KVM: s390/mm: support gmap page tables with less than 5 levels
Add an addressing limit to the gmap address spaces and only allocate
the page table levels that are needed for the given limit. The limit
is fixed and can not be changed after a gmap has been created.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
diff --git a/arch/s390/include/asm/pgtable.h b/arch/s390/include/asm/pgtable.h
index 9bfdbca..7705180e 100644
--- a/arch/s390/include/asm/pgtable.h
+++ b/arch/s390/include/asm/pgtable.h
@@ -808,6 +808,7 @@
spinlock_t guest_table_lock;
unsigned long *table;
unsigned long asce;
+ unsigned long asce_end;
void *private;
bool pfault_enabled;
};
@@ -844,7 +845,7 @@
void (*notifier_call)(struct gmap *gmap, unsigned long gaddr);
};
-struct gmap *gmap_alloc(struct mm_struct *mm);
+struct gmap *gmap_alloc(struct mm_struct *mm, unsigned long limit);
void gmap_free(struct gmap *gmap);
void gmap_enable(struct gmap *gmap);
void gmap_disable(struct gmap *gmap);
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 543c24b..82065dc 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -451,7 +451,7 @@
if (type & KVM_VM_S390_UCONTROL) {
kvm->arch.gmap = NULL;
} else {
- kvm->arch.gmap = gmap_alloc(current->mm);
+ kvm->arch.gmap = gmap_alloc(current->mm, -1UL);
if (!kvm->arch.gmap)
goto out_nogmap;
kvm->arch.gmap->private = kvm;
@@ -535,7 +535,7 @@
vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
kvm_clear_async_pf_completion_queue(vcpu);
if (kvm_is_ucontrol(vcpu->kvm)) {
- vcpu->arch.gmap = gmap_alloc(current->mm);
+ vcpu->arch.gmap = gmap_alloc(current->mm, -1UL);
if (!vcpu->arch.gmap)
return -ENOMEM;
vcpu->arch.gmap->private = vcpu->kvm;
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 74dfd9e..665714b 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -145,15 +145,34 @@
/**
* gmap_alloc - allocate a guest address space
* @mm: pointer to the parent mm_struct
+ * @limit: maximum size of the gmap address space
*
* Returns a guest address space structure.
*/
-struct gmap *gmap_alloc(struct mm_struct *mm)
+struct gmap *gmap_alloc(struct mm_struct *mm, unsigned long limit)
{
struct gmap *gmap;
struct page *page;
unsigned long *table;
+ unsigned long etype, atype;
+ if (limit < (1UL << 31)) {
+ limit = (1UL << 31) - 1;
+ atype = _ASCE_TYPE_SEGMENT;
+ etype = _SEGMENT_ENTRY_EMPTY;
+ } else if (limit < (1UL << 42)) {
+ limit = (1UL << 42) - 1;
+ atype = _ASCE_TYPE_REGION3;
+ etype = _REGION3_ENTRY_EMPTY;
+ } else if (limit < (1UL << 53)) {
+ limit = (1UL << 53) - 1;
+ atype = _ASCE_TYPE_REGION2;
+ etype = _REGION2_ENTRY_EMPTY;
+ } else {
+ limit = -1UL;
+ atype = _ASCE_TYPE_REGION1;
+ etype = _REGION1_ENTRY_EMPTY;
+ }
gmap = kzalloc(sizeof(struct gmap), GFP_KERNEL);
if (!gmap)
goto out;
@@ -168,10 +187,11 @@
page->index = 0;
list_add(&page->lru, &gmap->crst_list);
table = (unsigned long *) page_to_phys(page);
- crst_table_init(table, _REGION1_ENTRY_EMPTY);
+ crst_table_init(table, etype);
gmap->table = table;
- gmap->asce = _ASCE_TYPE_REGION1 | _ASCE_TABLE_LENGTH |
- _ASCE_USER_BITS | __pa(table);
+ gmap->asce = atype | _ASCE_TABLE_LENGTH |
+ _ASCE_USER_BITS | __pa(table);
+ gmap->asce_end = limit;
down_write(&mm->mmap_sem);
list_add(&gmap->list, &mm->context.gmap_list);
up_write(&mm->mmap_sem);
@@ -187,8 +207,7 @@
static void gmap_flush_tlb(struct gmap *gmap)
{
if (MACHINE_HAS_IDTE)
- __tlb_flush_asce(gmap->mm, (unsigned long) gmap->table |
- _ASCE_TYPE_REGION1);
+ __tlb_flush_asce(gmap->mm, gmap->asce);
else
__tlb_flush_global();
}
@@ -227,8 +246,7 @@
/* Flush tlb. */
if (MACHINE_HAS_IDTE)
- __tlb_flush_asce(gmap->mm, (unsigned long) gmap->table |
- _ASCE_TYPE_REGION1);
+ __tlb_flush_asce(gmap->mm, gmap->asce);
else
__tlb_flush_global();
@@ -394,8 +412,8 @@
if ((from | to | len) & (PMD_SIZE - 1))
return -EINVAL;
- if (len == 0 || from + len > TASK_MAX_SIZE ||
- from + len < from || to + len < to)
+ if (len == 0 || from + len < from || to + len < to ||
+ from + len > TASK_MAX_SIZE || to + len > gmap->asce_end)
return -EINVAL;
flush = 0;
@@ -501,25 +519,32 @@
int rc;
/* Create higher level tables in the gmap page table */
- table = gmap->table + ((gaddr >> 53) & 0x7ff);
- if ((*table & _REGION_ENTRY_INVALID) &&
- gmap_alloc_table(gmap, table, _REGION2_ENTRY_EMPTY,
- gaddr & 0xffe0000000000000))
- return -ENOMEM;
- table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
- table = table + ((gaddr >> 42) & 0x7ff);
- if ((*table & _REGION_ENTRY_INVALID) &&
- gmap_alloc_table(gmap, table, _REGION3_ENTRY_EMPTY,
- gaddr & 0xfffffc0000000000))
- return -ENOMEM;
- table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
- table = table + ((gaddr >> 31) & 0x7ff);
- if ((*table & _REGION_ENTRY_INVALID) &&
- gmap_alloc_table(gmap, table, _SEGMENT_ENTRY_EMPTY,
- gaddr & 0xffffffff80000000))
- return -ENOMEM;
- table = (unsigned long *) (*table & _REGION_ENTRY_ORIGIN);
- table = table + ((gaddr >> 20) & 0x7ff);
+ table = gmap->table;
+ if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION1) {
+ table += (gaddr >> 53) & 0x7ff;
+ if ((*table & _REGION_ENTRY_INVALID) &&
+ gmap_alloc_table(gmap, table, _REGION2_ENTRY_EMPTY,
+ gaddr & 0xffe0000000000000))
+ return -ENOMEM;
+ table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
+ }
+ if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION2) {
+ table += (gaddr >> 42) & 0x7ff;
+ if ((*table & _REGION_ENTRY_INVALID) &&
+ gmap_alloc_table(gmap, table, _REGION3_ENTRY_EMPTY,
+ gaddr & 0xfffffc0000000000))
+ return -ENOMEM;
+ table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
+ }
+ if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION3) {
+ table += (gaddr >> 31) & 0x7ff;
+ if ((*table & _REGION_ENTRY_INVALID) &&
+ gmap_alloc_table(gmap, table, _SEGMENT_ENTRY_EMPTY,
+ gaddr & 0xffffffff80000000))
+ return -ENOMEM;
+ table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
+ }
+ table += (gaddr >> 20) & 0x7ff;
/* Walk the parent mm page table */
mm = gmap->mm;
pgd = pgd_offset(mm, vmaddr);
