Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc

Pull sparc changes from David S. Miller:
 "This has the generic strncpy_from_user() implementation architectures
  can now use, which we've been developing on linux-arch over the past
  few days.

  For good measure I ran both a 32-bit and a 64-bit glibc testsuite run,
  and the latter of which pointed out an adjustment I needed to make to
  sparc's user_addr_max() definition.  Linus, you were right, STACK_TOP
  was not the right thing to use, even on sparc itself :-)

  From Sam Ravnborg, we have a conversion of sparc32 over to the common
  alloc_thread_info_node(), since the aspect which originally blocked
  our doing so (sun4c) has been removed."

Fix up trivial arch/sparc/Kconfig and lib/Makefile conflicts.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc: Fix user_addr_max() definition.
  lib: Sparc's strncpy_from_user is generic enough, move under lib/
  kernel: Move REPEAT_BYTE definition into linux/kernel.h
  sparc: Increase portability of strncpy_from_user() implementation.
  sparc: Optimize strncpy_from_user() zero byte search.
  sparc: Add full proper error handling to strncpy_from_user().
  sparc32: use the common implementation of alloc_thread_info_node()

arch/sparc/Kconfig

@@ -34,12 +34,12 @@ config SPARC
 	select GENERIC_SMP_IDLE_THREAD
 	select GENERIC_CMOS_UPDATE
 	select GENERIC_CLOCKEVENTS
+	select GENERIC_STRNCPY_FROM_USER
 
 config SPARC32
 	def_bool !64BIT
 	select GENERIC_ATOMIC64
 	select CLZ_TAB
-	select ARCH_THREAD_INFO_ALLOCATOR
 	select ARCH_USES_GETTIMEOFFSET
 
 config SPARC64

arch/sparc/include/asm/processor_64.h

@@ -42,7 +42,9 @@
 #define TASK_SIZE_OF(tsk) \
 	(test_tsk_thread_flag(tsk,TIF_32BIT) ? \
 	 (1UL << 32UL) : ((unsigned long)-VPTE_SIZE))
-#define TASK_SIZE	TASK_SIZE_OF(current)
+#define TASK_SIZE \
+	(test_thread_flag(TIF_32BIT) ? \
+	 (1UL << 32UL) : ((unsigned long)-VPTE_SIZE))
 #ifdef __KERNEL__
 
 #define STACK_TOP32	((1UL << 32UL) - PAGE_SIZE)

arch/sparc/include/asm/thread_info_32.h

@@ -77,18 +77,11 @@ register struct thread_info *current_thread_info_reg asm("g6");
 /*
  * thread information allocation
  */
-#define THREAD_INFO_ORDER  1
-
-struct thread_info * alloc_thread_info_node(struct task_struct *tsk, int node);
-void free_thread_info(struct thread_info *);
+#define THREAD_SIZE_ORDER  1
 
 #endif /* __ASSEMBLY__ */
 
-/*
- * Size of kernel stack for each process.
- * Observe the order of get_free_pages() in alloc_thread_info_node().
- * The sun4 has 8K stack too, because it's short on memory, and 16K is a waste.
- */
+/* Size of kernel stack for each process */
 #define THREAD_SIZE		(2 * PAGE_SIZE)
 
 /*

arch/sparc/include/asm/uaccess.h

@@ -5,4 +5,10 @@
 #else
 #include <asm/uaccess_32.h>
 #endif
+
+#define user_addr_max() \
+	(segment_eq(get_fs(), USER_DS) ? TASK_SIZE : ~0UL)
+
+extern long strncpy_from_user(char *dest, const char __user *src, long count);
+
 #endif

arch/sparc/include/asm/uaccess_32.h

@@ -304,16 +304,6 @@ static inline unsigned long clear_user(void __user *addr, unsigned long n)
 		return n;
 }
 
-extern long __strncpy_from_user(char *dest, const char __user *src, long count);
-
-static inline long strncpy_from_user(char *dest, const char __user *src, long count)
-{
-	if (__access_ok((unsigned long) src, count))
-		return __strncpy_from_user(dest, src, count);
-	else
-		return -EFAULT;
-}
-
 extern long __strlen_user(const char __user *);
 extern long __strnlen_user(const char __user *, long len);
 

arch/sparc/include/asm/uaccess_64.h

@@ -257,10 +257,6 @@ extern unsigned long __must_check __clear_user(void __user *, unsigned long);
 
 #define clear_user __clear_user
 
-extern long __must_check __strncpy_from_user(char *dest, const char __user *src, long count);
-
-#define strncpy_from_user __strncpy_from_user
-
 extern long __strlen_user(const char __user *);
 extern long __strnlen_user(const char __user *, long len);
 

arch/sparc/lib/Makefile

@@ -10,7 +10,7 @@ lib-y                 += strlen.o
 lib-y                 += checksum_$(BITS).o
 lib-$(CONFIG_SPARC32) += blockops.o
 lib-y                 += memscan_$(BITS).o memcmp.o strncmp_$(BITS).o
-lib-y                 += strncpy_from_user_$(BITS).o strlen_user_$(BITS).o
+lib-y                 += strlen_user_$(BITS).o
 lib-$(CONFIG_SPARC32) += divdi3.o udivdi3.o
 lib-$(CONFIG_SPARC32) += copy_user.o locks.o
 lib-$(CONFIG_SPARC64) += atomic_64.o

arch/sparc/lib/ksyms.c

@@ -33,9 +33,6 @@ EXPORT_SYMBOL(memset);
 EXPORT_SYMBOL(memmove);
 EXPORT_SYMBOL(__bzero);
 
-/* Moving data to/from/in userspace. */
-EXPORT_SYMBOL(__strncpy_from_user);
-
 /* Networking helper routines. */
 EXPORT_SYMBOL(csum_partial);
 

arch/sparc/lib/strncpy_from_user_32.S

-/* strncpy_from_user.S: Sparc strncpy from userspace.
- *
- *  Copyright(C) 1996 David S. Miller
- */
-
-#include <linux/linkage.h>
-#include <asm/ptrace.h>
-#include <asm/errno.h>
-
-	.text
-
-	/* Must return:
-	 *
-	 * -EFAULT		for an exception
-	 * count		if we hit the buffer limit
-	 * bytes copied		if we hit a null byte
-	 */
-
-ENTRY(__strncpy_from_user)
-	/* %o0=dest, %o1=src, %o2=count */
-	mov	%o2, %o3
-1:
-	subcc	%o2, 1, %o2
-	bneg	2f
-	 nop
-10:
-	ldub	[%o1], %o4
-	add	%o0, 1, %o0
-	cmp	%o4, 0
-	add	%o1, 1, %o1
-	bne	1b
-	 stb	%o4, [%o0 - 1]
-2:
-	add	%o2, 1, %o0
-	retl
-	 sub	%o3, %o0, %o0
-ENDPROC(__strncpy_from_user)
-
-	.section .fixup,#alloc,#execinstr
-	.align	4
-4:
-	retl
-	 mov	-EFAULT, %o0
-
-	.section __ex_table,#alloc
-	.align	4
-	.word	10b, 4b

arch/sparc/lib/strncpy_from_user_64.S

-/*
- * strncpy_from_user.S: Sparc64 strncpy from userspace.
- *
- *  Copyright (C) 1997, 1999 Jakub Jelinek (jj@ultra.linux.cz)
- */
-
-#include <linux/linkage.h>
-#include <asm/asi.h>
-#include <asm/errno.h>
-
-	.data
-	.align	8
-0:	.xword	0x0101010101010101
-
-	.text
-
-	/* Must return:
-	 *
-	 * -EFAULT		for an exception
-	 * count		if we hit the buffer limit
-	 * bytes copied		if we hit a null byte
-	 * (without the null byte)
-	 *
-	 * This implementation assumes:
-	 * %o1 is 8 aligned => !(%o2 & 7)
-	 * %o0 is 8 aligned (if not, it will be slooooow, but will work)
-	 *
-	 * This is optimized for the common case:
-	 * in my stats, 90% of src are 8 aligned (even on sparc32)
-	 * and average length is 18 or so.
-	 */
-
-ENTRY(__strncpy_from_user)
-	/* %o0=dest, %o1=src, %o2=count */
-	andcc	%o1, 7, %g0		! IEU1	Group
-	bne,pn	%icc, 30f		! CTI
-	 add	%o0, %o2, %g3		! IEU0
-60:	ldxa	[%o1] %asi, %g1		! Load	Group
-	brlez,pn %o2, 10f		! CTI
-	 mov	%o0, %o3		! IEU0
-50:	sethi	%hi(0b), %o4		! IEU0	Group
-	ldx	[%o4 + %lo(0b)], %o4	! Load
-	sllx	%o4, 7, %o5		! IEU1	Group
-1:	sub	%g1, %o4, %g2		! IEU0	Group
-	stx	%g1, [%o0]		! Store
-	add	%o0, 8, %o0		! IEU1
-	andcc	%g2, %o5, %g0		! IEU1	Group
-	bne,pn	%xcc, 5f		! CTI
-	 add	%o1, 8, %o1		! IEU0
-	cmp	%o0, %g3		! IEU1	Group
-	bl,a,pt %xcc, 1b		! CTI
-61:	 ldxa	[%o1] %asi, %g1		! Load
-10:	retl				! CTI	Group
-	 mov	%o2, %o0		! IEU0
-5:	srlx	%g2, 32, %g7		! IEU0	Group
-	sethi	%hi(0xff00), %o4	! IEU1
-	andcc	%g7, %o5, %g0		! IEU1	Group
-	be,pn	%icc, 2f		! CTI
-	 or	%o4, %lo(0xff00), %o4	! IEU0
-	srlx	%g1, 48, %g7		! IEU0	Group
-	andcc	%g7, %o4, %g0		! IEU1	Group
-	be,pn	%icc, 50f		! CTI
-	 andcc	%g7, 0xff, %g0		! IEU1	Group
-	be,pn	%icc, 51f		! CTI
-	 srlx	%g1, 32, %g7		! IEU0
-	andcc	%g7, %o4, %g0		! IEU1	Group
-	be,pn	%icc, 52f		! CTI
-	 andcc	%g7, 0xff, %g0		! IEU1	Group
-	be,pn	%icc, 53f		! CTI
-2:	 andcc	%g2, %o5, %g0		! IEU1	Group
-	be,pn	%icc, 2f		! CTI
-	 srl	%g1, 16, %g7		! IEU0
-	andcc	%g7, %o4, %g0		! IEU1	Group
-	be,pn	%icc, 54f		! CTI
-	 andcc	%g7, 0xff, %g0		! IEU1	Group
-	be,pn	%icc, 55f		! CTI
-	 andcc	%g1, %o4, %g0		! IEU1	Group
-	be,pn	%icc, 56f		! CTI
-	 andcc	%g1, 0xff, %g0		! IEU1	Group
-	be,a,pn	%icc, 57f		! CTI
-	 sub	%o0, %o3, %o0		! IEU0
-2:	cmp	%o0, %g3		! IEU1	Group
-	bl,a,pt	%xcc, 50b		! CTI
-62:	 ldxa	[%o1] %asi, %g1		! Load
-	retl				! CTI	Group
-	 mov	%o2, %o0		! IEU0
-50:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 8, %o0
-51:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 7, %o0
-52:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 6, %o0
-53:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 5, %o0
-54:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 4, %o0
-55:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 3, %o0
-56:	sub	%o0, %o3, %o0
-	retl
-	 sub	%o0, 2, %o0
-57:	retl
-	 sub	%o0, 1, %o0
-30:	brlez,pn %o2, 3f
-	 sub	%g0, %o2, %o3
-	add	%o0, %o2, %o0
-63:	lduba	[%o1] %asi, %o4
-1:	add	%o1, 1, %o1
-	brz,pn	%o4, 2f
-	 stb	%o4, [%o0 + %o3]
-	addcc	%o3, 1, %o3
-	bne,pt	%xcc, 1b
-64:	 lduba	[%o1] %asi, %o4
-3:	retl
-	 mov	%o2, %o0
-2:	retl
-	 add	%o2, %o3, %o0
-ENDPROC(__strncpy_from_user)
-
-	.section __ex_table,"a"
-	.align	4
-	.word	60b, __retl_efault
-	.word	61b, __retl_efault
-	.word	62b, __retl_efault
-	.word	63b, __retl_efault
-	.word	64b, __retl_efault
-	.previous

arch/sparc/lib/usercopy.c

 #include <linux/module.h>
+#include <linux/kernel.h>
 #include <linux/bug.h>
 
 void copy_from_user_overflow(void)

arch/sparc/mm/srmmu.c

@@ -467,33 +467,6 @@ void srmmu_unmapiorange(unsigned long virt_addr, unsigned int len)
 	flush_tlb_all();
 }
 
-/*
- * On the SRMMU we do not have the problems with limited tlb entries
- * for mapping kernel pages, so we just take things from the free page
- * pool.  As a side effect we are putting a little too much pressure
- * on the gfp() subsystem.  This setup also makes the logic of the
- * iommu mapping code a lot easier as we can transparently handle
- * mappings on the kernel stack without any special code.
- */
-struct thread_info *alloc_thread_info_node(struct task_struct *tsk, int node)
-{
-	struct thread_info *ret;
-
-	ret = (struct thread_info *)__get_free_pages(GFP_KERNEL,
-						     THREAD_INFO_ORDER);
-#ifdef CONFIG_DEBUG_STACK_USAGE
-	if (ret)
-		memset(ret, 0, PAGE_SIZE << THREAD_INFO_ORDER);
-#endif /* DEBUG_STACK_USAGE */
-
-	return ret;
-}
-
-void free_thread_info(struct thread_info *ti)
-{
-	free_pages((unsigned long)ti, THREAD_INFO_ORDER);
-}
-
 /* tsunami.S */
 extern void tsunami_flush_cache_all(void);
 extern void tsunami_flush_cache_mm(struct mm_struct *mm);

arch/x86/include/asm/word-at-a-time.h

 #ifndef _ASM_WORD_AT_A_TIME_H
 #define _ASM_WORD_AT_A_TIME_H
 
+#include <linux/kernel.h>
+
 /*
  * This is largely generic for little-endian machines, but the
  * optimal byte mask counting is probably going to be something
@@ -35,8 +37,6 @@ static inline long count_masked_bytes(long mask)
 
 #endif
 
-#define REPEAT_BYTE(x)	((~0ul / 0xff) * (x))
-
 /* Return the high bit set in the first byte that is a zero */
 static inline unsigned long has_zero(unsigned long a)
 {

fs/namei.c

@@ -16,6 +16,7 @@
 
 #include <linux/init.h>
 #include <linux/export.h>
+#include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/fs.h>
 #include <linux/namei.h>

include/linux/kernel.h

@@ -38,6 +38,8 @@
 
 #define STACK_MAGIC	0xdeadbeef
 
+#define REPEAT_BYTE(x)	((~0ul / 0xff) * (x))
+
 #define ALIGN(x, a)		__ALIGN_KERNEL((x), (a))
 #define __ALIGN_MASK(x, mask)	__ALIGN_KERNEL_MASK((x), (mask))
 #define PTR_ALIGN(p, a)		((typeof(p))ALIGN((unsigned long)(p), (a)))

lib/Kconfig

@@ -16,6 +16,9 @@ config BITREVERSE
 config RATIONAL
 	boolean
 
+config GENERIC_STRNCPY_FROM_USER
+	bool
+
 config GENERIC_FIND_FIRST_BIT
 	bool
 

lib/Makefile

@@ -125,6 +125,8 @@ obj-$(CONFIG_CLZ_TAB) += clz_tab.o
 
 obj-$(CONFIG_DDR) += jedec_ddr_data.o
 
+obj-$(CONFIG_GENERIC_STRNCPY_FROM_USER) += strncpy_from_user.o
+
 hostprogs-y	:= gen_crc32table
 clean-files	:= crc32table.h
 

lib/strncpy_from_user.c

+#include <linux/module.h>
+#include <linux/uaccess.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+
+#include <asm/byteorder.h>
+
+static inline long find_zero(unsigned long mask)
+{
+	long byte = 0;
+
+#ifdef __BIG_ENDIAN
+#ifdef CONFIG_64BIT
+	if (mask >> 32)
+		mask >>= 32;
+	else
+		byte = 4;
+#endif
+	if (mask >> 16)
+		mask >>= 16;
+	else
+		byte += 2;
+	return (mask >> 8) ? byte : byte + 1;
+#else
+#ifdef CONFIG_64BIT
+	if (!((unsigned int) mask)) {
+		mask >>= 32;
+		byte = 4;
+	}
+#endif
+	if (!(mask & 0xffff)) {
+		mask >>= 16;
+		byte += 2;
+	}
+	return (mask & 0xff) ? byte : byte + 1;
+#endif
+}
+
+#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+#define IS_UNALIGNED(src, dst)	0
+#else
+#define IS_UNALIGNED(src, dst)	\
+	(((long) dst | (long) src) & (sizeof(long) - 1))
+#endif
+
+/*
+ * Do a strncpy, return length of string without final '\0'.
+ * 'count' is the user-supplied count (return 'count' if we
+ * hit it), 'max' is the address space maximum (and we return
+ * -EFAULT if we hit it).
+ */
+static inline long do_strncpy_from_user(char *dst, const char __user *src, long count, unsigned long max)
+{
+	const unsigned long high_bits = REPEAT_BYTE(0xfe) + 1;
+	const unsigned long low_bits = REPEAT_BYTE(0x7f);
+	long res = 0;
+
+	/*
+	 * Truncate 'max' to the user-specified limit, so that
+	 * we only have one limit we need to check in the loop
+	 */
+	if (max > count)
+		max = count;
+
+	if (IS_UNALIGNED(src, dst))
+		goto byte_at_a_time;
+
+	while (max >= sizeof(unsigned long)) {
+		unsigned long c, v, rhs;
+
+		/* Fall back to byte-at-a-time if we get a page fault */
+		if (unlikely(__get_user(c,(unsigned long __user *)(src+res))))
+			break;
+		rhs = c | low_bits;
+		v = (c + high_bits) & ~rhs;
+		*(unsigned long *)(dst+res) = c;
+		if (v) {
+			v = (c & low_bits) + low_bits;
+			v = ~(v | rhs);
+			return res + find_zero(v);
+		}
+		res += sizeof(unsigned long);
+		max -= sizeof(unsigned long);
+	}
+
+byte_at_a_time:
+	while (max) {
+		char c;
+
+		if (unlikely(__get_user(c,src+res)))
+			return -EFAULT;
+		dst[res] = c;
+		if (!c)
+			return res;
+		res++;
+		max--;
+	}
+
+	/*
+	 * Uhhuh. We hit 'max'. But was that the user-specified maximum
+	 * too? If so, that's ok - we got as much as the user asked for.
+	 */
+	if (res >= count)
+		return res;
+
+	/*
+	 * Nope: we hit the address space limit, and we still had more
+	 * characters the caller would have wanted. That's an EFAULT.
+	 */
+	return -EFAULT;
+}
+
+/**
+ * strncpy_from_user: - Copy a NUL terminated string from userspace.
+ * @dst:   Destination address, in kernel space.  This buffer must be at
+ *         least @count bytes long.
+ * @src:   Source address, in user space.
+ * @count: Maximum number of bytes to copy, including the trailing NUL.
+ *
+ * Copies a NUL-terminated string from userspace to kernel space.
+ *
+ * On success, returns the length of the string (not including the trailing
+ * NUL).
+ *
+ * If access to userspace fails, returns -EFAULT (some data may have been
+ * copied).
+ *
+ * If @count is smaller than the length of the string, copies @count bytes
+ * and returns @count.
+ */
+long strncpy_from_user(char *dst, const char __user *src, long count)
+{
+	unsigned long max_addr, src_addr;
+
+	if (unlikely(count <= 0))
+		return 0;
+
+	max_addr = user_addr_max();
+	src_addr = (unsigned long)src;
+	if (likely(src_addr < max_addr)) {
+		unsigned long max = max_addr - src_addr;
+		return do_strncpy_from_user(dst, src, count, max);
+	}
+	return -EFAULT;
+}
+EXPORT_SYMBOL(strncpy_from_user);