2015-02-03 19:39:03 +07:00
|
|
|
/*
|
|
|
|
|
* Based on arch/arm/include/asm/atomic.h
|
|
|
|
|
*
|
|
|
|
|
* Copyright (C) 1996 Russell King.
|
|
|
|
|
* Copyright (C) 2002 Deep Blue Solutions Ltd.
|
|
|
|
|
* Copyright (C) 2012 ARM Ltd.
|
|
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License version 2 as
|
|
|
|
|
* published by the Free Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#ifndef __ASM_ATOMIC_LSE_H
|
|
|
|
|
#define __ASM_ATOMIC_LSE_H
|
|
|
|
|
|
|
|
|
|
#ifndef __ARM64_IN_ATOMIC_IMPL
|
|
|
|
|
#error "please don't include this file directly"
|
|
|
|
|
#endif
|
|
|
|
|
|
2015-02-03 23:14:13 +07:00
|
|
|
#define __LL_SC_ATOMIC(op) __LL_SC_CALL(atomic_##op)
|
|
|
|
|
|
|
|
|
|
static inline void atomic_andnot(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC(andnot),
|
|
|
|
|
" stclr %w[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic_or(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC(or),
|
|
|
|
|
" stset %w[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic_xor(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC(xor),
|
|
|
|
|
" steor %w[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic_add(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC(add),
|
|
|
|
|
" stadd %w[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#define ATOMIC_OP_ADD_RETURN(name, mb, cl...) \
|
|
|
|
|
static inline int atomic_add_return##name(int i, atomic_t *v) \
|
|
|
|
|
{ \
|
|
|
|
|
register int w0 asm ("w0") = i; \
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_ATOMIC(add_return##name), \
|
|
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" ldadd" #mb " %w[i], w30, %[v]\n" \
|
|
|
|
|
" add %w[i], %w[i], w30") \
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter) \
|
|
|
|
|
: "r" (x1) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, ##cl); \
|
2015-10-09 02:15:18 +07:00
|
|
|
\
|
|
|
|
|
return w0; \
|
|
|
|
|
}
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
ATOMIC_OP_ADD_RETURN(_relaxed, )
|
|
|
|
|
ATOMIC_OP_ADD_RETURN(_acquire, a, "memory")
|
|
|
|
|
ATOMIC_OP_ADD_RETURN(_release, l, "memory")
|
|
|
|
|
ATOMIC_OP_ADD_RETURN( , al, "memory")
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#undef ATOMIC_OP_ADD_RETURN
|
2015-02-03 23:14:13 +07:00
|
|
|
|
|
|
|
|
static inline void atomic_and(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(
|
|
|
|
|
/* LL/SC */
|
|
|
|
|
" nop\n"
|
|
|
|
|
__LL_SC_ATOMIC(and),
|
|
|
|
|
/* LSE atomics */
|
|
|
|
|
" mvn %w[i], %w[i]\n"
|
|
|
|
|
" stclr %w[i], %[v]")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic_sub(int i, atomic_t *v)
|
|
|
|
|
{
|
|
|
|
|
register int w0 asm ("w0") = i;
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(
|
|
|
|
|
/* LL/SC */
|
|
|
|
|
" nop\n"
|
|
|
|
|
__LL_SC_ATOMIC(sub),
|
|
|
|
|
/* LSE atomics */
|
|
|
|
|
" neg %w[i], %w[i]\n"
|
|
|
|
|
" stadd %w[i], %[v]")
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#define ATOMIC_OP_SUB_RETURN(name, mb, cl...) \
|
|
|
|
|
static inline int atomic_sub_return##name(int i, atomic_t *v) \
|
|
|
|
|
{ \
|
|
|
|
|
register int w0 asm ("w0") = i; \
|
|
|
|
|
register atomic_t *x1 asm ("x1") = v; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_ATOMIC(sub_return##name) \
|
|
|
|
|
" nop", \
|
|
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" neg %w[i], %w[i]\n" \
|
|
|
|
|
" ldadd" #mb " %w[i], w30, %[v]\n" \
|
|
|
|
|
" add %w[i], %w[i], w30") \
|
|
|
|
|
: [i] "+r" (w0), [v] "+Q" (v->counter) \
|
|
|
|
|
: "r" (x1) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS , ##cl); \
|
2015-10-09 02:15:18 +07:00
|
|
|
\
|
|
|
|
|
return w0; \
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
2015-02-03 19:39:03 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
ATOMIC_OP_SUB_RETURN(_relaxed, )
|
|
|
|
|
ATOMIC_OP_SUB_RETURN(_acquire, a, "memory")
|
|
|
|
|
ATOMIC_OP_SUB_RETURN(_release, l, "memory")
|
|
|
|
|
ATOMIC_OP_SUB_RETURN( , al, "memory")
|
|
|
|
|
|
|
|
|
|
#undef ATOMIC_OP_SUB_RETURN
|
2015-02-03 23:14:13 +07:00
|
|
|
#undef __LL_SC_ATOMIC
|
2015-02-03 19:39:03 +07:00
|
|
|
|
2015-02-03 23:14:13 +07:00
|
|
|
#define __LL_SC_ATOMIC64(op) __LL_SC_CALL(atomic64_##op)
|
|
|
|
|
|
|
|
|
|
static inline void atomic64_andnot(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC64(andnot),
|
|
|
|
|
" stclr %[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic64_or(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC64(or),
|
|
|
|
|
" stset %[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic64_xor(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC64(xor),
|
|
|
|
|
" steor %[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic64_add(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(__LL_SC_ATOMIC64(add),
|
|
|
|
|
" stadd %[i], %[v]\n")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#define ATOMIC64_OP_ADD_RETURN(name, mb, cl...) \
|
|
|
|
|
static inline long atomic64_add_return##name(long i, atomic64_t *v) \
|
|
|
|
|
{ \
|
|
|
|
|
register long x0 asm ("x0") = i; \
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_ATOMIC64(add_return##name), \
|
|
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" ldadd" #mb " %[i], x30, %[v]\n" \
|
|
|
|
|
" add %[i], %[i], x30") \
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter) \
|
|
|
|
|
: "r" (x1) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, ##cl); \
|
2015-10-09 02:15:18 +07:00
|
|
|
\
|
|
|
|
|
return x0; \
|
|
|
|
|
}
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
ATOMIC64_OP_ADD_RETURN(_relaxed, )
|
|
|
|
|
ATOMIC64_OP_ADD_RETURN(_acquire, a, "memory")
|
|
|
|
|
ATOMIC64_OP_ADD_RETURN(_release, l, "memory")
|
|
|
|
|
ATOMIC64_OP_ADD_RETURN( , al, "memory")
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#undef ATOMIC64_OP_ADD_RETURN
|
2015-02-03 23:14:13 +07:00
|
|
|
|
|
|
|
|
static inline void atomic64_and(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(
|
|
|
|
|
/* LL/SC */
|
|
|
|
|
" nop\n"
|
|
|
|
|
__LL_SC_ATOMIC64(and),
|
|
|
|
|
/* LSE atomics */
|
|
|
|
|
" mvn %[i], %[i]\n"
|
|
|
|
|
" stclr %[i], %[v]")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void atomic64_sub(long i, atomic64_t *v)
|
|
|
|
|
{
|
|
|
|
|
register long x0 asm ("x0") = i;
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v;
|
|
|
|
|
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(
|
|
|
|
|
/* LL/SC */
|
|
|
|
|
" nop\n"
|
|
|
|
|
__LL_SC_ATOMIC64(sub),
|
|
|
|
|
/* LSE atomics */
|
|
|
|
|
" neg %[i], %[i]\n"
|
|
|
|
|
" stadd %[i], %[v]")
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter)
|
|
|
|
|
: "r" (x1)
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS);
|
2015-02-03 23:14:13 +07:00
|
|
|
}
|
|
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#define ATOMIC64_OP_SUB_RETURN(name, mb, cl...) \
|
|
|
|
|
static inline long atomic64_sub_return##name(long i, atomic64_t *v) \
|
|
|
|
|
{ \
|
|
|
|
|
register long x0 asm ("x0") = i; \
|
|
|
|
|
register atomic64_t *x1 asm ("x1") = v; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_ATOMIC64(sub_return##name) \
|
|
|
|
|
" nop", \
|
|
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" neg %[i], %[i]\n" \
|
|
|
|
|
" ldadd" #mb " %[i], x30, %[v]\n" \
|
|
|
|
|
" add %[i], %[i], x30") \
|
|
|
|
|
: [i] "+r" (x0), [v] "+Q" (v->counter) \
|
|
|
|
|
: "r" (x1) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, ##cl); \
|
2015-10-09 02:15:18 +07:00
|
|
|
\
|
|
|
|
|
return x0; \
|
|
|
|
|
}
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
ATOMIC64_OP_SUB_RETURN(_relaxed, )
|
|
|
|
|
ATOMIC64_OP_SUB_RETURN(_acquire, a, "memory")
|
|
|
|
|
ATOMIC64_OP_SUB_RETURN(_release, l, "memory")
|
|
|
|
|
ATOMIC64_OP_SUB_RETURN( , al, "memory")
|
2015-02-03 23:14:13 +07:00
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
#undef ATOMIC64_OP_SUB_RETURN
|
2015-02-03 19:39:03 +07:00
|
|
|
|
|
|
|
|
static inline long atomic64_dec_if_positive(atomic64_t *v)
|
|
|
|
|
{
|
2015-02-03 23:14:13 +07:00
|
|
|
register long x0 asm ("x0") = (long)v;
|
2015-02-03 19:39:03 +07:00
|
|
|
|
2015-02-03 23:14:13 +07:00
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN(
|
|
|
|
|
/* LL/SC */
|
|
|
|
|
" nop\n"
|
|
|
|
|
__LL_SC_ATOMIC64(dec_if_positive)
|
|
|
|
|
" nop\n"
|
|
|
|
|
" nop\n"
|
|
|
|
|
" nop\n"
|
|
|
|
|
" nop\n"
|
|
|
|
|
" nop",
|
|
|
|
|
/* LSE atomics */
|
|
|
|
|
"1: ldr x30, %[v]\n"
|
|
|
|
|
" subs %[ret], x30, #1\n"
|
2015-05-29 20:44:06 +07:00
|
|
|
" b.lt 2f\n"
|
2015-02-03 23:14:13 +07:00
|
|
|
" casal x30, %[ret], %[v]\n"
|
|
|
|
|
" sub x30, x30, #1\n"
|
|
|
|
|
" sub x30, x30, %[ret]\n"
|
|
|
|
|
" cbnz x30, 1b\n"
|
|
|
|
|
"2:")
|
|
|
|
|
: [ret] "+&r" (x0), [v] "+Q" (v->counter)
|
2015-02-03 19:39:03 +07:00
|
|
|
:
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, "cc", "memory");
|
2015-02-03 19:39:03 +07:00
|
|
|
|
|
|
|
|
return x0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-02-03 23:14:13 +07:00
|
|
|
#undef __LL_SC_ATOMIC64
|
|
|
|
|
|
2015-04-24 02:08:49 +07:00
|
|
|
#define __LL_SC_CMPXCHG(op) __LL_SC_CALL(__cmpxchg_case_##op)
|
|
|
|
|
|
|
|
|
|
#define __CMPXCHG_CASE(w, sz, name, mb, cl...) \
|
|
|
|
|
static inline unsigned long __cmpxchg_case_##name(volatile void *ptr, \
|
|
|
|
|
unsigned long old, \
|
|
|
|
|
unsigned long new) \
|
|
|
|
|
{ \
|
|
|
|
|
register unsigned long x0 asm ("x0") = (unsigned long)ptr; \
|
|
|
|
|
register unsigned long x1 asm ("x1") = old; \
|
|
|
|
|
register unsigned long x2 asm ("x2") = new; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
2015-07-29 21:16:22 +07:00
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_CMPXCHG(name) \
|
|
|
|
|
" nop", \
|
2015-04-24 02:08:49 +07:00
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" mov " #w "30, %" #w "[old]\n" \
|
|
|
|
|
" cas" #mb #sz "\t" #w "30, %" #w "[new], %[v]\n" \
|
|
|
|
|
" mov %" #w "[ret], " #w "30") \
|
|
|
|
|
: [ret] "+r" (x0), [v] "+Q" (*(unsigned long *)ptr) \
|
|
|
|
|
: [old] "r" (x1), [new] "r" (x2) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, ##cl); \
|
2015-04-24 02:08:49 +07:00
|
|
|
\
|
|
|
|
|
return x0; \
|
|
|
|
|
}
|
|
|
|
|
|
2015-10-09 02:15:18 +07:00
|
|
|
__CMPXCHG_CASE(w, b, 1, )
|
|
|
|
|
__CMPXCHG_CASE(w, h, 2, )
|
|
|
|
|
__CMPXCHG_CASE(w, , 4, )
|
|
|
|
|
__CMPXCHG_CASE(x, , 8, )
|
|
|
|
|
__CMPXCHG_CASE(w, b, acq_1, a, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, h, acq_2, a, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, , acq_4, a, "memory")
|
|
|
|
|
__CMPXCHG_CASE(x, , acq_8, a, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, b, rel_1, l, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, h, rel_2, l, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, , rel_4, l, "memory")
|
|
|
|
|
__CMPXCHG_CASE(x, , rel_8, l, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, b, mb_1, al, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, h, mb_2, al, "memory")
|
|
|
|
|
__CMPXCHG_CASE(w, , mb_4, al, "memory")
|
|
|
|
|
__CMPXCHG_CASE(x, , mb_8, al, "memory")
|
2015-04-24 02:08:49 +07:00
|
|
|
|
|
|
|
|
#undef __LL_SC_CMPXCHG
|
|
|
|
|
#undef __CMPXCHG_CASE
|
|
|
|
|
|
2015-05-15 00:05:50 +07:00
|
|
|
#define __LL_SC_CMPXCHG_DBL(op) __LL_SC_CALL(__cmpxchg_double##op)
|
|
|
|
|
|
|
|
|
|
#define __CMPXCHG_DBL(name, mb, cl...) \
|
2015-11-05 21:00:56 +07:00
|
|
|
static inline long __cmpxchg_double##name(unsigned long old1, \
|
2015-05-15 00:05:50 +07:00
|
|
|
unsigned long old2, \
|
|
|
|
|
unsigned long new1, \
|
|
|
|
|
unsigned long new2, \
|
|
|
|
|
volatile void *ptr) \
|
|
|
|
|
{ \
|
|
|
|
|
unsigned long oldval1 = old1; \
|
|
|
|
|
unsigned long oldval2 = old2; \
|
|
|
|
|
register unsigned long x0 asm ("x0") = old1; \
|
|
|
|
|
register unsigned long x1 asm ("x1") = old2; \
|
|
|
|
|
register unsigned long x2 asm ("x2") = new1; \
|
|
|
|
|
register unsigned long x3 asm ("x3") = new2; \
|
|
|
|
|
register unsigned long x4 asm ("x4") = (unsigned long)ptr; \
|
|
|
|
|
\
|
|
|
|
|
asm volatile(ARM64_LSE_ATOMIC_INSN( \
|
|
|
|
|
/* LL/SC */ \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
" nop\n" \
|
|
|
|
|
__LL_SC_CMPXCHG_DBL(name), \
|
|
|
|
|
/* LSE atomics */ \
|
|
|
|
|
" casp" #mb "\t%[old1], %[old2], %[new1], %[new2], %[v]\n"\
|
|
|
|
|
" eor %[old1], %[old1], %[oldval1]\n" \
|
|
|
|
|
" eor %[old2], %[old2], %[oldval2]\n" \
|
|
|
|
|
" orr %[old1], %[old1], %[old2]") \
|
|
|
|
|
: [old1] "+r" (x0), [old2] "+r" (x1), \
|
|
|
|
|
[v] "+Q" (*(unsigned long *)ptr) \
|
|
|
|
|
: [new1] "r" (x2), [new2] "r" (x3), [ptr] "r" (x4), \
|
|
|
|
|
[oldval1] "r" (oldval1), [oldval2] "r" (oldval2) \
|
arm64: lse: deal with clobbered IP registers after branch via PLT
The LSE atomics implementation uses runtime patching to patch in calls
to out of line non-LSE atomics implementations on cores that lack hardware
support for LSE. To avoid paying the overhead cost of a function call even
if no call ends up being made, the bl instruction is kept invisible to the
compiler, and the out of line implementations preserve all registers, not
just the ones that they are required to preserve as per the AAPCS64.
However, commit fd045f6cd98e ("arm64: add support for module PLTs") added
support for routing branch instructions via veneers if the branch target
offset exceeds the range of the ordinary relative branch instructions.
Since this deals with jump and call instructions that are exposed to ELF
relocations, the PLT code uses x16 to hold the address of the branch target
when it performs an indirect branch-to-register, something which is
explicitly allowed by the AAPCS64 (and ordinary compiler generated code
does not expect register x16 or x17 to retain their values across a bl
instruction).
Since the lse runtime patched bl instructions don't adhere to the AAPCS64,
they don't deal with this clobbering of registers x16 and x17. So add them
to the clobber list of the asm() statements that perform the call
instructions, and drop x16 and x17 from the list of registers that are
callee saved in the out of line non-LSE implementations.
In addition, since we have given these functions two scratch registers,
they no longer need to stack/unstack temp registers.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[will: factored clobber list into #define, updated Makefile comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-02-26 02:48:53 +07:00
|
|
|
: __LL_SC_CLOBBERS, ##cl); \
|
2015-05-15 00:05:50 +07:00
|
|
|
\
|
|
|
|
|
return x0; \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
__CMPXCHG_DBL( , )
|
|
|
|
|
__CMPXCHG_DBL(_mb, al, "memory")
|
|
|
|
|
|
|
|
|
|
#undef __LL_SC_CMPXCHG_DBL
|
|
|
|
|
#undef __CMPXCHG_DBL
|
|
|
|
|
|
2015-02-03 19:39:03 +07:00
|
|
|
#endif /* __ASM_ATOMIC_LSE_H */
|