summaryrefslogtreecommitdiffstats
path: root/include/asm-s390/pgtable.h
diff options
context:
space:
mode:
authorGerald Schaefer <geraldsc@de.ibm.com>2007-02-05 21:18:17 +0100
committerMartin Schwidefsky <schwidefsky@de.ibm.com>2007-02-05 21:18:17 +0100
commitc1821c2e9711adc3cd298a16b7237c92a2cee78d (patch)
tree9155b089db35a37d95863125ea4c5f918bd7801b /include/asm-s390/pgtable.h
parent86aa9fc2456d8a662f299a70bdb70987209170f0 (diff)
downloadop-kernel-dev-c1821c2e9711adc3cd298a16b7237c92a2cee78d.zip
op-kernel-dev-c1821c2e9711adc3cd298a16b7237c92a2cee78d.tar.gz
[S390] noexec protection
This provides a noexec protection on s390 hardware. Our hardware does not have any bits left in the pte for a hw noexec bit, so this is a different approach using shadow page tables and a special addressing mode that allows separate address spaces for code and data. As a special feature of our "secondary-space" addressing mode, separate page tables can be specified for the translation of data addresses (storage operands) and instruction addresses. The shadow page table is used for the instruction addresses and the standard page table for the data addresses. The shadow page table is linked to the standard page table by a pointer in page->lru.next of the struct page corresponding to the page that contains the standard page table (since page->private is not really private with the pte_lock and the page table pages are not in the LRU list). Depending on the software bits of a pte, it is either inserted into both page tables or just into the standard (data) page table. Pages of a vma that does not have the VM_EXEC bit set get mapped only in the data address space. Any try to execute code on such a page will cause a page translation exception. The standard reaction to this is a SIGSEGV with two exceptions: the two system call opcodes 0x0a77 (sys_sigreturn) and 0x0aad (sys_rt_sigreturn) are allowed. They are stored by the kernel to the signal stack frame. Unfortunately, the signal return mechanism cannot be modified to use an SA_RESTORER because the exception unwinding code depends on the system call opcode stored behind the signal stack frame. This feature requires that user space is executed in secondary-space mode and the kernel in home-space mode, which means that the addressing modes need to be switched and that the noexec protection only works for user space. After switching the addressing modes, we cannot use the mvcp/mvcs instructions anymore to copy between kernel and user space. A new mvcos instruction has been added to the z9 EC/BC hardware which allows to copy between arbitrary address spaces, but on older hardware the page tables need to be walked manually. Signed-off-by: Gerald Schaefer <geraldsc@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Diffstat (limited to 'include/asm-s390/pgtable.h')
-rw-r--r--include/asm-s390/pgtable.h146
1 files changed, 130 insertions, 16 deletions
diff --git a/include/asm-s390/pgtable.h b/include/asm-s390/pgtable.h
index 304ee77..13c1654 100644
--- a/include/asm-s390/pgtable.h
+++ b/include/asm-s390/pgtable.h
@@ -224,6 +224,8 @@ extern unsigned long vmalloc_end;
#define _PAGE_TYPE_FILE 0x601 /* bit 0x002 is used for offset !! */
#define _PAGE_TYPE_RO 0x200
#define _PAGE_TYPE_RW 0x000
+#define _PAGE_TYPE_EX_RO 0x202
+#define _PAGE_TYPE_EX_RW 0x002
/*
* PTE type bits are rather complicated. handle_pte_fault uses pte_present,
@@ -244,11 +246,13 @@ extern unsigned long vmalloc_end;
* _PAGE_TYPE_FILE 11?1 -> 11?1
* _PAGE_TYPE_RO 0100 -> 1100
* _PAGE_TYPE_RW 0000 -> 1000
+ * _PAGE_TYPE_EX_RO 0110 -> 1110
+ * _PAGE_TYPE_EX_RW 0010 -> 1010
*
- * pte_none is true for bits combinations 1000, 1100
+ * pte_none is true for bits combinations 1000, 1010, 1100, 1110
* pte_present is true for bits combinations 0000, 0010, 0100, 0110, 1001
* pte_file is true for bits combinations 1101, 1111
- * swap pte is 1011 and 0001, 0011, 0101, 0111, 1010 and 1110 are invalid.
+ * swap pte is 1011 and 0001, 0011, 0101, 0111 are invalid.
*/
#ifndef __s390x__
@@ -313,33 +317,100 @@ extern unsigned long vmalloc_end;
#define PAGE_NONE __pgprot(_PAGE_TYPE_NONE)
#define PAGE_RO __pgprot(_PAGE_TYPE_RO)
#define PAGE_RW __pgprot(_PAGE_TYPE_RW)
+#define PAGE_EX_RO __pgprot(_PAGE_TYPE_EX_RO)
+#define PAGE_EX_RW __pgprot(_PAGE_TYPE_EX_RW)
#define PAGE_KERNEL PAGE_RW
#define PAGE_COPY PAGE_RO
/*
- * The S390 can't do page protection for execute, and considers that the
- * same are read. Also, write permissions imply read permissions. This is
- * the closest we can get..
+ * Dependent on the EXEC_PROTECT option s390 can do execute protection.
+ * Write permission always implies read permission. In theory with a
+ * primary/secondary page table execute only can be implemented but
+ * it would cost an additional bit in the pte to distinguish all the
+ * different pte types. To avoid that execute permission currently
+ * implies read permission as well.
*/
/*xwr*/
#define __P000 PAGE_NONE
#define __P001 PAGE_RO
#define __P010 PAGE_RO
#define __P011 PAGE_RO
-#define __P100 PAGE_RO
-#define __P101 PAGE_RO
-#define __P110 PAGE_RO
-#define __P111 PAGE_RO
+#define __P100 PAGE_EX_RO
+#define __P101 PAGE_EX_RO
+#define __P110 PAGE_EX_RO
+#define __P111 PAGE_EX_RO
#define __S000 PAGE_NONE
#define __S001 PAGE_RO
#define __S010 PAGE_RW
#define __S011 PAGE_RW
-#define __S100 PAGE_RO
-#define __S101 PAGE_RO
-#define __S110 PAGE_RW
-#define __S111 PAGE_RW
+#define __S100 PAGE_EX_RO
+#define __S101 PAGE_EX_RO
+#define __S110 PAGE_EX_RW
+#define __S111 PAGE_EX_RW
+
+#ifndef __s390x__
+# define PMD_SHADOW_SHIFT 1
+# define PGD_SHADOW_SHIFT 1
+#else /* __s390x__ */
+# define PMD_SHADOW_SHIFT 2
+# define PGD_SHADOW_SHIFT 2
+#endif /* __s390x__ */
+
+static inline struct page *get_shadow_page(struct page *page)
+{
+ if (s390_noexec && !list_empty(&page->lru))
+ return virt_to_page(page->lru.next);
+ return NULL;
+}
+
+static inline pte_t *get_shadow_pte(pte_t *ptep)
+{
+ unsigned long pteptr = (unsigned long) (ptep);
+
+ if (s390_noexec) {
+ unsigned long offset = pteptr & (PAGE_SIZE - 1);
+ void *addr = (void *) (pteptr ^ offset);
+ struct page *page = virt_to_page(addr);
+ if (!list_empty(&page->lru))
+ return (pte_t *) ((unsigned long) page->lru.next |
+ offset);
+ }
+ return NULL;
+}
+
+static inline pmd_t *get_shadow_pmd(pmd_t *pmdp)
+{
+ unsigned long pmdptr = (unsigned long) (pmdp);
+
+ if (s390_noexec) {
+ unsigned long offset = pmdptr &
+ ((PAGE_SIZE << PMD_SHADOW_SHIFT) - 1);
+ void *addr = (void *) (pmdptr ^ offset);
+ struct page *page = virt_to_page(addr);
+ if (!list_empty(&page->lru))
+ return (pmd_t *) ((unsigned long) page->lru.next |
+ offset);
+ }
+ return NULL;
+}
+
+static inline pgd_t *get_shadow_pgd(pgd_t *pgdp)
+{
+ unsigned long pgdptr = (unsigned long) (pgdp);
+
+ if (s390_noexec) {
+ unsigned long offset = pgdptr &
+ ((PAGE_SIZE << PGD_SHADOW_SHIFT) - 1);
+ void *addr = (void *) (pgdptr ^ offset);
+ struct page *page = virt_to_page(addr);
+ if (!list_empty(&page->lru))
+ return (pgd_t *) ((unsigned long) page->lru.next |
+ offset);
+ }
+ return NULL;
+}
/*
* Certain architectures need to do special things when PTEs
@@ -348,7 +419,16 @@ extern unsigned long vmalloc_end;
*/
static inline void set_pte(pte_t *pteptr, pte_t pteval)
{
+ pte_t *shadow_pte = get_shadow_pte(pteptr);
+
*pteptr = pteval;
+ if (shadow_pte) {
+ if (!(pte_val(pteval) & _PAGE_INVALID) &&
+ (pte_val(pteval) & _PAGE_SWX))
+ pte_val(*shadow_pte) = pte_val(pteval) | _PAGE_RO;
+ else
+ pte_val(*shadow_pte) = _PAGE_TYPE_EMPTY;
+ }
}
#define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval)
@@ -466,7 +546,7 @@ static inline int pte_read(pte_t pte)
static inline void pgd_clear(pgd_t * pgdp) { }
-static inline void pmd_clear(pmd_t * pmdp)
+static inline void pmd_clear_kernel(pmd_t * pmdp)
{
pmd_val(pmdp[0]) = _PAGE_TABLE_INV;
pmd_val(pmdp[1]) = _PAGE_TABLE_INV;
@@ -474,24 +554,55 @@ static inline void pmd_clear(pmd_t * pmdp)
pmd_val(pmdp[3]) = _PAGE_TABLE_INV;
}
+static inline void pmd_clear(pmd_t * pmdp)
+{
+ pmd_t *shadow_pmd = get_shadow_pmd(pmdp);
+
+ pmd_clear_kernel(pmdp);
+ if (shadow_pmd)
+ pmd_clear_kernel(shadow_pmd);
+}
+
#else /* __s390x__ */
-static inline void pgd_clear(pgd_t * pgdp)
+static inline void pgd_clear_kernel(pgd_t * pgdp)
{
pgd_val(*pgdp) = _PGD_ENTRY_INV | _PGD_ENTRY;
}
-static inline void pmd_clear(pmd_t * pmdp)
+static inline void pgd_clear(pgd_t * pgdp)
+{
+ pgd_t *shadow_pgd = get_shadow_pgd(pgdp);
+
+ pgd_clear_kernel(pgdp);
+ if (shadow_pgd)
+ pgd_clear_kernel(shadow_pgd);
+}
+
+static inline void pmd_clear_kernel(pmd_t * pmdp)
{
pmd_val(*pmdp) = _PMD_ENTRY_INV | _PMD_ENTRY;
pmd_val1(*pmdp) = _PMD_ENTRY_INV | _PMD_ENTRY;
}
+static inline void pmd_clear(pmd_t * pmdp)
+{
+ pmd_t *shadow_pmd = get_shadow_pmd(pmdp);
+
+ pmd_clear_kernel(pmdp);
+ if (shadow_pmd)
+ pmd_clear_kernel(shadow_pmd);
+}
+
#endif /* __s390x__ */
static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
{
+ pte_t *shadow_pte = get_shadow_pte(ptep);
+
pte_val(*ptep) = _PAGE_TYPE_EMPTY;
+ if (shadow_pte)
+ pte_val(*shadow_pte) = _PAGE_TYPE_EMPTY;
}
/*
@@ -609,8 +720,11 @@ ptep_clear_flush(struct vm_area_struct *vma,
unsigned long address, pte_t *ptep)
{
pte_t pte = *ptep;
+ pte_t *shadow_pte = get_shadow_pte(ptep);
__ptep_ipte(address, ptep);
+ if (shadow_pte)
+ __ptep_ipte(address, shadow_pte);
return pte;
}
OpenPOWER on IntegriCloud