summaryrefslogtreecommitdiff
path: root/arch/s390/mm
diff options
context:
space:
mode:
Diffstat (limited to 'arch/s390/mm')
-rw-r--r--arch/s390/mm/Makefile3
-rw-r--r--arch/s390/mm/dump_pagetables.c226
-rw-r--r--arch/s390/mm/extable.c81
-rw-r--r--arch/s390/mm/fault.c8
-rw-r--r--arch/s390/mm/gup.c48
-rw-r--r--arch/s390/mm/init.c4
-rw-r--r--arch/s390/mm/pageattr.c40
-rw-r--r--arch/s390/mm/pgtable.c114
-rw-r--r--arch/s390/mm/vmem.c47
9 files changed, 526 insertions, 45 deletions
diff --git a/arch/s390/mm/Makefile b/arch/s390/mm/Makefile
index d98fe9004a52..1bea6d1f55ab 100644
--- a/arch/s390/mm/Makefile
+++ b/arch/s390/mm/Makefile
@@ -3,7 +3,8 @@
#
obj-y := init.o fault.o extmem.o mmap.o vmem.o pgtable.o maccess.o \
- page-states.o gup.o
+ page-states.o gup.o extable.o
obj-$(CONFIG_CMM) += cmm.o
obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
obj-$(CONFIG_DEBUG_SET_MODULE_RONX) += pageattr.o
+obj-$(CONFIG_S390_PTDUMP) += dump_pagetables.o
diff --git a/arch/s390/mm/dump_pagetables.c b/arch/s390/mm/dump_pagetables.c
new file mode 100644
index 000000000000..cbc6668acb85
--- /dev/null
+++ b/arch/s390/mm/dump_pagetables.c
@@ -0,0 +1,226 @@
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <asm/sections.h>
+#include <asm/pgtable.h>
+
+static unsigned long max_addr;
+
+struct addr_marker {
+ unsigned long start_address;
+ const char *name;
+};
+
+enum address_markers_idx {
+ IDENTITY_NR = 0,
+ KERNEL_START_NR,
+ KERNEL_END_NR,
+ VMEMMAP_NR,
+ VMALLOC_NR,
+#ifdef CONFIG_64BIT
+ MODULES_NR,
+#endif
+};
+
+static struct addr_marker address_markers[] = {
+ [IDENTITY_NR] = {0, "Identity Mapping"},
+ [KERNEL_START_NR] = {(unsigned long)&_stext, "Kernel Image Start"},
+ [KERNEL_END_NR] = {(unsigned long)&_end, "Kernel Image End"},
+ [VMEMMAP_NR] = {0, "vmemmap Area"},
+ [VMALLOC_NR] = {0, "vmalloc Area"},
+#ifdef CONFIG_64BIT
+ [MODULES_NR] = {0, "Modules Area"},
+#endif
+ { -1, NULL }
+};
+
+struct pg_state {
+ int level;
+ unsigned int current_prot;
+ unsigned long start_address;
+ unsigned long current_address;
+ const struct addr_marker *marker;
+};
+
+static void print_prot(struct seq_file *m, unsigned int pr, int level)
+{
+ static const char * const level_name[] =
+ { "ASCE", "PGD", "PUD", "PMD", "PTE" };
+
+ seq_printf(m, "%s ", level_name[level]);
+ if (pr & _PAGE_INVALID)
+ seq_printf(m, "I\n");
+ else
+ seq_printf(m, "%s\n", pr & _PAGE_RO ? "RO" : "RW");
+}
+
+static void note_page(struct seq_file *m, struct pg_state *st,
+ unsigned int new_prot, int level)
+{
+ static const char units[] = "KMGTPE";
+ int width = sizeof(unsigned long) * 2;
+ const char *unit = units;
+ unsigned int prot, cur;
+ unsigned long delta;
+
+ /*
+ * If we have a "break" in the series, we need to flush the state
+ * that we have now. "break" is either changing perms, levels or
+ * address space marker.
+ */
+ prot = new_prot;
+ cur = st->current_prot;
+
+ if (!st->level) {
+ /* First entry */
+ st->current_prot = new_prot;
+ st->level = level;
+ st->marker = address_markers;
+ seq_printf(m, "---[ %s ]---\n", st->marker->name);
+ } else if (prot != cur || level != st->level ||
+ st->current_address >= st->marker[1].start_address) {
+ /* Print the actual finished series */
+ seq_printf(m, "0x%0*lx-0x%0*lx",
+ width, st->start_address,
+ width, st->current_address);
+ delta = (st->current_address - st->start_address) >> 10;
+ while (!(delta & 0x3ff) && unit[1]) {
+ delta >>= 10;
+ unit++;
+ }
+ seq_printf(m, "%9lu%c ", delta, *unit);
+ print_prot(m, st->current_prot, st->level);
+ if (st->current_address >= st->marker[1].start_address) {
+ st->marker++;
+ seq_printf(m, "---[ %s ]---\n", st->marker->name);
+ }
+ st->start_address = st->current_address;
+ st->current_prot = new_prot;
+ st->level = level;
+ }
+}
+
+/*
+ * The actual page table walker functions. In order to keep the implementation
+ * of print_prot() short, we only check and pass _PAGE_INVALID and _PAGE_RO
+ * flags to note_page() if a region, segment or page table entry is invalid or
+ * read-only.
+ * After all it's just a hint that the current level being walked contains an
+ * invalid or read-only entry.
+ */
+static void walk_pte_level(struct seq_file *m, struct pg_state *st,
+ pmd_t *pmd, unsigned long addr)
+{
+ unsigned int prot;
+ pte_t *pte;
+ int i;
+
+ for (i = 0; i < PTRS_PER_PTE && addr < max_addr; i++) {
+ st->current_address = addr;
+ pte = pte_offset_kernel(pmd, addr);
+ prot = pte_val(*pte) & (_PAGE_RO | _PAGE_INVALID);
+ note_page(m, st, prot, 4);
+ addr += PAGE_SIZE;
+ }
+}
+
+static void walk_pmd_level(struct seq_file *m, struct pg_state *st,
+ pud_t *pud, unsigned long addr)
+{
+ unsigned int prot;
+ pmd_t *pmd;
+ int i;
+
+ for (i = 0; i < PTRS_PER_PMD && addr < max_addr; i++) {
+ st->current_address = addr;
+ pmd = pmd_offset(pud, addr);
+ if (!pmd_none(*pmd)) {
+ if (pmd_large(*pmd)) {
+ prot = pmd_val(*pmd) & _SEGMENT_ENTRY_RO;
+ note_page(m, st, prot, 3);
+ } else
+ walk_pte_level(m, st, pmd, addr);
+ } else
+ note_page(m, st, _PAGE_INVALID, 3);
+ addr += PMD_SIZE;
+ }
+}
+
+static void walk_pud_level(struct seq_file *m, struct pg_state *st,
+ pgd_t *pgd, unsigned long addr)
+{
+ pud_t *pud;
+ int i;
+
+ for (i = 0; i < PTRS_PER_PUD && addr < max_addr; i++) {
+ st->current_address = addr;
+ pud = pud_offset(pgd, addr);
+ if (!pud_none(*pud))
+ walk_pmd_level(m, st, pud, addr);
+ else
+ note_page(m, st, _PAGE_INVALID, 2);
+ addr += PUD_SIZE;
+ }
+}
+
+static void walk_pgd_level(struct seq_file *m)
+{
+ unsigned long addr = 0;
+ struct pg_state st;
+ pgd_t *pgd;
+ int i;
+
+ memset(&st, 0, sizeof(st));
+ for (i = 0; i < PTRS_PER_PGD && addr < max_addr; i++) {
+ st.current_address = addr;
+ pgd = pgd_offset_k(addr);
+ if (!pgd_none(*pgd))
+ walk_pud_level(m, &st, pgd, addr);
+ else
+ note_page(m, &st, _PAGE_INVALID, 1);
+ addr += PGDIR_SIZE;
+ }
+ /* Flush out the last page */
+ st.current_address = max_addr;
+ note_page(m, &st, 0, 0);
+}
+
+static int ptdump_show(struct seq_file *m, void *v)
+{
+ walk_pgd_level(m);
+ return 0;
+}
+
+static int ptdump_open(struct inode *inode, struct file *filp)
+{
+ return single_open(filp, ptdump_show, NULL);
+}
+
+static const struct file_operations ptdump_fops = {
+ .open = ptdump_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+static int pt_dump_init(void)
+{
+ /*
+ * Figure out the maximum virtual address being accessible with the
+ * kernel ASCE. We need this to keep the page table walker functions
+ * from accessing non-existent entries.
+ */
+#ifdef CONFIG_32BIT
+ max_addr = 1UL << 31;
+#else
+ max_addr = (S390_lowcore.kernel_asce & _REGION_ENTRY_TYPE_MASK) >> 2;
+ max_addr = 1UL << (max_addr * 11 + 31);
+ address_markers[MODULES_NR].start_address = MODULES_VADDR;
+#endif
+ address_markers[VMEMMAP_NR].start_address = (unsigned long) vmemmap;
+ address_markers[VMALLOC_NR].start_address = VMALLOC_START;
+ debugfs_create_file("kernel_page_tables", 0400, NULL, NULL, &ptdump_fops);
+ return 0;
+}
+device_initcall(pt_dump_init);
diff --git a/arch/s390/mm/extable.c b/arch/s390/mm/extable.c
new file mode 100644
index 000000000000..4d1ee88864e8
--- /dev/null
+++ b/arch/s390/mm/extable.c
@@ -0,0 +1,81 @@
+#include <linux/module.h>
+#include <linux/sort.h>
+#include <asm/uaccess.h>
+
+/*
+ * Search one exception table for an entry corresponding to the
+ * given instruction address, and return the address of the entry,
+ * or NULL if none is found.
+ * We use a binary search, and thus we assume that the table is
+ * already sorted.
+ */
+const struct exception_table_entry *
+search_extable(const struct exception_table_entry *first,
+ const struct exception_table_entry *last,
+ unsigned long value)
+{
+ const struct exception_table_entry *mid;
+ unsigned long addr;
+
+ while (first <= last) {
+ mid = ((last - first) >> 1) + first;
+ addr = extable_insn(mid);
+ if (addr < value)
+ first = mid + 1;
+ else if (addr > value)
+ last = mid - 1;
+ else
+ return mid;
+ }
+ return NULL;
+}
+
+/*
+ * The exception table needs to be sorted so that the binary
+ * search that we use to find entries in it works properly.
+ * This is used both for the kernel exception table and for
+ * the exception tables of modules that get loaded.
+ *
+ */
+static int cmp_ex(const void *a, const void *b)
+{
+ const struct exception_table_entry *x = a, *y = b;
+
+ /* This compare is only valid after normalization. */
+ return x->insn - y->insn;
+}
+
+void sort_extable(struct exception_table_entry *start,
+ struct exception_table_entry *finish)
+{
+ struct exception_table_entry *p;
+ int i;
+
+ /* Normalize entries to being relative to the start of the section */
+ for (p = start, i = 0; p < finish; p++, i += 8)
+ p->insn += i;
+ sort(start, finish - start, sizeof(*start), cmp_ex, NULL);
+ /* Denormalize all entries */
+ for (p = start, i = 0; p < finish; p++, i += 8)
+ p->insn -= i;
+}
+
+#ifdef CONFIG_MODULES
+/*
+ * If the exception table is sorted, any referring to the module init
+ * will be at the beginning or the end.
+ */
+void trim_init_extable(struct module *m)
+{
+ /* Trim the beginning */
+ while (m->num_exentries &&
+ within_module_init(extable_insn(&m->extable[0]), m)) {
+ m->extable++;
+ m->num_exentries--;
+ }
+ /* Trim the end */
+ while (m->num_exentries &&
+ within_module_init(extable_insn(&m->extable[m->num_exentries-1]), m))
+ m->num_exentries--;
+}
+#endif /* CONFIG_MODULES */
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 6c013f544146..04ad4001a289 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -111,7 +111,7 @@ static inline int user_space_fault(unsigned long trans_exc_code)
if (trans_exc_code == 2)
/* Access via secondary space, set_fs setting decides */
return current->thread.mm_segment.ar4;
- if (addressing_mode == HOME_SPACE_MODE)
+ if (s390_user_mode == HOME_SPACE_MODE)
/* User space if the access has been done via home space. */
return trans_exc_code == 3;
/*
@@ -163,7 +163,7 @@ static noinline void do_no_context(struct pt_regs *regs)
/* Are we prepared to handle this kernel fault? */
fixup = search_exception_tables(regs->psw.addr & PSW_ADDR_INSN);
if (fixup) {
- regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE;
+ regs->psw.addr = extable_fixup(fixup) | PSW_ADDR_AMODE;
return;
}
@@ -367,6 +367,7 @@ retry:
/* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
* of starvation. */
flags &= ~FAULT_FLAG_ALLOW_RETRY;
+ flags |= FAULT_FLAG_TRIED;
down_read(&mm->mmap_sem);
goto retry;
}
@@ -628,9 +629,8 @@ static int __cpuinit pfault_cpu_notify(struct notifier_block *self,
struct thread_struct *thread, *next;
struct task_struct *tsk;
- switch (action) {
+ switch (action & ~CPU_TASKS_FROZEN) {
case CPU_DEAD:
- case CPU_DEAD_FROZEN:
spin_lock_irq(&pfault_lock);
list_for_each_entry_safe(thread, next, &pfault_list, list) {
thread->pfault_wait = 0;
diff --git a/arch/s390/mm/gup.c b/arch/s390/mm/gup.c
index 65cb06e2af4e..60acb93a4680 100644
--- a/arch/s390/mm/gup.c
+++ b/arch/s390/mm/gup.c
@@ -115,7 +115,16 @@ static inline int gup_pmd_range(pud_t *pudp, pud_t pud, unsigned long addr,
pmd = *pmdp;
barrier();
next = pmd_addr_end(addr, end);
- if (pmd_none(pmd))
+ /*
+ * The pmd_trans_splitting() check below explains why
+ * pmdp_splitting_flush() has to serialize with
+ * smp_call_function() against our disabled IRQs, to stop
+ * this gup-fast code from running while we set the
+ * splitting bit in the pmd. Returning zero will take
+ * the slow path that will call wait_split_huge_page()
+ * if the pmd is still in splitting state.
+ */
+ if (pmd_none(pmd) || pmd_trans_splitting(pmd))
return 0;
if (unlikely(pmd_huge(pmd))) {
if (!gup_huge_pmd(pmdp, pmd, addr, next,
@@ -154,6 +163,43 @@ static inline int gup_pud_range(pgd_t *pgdp, pgd_t pgd, unsigned long addr,
return 1;
}
+/*
+ * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
+ * back to the regular GUP.
+ */
+int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
+ struct page **pages)
+{
+ struct mm_struct *mm = current->mm;
+ unsigned long addr, len, end;
+ unsigned long next, flags;
+ pgd_t *pgdp, pgd;
+ int nr = 0;
+
+ start &= PAGE_MASK;
+ addr = start;
+ len = (unsigned long) nr_pages << PAGE_SHIFT;
+ end = start + len;
+ if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
+ (void __user *)start, len)))
+ return 0;
+
+ local_irq_save(flags);
+ pgdp = pgd_offset(mm, addr);
+ do {
+ pgd = *pgdp;
+ barrier();
+ next = pgd_addr_end(addr, end);
+ if (pgd_none(pgd))
+ break;
+ if (!gup_pud_range(pgdp, pgd, addr, next, write, pages, &nr))
+ break;
+ } while (pgdp++, addr = next, addr != end);
+ local_irq_restore(flags);
+
+ return nr;
+}
+
/**
* get_user_pages_fast() - pin user pages in memory
* @start: starting user address
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index 6adbc082618a..81e596c65dee 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -42,7 +42,7 @@ pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__((__aligned__(PAGE_SIZE)));
unsigned long empty_zero_page, zero_page_mask;
EXPORT_SYMBOL(empty_zero_page);
-static unsigned long setup_zero_pages(void)
+static unsigned long __init setup_zero_pages(void)
{
struct cpuid cpu_id;
unsigned int order;
@@ -212,7 +212,7 @@ void free_initmem(void)
}
#ifdef CONFIG_BLK_DEV_INITRD
-void free_initrd_mem(unsigned long start, unsigned long end)
+void __init free_initrd_mem(unsigned long start, unsigned long end)
{
free_init_pages("initrd memory", start, end);
}
diff --git a/arch/s390/mm/pageattr.c b/arch/s390/mm/pageattr.c
index b36537a5f43e..00be01c4b4f3 100644
--- a/arch/s390/mm/pageattr.c
+++ b/arch/s390/mm/pageattr.c
@@ -8,25 +8,38 @@
#include <asm/cacheflush.h>
#include <asm/pgtable.h>
+static pte_t *walk_page_table(unsigned long addr)
+{
+ pgd_t *pgdp;
+ pud_t *pudp;
+ pmd_t *pmdp;
+ pte_t *ptep;
+
+ pgdp = pgd_offset_k(addr);
+ if (pgd_none(*pgdp))
+ return NULL;
+ pudp = pud_offset(pgdp, addr);
+ if (pud_none(*pudp))
+ return NULL;
+ pmdp = pmd_offset(pudp, addr);
+ if (pmd_none(*pmdp) || pmd_large(*pmdp))
+ return NULL;
+ ptep = pte_offset_kernel(pmdp, addr);
+ if (pte_none(*ptep))
+ return NULL;
+ return ptep;
+}
+
static void change_page_attr(unsigned long addr, int numpages,
pte_t (*set) (pte_t))
{
pte_t *ptep, pte;
- pmd_t *pmdp;
- pud_t *pudp;
- pgd_t *pgdp;
int i;
for (i = 0; i < numpages; i++) {
- pgdp = pgd_offset(&init_mm, addr);
- pudp = pud_offset(pgdp, addr);
- pmdp = pmd_offset(pudp, addr);
- if (pmd_huge(*pmdp)) {
- WARN_ON_ONCE(1);
- continue;
- }
- ptep = pte_offset_kernel(pmdp, addr);
-
+ ptep = walk_page_table(addr);
+ if (WARN_ON_ONCE(!ptep))
+ break;
pte = *ptep;
pte = set(pte);
__ptep_ipte(addr, ptep);
@@ -40,21 +53,18 @@ int set_memory_ro(unsigned long addr, int numpages)
change_page_attr(addr, numpages, pte_wrprotect);
return 0;
}
-EXPORT_SYMBOL_GPL(set_memory_ro);
int set_memory_rw(unsigned long addr, int numpages)
{
change_page_attr(addr, numpages, pte_mkwrite);
return 0;
}
-EXPORT_SYMBOL_GPL(set_memory_rw);
/* not possible */
int set_memory_nx(unsigned long addr, int numpages)
{
return 0;
}
-EXPORT_SYMBOL_GPL(set_memory_nx);
int set_memory_x(unsigned long addr, int numpages)
{
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 18df31d1f2c9..c8188a18af05 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -609,8 +609,8 @@ static inline unsigned int atomic_xor_bits(atomic_t *v, unsigned int bits)
*/
unsigned long *page_table_alloc(struct mm_struct *mm, unsigned long vmaddr)
{
- struct page *page;
- unsigned long *table;
+ unsigned long *uninitialized_var(table);
+ struct page *uninitialized_var(page);
unsigned int mask, bit;
if (mm_has_pgste(mm))
@@ -787,6 +787,30 @@ void tlb_remove_table(struct mmu_gather *tlb, void *table)
tlb_table_flush(tlb);
}
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+void thp_split_vma(struct vm_area_struct *vma)
+{
+ unsigned long addr;
+ struct page *page;
+
+ for (addr = vma->vm_start; addr < vma->vm_end; addr += PAGE_SIZE) {
+ page = follow_page(vma, addr, FOLL_SPLIT);
+ }
+}
+
+void thp_split_mm(struct mm_struct *mm)
+{
+ struct vm_area_struct *vma = mm->mmap;
+
+ while (vma != NULL) {
+ thp_split_vma(vma);
+ vma->vm_flags &= ~VM_HUGEPAGE;
+ vma->vm_flags |= VM_NOHUGEPAGE;
+ vma = vma->vm_next;
+ }
+}
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+
/*
* switch on pgstes for its userspace process (for kvm)
*/
@@ -796,7 +820,7 @@ int s390_enable_sie(void)
struct mm_struct *mm, *old_mm;
/* Do we have switched amode? If no, we cannot do sie */
- if (addressing_mode == HOME_SPACE_MODE)
+ if (s390_user_mode == HOME_SPACE_MODE)
return -EINVAL;
/* Do we have pgstes? if yes, we are done */
@@ -824,6 +848,12 @@ int s390_enable_sie(void)
if (!mm)
return -ENOMEM;
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+ /* split thp mappings and disable thp for future mappings */
+ thp_split_mm(mm);
+ mm->def_flags |= VM_NOHUGEPAGE;
+#endif
+
/* Now lets check again if something happened */
task_lock(tsk);
if (!tsk->mm || atomic_read(&tsk->mm->mm_users) > 1 ||
@@ -866,3 +896,81 @@ bool kernel_page_present(struct page *page)
return cc == 0;
}
#endif /* CONFIG_HIBERNATION && CONFIG_DEBUG_PAGEALLOC */
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+int pmdp_clear_flush_young(struct vm_area_struct *vma, unsigned long address,
+ pmd_t *pmdp)
+{
+ VM_BUG_ON(address & ~HPAGE_PMD_MASK);
+ /* No need to flush TLB
+ * On s390 reference bits are in storage key and never in TLB */
+ return pmdp_test_and_clear_young(vma, address, pmdp);
+}
+
+int pmdp_set_access_flags(struct vm_area_struct *vma,
+ unsigned long address, pmd_t *pmdp,
+ pmd_t entry, int dirty)
+{
+ VM_BUG_ON(address & ~HPAGE_PMD_MASK);
+
+ if (pmd_same(*pmdp, entry))
+ return 0;
+ pmdp_invalidate(vma, address, pmdp);
+ set_pmd_at(vma->vm_mm, address, pmdp, entry);
+ return 1;
+}
+
+static void pmdp_splitting_flush_sync(void *arg)
+{
+ /* Simply deliver the interrupt */
+}
+
+void pmdp_splitting_flush(struct vm_area_struct *vma, unsigned long address,
+ pmd_t *pmdp)
+{
+ VM_BUG_ON(address & ~HPAGE_PMD_MASK);
+ if (!test_and_set_bit(_SEGMENT_ENTRY_SPLIT_BIT,
+ (unsigned long *) pmdp)) {
+ /* need to serialize against gup-fast (IRQ disabled) */
+ smp_call_function(pmdp_splitting_flush_sync, NULL, 1);
+ }
+}
+
+void pgtable_trans_huge_deposit(struct mm_struct *mm, pgtable_t pgtable)
+{
+ struct list_head *lh = (struct list_head *) pgtable;
+
+ assert_spin_locked(&mm->page_table_lock);
+
+ /* FIFO */
+ if (!mm->pmd_huge_pte)
+ INIT_LIST_HEAD(lh);
+ else
+ list_add(lh, (struct list_head *) mm->pmd_huge_pte);
+ mm->pmd_huge_pte = pgtable;
+}
+
+pgtable_t pgtable_trans_huge_withdraw(struct mm_struct *mm)
+{
+ struct list_head *lh;
+ pgtable_t pgtable;
+ pte_t *ptep;
+
+ assert_spin_locked(&mm->page_table_lock);
+
+ /* FIFO */
+ pgtable = mm->pmd_huge_pte;
+ lh = (struct list_head *) pgtable;
+ if (list_empty(lh))
+ mm->pmd_huge_pte = NULL;
+ else {
+ mm->pmd_huge_pte = (pgtable_t) lh->next;
+ list_del(lh);
+ }
+ ptep = (pte_t *) pgtable;
+ pte_val(*ptep) = _PAGE_TYPE_EMPTY;
+ ptep++;
+ pte_val(*ptep) = _PAGE_TYPE_EMPTY;
+ return pgtable;
+}
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index 6f896e75ab49..387c7c60b5b8 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -79,7 +79,8 @@ static pte_t __ref *vmem_pte_alloc(unsigned long address)
*/
static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
{
- unsigned long address;
+ unsigned long end = start + size;
+ unsigned long address = start;
pgd_t *pg_dir;
pud_t *pu_dir;
pmd_t *pm_dir;
@@ -87,7 +88,7 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
pte_t pte;
int ret = -ENOMEM;
- for (address = start; address < start + size; address += PAGE_SIZE) {
+ while (address < end) {
pg_dir = pgd_offset_k(address);
if (pgd_none(*pg_dir)) {
pu_dir = vmem_pud_alloc();
@@ -107,13 +108,12 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
pte = mk_pte_phys(address, __pgprot(ro ? _PAGE_RO : 0));
pm_dir = pmd_offset(pu_dir, address);
-#ifdef CONFIG_64BIT
- if (MACHINE_HAS_HPAGE && !(address & ~HPAGE_MASK) &&
- (address + HPAGE_SIZE <= start + size) &&
- (address >= HPAGE_SIZE)) {
+#if defined(CONFIG_64BIT) && !defined(CONFIG_DEBUG_PAGEALLOC)
+ if (MACHINE_HAS_EDAT1 && pmd_none(*pm_dir) && address &&
+ !(address & ~PMD_MASK) && (address + PMD_SIZE <= end)) {
pte_val(pte) |= _SEGMENT_ENTRY_LARGE;
pmd_val(*pm_dir) = pte_val(pte);
- address += HPAGE_SIZE - PAGE_SIZE;
+ address += PMD_SIZE;
continue;
}
#endif
@@ -126,10 +126,11 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
pt_dir = pte_offset_kernel(pm_dir, address);
*pt_dir = pte;
+ address += PAGE_SIZE;
}
ret = 0;
out:
- flush_tlb_kernel_range(start, start + size);
+ flush_tlb_kernel_range(start, end);
return ret;
}
@@ -139,7 +140,8 @@ out:
*/
static void vmem_remove_range(unsigned long start, unsigned long size)
{
- unsigned long address;
+ unsigned long end = start + size;
+ unsigned long address = start;
pgd_t *pg_dir;
pud_t *pu_dir;
pmd_t *pm_dir;
@@ -147,25 +149,32 @@ static void vmem_remove_range(unsigned long start, unsigned long size)
pte_t pte;
pte_val(pte) = _PAGE_TYPE_EMPTY;
- for (address = start; address < start + size; address += PAGE_SIZE) {
+ while (address < end) {
pg_dir = pgd_offset_k(address);
+ if (pgd_none(*pg_dir)) {
+ address += PGDIR_SIZE;
+ continue;
+ }
pu_dir = pud_offset(pg_dir, address);
- if (pud_none(*pu_dir))
+ if (pud_none(*pu_dir)) {
+ address += PUD_SIZE;
continue;
+ }
pm_dir = pmd_offset(pu_dir, address);
- if (pmd_none(*pm_dir))
+ if (pmd_none(*pm_dir)) {
+ address += PMD_SIZE;
continue;
-
- if (pmd_huge(*pm_dir)) {
+ }
+ if (pmd_large(*pm_dir)) {
pmd_clear(pm_dir);
- address += HPAGE_SIZE - PAGE_SIZE;
+ address += PMD_SIZE;
continue;
}
-
pt_dir = pte_offset_kernel(pm_dir, address);
*pt_dir = pte;
+ address += PAGE_SIZE;
}
- flush_tlb_kernel_range(start, start + size);
+ flush_tlb_kernel_range(start, end);
}
/*
@@ -330,8 +339,8 @@ void __init vmem_map_init(void)
unsigned long start, end;
int i;
- ro_start = ((unsigned long)&_stext) & PAGE_MASK;
- ro_end = PFN_ALIGN((unsigned long)&_eshared);
+ ro_start = PFN_ALIGN((unsigned long)&_stext);
+ ro_end = (unsigned long)&_eshared & PAGE_MASK;
for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++) {
if (memory_chunk[i].type == CHUNK_CRASHK ||
memory_chunk[i].type == CHUNK_OLDMEM)
generated by cgit v1.2.3 (git 2.39.0) at 2025-01-10 21:56:52 +0300