freebsd-nq/sys/sun4v/include/cpufunc.h
Kip Macy 7c0435b933 MUTEX_PROFILING has been generalized to LOCK_PROFILING. We now profile
wait (time waited to acquire) and hold times for *all* kernel locks. If
the architecture has a system synchronized TSC, the profiling code will
use that - thereby minimizing profiling overhead. Large chunks of profiling
code have been moved out of line, the overhead measured on the T1 for when
it is compiled in but not enabled is < 1%.

Approved by: scottl (standing in for mentor rwatson)
Reviewed by: des and jhb
2006-11-11 03:18:07 +00:00

259 lines
7.5 KiB
C

/*-
* Copyright (c) 2001 Jake Burkholder.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef _MACHINE_CPUFUNC_H_
#define _MACHINE_CPUFUNC_H_
#include <machine/asi.h>
#include <machine/pstate.h>
struct thread;
/*
* membar operand macros for use in other macros when # is a special
* character. Keep these in sync with what the hardware expects.
*/
#define C_Lookaside (0)
#define C_MemIssue (1)
#define C_Sync (2)
#define M_LoadLoad (0)
#define M_StoreLoad (1)
#define M_LoadStore (2)
#define M_StoreStore (3)
#define CMASK_SHIFT (4)
#define MMASK_SHIFT (0)
#define CMASK_GEN(bit) ((1 << (bit)) << CMASK_SHIFT)
#define MMASK_GEN(bit) ((1 << (bit)) << MMASK_SHIFT)
#define Lookaside CMASK_GEN(C_Lookaside)
#define MemIssue CMASK_GEN(C_MemIssue)
#define Sync CMASK_GEN(C_Sync)
#define LoadLoad MMASK_GEN(M_LoadLoad)
#define StoreLoad MMASK_GEN(M_StoreLoad)
#define LoadStore MMASK_GEN(M_LoadStore)
#define StoreStore MMASK_GEN(M_StoreStore)
#define casa(rs1, rs2, rd, asi) ({ \
u_int __rd = (uint32_t)(rd); \
__asm __volatile("casa [%2] %3, %4, %0" \
: "+r" (__rd), "=m" (*rs1) \
: "r" (rs1), "n" (asi), "r" (rs2), "m" (*rs1)); \
__rd; \
})
#define casxa(rs1, rs2, rd, asi) ({ \
u_long __rd = (uint64_t)(rd); \
__asm __volatile("casxa [%2] %3, %4, %0" \
: "+r" (__rd), "=m" (*rs1) \
: "r" (rs1), "n" (asi), "r" (rs2), "m" (*rs1)); \
__rd; \
})
#define flush(va) do { \
__asm __volatile("flush %0" : : "r" (va)); \
} while (0)
#define flushw() do { \
__asm __volatile("flushw" : :); \
} while (0)
#define mov(val, reg) do { \
__asm __volatile("mov %0, %" __XSTRING(reg) : : "r" (val)); \
} while (0)
/* Generate ld*a/st*a functions for non-constant ASI's. */
#define LDNC_GEN(tp, o) \
static __inline tp \
o ## _nc(caddr_t va, int asi) \
{ \
tp r; \
__asm __volatile("wr %2, 0, %%asi;" #o " [%1] %%asi, %0"\
: "=r" (r) : "r" (va), "r" (asi)); \
return (r); \
}
LDNC_GEN(u_char, lduba);
LDNC_GEN(u_short, lduha);
LDNC_GEN(u_int, lduwa);
LDNC_GEN(u_long, ldxa);
#define LD_GENERIC(va, asi, op, type) ({ \
type __r; \
__asm __volatile(#op " [%1] %2, %0" \
: "=r" (__r) : "r" (va), "n" (asi)); \
__r; \
})
#define lduba(va, asi) LD_GENERIC(va, asi, lduba, u_char)
#define lduha(va, asi) LD_GENERIC(va, asi, lduha, u_short)
#define lduwa(va, asi) LD_GENERIC(va, asi, lduwa, u_int)
#define ldxa(va, asi) LD_GENERIC(va, asi, ldxa, u_long)
#if 0
#define STNC_GEN(tp, o) \
static __inline void \
o ## _nc(caddr_t va, int asi, tp val) \
{ \
__asm __volatile(#o " %0, [%g0 + %1] %2"\
: : "r" (val), "r" (va), "r" (asi)); \
}
#else
#define STNC_GEN(tp, o) \
static __inline void \
o ## _nc(caddr_t va, int asi, tp val) \
{ \
__asm __volatile("wr %2, 0, %%asi;" #o " %0, [%1] %%asi"\
: : "r" (val), "r" (va), "r" (asi)); \
}
#endif
STNC_GEN(u_char, stba);
STNC_GEN(u_short, stha);
STNC_GEN(u_int, stwa);
STNC_GEN(u_long, stxa);
#define ST_GENERIC(va, asi, val, op) \
__asm __volatile(#op " %0, [%1] %2" \
: : "r" (val), "r" (va), "n" (asi)); \
#define stba(va, asi, val) ST_GENERIC(va, asi, val, stba)
#define stha(va, asi, val) ST_GENERIC(va, asi, val, stha)
#define stwa(va, asi, val) ST_GENERIC(va, asi, val, stwa)
#define stxa(va, asi, val) ST_GENERIC(va, asi, val, stxa)
/*
* Attempt to read from addr, val. If a Data Access Error trap happens,
* they return -1 and the contents of val is undefined. A return of 0
* means no trap happened, and the contents of val is valid.
*/
int fasword8(u_long asi, void *addr, uint8_t *val);
int fasword16(u_long asi, void *addr, uint16_t *val);
int fasword32(u_long asi, void *addr, uint32_t *val);
#define membar(mask) do { \
__asm __volatile("membar %0" : : "n" (mask) : "memory"); \
} while (0)
#define rd(name) ({ \
uint64_t __sr; \
__asm __volatile("rd %%" #name ", %0" : "=r" (__sr) :); \
__sr; \
})
#define wr(name, val, xor) do { \
__asm __volatile("wr %0, %1, %%" #name \
: : "r" (val), "rI" (xor)); \
} while (0)
#define rdpr(name) ({ \
uint64_t __pr; \
__asm __volatile("rdpr %%" #name", %0" : "=r" (__pr) :); \
__pr; \
})
#define wrpr(name, val, xor) do { \
__asm __volatile("wrpr %0, %1, %%" #name \
: : "r" (val), "rI" (xor)); \
} while (0)
/*
* Macro intended to be used instead of wr(asr23, val, xor) for writing to
* the TICK_CMPR register in order to avoid a bug in BlackBird CPUs that
* can cause these writes to fail under certain condidtions which in turn
* causes the hardclock to stop. The workaround is to perform the write
* at the beginning of an I-Cache line directly followed by a dummy read.
*/
#define wrtickcmpr(val, xor) ({ \
__asm __volatile( \
" ba,pt %%xcc, 1f ; " \
" nop ; " \
" .align 64 ; " \
"1: wr %0, %1, %%asr23 ; " \
" rd %%asr23, %%g0 ; " \
: : "r" (val), "rI" (xor)); \
})
static __inline void
breakpoint(void)
{
__asm __volatile("ta %%xcc, 1" : :);
}
static __inline register_t
intr_disable_all(void)
{
u_long s;
s = rdpr(pstate);
wrpr(pstate, s & ~PSTATE_IE, 0);
return (s);
}
#define intr_restore_all(s) wrpr(pstate, (s), 0)
static __inline register_t
intr_disable(void)
{
u_long s;
s = rdpr(pil);
wrpr(pil, 14, 0);
return (s);
}
#define intr_restore(s) wrpr(pil, (s), 0)
/*
* In some places, it is required that the store is directly followed by a
* membar #Sync. Don't trust the compiler to not insert instructions in
* between. We also need to disable interrupts completely.
*/
#define stxa_sync(va, asi, val) do { \
u_long s; \
s = intr_disable_all(); \
__asm __volatile("stxa %0, [%1] %2; membar #Sync" \
: : "r" (val), "r" (va), "n" (asi)); \
intr_restore_all(s); \
} while (0)
void ascopy(u_long asi, vm_offset_t src, vm_offset_t dst, size_t len);
void ascopyfrom(u_long sasi, vm_offset_t src, caddr_t dst, size_t len);
void ascopyto(caddr_t src, u_long dasi, vm_offset_t dst, size_t len);
void aszero(u_long asi, vm_offset_t dst, size_t len);
#include <machine/sun4v_cpufunc.h>
#define USE_CPU_NANOSECONDS
#define nanoseconds() rd(tick)
#undef LDNC_GEN
#undef STNC_GEN
#endif /* !_MACHINE_CPUFUNC_H_ */