pte);
pgprot_val(new_prot) &= ~pgprot_val(cpa->mask_clr);
pgprot_val(new_prot) |= pgprot_val(cpa->mask_set);
- new_prot = static_protections(new_prot, address);
+ new_prot = static_protections(new_prot, address, pfn);
/*
* We need to keep the pfn from the existing PTE,
* after all we're only going to change it's attributes
* not the memory it points to
*/
- new_pte = pfn_pte(pte_pfn(old_pte), canon_pgprot(new_prot));
-
+ new_pte = pfn_pte(pfn, canon_pgprot(new_prot));
+ cpa->pfn = pfn;
/*
* Do we really change anything ?
*/
if (pte_val(old_pte) != pte_val(new_pte)) {
set_pte_atomic(kpte, new_pte);
- cpa->flushtlb = 1;
+ cpa->flags |= CPA_FLUSHTLB;
}
cpa->numpages = 1;
return 0;
*/
err = split_large_page(kpte, address);
if (!err) {
- cpa->flushtlb = 1;
+ /*
+ * Do a global flush tlb after splitting the large page
+ * and before we do the actual change page attribute in the PTE.
+ *
+ * With out this, we violate the TLB application note, that says
+ * "The TLBs may contain both ordinary and large-page
+ * translations for a 4-KByte range of linear addresses. This
+ * may occur if software modifies the paging structures so that
+ * the page size used for the address range changes. If the two
+ * translations differ with respect to page frame or attributes
+ * (e.g., permissions), processor behavior is undefined and may
+ * be implementation-specific."
+ *
+ * We do this global tlb flush inside the cpa_lock, so that we
+ * don't allow any other cpu, with stale tlb entries change the
+ * page attribute in parallel, that also falls into the
+ * just split large page entry.
+ */
+ flush_tlb_all();
goto repeat;
}
return err;
}
-/**
- * change_page_attr_addr - Change page table attributes in linear mapping
- * @address: Virtual address in linear mapping.
- * @prot: New page table attribute (PAGE_*)
- *
- * Change page attributes of a page in the direct mapping. This is a variant
- * of change_page_attr() that also works on memory holes that do not have
- * mem_map entry (pfn_valid() is false).
- *
- * See change_page_attr() documentation for more details.
- *
- * Modules and drivers should use the set_memory_* APIs instead.
- */
-static int change_page_attr_addr(struct cpa_data *cpa)
+static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias);
+
+static int cpa_process_alias(struct cpa_data *cpa)
{
- int err;
- unsigned long address = cpa->vaddr;
+ struct cpa_data alias_cpa;
+ int ret = 0;
+ unsigned long temp_cpa_vaddr, vaddr;
-#ifdef CONFIG_X86_64
- unsigned long phys_addr = __pa(address);
+ if (cpa->pfn >= max_pfn_mapped)
+ return 0;
+#ifdef CONFIG_X86_64
+ if (cpa->pfn >= max_low_pfn_mapped && cpa->pfn < (1UL<<(32-PAGE_SHIFT)))
+ return 0;
+#endif
/*
- * If we are inside the high mapped kernel range, then we
- * fixup the low mapping first. __va() returns the virtual
- * address in the linear mapping:
+ * No need to redo, when the primary call touched the direct
+ * mapping already:
*/
- if (within(address, HIGH_MAP_START, HIGH_MAP_END))
- address = (unsigned long) __va(phys_addr);
-#endif
+ if (cpa->flags & CPA_ARRAY)
+ vaddr = cpa->vaddr[cpa->curpage];
+ else
+ vaddr = *cpa->vaddr;
- err = __change_page_attr(address, cpa);
- if (err)
- return err;
+ if (!(within(vaddr, PAGE_OFFSET,
+ PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT)))) {
+
+ alias_cpa = *cpa;
+ temp_cpa_vaddr = (unsigned long) __va(cpa->pfn << PAGE_SHIFT);
+ alias_cpa.vaddr = &temp_cpa_vaddr;
+ alias_cpa.flags &= ~CPA_ARRAY;
+
+
+ ret = __change_page_attr_set_clr(&alias_cpa, 0);
+ }
#ifdef CONFIG_X86_64
+ if (ret)
+ return ret;
+ /*
+ * No need to redo, when the primary call touched the high
+ * mapping already:
+ */
+ if (within(vaddr, (unsigned long) _text, (unsigned long) _end))
+ return 0;
+
/*
* If the physical address is inside the kernel map, we need
* to touch the high mapped kernel as well:
*/
- if (within(phys_addr, 0, KERNEL_TEXT_SIZE)) {
- /*
- * Calc the high mapping address. See __phys_addr()
- * for the non obvious details.
- *
- * Note that NX and other required permissions are
- * checked in static_protections().
- */
- address = phys_addr + HIGH_MAP_START - phys_base;
+ if (!within(cpa->pfn, highmap_start_pfn(), highmap_end_pfn()))
+ return 0;
- /*
- * Our high aliases are imprecise, because we check
- * everything between 0 and KERNEL_TEXT_SIZE, so do
- * not propagate lookup failures back to users:
- */
- __change_page_attr(address, cpa);
- }
+ alias_cpa = *cpa;
+ temp_cpa_vaddr = (cpa->pfn << PAGE_SHIFT) + __START_KERNEL_map - phys_base;
+ alias_cpa.vaddr = &temp_cpa_vaddr;
+ alias_cpa.flags &= ~CPA_ARRAY;
+
+ /*
+ * The high mapping range is imprecise, so ignore the return value.
+ */
+ __change_page_attr_set_clr(&alias_cpa, 0);
#endif
- return err;
+ return ret;
}
-static int __change_page_attr_set_clr(struct cpa_data *cpa)
+static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias)
{
int ret, numpages = cpa->numpages;
* preservation check.
*/
cpa->numpages = numpages;
- ret = change_page_attr_addr(cpa);
+ /* for array changes, we can't use large page */
+ if (cpa->flags & CPA_ARRAY)
+ cpa->numpages = 1;
+
+ if (!debug_pagealloc)
+ spin_lock(&cpa_lock);
+ ret = __change_page_attr(cpa, checkalias);
+ if (!debug_pagealloc)
+ spin_unlock(&cpa_lock);
if (ret)
return ret;
+ if (checkalias) {
+ ret = cpa_process_alias(cpa);
+ if (ret)
+ return ret;
+ }
+
/*
* Adjust the number of pages with the result of the
* CPA operation. Either a large page has been
*/
BUG_ON(cpa->numpages > numpages);
numpages -= cpa->numpages;
- cpa->vaddr += cpa->numpages * PAGE_SIZE;
+ if (cpa->flags & CPA_ARRAY)
+ cpa->curpage++;
+ else
+ *cpa->vaddr += cpa->numpages * PAGE_SIZE;
+
}
return 0;
}
(_PAGE_PAT | _PAGE_PAT_LARGE | _PAGE_PWT | _PAGE_PCD);
}
-static int change_page_attr_set_clr(unsigned long addr, int numpages,
- pgprot_t mask_set, pgprot_t mask_clr)
+static int change_page_attr_set_clr(unsigned long *addr, int numpages,
+ pgprot_t mask_set, pgprot_t mask_clr,
+ int force_split, int array)
{
struct cpa_data cpa;
- int ret, cache;
+ int ret, cache, checkalias;
/*
* Check, if we are requested to change a not supported
*/
mask_set = canon_pgprot(mask_set);
mask_clr = canon_pgprot(mask_clr);
- if (!pgprot_val(mask_set) && !pgprot_val(mask_clr))
+ if (!pgprot_val(mask_set) && !pgprot_val(mask_clr) && !force_split)
return 0;
+ /* Ensure we are PAGE_SIZE aligned */
+ if (!array) {
+ if (*addr & ~PAGE_MASK) {
+ *addr &= PAGE_MASK;
+ /*
+ * People should not be passing in unaligned addresses:
+ */
+ WARN_ON_ONCE(1);
+ }
+ } else {
+ int i;
+ for (i = 0; i < numpages; i++) {
+ if (addr[i] & ~PAGE_MASK) {
+ addr[i] &= PAGE_MASK;
+ WARN_ON_ONCE(1);
+ }
+ }
+ }
+
+ /* Must avoid aliasing mappings in the highmem code */
+ kmap_flush_unused();
+
+ vm_unmap_aliases();
+
cpa.vaddr = addr;
cpa.numpages = numpages;
cpa.mask_set = mask_set;
cpa.mask_clr = mask_clr;
- cpa.flushtlb = 0;
+ cpa.flags = 0;
+ cpa.curpage = 0;
+ cpa.force_split = force_split;
- ret = __change_page_attr_set_clr(&cpa);
+ if (array)
+ cpa.flags |= CPA_ARRAY;
+
+ /* No alias checking for _NX bit modifications */
+ checkalias = (pgprot_val(mask_set) | pgprot_val(mask_clr)) != _PAGE_NX;
+
+ ret = __change_page_attr_set_clr(&cpa, checkalias);
/*
* Check whether we really changed something:
*/
- if (!cpa.flushtlb)
+ if (!(cpa.flags & CPA_FLUSHTLB))
goto out;
/*
* error case we fall back to cpa_flush_all (which uses
* wbindv):
*/
- if (!ret && cpu_has_clflush)
- cpa_flush_range(addr, numpages, cache);
- else
+ if (!ret && cpu_has_clflush) {
+ if (cpa.flags & CPA_ARRAY)
+ cpa_flush_array(addr, numpages, cache);
+ else
+ cpa_flush_range(*addr, numpages, cache);
+ } else
cpa_flush_all(cache);
out:
- cpa_fill_pool();
return ret;
}
-static inline int change_page_attr_set(unsigned long addr, int numpages,
- pgprot_t mask)
+static inline int change_page_attr_set(unsigned long *addr, int numpages,
+ pgprot_t mask, int array)
+{
+ return change_page_attr_set_clr(addr, numpages, mask, __pgprot(0), 0,
+ array);
+}
+
+static inline int change_page_attr_clear(unsigned long *addr, int numpages,
+ pgprot_t mask, int array)
{
- return change_page_attr_set_clr(addr, numpages, mask, __pgprot(0));
+ return change_page_attr_set_clr(addr, numpages, __pgprot(0), mask, 0,
+ array);
}
-static inline int change_page_attr_clear(unsigned long addr, int numpages,
- pgprot_t mask)
+int _set_memory_uc(unsigned long addr, int numpages)
{
- return change_page_attr_set_clr(addr, numpages, __pgprot(0), mask);
+ /*
+ * for now UC MINUS. see comments in ioremap_nocache()
+ */
+ return change_page_attr_set(&addr, numpages,
+ __pgprot(_PAGE_CACHE_UC_MINUS), 0);
}
int set_memory_uc(unsigned long addr, int numpages)
{
- return change_page_attr_set(addr, numpages,
- __pgprot(_PAGE_PCD | _PAGE_PWT));
+ /*
+ * for now UC MINUS. see comments in ioremap_nocache()
+ */
+ if (reserve_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
+ _PAGE_CACHE_UC_MINUS, NULL))
+ return -EINVAL;
+
+ return _set_memory_uc(addr, numpages);
}
EXPORT_SYMBOL(set_memory_uc);
+int set_memory_array_uc(unsigned long *addr, int addrinarray)
+{
+ unsigned long start;
+ unsigned long end;
+ int i;
+ /*
+ * for now UC MINUS. see comments in ioremap_nocache()
+ */
+ for (i = 0; i < addrinarray; i++) {
+ start = __pa(addr[i]);
+ for (end = start + PAGE_SIZE; i < addrinarray - 1; end += PAGE_SIZE) {
+ if (end != __pa(addr[i + 1]))
+ break;
+ i++;
+ }
+ if (reserve_memtype(start, end, _PAGE_CACHE_UC_MINUS, NULL))
+ goto out;
+ }
+
+ return change_page_attr_set(addr, addrinarray,
+ __pgprot(_PAGE_CACHE_UC_MINUS), 1);
+out:
+ for (i = 0; i < addrinarray; i++) {
+ unsigned long tmp = __pa(addr[i]);
+
+ if (tmp == start)
+ break;
+ for (end = tmp + PAGE_SIZE; i < addrinarray - 1; end += PAGE_SIZE) {
+ if (end != __pa(addr[i + 1]))
+ break;
+ i++;
+ }
+ free_memtype(tmp, end);
+ }
+ return -EINVAL;
+}
+EXPORT_SYMBOL(set_memory_array_uc);
+
+int _set_memory_wc(unsigned long addr, int numpages)
+{
+ return change_page_attr_set(&addr, numpages,
+ __pgprot(_PAGE_CACHE_WC), 0);
+}
+
+int set_memory_wc(unsigned long addr, int numpages)
+{
+ if (!pat_enabled)
+ return set_memory_uc(addr, numpages);
+
+ if (reserve_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
+ _PAGE_CACHE_WC, NULL))
+ return -EINVAL;
+
+ return _set_memory_wc(addr, numpages);
+}
+EXPORT_SYMBOL(set_memory_wc);
+
+int _set_memory_wb(unsigned long addr, int numpages)
+{
+ return change_page_attr_clear(&addr, numpages,
+ __pgprot(_PAGE_CACHE_MASK), 0);
+}
+
int set_memory_wb(unsigned long addr, int numpages)
{
- return change_page_attr_clear(addr, numpages,
- __pgprot(_PAGE_PCD | _PAGE_PWT));
+ free_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
+
+ return _set_memory_wb(addr, numpages);
}
EXPORT_SYMBOL(set_memory_wb);
+int set_memory_array_wb(unsigned long *addr, int addrinarray)
+{
+ int i;
+
+ for (i = 0; i < addrinarray; i++) {
+ unsigned long start = __pa(addr[i]);
+ unsigned long end;
+
+ for (end = start + PAGE_SIZE; i < addrinarray - 1; end += PAGE_SIZE) {
+ if (end != __pa(addr[i + 1]))
+ break;
+ i++;
+ }
+ free_memtype(start, end);
+ }
+ return change_page_attr_clear(addr, addrinarray,
+ __pgprot(_PAGE_CACHE_MASK), 1);
+}
+EXPORT_SYMBOL(set_memory_array_wb);
+
int set_memory_x(unsigned long addr, int numpages)
{
- return change_page_attr_clear(addr, numpages, __pgprot(_PAGE_NX));
+ return change_page_attr_clear(&addr, numpages, __pgprot(_PAGE_NX), 0);
}
EXPORT_SYMBOL(set_memory_x);
int set_memory_nx(unsigned long addr, int numpages)
{
- return change_page_attr_set(addr, numpages, __pgprot(_PAGE_NX));
+ return change_page_attr_set(&addr, numpages, __pgprot(_PAGE_NX), 0);
}
EXPORT_SYMBOL(set_memory_nx);
int set_memory_ro(unsigned long addr, int numpages)
{
- return change_page_attr_clear(addr, numpages, __pgprot(_PAGE_RW));
+ return change_page_attr_clear(&addr, numpages, __pgprot(_PAGE_RW), 0);
}
+EXPORT_SYMBOL_GPL(set_memory_ro);
int set_memory_rw(unsigned long addr, int numpages)
{
- return change_page_attr_set(addr, numpages, __pgprot(_PAGE_RW));
+ return change_page_attr_set(&addr, numpages, __pgprot(_PAGE_RW), 0);
}
+EXPORT_SYMBOL_GPL(set_memory_rw);
int set_memory_np(unsigned long addr, int numpages)
{
- return change_page_attr_clear(addr, numpages, __pgprot(_PAGE_PRESENT));
+ return change_page_attr_clear(&addr, numpages, __pgprot(_PAGE_PRESENT), 0);
+}
+
+int set_memory_4k(unsigned long addr, int numpages)
+{
+ return change_page_attr_set_clr(&addr, numpages, __pgprot(0),
+ __pgprot(0), 1, 0);
}
int set_pages_uc(struct page *page, int numpages)
static int __set_pages_p(struct page *page, int numpages)
{
- struct cpa_data cpa = { .vaddr = (unsigned long) page_address(page),
+ unsigned long tempaddr = (unsigned long) page_address(page);
+ struct cpa_data cpa = { .vaddr = &tempaddr,
.numpages = numpages,
.mask_set = __pgprot(_PAGE_PRESENT | _PAGE_RW),
- .mask_clr = __pgprot(0)};
+ .mask_clr = __pgprot(0),
+ .flags = 0};
- return __change_page_attr_set_clr(&cpa);
+ /*
+ * No alias checking needed for setting present flag. otherwise,
+ * we may need to break large pages for 64-bit kernel text
+ * mappings (this adds to complexity if we want to do this from
+ * atomic context especially). Let's keep it simple!
+ */
+ return __change_page_attr_set_clr(&cpa, 0);
}
static int __set_pages_np(struct page *page, int numpages)
{
- struct cpa_data cpa = { .vaddr = (unsigned long) page_address(page),
+ unsigned long tempaddr = (unsigned long) page_address(page);
+ struct cpa_data cpa = { .vaddr = &tempaddr,
.numpages = numpages,
.mask_set = __pgprot(0),
- .mask_clr = __pgprot(_PAGE_PRESENT | _PAGE_RW)};
+ .mask_clr = __pgprot(_PAGE_PRESENT | _PAGE_RW),
+ .flags = 0};
- return __change_page_attr_set_clr(&cpa);
+ /*
+ * No alias checking needed for setting not present flag. otherwise,
+ * we may need to break large pages for 64-bit kernel text
+ * mappings (this adds to complexity if we want to do this from
+ * atomic context especially). Let's keep it simple!
+ */
+ return __change_page_attr_set_clr(&cpa, 0);
}
void kernel_map_pages(struct page *page, int numpages, int enable)
return;
/*
- * The return value is ignored - the calls cannot fail,
- * large pages are disabled at boot time:
+ * The return value is ignored as the calls cannot fail.
+ * Large pages for identity mappings are not used at boot time
+ * and hence no memory allocations during large page split.
*/
if (enable)
__set_pages_p(page, numpages);
* but that can deadlock->flush only current cpu:
*/
__flush_tlb_all();
+}
- /*
- * Try to refill the page pool here. We can do this only after
- * the tlb flush.
- */
- cpa_fill_pool();
+#ifdef CONFIG_HIBERNATION
+
+bool kernel_page_present(struct page *page)
+{
+ unsigned int level;
+ pte_t *pte;
+
+ if (PageHighMem(page))
+ return false;
+
+ pte = lookup_address((unsigned long)page_address(page), &level);
+ return (pte_val(*pte) & _PAGE_PRESENT);
}
-#endif
+
+#endif /* CONFIG_HIBERNATION */
+
+#endif /* CONFIG_DEBUG_PAGEALLOC */
/*
* The testcases use internal knowledge of the implementation that shouldn't