summaryrefslogtreecommitdiff
path: root/arch/arm64/lib
diff options
context:
space:
mode:
Diffstat (limited to 'arch/arm64/lib')
-rw-r--r--arch/arm64/lib/Makefile13
-rw-r--r--arch/arm64/lib/atomic_ll_sc.c3
-rw-r--r--arch/arm64/lib/bitops.S45
-rw-r--r--arch/arm64/lib/clear_user.S8
-rw-r--r--arch/arm64/lib/copy_from_user.S25
-rw-r--r--arch/arm64/lib/copy_in_user.S25
-rw-r--r--arch/arm64/lib/copy_to_user.S25
7 files changed, 107 insertions, 37 deletions
diff --git a/arch/arm64/lib/Makefile b/arch/arm64/lib/Makefile
index d98d3e39879e..1a811ecf71da 100644
--- a/arch/arm64/lib/Makefile
+++ b/arch/arm64/lib/Makefile
@@ -3,3 +3,16 @@ lib-y := bitops.o clear_user.o delay.o copy_from_user.o \
clear_page.o memchr.o memcpy.o memmove.o memset.o \
memcmp.o strcmp.o strncmp.o strlen.o strnlen.o \
strchr.o strrchr.o
+
+# Tell the compiler to treat all general purpose registers as
+# callee-saved, which allows for efficient runtime patching of the bl
+# instruction in the caller with an atomic instruction when supported by
+# the CPU. Result and argument registers are handled correctly, based on
+# the function prototype.
+lib-$(CONFIG_ARM64_LSE_ATOMICS) += atomic_ll_sc.o
+CFLAGS_atomic_ll_sc.o := -fcall-used-x0 -ffixed-x1 -ffixed-x2 \
+ -ffixed-x3 -ffixed-x4 -ffixed-x5 -ffixed-x6 \
+ -ffixed-x7 -fcall-saved-x8 -fcall-saved-x9 \
+ -fcall-saved-x10 -fcall-saved-x11 -fcall-saved-x12 \
+ -fcall-saved-x13 -fcall-saved-x14 -fcall-saved-x15 \
+ -fcall-saved-x16 -fcall-saved-x17 -fcall-saved-x18
diff --git a/arch/arm64/lib/atomic_ll_sc.c b/arch/arm64/lib/atomic_ll_sc.c
new file mode 100644
index 000000000000..b0c538b0da28
--- /dev/null
+++ b/arch/arm64/lib/atomic_ll_sc.c
@@ -0,0 +1,3 @@
+#include <asm/atomic.h>
+#define __ARM64_IN_ATOMIC_IMPL
+#include <asm/atomic_ll_sc.h>
diff --git a/arch/arm64/lib/bitops.S b/arch/arm64/lib/bitops.S
index 7dac371cc9a2..43ac736baa5b 100644
--- a/arch/arm64/lib/bitops.S
+++ b/arch/arm64/lib/bitops.S
@@ -18,52 +18,59 @@
#include <linux/linkage.h>
#include <asm/assembler.h>
+#include <asm/lse.h>
/*
* x0: bits 5:0 bit offset
* bits 31:6 word offset
* x1: address
*/
- .macro bitop, name, instr
+ .macro bitop, name, llsc, lse
ENTRY( \name )
and w3, w0, #63 // Get bit offset
eor w0, w0, w3 // Clear low bits
mov x2, #1
add x1, x1, x0, lsr #3 // Get word offset
+alt_lse " prfm pstl1strm, [x1]", "nop"
lsl x3, x2, x3 // Create mask
-1: ldxr x2, [x1]
- \instr x2, x2, x3
- stxr w0, x2, [x1]
- cbnz w0, 1b
+
+alt_lse "1: ldxr x2, [x1]", "\lse x3, [x1]"
+alt_lse " \llsc x2, x2, x3", "nop"
+alt_lse " stxr w0, x2, [x1]", "nop"
+alt_lse " cbnz w0, 1b", "nop"
+
ret
ENDPROC(\name )
.endm
- .macro testop, name, instr
+ .macro testop, name, llsc, lse
ENTRY( \name )
and w3, w0, #63 // Get bit offset
eor w0, w0, w3 // Clear low bits
mov x2, #1
add x1, x1, x0, lsr #3 // Get word offset
+alt_lse " prfm pstl1strm, [x1]", "nop"
lsl x4, x2, x3 // Create mask
-1: ldxr x2, [x1]
- lsr x0, x2, x3 // Save old value of bit
- \instr x2, x2, x4 // toggle bit
- stlxr w5, x2, [x1]
- cbnz w5, 1b
- dmb ish
+
+alt_lse "1: ldxr x2, [x1]", "\lse x4, x2, [x1]"
+ lsr x0, x2, x3
+alt_lse " \llsc x2, x2, x4", "nop"
+alt_lse " stlxr w5, x2, [x1]", "nop"
+alt_lse " cbnz w5, 1b", "nop"
+alt_lse " dmb ish", "nop"
+
and x0, x0, #1
-3: ret
+ ret
ENDPROC(\name )
.endm
/*
* Atomic bit operations.
*/
- bitop change_bit, eor
- bitop clear_bit, bic
- bitop set_bit, orr
+ bitop change_bit, eor, steor
+ bitop clear_bit, bic, stclr
+ bitop set_bit, orr, stset
- testop test_and_change_bit, eor
- testop test_and_clear_bit, bic
- testop test_and_set_bit, orr
+ testop test_and_change_bit, eor, ldeoral
+ testop test_and_clear_bit, bic, ldclral
+ testop test_and_set_bit, orr, ldsetal
diff --git a/arch/arm64/lib/clear_user.S b/arch/arm64/lib/clear_user.S
index c17967fdf5f6..a9723c71c52b 100644
--- a/arch/arm64/lib/clear_user.S
+++ b/arch/arm64/lib/clear_user.S
@@ -16,7 +16,11 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <linux/linkage.h>
+
+#include <asm/alternative.h>
#include <asm/assembler.h>
+#include <asm/cpufeature.h>
+#include <asm/sysreg.h>
.text
@@ -29,6 +33,8 @@
* Alignment fixed up by hardware.
*/
ENTRY(__clear_user)
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
mov x2, x1 // save the size for fixup return
subs x1, x1, #8
b.mi 2f
@@ -48,6 +54,8 @@ USER(9f, strh wzr, [x0], #2 )
b.mi 5f
USER(9f, strb wzr, [x0] )
5: mov x0, #0
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
ret
ENDPROC(__clear_user)
diff --git a/arch/arm64/lib/copy_from_user.S b/arch/arm64/lib/copy_from_user.S
index 5e27add9d362..1be9ef27be97 100644
--- a/arch/arm64/lib/copy_from_user.S
+++ b/arch/arm64/lib/copy_from_user.S
@@ -15,7 +15,11 @@
*/
#include <linux/linkage.h>
+
+#include <asm/alternative.h>
#include <asm/assembler.h>
+#include <asm/cpufeature.h>
+#include <asm/sysreg.h>
/*
* Copy from user space to a kernel buffer (alignment handled by the hardware)
@@ -28,14 +32,21 @@
* x0 - bytes not copied
*/
ENTRY(__copy_from_user)
- add x4, x1, x2 // upper user buffer boundary
- subs x2, x2, #8
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
+ add x5, x1, x2 // upper user buffer boundary
+ subs x2, x2, #16
+ b.mi 1f
+0:
+USER(9f, ldp x3, x4, [x1], #16)
+ subs x2, x2, #16
+ stp x3, x4, [x0], #16
+ b.pl 0b
+1: adds x2, x2, #8
b.mi 2f
-1:
USER(9f, ldr x3, [x1], #8 )
- subs x2, x2, #8
+ sub x2, x2, #8
str x3, [x0], #8
- b.pl 1b
2: adds x2, x2, #4
b.mi 3f
USER(9f, ldr w3, [x1], #4 )
@@ -51,12 +62,14 @@ USER(9f, ldrh w3, [x1], #2 )
USER(9f, ldrb w3, [x1] )
strb w3, [x0]
5: mov x0, #0
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
ret
ENDPROC(__copy_from_user)
.section .fixup,"ax"
.align 2
-9: sub x2, x4, x1
+9: sub x2, x5, x1
mov x3, x2
10: strb wzr, [x0], #1 // zero remaining buffer space
subs x3, x3, #1
diff --git a/arch/arm64/lib/copy_in_user.S b/arch/arm64/lib/copy_in_user.S
index 84b6c9bb9b93..1b94661e22b3 100644
--- a/arch/arm64/lib/copy_in_user.S
+++ b/arch/arm64/lib/copy_in_user.S
@@ -17,7 +17,11 @@
*/
#include <linux/linkage.h>
+
+#include <asm/alternative.h>
#include <asm/assembler.h>
+#include <asm/cpufeature.h>
+#include <asm/sysreg.h>
/*
* Copy from user space to user space (alignment handled by the hardware)
@@ -30,14 +34,21 @@
* x0 - bytes not copied
*/
ENTRY(__copy_in_user)
- add x4, x0, x2 // upper user buffer boundary
- subs x2, x2, #8
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
+ add x5, x0, x2 // upper user buffer boundary
+ subs x2, x2, #16
+ b.mi 1f
+0:
+USER(9f, ldp x3, x4, [x1], #16)
+ subs x2, x2, #16
+USER(9f, stp x3, x4, [x0], #16)
+ b.pl 0b
+1: adds x2, x2, #8
b.mi 2f
-1:
USER(9f, ldr x3, [x1], #8 )
- subs x2, x2, #8
+ sub x2, x2, #8
USER(9f, str x3, [x0], #8 )
- b.pl 1b
2: adds x2, x2, #4
b.mi 3f
USER(9f, ldr w3, [x1], #4 )
@@ -53,11 +64,13 @@ USER(9f, strh w3, [x0], #2 )
USER(9f, ldrb w3, [x1] )
USER(9f, strb w3, [x0] )
5: mov x0, #0
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
ret
ENDPROC(__copy_in_user)
.section .fixup,"ax"
.align 2
-9: sub x0, x4, x0 // bytes not copied
+9: sub x0, x5, x0 // bytes not copied
ret
.previous
diff --git a/arch/arm64/lib/copy_to_user.S b/arch/arm64/lib/copy_to_user.S
index a0aeeb9b7a28..a257b47e2dc4 100644
--- a/arch/arm64/lib/copy_to_user.S
+++ b/arch/arm64/lib/copy_to_user.S
@@ -15,7 +15,11 @@
*/
#include <linux/linkage.h>
+
+#include <asm/alternative.h>
#include <asm/assembler.h>
+#include <asm/cpufeature.h>
+#include <asm/sysreg.h>
/*
* Copy to user space from a kernel buffer (alignment handled by the hardware)
@@ -28,14 +32,21 @@
* x0 - bytes not copied
*/
ENTRY(__copy_to_user)
- add x4, x0, x2 // upper user buffer boundary
- subs x2, x2, #8
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
+ add x5, x0, x2 // upper user buffer boundary
+ subs x2, x2, #16
+ b.mi 1f
+0:
+ ldp x3, x4, [x1], #16
+ subs x2, x2, #16
+USER(9f, stp x3, x4, [x0], #16)
+ b.pl 0b
+1: adds x2, x2, #8
b.mi 2f
-1:
ldr x3, [x1], #8
- subs x2, x2, #8
+ sub x2, x2, #8
USER(9f, str x3, [x0], #8 )
- b.pl 1b
2: adds x2, x2, #4
b.mi 3f
ldr w3, [x1], #4
@@ -51,11 +62,13 @@ USER(9f, strh w3, [x0], #2 )
ldrb w3, [x1]
USER(9f, strb w3, [x0] )
5: mov x0, #0
+ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_HAS_PAN, \
+ CONFIG_ARM64_PAN)
ret
ENDPROC(__copy_to_user)
.section .fixup,"ax"
.align 2
-9: sub x0, x4, x0 // bytes not copied
+9: sub x0, x5, x0 // bytes not copied
ret
.previous
generated by cgit v1.2.3 (git 2.39.0) at 2024-12-23 00:31:38 +0300