summaryrefslogtreecommitdiff
path: root/arch/s390/mm
diff options
context:
space:
mode:
Diffstat (limited to 'arch/s390/mm')
-rw-r--r--arch/s390/mm/fault.c62
-rw-r--r--arch/s390/mm/init.c2
-rw-r--r--arch/s390/mm/maccess.c4
-rw-r--r--arch/s390/mm/pgtable.c281
4 files changed, 159 insertions, 190 deletions
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index a0f9e730f26a..fe103e891e7a 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -34,7 +34,7 @@
#include <asm/asm-offsets.h>
#include <asm/system.h>
#include <asm/pgtable.h>
-#include <asm/s390_ext.h>
+#include <asm/irq.h>
#include <asm/mmu_context.h>
#include <asm/compat.h>
#include "../kernel/entry.h"
@@ -245,9 +245,12 @@ static noinline void do_fault_error(struct pt_regs *regs, long int_code,
do_no_context(regs, int_code, trans_exc_code);
break;
default: /* fault & VM_FAULT_ERROR */
- if (fault & VM_FAULT_OOM)
- pagefault_out_of_memory();
- else if (fault & VM_FAULT_SIGBUS) {
+ if (fault & VM_FAULT_OOM) {
+ if (!(regs->psw.mask & PSW_MASK_PSTATE))
+ do_no_context(regs, int_code, trans_exc_code);
+ else
+ pagefault_out_of_memory();
+ } else if (fault & VM_FAULT_SIGBUS) {
/* Kernel mode? Handle exceptions or die */
if (!(regs->psw.mask & PSW_MASK_PSTATE))
do_no_context(regs, int_code, trans_exc_code);
@@ -277,7 +280,8 @@ static inline int do_exception(struct pt_regs *regs, int access,
struct mm_struct *mm;
struct vm_area_struct *vma;
unsigned long address;
- int fault, write;
+ unsigned int flags;
+ int fault;
if (notify_page_fault(regs))
return 0;
@@ -296,6 +300,10 @@ static inline int do_exception(struct pt_regs *regs, int access,
address = trans_exc_code & __FAIL_ADDR_MASK;
perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+ flags = FAULT_FLAG_ALLOW_RETRY;
+ if (access == VM_WRITE || (trans_exc_code & store_indication) == 0x400)
+ flags |= FAULT_FLAG_WRITE;
+retry:
down_read(&mm->mmap_sem);
fault = VM_FAULT_BADMAP;
@@ -325,21 +333,31 @@ static inline int do_exception(struct pt_regs *regs, int access,
* make sure we exit gracefully rather than endlessly redo
* the fault.
*/
- write = (access == VM_WRITE ||
- (trans_exc_code & store_indication) == 0x400) ?
- FAULT_FLAG_WRITE : 0;
- fault = handle_mm_fault(mm, vma, address, write);
+ fault = handle_mm_fault(mm, vma, address, flags);
if (unlikely(fault & VM_FAULT_ERROR))
goto out_up;
- if (fault & VM_FAULT_MAJOR) {
- tsk->maj_flt++;
- perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
- regs, address);
- } else {
- tsk->min_flt++;
- perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
- regs, address);
+ /*
+ * Major/minor page fault accounting is only done on the
+ * initial attempt. If we go through a retry, it is extremely
+ * likely that the page will be found in page cache at that point.
+ */
+ if (flags & FAULT_FLAG_ALLOW_RETRY) {
+ if (fault & VM_FAULT_MAJOR) {
+ tsk->maj_flt++;
+ perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+ regs, address);
+ } else {
+ tsk->min_flt++;
+ perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+ regs, address);
+ }
+ if (fault & VM_FAULT_RETRY) {
+ /* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
+ * of starvation. */
+ flags &= ~FAULT_FLAG_ALLOW_RETRY;
+ goto retry;
+ }
}
/*
* The instruction that caused the program check will
@@ -429,10 +447,9 @@ int __handle_fault(unsigned long uaddr, unsigned long pgm_int_code, int write)
access = write ? VM_WRITE : VM_READ;
fault = do_exception(&regs, access, uaddr | 2);
if (unlikely(fault)) {
- if (fault & VM_FAULT_OOM) {
- pagefault_out_of_memory();
- fault = 0;
- } else if (fault & VM_FAULT_SIGBUS)
+ if (fault & VM_FAULT_OOM)
+ return -EFAULT;
+ else if (fault & VM_FAULT_SIGBUS)
do_sigbus(&regs, pgm_int_code, uaddr);
}
return fault ? -EFAULT : 0;
@@ -485,7 +502,6 @@ int pfault_init(void)
"2:\n"
EX_TABLE(0b,1b)
: "=d" (rc) : "a" (&refbk), "m" (refbk) : "cc");
- __ctl_set_bit(0, 9);
return rc;
}
@@ -500,7 +516,6 @@ void pfault_fini(void)
if (!MACHINE_IS_VM || pfault_disable)
return;
- __ctl_clear_bit(0,9);
asm volatile(
" diag %0,0,0x258\n"
"0:\n"
@@ -615,6 +630,7 @@ static int __init pfault_irq_init(void)
rc = pfault_init() == 0 ? 0 : -EOPNOTSUPP;
if (rc)
goto out_pfault;
+ service_subclass_irq_register();
hotcpu_notifier(pfault_cpu_notify, 0);
return 0;
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index dfefc2171691..59b663109d90 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -119,9 +119,7 @@ void __init paging_init(void)
sparse_memory_present_with_active_regions(MAX_NUMNODES);
sparse_init();
memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
-#ifdef CONFIG_ZONE_DMA
max_zone_pfns[ZONE_DMA] = PFN_DOWN(MAX_DMA_ADDRESS);
-#endif
max_zone_pfns[ZONE_NORMAL] = max_low_pfn;
free_area_init_nodes(max_zone_pfns);
fault_init();
diff --git a/arch/s390/mm/maccess.c b/arch/s390/mm/maccess.c
index 71a4b0d34be0..51e5cd9b906a 100644
--- a/arch/s390/mm/maccess.c
+++ b/arch/s390/mm/maccess.c
@@ -19,7 +19,7 @@
* using the stura instruction.
* Returns the number of bytes copied or -EFAULT.
*/
-static long probe_kernel_write_odd(void *dst, void *src, size_t size)
+static long probe_kernel_write_odd(void *dst, const void *src, size_t size)
{
unsigned long count, aligned;
int offset, mask;
@@ -45,7 +45,7 @@ static long probe_kernel_write_odd(void *dst, void *src, size_t size)
return rc ? rc : count;
}
-long probe_kernel_write(void *dst, void *src, size_t size)
+long probe_kernel_write(void *dst, const void *src, size_t size)
{
long copied = 0;
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 14c6fae6fe6b..37a23c223705 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -24,91 +24,12 @@
#include <asm/tlbflush.h>
#include <asm/mmu_context.h>
-struct rcu_table_freelist {
- struct rcu_head rcu;
- struct mm_struct *mm;
- unsigned int pgt_index;
- unsigned int crst_index;
- unsigned long *table[0];
-};
-
-#define RCU_FREELIST_SIZE \
- ((PAGE_SIZE - sizeof(struct rcu_table_freelist)) \
- / sizeof(unsigned long))
-
-static DEFINE_PER_CPU(struct rcu_table_freelist *, rcu_table_freelist);
-
-static void __page_table_free(struct mm_struct *mm, unsigned long *table);
-
-static struct rcu_table_freelist *rcu_table_freelist_get(struct mm_struct *mm)
-{
- struct rcu_table_freelist **batchp = &__get_cpu_var(rcu_table_freelist);
- struct rcu_table_freelist *batch = *batchp;
-
- if (batch)
- return batch;
- batch = (struct rcu_table_freelist *) __get_free_page(GFP_ATOMIC);
- if (batch) {
- batch->mm = mm;
- batch->pgt_index = 0;
- batch->crst_index = RCU_FREELIST_SIZE;
- *batchp = batch;
- }
- return batch;
-}
-
-static void rcu_table_freelist_callback(struct rcu_head *head)
-{
- struct rcu_table_freelist *batch =
- container_of(head, struct rcu_table_freelist, rcu);
-
- while (batch->pgt_index > 0)
- __page_table_free(batch->mm, batch->table[--batch->pgt_index]);
- while (batch->crst_index < RCU_FREELIST_SIZE)
- crst_table_free(batch->mm, batch->table[batch->crst_index++]);
- free_page((unsigned long) batch);
-}
-
-void rcu_table_freelist_finish(void)
-{
- struct rcu_table_freelist *batch = __get_cpu_var(rcu_table_freelist);
-
- if (!batch)
- return;
- call_rcu(&batch->rcu, rcu_table_freelist_callback);
- __get_cpu_var(rcu_table_freelist) = NULL;
-}
-
-static void smp_sync(void *arg)
-{
-}
-
#ifndef CONFIG_64BIT
#define ALLOC_ORDER 1
-#define TABLES_PER_PAGE 4
-#define FRAG_MASK 15UL
-#define SECOND_HALVES 10UL
-
-void clear_table_pgstes(unsigned long *table)
-{
- clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
- memset(table + 256, 0, PAGE_SIZE/4);
- clear_table(table + 512, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
- memset(table + 768, 0, PAGE_SIZE/4);
-}
-
+#define FRAG_MASK 0x0f
#else
#define ALLOC_ORDER 2
-#define TABLES_PER_PAGE 2
-#define FRAG_MASK 3UL
-#define SECOND_HALVES 2UL
-
-void clear_table_pgstes(unsigned long *table)
-{
- clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/2);
- memset(table + 256, 0, PAGE_SIZE/2);
-}
-
+#define FRAG_MASK 0x03
#endif
unsigned long VMALLOC_START = VMALLOC_END - VMALLOC_SIZE;
@@ -137,26 +58,6 @@ void crst_table_free(struct mm_struct *mm, unsigned long *table)
free_pages((unsigned long) table, ALLOC_ORDER);
}
-void crst_table_free_rcu(struct mm_struct *mm, unsigned long *table)
-{
- struct rcu_table_freelist *batch;
-
- if (atomic_read(&mm->mm_users) < 2 &&
- cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
- crst_table_free(mm, table);
- return;
- }
- batch = rcu_table_freelist_get(mm);
- if (!batch) {
- smp_call_function(smp_sync, NULL, 1);
- crst_table_free(mm, table);
- return;
- }
- batch->table[--batch->crst_index] = table;
- if (batch->pgt_index >= batch->crst_index)
- rcu_table_freelist_finish();
-}
-
#ifdef CONFIG_64BIT
int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
{
@@ -232,121 +133,175 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
}
#endif
+static inline unsigned int atomic_xor_bits(atomic_t *v, unsigned int bits)
+{
+ unsigned int old, new;
+
+ do {
+ old = atomic_read(v);
+ new = old ^ bits;
+ } while (atomic_cmpxchg(v, old, new) != old);
+ return new;
+}
+
/*
* page table entry allocation/free routines.
*/
+#ifdef CONFIG_PGSTE
+static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm)
+{
+ struct page *page;
+ unsigned long *table;
+
+ page = alloc_page(GFP_KERNEL|__GFP_REPEAT);
+ if (!page)
+ return NULL;
+ pgtable_page_ctor(page);
+ atomic_set(&page->_mapcount, 3);
+ table = (unsigned long *) page_to_phys(page);
+ clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/2);
+ clear_table(table + PTRS_PER_PTE, 0, PAGE_SIZE/2);
+ return table;
+}
+
+static inline void page_table_free_pgste(unsigned long *table)
+{
+ struct page *page;
+
+ page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
+ pgtable_page_ctor(page);
+ atomic_set(&page->_mapcount, -1);
+ __free_page(page);
+}
+#endif
+
unsigned long *page_table_alloc(struct mm_struct *mm)
{
struct page *page;
unsigned long *table;
- unsigned long bits;
+ unsigned int mask, bit;
- bits = (mm->context.has_pgste) ? 3UL : 1UL;
+#ifdef CONFIG_PGSTE
+ if (mm_has_pgste(mm))
+ return page_table_alloc_pgste(mm);
+#endif
+ /* Allocate fragments of a 4K page as 1K/2K page table */
spin_lock_bh(&mm->context.list_lock);
- page = NULL;
+ mask = FRAG_MASK;
if (!list_empty(&mm->context.pgtable_list)) {
page = list_first_entry(&mm->context.pgtable_list,
struct page, lru);
- if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
- page = NULL;
+ table = (unsigned long *) page_to_phys(page);
+ mask = atomic_read(&page->_mapcount);
+ mask = mask | (mask >> 4);
}
- if (!page) {
+ if ((mask & FRAG_MASK) == FRAG_MASK) {
spin_unlock_bh(&mm->context.list_lock);
page = alloc_page(GFP_KERNEL|__GFP_REPEAT);
if (!page)
return NULL;
pgtable_page_ctor(page);
- page->flags &= ~FRAG_MASK;
+ atomic_set(&page->_mapcount, 1);
table = (unsigned long *) page_to_phys(page);
- if (mm->context.has_pgste)
- clear_table_pgstes(table);
- else
- clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
+ clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
spin_lock_bh(&mm->context.list_lock);
list_add(&page->lru, &mm->context.pgtable_list);
+ } else {
+ for (bit = 1; mask & bit; bit <<= 1)
+ table += PTRS_PER_PTE;
+ mask = atomic_xor_bits(&page->_mapcount, bit);
+ if ((mask & FRAG_MASK) == FRAG_MASK)
+ list_del(&page->lru);
}
- table = (unsigned long *) page_to_phys(page);
- while (page->flags & bits) {
- table += 256;
- bits <<= 1;
- }
- page->flags |= bits;
- if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
- list_move_tail(&page->lru, &mm->context.pgtable_list);
spin_unlock_bh(&mm->context.list_lock);
return table;
}
-static void __page_table_free(struct mm_struct *mm, unsigned long *table)
+void page_table_free(struct mm_struct *mm, unsigned long *table)
{
struct page *page;
- unsigned long bits;
+ unsigned int bit, mask;
- bits = ((unsigned long) table) & 15;
- table = (unsigned long *)(((unsigned long) table) ^ bits);
+#ifdef CONFIG_PGSTE
+ if (mm_has_pgste(mm))
+ return page_table_free_pgste(table);
+#endif
+ /* Free 1K/2K page table fragment of a 4K page */
page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
- page->flags ^= bits;
- if (!(page->flags & FRAG_MASK)) {
+ bit = 1 << ((__pa(table) & ~PAGE_MASK)/(PTRS_PER_PTE*sizeof(pte_t)));
+ spin_lock_bh(&mm->context.list_lock);
+ if ((atomic_read(&page->_mapcount) & FRAG_MASK) != FRAG_MASK)
+ list_del(&page->lru);
+ mask = atomic_xor_bits(&page->_mapcount, bit);
+ if (mask & FRAG_MASK)
+ list_add(&page->lru, &mm->context.pgtable_list);
+ spin_unlock_bh(&mm->context.list_lock);
+ if (mask == 0) {
pgtable_page_dtor(page);
+ atomic_set(&page->_mapcount, -1);
__free_page(page);
}
}
-void page_table_free(struct mm_struct *mm, unsigned long *table)
+#ifdef CONFIG_HAVE_RCU_TABLE_FREE
+
+static void __page_table_free_rcu(void *table, unsigned bit)
{
struct page *page;
- unsigned long bits;
- bits = (mm->context.has_pgste) ? 3UL : 1UL;
- bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
+#ifdef CONFIG_PGSTE
+ if (bit == FRAG_MASK)
+ return page_table_free_pgste(table);
+#endif
+ /* Free 1K/2K page table fragment of a 4K page */
page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
- spin_lock_bh(&mm->context.list_lock);
- page->flags ^= bits;
- if (page->flags & FRAG_MASK) {
- /* Page now has some free pgtable fragments. */
- if (!list_empty(&page->lru))
- list_move(&page->lru, &mm->context.pgtable_list);
- page = NULL;
- } else
- /* All fragments of the 4K page have been freed. */
- list_del(&page->lru);
- spin_unlock_bh(&mm->context.list_lock);
- if (page) {
+ if (atomic_xor_bits(&page->_mapcount, bit) == 0) {
pgtable_page_dtor(page);
+ atomic_set(&page->_mapcount, -1);
__free_page(page);
}
}
-void page_table_free_rcu(struct mm_struct *mm, unsigned long *table)
+void page_table_free_rcu(struct mmu_gather *tlb, unsigned long *table)
{
- struct rcu_table_freelist *batch;
+ struct mm_struct *mm;
struct page *page;
- unsigned long bits;
+ unsigned int bit, mask;
- if (atomic_read(&mm->mm_users) < 2 &&
- cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
- page_table_free(mm, table);
- return;
- }
- batch = rcu_table_freelist_get(mm);
- if (!batch) {
- smp_call_function(smp_sync, NULL, 1);
- page_table_free(mm, table);
+ mm = tlb->mm;
+#ifdef CONFIG_PGSTE
+ if (mm_has_pgste(mm)) {
+ table = (unsigned long *) (__pa(table) | FRAG_MASK);
+ tlb_remove_table(tlb, table);
return;
}
- bits = (mm->context.has_pgste) ? 3UL : 1UL;
- bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
+#endif
+ bit = 1 << ((__pa(table) & ~PAGE_MASK) / (PTRS_PER_PTE*sizeof(pte_t)));
page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
spin_lock_bh(&mm->context.list_lock);
- /* Delayed freeing with rcu prevents reuse of pgtable fragments */
- list_del_init(&page->lru);
+ if ((atomic_read(&page->_mapcount) & FRAG_MASK) != FRAG_MASK)
+ list_del(&page->lru);
+ mask = atomic_xor_bits(&page->_mapcount, bit | (bit << 4));
+ if (mask & FRAG_MASK)
+ list_add_tail(&page->lru, &mm->context.pgtable_list);
spin_unlock_bh(&mm->context.list_lock);
- table = (unsigned long *)(((unsigned long) table) | bits);
- batch->table[batch->pgt_index++] = table;
- if (batch->pgt_index >= batch->crst_index)
- rcu_table_freelist_finish();
+ table = (unsigned long *) (__pa(table) | (bit << 4));
+ tlb_remove_table(tlb, table);
}
+void __tlb_remove_table(void *_table)
+{
+ void *table = (void *)((unsigned long) _table & PAGE_MASK);
+ unsigned type = (unsigned long) _table & ~PAGE_MASK;
+
+ if (type)
+ __page_table_free_rcu(table, type);
+ else
+ free_pages((unsigned long) table, ALLOC_ORDER);
+}
+
+#endif
+
/*
* switch on pgstes for its userspace process (for kvm)
*/
@@ -360,7 +315,7 @@ int s390_enable_sie(void)
return -EINVAL;
/* Do we have pgstes? if yes, we are done */
- if (tsk->mm->context.has_pgste)
+ if (mm_has_pgste(tsk->mm))
return 0;
/* lets check if we are allowed to replace the mm */
generated by cgit v1.2.3 (git 2.39.0) at 2025-02-27 03:00:47 +0300