Here's a set of minor updates for arch/alpha that should not
be controversial.
r~
The following changes since commit 47188d39b5deeebf41f87a02af1b3935866364cf:
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4 (2013-07-14 21:47:51 -0700)
are available in the git repository at:
git://github.com/rth7680/linux.git axp-next
for you to fetch changes up to fd60fe7052eceab1b1e7c8691263ec2b2634ace0:
alpha: Fix type compatibility warning for marvel_map_irq (2013-07-16 09:55:06 -0700)
----------------------------------------------------------------
Richard Henderson (7):
alpha: Add kcmp and finit_module syscalls
alpha: Eliminate compiler warning from memset macro
alpha: Modernize lib/mpi/longlong.h
alpha: Improve atomic_add_unless
alpha: Implement atomic64_dec_if_positive
alpha: Generate dwarf2 unwind info for various kernel entry points.
alpha: Fix type compatibility warning for marvel_map_irq
arch/alpha/Kconfig | 1 +
arch/alpha/include/asm/atomic.h | 88 ++++++--
arch/alpha/include/asm/string.h | 18 +-
arch/alpha/include/asm/unistd.h | 3 +-
arch/alpha/include/uapi/asm/unistd.h | 2 +
arch/alpha/kernel/entry.S | 399 +++++++++++++++++++++++++----------
arch/alpha/kernel/sys_marvel.c | 3 +-
arch/alpha/kernel/systbls.S | 2 +
lib/mpi/longlong.h | 17 +-
9 files changed, 380 insertions(+), 153 deletions(-)
Compiling with GCC 4.8 yields several instances of
crypto/vmac.c: In function ‘vmac_final’:
crypto/vmac.c:616:9: warning: value computed is not used [-Wunused-value]
memset(&mac, 0, sizeof(vmac_t));
^
arch/alpha/include/asm/string.h:31:25: note: in definition of macro ‘memset’
? __builtin_memset((s),0,(n)) \
^
Converting the macro to an inline function eliminates this problem.
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/include/asm/string.h | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/arch/alpha/include/asm/string.h b/arch/alpha/include/asm/string.h
index b02b8a2..8b478ae 100644
--- a/arch/alpha/include/asm/string.h
+++ b/arch/alpha/include/asm/string.h
@@ -25,12 +25,18 @@ extern void * __constant_c_memset(void *, unsigned long, size_t);
extern void * __memset(void *, int, size_t);
extern void * memset(void *, int, size_t);
-#define memset(s, c, n) \
-(__builtin_constant_p(c) \
- ? (__builtin_constant_p(n) && (c) == 0 \
- ? __builtin_memset((s),0,(n)) \
- : __constant_c_memset((s),0x0101010101010101UL*(unsigned char)(c),(n))) \
- : __memset((s),(c),(n)))
+extern inline void *memset(void *s, int c, size_t n)
+{
+ if (__builtin_constant_p(c)) {
+ if (__builtin_constant_p(n)) {
+ return __builtin_memset(s, c, n);
+ } else {
+ unsigned long c8 = (c & 0xff) * 0x0101010101010101UL;
+ return __constant_c_memset(s, c8, n);
+ }
+ }
+ return __memset(s, c, n);
+}
#define __HAVE_ARCH_STRCPY
extern char * strcpy(char *,const char *);
--
1.8.1.4
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/Kconfig | 1 +
arch/alpha/include/asm/atomic.h | 28 ++++++++++++++++++++++++++++
2 files changed, 29 insertions(+)
diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig
index 837a1f2..082d9b4 100644
--- a/arch/alpha/Kconfig
+++ b/arch/alpha/Kconfig
@@ -15,6 +15,7 @@ config ALPHA
select ARCH_WANT_OPTIONAL_GPIOLIB
select ARCH_WANT_IPC_PARSE_VERSION
select ARCH_HAVE_NMI_SAFE_CMPXCHG
+ select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE
select GENERIC_SMP_IDLE_THREAD
select GENERIC_CMOS_UPDATE
select GENERIC_STRNCPY_FROM_USER
diff --git a/arch/alpha/include/asm/atomic.h b/arch/alpha/include/asm/atomic.h
index 0dc18fc..78b03ef 100644
--- a/arch/alpha/include/asm/atomic.h
+++ b/arch/alpha/include/asm/atomic.h
@@ -238,6 +238,34 @@ static __inline__ int atomic64_add_unless(atomic64_t *v, long a, long u)
return !c;
}
+/*
+ * atomic64_dec_if_positive - decrement by 1 if old value positive
+ * @v: pointer of type atomic_t
+ *
+ * The function returns the old value of *v minus 1, even if
+ * the atomic variable, v, was not decremented.
+ */
+static inline long atomic64_dec_if_positive(atomic64_t *v)
+{
+ long old, tmp;
+ smp_mb();
+ __asm__ __volatile__(
+ "1: ldq_l %[old],%[mem]\n"
+ " subq %[old],1,%[tmp]\n"
+ " ble %[old],2f\n"
+ " stq_c %[tmp],%[mem]\n"
+ " beq %[tmp],3f\n"
+ "2:\n"
+ ".subsection 2\n"
+ "3: br 1b\n"
+ ".previous"
+ : [old] "=&r"(old), [tmp] "=&r"(tmp)
+ : [mem] "m"(*v)
+ : "memory");
+ smp_mb();
+ return old - 1;
+}
+
#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
#define atomic_add_negative(a, v) (atomic_add_return((a), (v)) < 0)
--
1.8.1.4
Use ll/sc loops instead of C loops around cmpxchg.
Update the atomic64_add_unless block comment to match the code.
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/include/asm/atomic.h | 60 +++++++++++++++++++++++++----------------
1 file changed, 37 insertions(+), 23 deletions(-)
diff --git a/arch/alpha/include/asm/atomic.h b/arch/alpha/include/asm/atomic.h
index c2cbe4f..0dc18fc 100644
--- a/arch/alpha/include/asm/atomic.h
+++ b/arch/alpha/include/asm/atomic.h
@@ -186,17 +186,24 @@ static __inline__ long atomic64_sub_return(long i, atomic64_t * v)
*/
static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u)
{
- int c, old;
- c = atomic_read(v);
- for (;;) {
- if (unlikely(c == (u)))
- break;
- old = atomic_cmpxchg((v), c, c + (a));
- if (likely(old == c))
- break;
- c = old;
- }
- return c;
+ int c, new, old;
+ smp_mb();
+ __asm__ __volatile__(
+ "1: ldl_l %[old],%[mem]\n"
+ " cmpeq %[old],%[u],%[c]\n"
+ " addl %[old],%[a],%[new]\n"
+ " bne %[c],2f\n"
+ " stl_c %[new],%[mem]\n"
+ " beq %[new],3f\n"
+ "2:\n"
+ ".subsection 2\n"
+ "3: br 1b\n"
+ ".previous"
+ : [old] "=&r"(old), [new] "=&r"(new), [c] "=&r"(c)
+ : [mem] "m"(*v), [a] "rI"(a), [u] "rI"((long)u)
+ : "memory");
+ smp_mb();
+ return old;
}
@@ -207,21 +214,28 @@ static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u)
* @u: ...unless v is equal to u.
*
* Atomically adds @a to @v, so long as it was not @u.
- * Returns the old value of @v.
+ * Returns true iff @v was not @u.
*/
static __inline__ int atomic64_add_unless(atomic64_t *v, long a, long u)
{
- long c, old;
- c = atomic64_read(v);
- for (;;) {
- if (unlikely(c == (u)))
- break;
- old = atomic64_cmpxchg((v), c, c + (a));
- if (likely(old == c))
- break;
- c = old;
- }
- return c != (u);
+ long c, tmp;
+ smp_mb();
+ __asm__ __volatile__(
+ "1: ldq_l %[tmp],%[mem]\n"
+ " cmpeq %[tmp],%[u],%[c]\n"
+ " addq %[tmp],%[a],%[tmp]\n"
+ " bne %[c],2f\n"
+ " stq_c %[tmp],%[mem]\n"
+ " beq %[tmp],3f\n"
+ "2:\n"
+ ".subsection 2\n"
+ "3: br 1b\n"
+ ".previous"
+ : [tmp] "=&r"(tmp), [c] "=&r"(c)
+ : [mem] "m"(*v), [a] "rI"(a), [u] "rI"(u)
+ : "memory");
+ smp_mb();
+ return !c;
}
#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
--
1.8.1.4
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/kernel/sys_marvel.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/alpha/kernel/sys_marvel.c b/arch/alpha/kernel/sys_marvel.c
index 407accc..53d6e4a 100644
--- a/arch/alpha/kernel/sys_marvel.c
+++ b/arch/alpha/kernel/sys_marvel.c
@@ -317,8 +317,9 @@ marvel_init_irq(void)
}
static int
-marvel_map_irq(struct pci_dev *dev, u8 slot, u8 pin)
+marvel_map_irq(const struct pci_dev *cdev, u8 slot, u8 pin)
{
+ struct pci_dev *dev = (struct pci_dev *)dev;
struct pci_controller *hose = dev->sysdata;
struct io7_port *io7_port = hose->sysdata;
struct io7 *io7 = io7_port->io7;
--
1.8.1.4
Having unwind info past the PALcode generated stack frame makes
debugging the kernel significantly easier.
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/kernel/entry.S | 399 +++++++++++++++++++++++++++++++++-------------
1 file changed, 288 insertions(+), 111 deletions(-)
diff --git a/arch/alpha/kernel/entry.S b/arch/alpha/kernel/entry.S
index f62a994..a969b95 100644
--- a/arch/alpha/kernel/entry.S
+++ b/arch/alpha/kernel/entry.S
@@ -12,11 +12,32 @@
.text
.set noat
+ .cfi_sections .debug_frame
/* Stack offsets. */
#define SP_OFF 184
#define SWITCH_STACK_SIZE 320
+.macro CFI_START_OSF_FRAME func
+ .align 4
+ .globl \func
+ .type \func,@function
+\func:
+ .cfi_startproc simple
+ .cfi_return_column 64
+ .cfi_def_cfa $sp, 48
+ .cfi_rel_offset 64, 8
+ .cfi_rel_offset $gp, 16
+ .cfi_rel_offset $16, 24
+ .cfi_rel_offset $17, 32
+ .cfi_rel_offset $18, 40
+.endm
+
+.macro CFI_END_OSF_FRAME func
+ .cfi_endproc
+ .size \func, . - \func
+.endm
+
/*
* This defines the normal kernel pt-regs layout.
*
@@ -27,100 +48,158 @@
* the palcode-provided values are available to the signal handler.
*/
-#define SAVE_ALL \
- subq $sp, SP_OFF, $sp; \
- stq $0, 0($sp); \
- stq $1, 8($sp); \
- stq $2, 16($sp); \
- stq $3, 24($sp); \
- stq $4, 32($sp); \
- stq $28, 144($sp); \
- lda $2, alpha_mv; \
- stq $5, 40($sp); \
- stq $6, 48($sp); \
- stq $7, 56($sp); \
- stq $8, 64($sp); \
- stq $19, 72($sp); \
- stq $20, 80($sp); \
- stq $21, 88($sp); \
- ldq $2, HAE_CACHE($2); \
- stq $22, 96($sp); \
- stq $23, 104($sp); \
- stq $24, 112($sp); \
- stq $25, 120($sp); \
- stq $26, 128($sp); \
- stq $27, 136($sp); \
- stq $2, 152($sp); \
- stq $16, 160($sp); \
- stq $17, 168($sp); \
+.macro SAVE_ALL
+ subq $sp, SP_OFF, $sp
+ .cfi_adjust_cfa_offset SP_OFF
+ stq $0, 0($sp)
+ stq $1, 8($sp)
+ stq $2, 16($sp)
+ stq $3, 24($sp)
+ stq $4, 32($sp)
+ stq $28, 144($sp)
+ .cfi_rel_offset $0, 0
+ .cfi_rel_offset $1, 8
+ .cfi_rel_offset $2, 16
+ .cfi_rel_offset $3, 24
+ .cfi_rel_offset $4, 32
+ .cfi_rel_offset $28, 144
+ lda $2, alpha_mv
+ stq $5, 40($sp)
+ stq $6, 48($sp)
+ stq $7, 56($sp)
+ stq $8, 64($sp)
+ stq $19, 72($sp)
+ stq $20, 80($sp)
+ stq $21, 88($sp)
+ ldq $2, HAE_CACHE($2)
+ stq $22, 96($sp)
+ stq $23, 104($sp)
+ stq $24, 112($sp)
+ stq $25, 120($sp)
+ stq $26, 128($sp)
+ stq $27, 136($sp)
+ stq $2, 152($sp)
+ stq $16, 160($sp)
+ stq $17, 168($sp)
stq $18, 176($sp)
+ .cfi_rel_offset $5, 40
+ .cfi_rel_offset $6, 48
+ .cfi_rel_offset $7, 56
+ .cfi_rel_offset $8, 64
+ .cfi_rel_offset $19, 72
+ .cfi_rel_offset $20, 80
+ .cfi_rel_offset $21, 88
+ .cfi_rel_offset $22, 96
+ .cfi_rel_offset $23, 104
+ .cfi_rel_offset $24, 112
+ .cfi_rel_offset $25, 120
+ .cfi_rel_offset $26, 128
+ .cfi_rel_offset $27, 136
+.endm
-#define RESTORE_ALL \
- lda $19, alpha_mv; \
- ldq $0, 0($sp); \
- ldq $1, 8($sp); \
- ldq $2, 16($sp); \
- ldq $3, 24($sp); \
- ldq $21, 152($sp); \
- ldq $20, HAE_CACHE($19); \
- ldq $4, 32($sp); \
- ldq $5, 40($sp); \
- ldq $6, 48($sp); \
- ldq $7, 56($sp); \
- subq $20, $21, $20; \
- ldq $8, 64($sp); \
- beq $20, 99f; \
- ldq $20, HAE_REG($19); \
- stq $21, HAE_CACHE($19); \
- stq $21, 0($20); \
-99:; \
- ldq $19, 72($sp); \
- ldq $20, 80($sp); \
- ldq $21, 88($sp); \
- ldq $22, 96($sp); \
- ldq $23, 104($sp); \
- ldq $24, 112($sp); \
- ldq $25, 120($sp); \
- ldq $26, 128($sp); \
- ldq $27, 136($sp); \
- ldq $28, 144($sp); \
+.macro RESTORE_ALL
+ lda $19, alpha_mv
+ ldq $0, 0($sp)
+ ldq $1, 8($sp)
+ ldq $2, 16($sp)
+ ldq $3, 24($sp)
+ ldq $21, 152($sp)
+ ldq $20, HAE_CACHE($19)
+ ldq $4, 32($sp)
+ ldq $5, 40($sp)
+ ldq $6, 48($sp)
+ ldq $7, 56($sp)
+ subq $20, $21, $20
+ ldq $8, 64($sp)
+ beq $20, 99f
+ ldq $20, HAE_REG($19)
+ stq $21, HAE_CACHE($19)
+ stq $21, 0($20)
+99: ldq $19, 72($sp)
+ ldq $20, 80($sp)
+ ldq $21, 88($sp)
+ ldq $22, 96($sp)
+ ldq $23, 104($sp)
+ ldq $24, 112($sp)
+ ldq $25, 120($sp)
+ ldq $26, 128($sp)
+ ldq $27, 136($sp)
+ ldq $28, 144($sp)
addq $sp, SP_OFF, $sp
+ .cfi_restore $0
+ .cfi_restore $1
+ .cfi_restore $2
+ .cfi_restore $3
+ .cfi_restore $4
+ .cfi_restore $5
+ .cfi_restore $6
+ .cfi_restore $7
+ .cfi_restore $8
+ .cfi_restore $19
+ .cfi_restore $20
+ .cfi_restore $21
+ .cfi_restore $22
+ .cfi_restore $23
+ .cfi_restore $24
+ .cfi_restore $25
+ .cfi_restore $26
+ .cfi_restore $27
+ .cfi_restore $28
+ .cfi_adjust_cfa_offset -SP_OFF
+.endm
+
+.macro DO_SWITCH_STACK
+ bsr $1, do_switch_stack
+ .cfi_adjust_cfa_offset SWITCH_STACK_SIZE
+ .cfi_rel_offset $9, 0
+ .cfi_rel_offset $10, 8
+ .cfi_rel_offset $11, 16
+ .cfi_rel_offset $12, 24
+ .cfi_rel_offset $13, 32
+ .cfi_rel_offset $14, 40
+ .cfi_rel_offset $15, 48
+ /* We don't really care about the FP registers for debugging. */
+.endm
+
+.macro UNDO_SWITCH_STACK
+ bsr $1, undo_switch_stack
+ .cfi_restore $9
+ .cfi_restore $10
+ .cfi_restore $11
+ .cfi_restore $12
+ .cfi_restore $13
+ .cfi_restore $14
+ .cfi_restore $15
+ .cfi_adjust_cfa_offset -SWITCH_STACK_SIZE
+.endm
/*
* Non-syscall kernel entry points.
*/
- .align 4
- .globl entInt
- .ent entInt
-entInt:
+CFI_START_OSF_FRAME entInt
SAVE_ALL
lda $8, 0x3fff
lda $26, ret_from_sys_call
bic $sp, $8, $8
mov $sp, $19
jsr $31, do_entInt
-.end entInt
+CFI_END_OSF_FRAME entInt
- .align 4
- .globl entArith
- .ent entArith
-entArith:
+CFI_START_OSF_FRAME entArith
SAVE_ALL
lda $8, 0x3fff
lda $26, ret_from_sys_call
bic $sp, $8, $8
mov $sp, $18
jsr $31, do_entArith
-.end entArith
+CFI_END_OSF_FRAME entArith
- .align 4
- .globl entMM
- .ent entMM
-entMM:
+CFI_START_OSF_FRAME entMM
SAVE_ALL
/* save $9 - $15 so the inline exception code can manipulate them. */
subq $sp, 56, $sp
+ .cfi_adjust_cfa_offset 56
stq $9, 0($sp)
stq $10, 8($sp)
stq $11, 16($sp)
@@ -128,6 +207,13 @@ entMM:
stq $13, 32($sp)
stq $14, 40($sp)
stq $15, 48($sp)
+ .cfi_rel_offset $9, 0
+ .cfi_rel_offset $10, 8
+ .cfi_rel_offset $11, 16
+ .cfi_rel_offset $12, 24
+ .cfi_rel_offset $13, 32
+ .cfi_rel_offset $14, 40
+ .cfi_rel_offset $15, 48
addq $sp, 56, $19
/* handle the fault */
lda $8, 0x3fff
@@ -142,28 +228,33 @@ entMM:
ldq $14, 40($sp)
ldq $15, 48($sp)
addq $sp, 56, $sp
+ .cfi_restore $9
+ .cfi_restore $10
+ .cfi_restore $11
+ .cfi_restore $12
+ .cfi_restore $13
+ .cfi_restore $14
+ .cfi_restore $15
+ .cfi_adjust_cfa_offset -56
/* finish up the syscall as normal. */
br ret_from_sys_call
-.end entMM
+CFI_END_OSF_FRAME entMM
- .align 4
- .globl entIF
- .ent entIF
-entIF:
+CFI_START_OSF_FRAME entIF
SAVE_ALL
lda $8, 0x3fff
lda $26, ret_from_sys_call
bic $sp, $8, $8
mov $sp, $17
jsr $31, do_entIF
-.end entIF
+CFI_END_OSF_FRAME entIF
- .align 4
- .globl entUna
- .ent entUna
-entUna:
+CFI_START_OSF_FRAME entUna
lda $sp, -256($sp)
+ .cfi_adjust_cfa_offset 256
stq $0, 0($sp)
+ .cfi_rel_offset $0, 0
+ .cfi_remember_state
ldq $0, 256($sp) /* get PS */
stq $1, 8($sp)
stq $2, 16($sp)
@@ -195,6 +286,32 @@ entUna:
stq $28, 224($sp)
mov $sp, $19
stq $gp, 232($sp)
+ .cfi_rel_offset $1, 1*8
+ .cfi_rel_offset $2, 2*8
+ .cfi_rel_offset $3, 3*8
+ .cfi_rel_offset $4, 4*8
+ .cfi_rel_offset $5, 5*8
+ .cfi_rel_offset $6, 6*8
+ .cfi_rel_offset $7, 7*8
+ .cfi_rel_offset $8, 8*8
+ .cfi_rel_offset $9, 9*8
+ .cfi_rel_offset $10, 10*8
+ .cfi_rel_offset $11, 11*8
+ .cfi_rel_offset $12, 12*8
+ .cfi_rel_offset $13, 13*8
+ .cfi_rel_offset $14, 14*8
+ .cfi_rel_offset $15, 15*8
+ .cfi_rel_offset $19, 19*8
+ .cfi_rel_offset $20, 20*8
+ .cfi_rel_offset $21, 21*8
+ .cfi_rel_offset $22, 22*8
+ .cfi_rel_offset $23, 23*8
+ .cfi_rel_offset $24, 24*8
+ .cfi_rel_offset $25, 25*8
+ .cfi_rel_offset $26, 26*8
+ .cfi_rel_offset $27, 27*8
+ .cfi_rel_offset $28, 28*8
+ .cfi_rel_offset $29, 29*8
lda $8, 0x3fff
stq $31, 248($sp)
bic $sp, $8, $8
@@ -228,16 +345,45 @@ entUna:
ldq $28, 224($sp)
ldq $gp, 232($sp)
lda $sp, 256($sp)
+ .cfi_restore $1
+ .cfi_restore $2
+ .cfi_restore $3
+ .cfi_restore $4
+ .cfi_restore $5
+ .cfi_restore $6
+ .cfi_restore $7
+ .cfi_restore $8
+ .cfi_restore $9
+ .cfi_restore $10
+ .cfi_restore $11
+ .cfi_restore $12
+ .cfi_restore $13
+ .cfi_restore $14
+ .cfi_restore $15
+ .cfi_restore $19
+ .cfi_restore $20
+ .cfi_restore $21
+ .cfi_restore $22
+ .cfi_restore $23
+ .cfi_restore $24
+ .cfi_restore $25
+ .cfi_restore $26
+ .cfi_restore $27
+ .cfi_restore $28
+ .cfi_restore $29
+ .cfi_adjust_cfa_offset -256
call_pal PAL_rti
-.end entUna
.align 4
- .ent entUnaUser
entUnaUser:
+ .cfi_restore_state
ldq $0, 0($sp) /* restore original $0 */
lda $sp, 256($sp) /* pop entUna's stack frame */
+ .cfi_restore $0
+ .cfi_adjust_cfa_offset -256
SAVE_ALL /* setup normal kernel stack */
lda $sp, -56($sp)
+ .cfi_adjust_cfa_offset 56
stq $9, 0($sp)
stq $10, 8($sp)
stq $11, 16($sp)
@@ -245,6 +391,13 @@ entUnaUser:
stq $13, 32($sp)
stq $14, 40($sp)
stq $15, 48($sp)
+ .cfi_rel_offset $9, 0
+ .cfi_rel_offset $10, 8
+ .cfi_rel_offset $11, 16
+ .cfi_rel_offset $12, 24
+ .cfi_rel_offset $13, 32
+ .cfi_rel_offset $14, 40
+ .cfi_rel_offset $15, 48
lda $8, 0x3fff
addq $sp, 56, $19
bic $sp, $8, $8
@@ -257,20 +410,25 @@ entUnaUser:
ldq $14, 40($sp)
ldq $15, 48($sp)
lda $sp, 56($sp)
+ .cfi_restore $9
+ .cfi_restore $10
+ .cfi_restore $11
+ .cfi_restore $12
+ .cfi_restore $13
+ .cfi_restore $14
+ .cfi_restore $15
+ .cfi_adjust_cfa_offset -56
br ret_from_sys_call
-.end entUnaUser
+CFI_END_OSF_FRAME entUna
- .align 4
- .globl entDbg
- .ent entDbg
-entDbg:
+CFI_START_OSF_FRAME entDbg
SAVE_ALL
lda $8, 0x3fff
lda $26, ret_from_sys_call
bic $sp, $8, $8
mov $sp, $16
jsr $31, do_entDbg
-.end entDbg
+CFI_END_OSF_FRAME entDbg
/*
* The system call entry point is special. Most importantly, it looks
@@ -285,8 +443,12 @@ entDbg:
.align 4
.globl entSys
- .globl ret_from_sys_call
- .ent entSys
+ .type entSys, @function
+ .cfi_startproc simple
+ .cfi_return_column 64
+ .cfi_def_cfa $sp, 48
+ .cfi_rel_offset 64, 8
+ .cfi_rel_offset $gp, 16
entSys:
SAVE_ALL
lda $8, 0x3fff
@@ -300,6 +462,9 @@ entSys:
stq $17, SP_OFF+32($sp)
s8addq $0, $5, $5
stq $18, SP_OFF+40($sp)
+ .cfi_rel_offset $16, SP_OFF+24
+ .cfi_rel_offset $17, SP_OFF+32
+ .cfi_rel_offset $18, SP_OFF+40
blbs $3, strace
beq $4, 1f
ldq $27, 0($5)
@@ -310,6 +475,7 @@ entSys:
stq $31, 72($sp) /* a3=0 => no error */
.align 4
+ .globl ret_from_sys_call
ret_from_sys_call:
cmovne $26, 0, $18 /* $18 = 0 => non-restartable */
ldq $0, SP_OFF($sp)
@@ -324,10 +490,12 @@ ret_to_user:
and $17, _TIF_WORK_MASK, $2
bne $2, work_pending
restore_all:
+ .cfi_remember_state
RESTORE_ALL
call_pal PAL_rti
ret_to_kernel:
+ .cfi_restore_state
lda $16, 7
call_pal PAL_swpipl
br restore_all
@@ -356,7 +524,6 @@ $ret_success:
stq $0, 0($sp)
stq $31, 72($sp) /* a3=0 => no error */
br ret_from_sys_call
-.end entSys
/*
* Do all cleanup when returning from all interrupts and system calls.
@@ -370,7 +537,7 @@ $ret_success:
*/
.align 4
- .ent work_pending
+ .type work_pending, @function
work_pending:
and $17, _TIF_NOTIFY_RESUME | _TIF_SIGPENDING, $2
bne $2, $work_notifysig
@@ -387,23 +554,22 @@ $work_resched:
$work_notifysig:
mov $sp, $16
- bsr $1, do_switch_stack
+ DO_SWITCH_STACK
jsr $26, do_work_pending
- bsr $1, undo_switch_stack
+ UNDO_SWITCH_STACK
br restore_all
-.end work_pending
/*
* PTRACE syscall handler
*/
.align 4
- .ent strace
+ .type strace, @function
strace:
/* set up signal stack, call syscall_trace */
- bsr $1, do_switch_stack
+ DO_SWITCH_STACK
jsr $26, syscall_trace_enter /* returns the syscall number */
- bsr $1, undo_switch_stack
+ UNDO_SWITCH_STACK
/* get the arguments back.. */
ldq $16, SP_OFF+24($sp)
@@ -431,9 +597,9 @@ ret_from_straced:
$strace_success:
stq $0, 0($sp) /* save return value */
- bsr $1, do_switch_stack
+ DO_SWITCH_STACK
jsr $26, syscall_trace_leave
- bsr $1, undo_switch_stack
+ UNDO_SWITCH_STACK
br $31, ret_from_sys_call
.align 3
@@ -447,26 +613,31 @@ $strace_error:
stq $0, 0($sp)
stq $1, 72($sp) /* a3 for return */
- bsr $1, do_switch_stack
+ DO_SWITCH_STACK
mov $18, $9 /* save old syscall number */
mov $19, $10 /* save old a3 */
jsr $26, syscall_trace_leave
mov $9, $18
mov $10, $19
- bsr $1, undo_switch_stack
+ UNDO_SWITCH_STACK
mov $31, $26 /* tell "ret_from_sys_call" we can restart */
br ret_from_sys_call
-.end strace
+CFI_END_OSF_FRAME entSys
/*
* Save and restore the switch stack -- aka the balance of the user context.
*/
.align 4
- .ent do_switch_stack
+ .type do_switch_stack, @function
+ .cfi_startproc simple
+ .cfi_return_column 64
+ .cfi_def_cfa $sp, 0
+ .cfi_register 64, $1
do_switch_stack:
lda $sp, -SWITCH_STACK_SIZE($sp)
+ .cfi_adjust_cfa_offset SWITCH_STACK_SIZE
stq $9, 0($sp)
stq $10, 8($sp)
stq $11, 16($sp)
@@ -510,10 +681,14 @@ do_switch_stack:
stt $f0, 312($sp) # save fpcr in slot of $f31
ldt $f0, 64($sp) # dont let "do_switch_stack" change fp state.
ret $31, ($1), 1
-.end do_switch_stack
+ .cfi_endproc
+ .size do_switch_stack, .-do_switch_stack
.align 4
- .ent undo_switch_stack
+ .type undo_switch_stack, @function
+ .cfi_startproc simple
+ .cfi_def_cfa $sp, 0
+ .cfi_register 64, $1
undo_switch_stack:
ldq $9, 0($sp)
ldq $10, 8($sp)
@@ -558,7 +733,8 @@ undo_switch_stack:
ldt $f30, 304($sp)
lda $sp, SWITCH_STACK_SIZE($sp)
ret $31, ($1), 1
-.end undo_switch_stack
+ .cfi_endproc
+ .size undo_switch_stack, .-undo_switch_stack
/*
* The meat of the context switch code.
@@ -566,17 +742,18 @@ undo_switch_stack:
.align 4
.globl alpha_switch_to
- .ent alpha_switch_to
+ .type alpha_switch_to, @function
+ .cfi_startproc
alpha_switch_to:
- .prologue 0
- bsr $1, do_switch_stack
+ DO_SWITCH_STACK
call_pal PAL_swpctx
lda $8, 0x3fff
- bsr $1, undo_switch_stack
+ UNDO_SWITCH_STACK
bic $sp, $8, $8
mov $17, $0
ret
-.end alpha_switch_to
+ .cfi_endproc
+ .size alpha_switch_to, .-alpha_switch_to
/*
* New processes begin life here.
--
1.8.1.4
Remove the compile warning for __udiv_qrnnd not having a prototype.
Use the __builtin_alpha_umulh introduced in gcc 4.0.
Signed-off-by: Richard Henderson <[email protected]>
---
lib/mpi/longlong.h | 17 +++++++----------
1 file changed, 7 insertions(+), 10 deletions(-)
diff --git a/lib/mpi/longlong.h b/lib/mpi/longlong.h
index d411355..aac5114 100644
--- a/lib/mpi/longlong.h
+++ b/lib/mpi/longlong.h
@@ -151,15 +151,12 @@ do { \
#endif /* __a29k__ */
#if defined(__alpha) && W_TYPE_SIZE == 64
-#define umul_ppmm(ph, pl, m0, m1) \
-do { \
- UDItype __m0 = (m0), __m1 = (m1); \
- __asm__ ("umulh %r1,%2,%0" \
- : "=r" ((UDItype) ph) \
- : "%rJ" (__m0), \
- "rI" (__m1)); \
- (pl) = __m0 * __m1; \
- } while (0)
+#define umul_ppmm(ph, pl, m0, m1) \
+do { \
+ UDItype __m0 = (m0), __m1 = (m1); \
+ (ph) = __builtin_alpha_umulh(__m0, __m1); \
+ (pl) = __m0 * __m1; \
+} while (0)
#define UMUL_TIME 46
#ifndef LONGLONG_STANDALONE
#define udiv_qrnnd(q, r, n1, n0, d) \
@@ -167,7 +164,7 @@ do { UDItype __r; \
(q) = __udiv_qrnnd(&__r, (n1), (n0), (d)); \
(r) = __r; \
} while (0)
-extern UDItype __udiv_qrnnd();
+extern UDItype __udiv_qrnnd(UDItype *, UDItype, UDItype, UDItype);
#define UDIV_TIME 220
#endif /* LONGLONG_STANDALONE */
#endif /* __alpha */
--
1.8.1.4
Signed-off-by: Richard Henderson <[email protected]>
---
arch/alpha/include/asm/unistd.h | 3 +--
arch/alpha/include/uapi/asm/unistd.h | 2 ++
arch/alpha/kernel/systbls.S | 2 ++
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/alpha/include/asm/unistd.h b/arch/alpha/include/asm/unistd.h
index 43baee1..f2c9440 100644
--- a/arch/alpha/include/asm/unistd.h
+++ b/arch/alpha/include/asm/unistd.h
@@ -3,8 +3,7 @@
#include <uapi/asm/unistd.h>
-
-#define NR_SYSCALLS 506
+#define NR_SYSCALLS 508
#define __ARCH_WANT_OLD_READDIR
#define __ARCH_WANT_STAT64
diff --git a/arch/alpha/include/uapi/asm/unistd.h b/arch/alpha/include/uapi/asm/unistd.h
index 801d28b..53ae7bb 100644
--- a/arch/alpha/include/uapi/asm/unistd.h
+++ b/arch/alpha/include/uapi/asm/unistd.h
@@ -467,5 +467,7 @@
#define __NR_sendmmsg 503
#define __NR_process_vm_readv 504
#define __NR_process_vm_writev 505
+#define __NR_kcmp 506
+#define __NR_finit_module 507
#endif /* _UAPI_ALPHA_UNISTD_H */
diff --git a/arch/alpha/kernel/systbls.S b/arch/alpha/kernel/systbls.S
index 4284ec7..dca9b3f 100644
--- a/arch/alpha/kernel/systbls.S
+++ b/arch/alpha/kernel/systbls.S
@@ -524,6 +524,8 @@ sys_call_table:
.quad sys_sendmmsg
.quad sys_process_vm_readv
.quad sys_process_vm_writev /* 505 */
+ .quad sys_kcmp
+ .quad sys_finit_module
.size sys_call_table, . - sys_call_table
.type sys_call_table, @object
--
1.8.1.4
On Tue, 2013-07-16 at 10:04 -0700, Richard Henderson wrote:
> Remove the compile warning for __udiv_qrnnd not having a prototype.
> Use the __builtin_alpha_umulh introduced in gcc 4.0.
Isn't gcc 3.x still a supported compiler?
> diff --git a/lib/mpi/longlong.h b/lib/mpi/longlong.h
[]
> @@ -151,15 +151,12 @@ do { \
> #endif /* __a29k__ */
>
> #if defined(__alpha) && W_TYPE_SIZE == 64
> -#define umul_ppmm(ph, pl, m0, m1) \
> -do { \
> - UDItype __m0 = (m0), __m1 = (m1); \
> - __asm__ ("umulh %r1,%2,%0" \
> - : "=r" ((UDItype) ph) \
> - : "%rJ" (__m0), \
> - "rI" (__m1)); \
> - (pl) = __m0 * __m1; \
> - } while (0)
> +#define umul_ppmm(ph, pl, m0, m1) \
> +do { \
> + UDItype __m0 = (m0), __m1 = (m1); \
> + (ph) = __builtin_alpha_umulh(__m0, __m1); \
> + (pl) = __m0 * __m1; \
> +} while (0)
> #define UMUL_TIME 46
> #ifndef LONGLONG_STANDALONE
> #define udiv_qrnnd(q, r, n1, n0, d) \
> @@ -167,7 +164,7 @@ do { UDItype __r; \
> (q) = __udiv_qrnnd(&__r, (n1), (n0), (d)); \
> (r) = __r; \
> } while (0)
> -extern UDItype __udiv_qrnnd();
> +extern UDItype __udiv_qrnnd(UDItype *, UDItype, UDItype, UDItype);
> #define UDIV_TIME 220
> #endif /* LONGLONG_STANDALONE */
> #endif /* __alpha */
On 07/16/2013 10:19 AM, Joe Perches wrote:
> On Tue, 2013-07-16 at 10:04 -0700, Richard Henderson wrote:
>> Remove the compile warning for __udiv_qrnnd not having a prototype.
>> Use the __builtin_alpha_umulh introduced in gcc 4.0.
>
> Isn't gcc 3.x still a supported compiler?
Further investigation shows the builtin was added for gcc 3.2.
I just failed to update gcc's own longlong.h til 3 years later.
r~
On 07/16/2013 12:04:33 PM, Richard Henderson wrote:
> Here's a set of minor updates for arch/alpha that should not
> be controversial.
I also note that I had to do this to get busybox to build against
uClibc:
diff --git a/arch/alpha/include/uapi/asm/unistd.h
b/arch/alpha/include/uapi/asm/unistd.h
index 801d28b..1146e78 100644
--- a/arch/alpha/include/uapi/asm/unistd.h
+++ b/arch/alpha/include/uapi/asm/unistd.h
@@ -23,7 +23,7 @@
#define __NR_lseek 19
#define __NR_getxpid 20
#define __NR_osf_mount 21
-#define __NR_umount 22
+#define __NR_umount2 22
#define __NR_setuid 23
#define __NR_getxuid 24
#define __NR_exec_with_loader 25 /* not implemented */
@@ -253,7 +253,6 @@
#define __IGNORE_pause
#define __IGNORE_time
#define __IGNORE_utime
-#define __IGNORE_umount2
/*
* Linux-specific system calls begin at 300
@@ -279,7 +278,7 @@
#define __NR_sysinfo 318
#define __NR__sysctl 319
/* 320 was sys_idle. */
-#define __NR_oldumount 321
+#define __NR_umount 321
#define __NR_swapon 322
#define __NR_times 323
#define __NR_personality 324
(Because busybox uses umount2() directly. There's a man page for it and
everything.)
Rob-
On Tue, Jul 16, 2013 at 06:35:07PM -0500, Rob Landley wrote:
> On 07/16/2013 12:04:33 PM, Richard Henderson wrote:
> >Here's a set of minor updates for arch/alpha that should not
> >be controversial.
>
> I also note that I had to do this to get busybox to build against
> uClibc:
> -#define __NR_umount 22
> +#define __NR_umount2 22
> -#define __NR_oldumount 321
> +#define __NR_umount 321
I anticipate that this will likely break userspace.
busybox should be fixed to test for __NR_oldumount and then call the
correct functions, namely oldumount and umount if __NR_oldumount is defined
and umount and umount2 if it is not defined.
Cheers
Michael.
On Tue, Jul 16, 2013 at 10:04 AM, Richard Henderson <[email protected]> wrote:
> Here's a set of minor updates for arch/alpha that should not
> be controversial.
Patches 1-5 and 7 (with the typo fixed) are
Reviewed-and-Tested-by: Matt Turner <[email protected]>
I don't know enough patch 6 to make any kind of meaningful review, so it's
Acked-by: Matt Turner <[email protected]>
I've got five patches in a for-linus branch on kernel.org that I was
planning to send a pull request for soon. Would you like me to add
these to that branch?
Thanks!
Matt
On 07/16/2013 09:29 PM, Matt Turner wrote:
> I've got five patches in a for-linus branch on kernel.org that I was
> planning to send a pull request for soon. Would you like me to add
> these to that branch?
Yes, please.
r~
On Tue, Jul 16, 2013 at 10:04 AM, Richard Henderson <[email protected]> wrote:
> Compiling with GCC 4.8 yields several instances of
>
> crypto/vmac.c: In function ?vmac_final?:
> crypto/vmac.c:616:9: warning: value computed is not used [-Wunused-value]
> memset(&mac, 0, sizeof(vmac_t));
> ^
> arch/alpha/include/asm/string.h:31:25: note: in definition of macro ?memset?
> ? __builtin_memset((s),0,(n)) \
> ^
> Converting the macro to an inline function eliminates this problem.
>
> Signed-off-by: Richard Henderson <[email protected]>
> ---
On IRC we found out that this breaks the breaks kernel build with
gcc-3.4 with "sorry, unimplemented: inlining failed in call to
'memset': recursive inlining" message on some file. I'll drop this
patch for now.
On 07/16/2013 07:03:47 PM, Michael Cree wrote:
> On Tue, Jul 16, 2013 at 06:35:07PM -0500, Rob Landley wrote:
> > On 07/16/2013 12:04:33 PM, Richard Henderson wrote:
> > >Here's a set of minor updates for arch/alpha that should not
> > >be controversial.
> >
> > I also note that I had to do this to get busybox to build against
> > uClibc:
> > -#define __NR_umount 22
> > +#define __NR_umount2 22
> > -#define __NR_oldumount 321
> > +#define __NR_umount 321
>
> I anticipate that this will likely break userspace.
Haven't seen it so far. It's the same semantics all the other targets
have. Haven't built the whole of linux from scratch against it yet
though. (Most of my package builds are native and I'm still tweaking my
build environment to get a native toolchain built.)
> busybox should be fixed to test for __NR_oldumount and then call the
> correct functions, namely oldumount and umount if __NR_oldumount is
> defined
> and umount and umount2 if it is not defined.
The man page for umount and umount2 doesn't mention "oldumount":
$ man 2 umount | grep old
$
Only three kernel targets mention it, alpha, powerpc, and x86:
$ find arch/*/include -type f | xargs grep oldumount
arch/alpha/include/uapi/asm/unistd.h:#define __NR_oldumount
321
arch/powerpc/include/asm/systbl.h:SYSX(sys_ni_syscall,sys_oldumount,sys_oldumount)
arch/x86/include/generated/asm/syscalls_32.h:__SYSCALL_I386(22,
sys_oldumount, sys_oldumount)
And the other two export umount2:
$ find arch/*/include -type f | xargs grep umount2 | egrep "powerpc|x86"
arch/powerpc/include/uapi/asm/unistd.h:#define
__NR_umount2 52
arch/x86/include/generated/asm/unistd_32_ia32.h:#define
__NR_ia32_umount2 52
arch/x86/include/generated/uapi/asm/unistd_x32.h:#define __NR_umount2
(__X32_SYSCALL_BIT + 166)
arch/x86/include/generated/uapi/asm/unistd_32.h:#define __NR_umount2 52
arch/x86/include/generated/uapi/asm/unistd_64.h:#define __NR_umount2 166
So really what you're saying is busybox should #ifdef alpha/workaround
target-specific breakage/#unifdef, and what this says to _me_ is that
nobody's tried to build busybox on Alpha since it started using
umount2(). How long ago was that?
commit 6c5f2c602174c7fe0830a1fc4fe5b3dde5ed7068
Author: Erik Andersen <[email protected]>
Date: Fri May 5 19:49:33 2000 +0000
I.E. nobody's tried to build busybox umount for Alpha (except me) for
thirteen years.
There are waaaaay more busybox installations out there than even
_emulated_ alpha systems, and this is trivial to fix with a local patch
to the kernel. So I'll just do that. Your idea of the "correct" thing
to do to "fix" this seems entirely backwards to me.
Rob-
On Mon, Jul 22, 2013 at 09:31:41PM -0500, Rob Landley wrote:
> On 07/16/2013 07:03:47 PM, Michael Cree wrote:
> >On Tue, Jul 16, 2013 at 06:35:07PM -0500, Rob Landley wrote:
> >> On 07/16/2013 12:04:33 PM, Richard Henderson wrote:
> >> >Here's a set of minor updates for arch/alpha that should not
> >> >be controversial.
> >>
> >> I also note that I had to do this to get busybox to build against
> >> uClibc:
> >> -#define __NR_umount 22
> >> +#define __NR_umount2 22
> >> -#define __NR_oldumount 321
> >> +#define __NR_umount 321
> >
> >I anticipate that this will likely break userspace.
>
> Haven't seen it so far. It's the same semantics all the other
> targets have. Haven't built the whole of linux from scratch against
> it yet though. (Most of my package builds are native and I'm still
> tweaking my build environment to get a native toolchain built.)
[snip]
> I.E. nobody's tried to build busybox umount for Alpha (except me)
> for thirteen years.
That appears false to me. Busybox builds OK on Debian-Ports [1], and it looks
like in the build log that umount is built. A search in Debian busybox
sources does not return any hits for __NR_umount or __NR_umount2, but does
for umount2 which are library calls (admittedly ones that are just wrappers
for the kernel syscall). I also don't see any Debian patches in busybox that
change behaviour of busybox specifically for Alpha. But IIRC Debian links
against glibc even for the installer where busybox is used.
So maybe the problem is in uClibc? Yes, a quick look at the source shows
that uClibc does not test __NR_oldumount so presumably does not compile
correct umount and umount2 library calls on Alpha.
Maybe it's the case noone has compiled uClibc on Alpha until now?
> There are waaaaay more busybox installations out there than even
> _emulated_ alpha systems, and this is trivial to fix with a local
> patch to the kernel. So I'll just do that. Your idea of the
> "correct" thing to do to "fix" this seems entirely backwards to me.
I wondered if your proposal will break glibc as glibc checks for
__NR_oldumount and does different things based on that. But maybe your
fix will not adversely affect glibc (I did not look particularly closely
to see if so), but even so, there is no guarantee that other software does
not directly access the oldumount syscall when compiled on Alpha, and your
change would likely break any such software.
Cheers
Michael
[1] http://buildd.debian-ports.org/status/package.php?p=busybox
On 07/22/2013 07:25 PM, Michael Cree wrote:
> I wondered if your proposal will break glibc as glibc checks for
> __NR_oldumount and does different things based on that. But maybe your
> fix will not adversely affect glibc (I did not look particularly closely
> to see if so), but even so, there is no guarantee that other software does
> not directly access the oldumount syscall when compiled on Alpha, and your
> change would likely break any such software.
It won't break glibc. While there are conditionals for oldumount,
they do pretty much exactly the umount/umount2 dance you'd expect.
I'm for the patch, because anything that makes us match x86 more
closely has got to be a good thing from a portability standpoint.
r~
On Tue, Jul 23, 2013 at 07:20:22AM -0800, Richard Henderson wrote:
> On 07/22/2013 07:25 PM, Michael Cree wrote:
> > I wondered if your proposal will break glibc as glibc checks for
> > __NR_oldumount and does different things based on that. But maybe your
> > fix will not adversely affect glibc (I did not look particularly closely
> > to see if so), but even so, there is no guarantee that other software does
> > not directly access the oldumount syscall when compiled on Alpha, and your
> > change would likely break any such software.
>
> It won't break glibc. While there are conditionals for oldumount,
> they do pretty much exactly the umount/umount2 dance you'd expect.
That's good to hear.
> I'm for the patch, because anything that makes us match x86 more
> closely has got to be a good thing from a portability standpoint.
OK, but I think that behoves us to make an effort to check user space
applications are compliant with the change. Fortunately we can easily
scheck the vast majority of opensource programs now that Debian
provides a web interface for searching the source code of all their
packages.
Searching for NR_umount finds the following packages that compile on
Alpha:
eglibc
dietlibc
uclibc
qemu
radare
linux-tools
ns3
skyeye
(and a few others that are not buildable on Alpha and tons of
hits for every version of the linux kernel that result from every arch
directory therein...)
I'm happy to take a closer look at them and submit a patch to their upstreams
if need be.
Cheers
Michael.
On 07/24/2013 12:25 AM, Michael Cree wrote:
> Searching for NR_umount finds the following packages that compile on
> Alpha:
>
> eglibc
Fork of glibc, which I did examine; I'll assume it's the same code.
> dietlibc
> uclibc
Rob reports that uclibc is fixed by the rename. From what I see in
the dietlibc source, the same is true of Alpha, since the only mention
of "oldumount" is in alpha/syscalls.h.
> qemu
Definitely would be fixed by the rename.
Of course in both dietlibc and qemu, the packages need to be patched
to match the kernel rename, as they don't use the kernel headers.
> radare
Repository has vanished? There's a radare2 that apparently superceeds
it, but doesn't mention umount in any relevant context.
> linux-tools
> ns3
> skyeye
I can't find source for these to browse.
r~