2015-03-08 07:20:20 -06:00
|
|
|
// Copyright 2015 The Go Authors. All rights reserved.
|
|
|
|
|
// Use of this source code is governed by a BSD-style
|
|
|
|
|
// license that can be found in the LICENSE file.
|
|
|
|
|
|
|
|
|
|
#include "go_asm.h"
|
|
|
|
|
#include "go_tls.h"
|
2015-04-03 02:37:22 -06:00
|
|
|
#include "tls_arm64.h"
|
2015-03-08 07:20:20 -06:00
|
|
|
#include "funcdata.h"
|
|
|
|
|
#include "textflag.h"
|
|
|
|
|
|
|
|
|
|
TEXT runtime·rt0_go(SB),NOSPLIT,$0
|
|
|
|
|
// SP = stack; R0 = argc; R1 = argv
|
|
|
|
|
|
|
|
|
|
// initialize essential registers
|
|
|
|
|
BL runtime·reginit(SB)
|
|
|
|
|
|
|
|
|
|
SUB $32, RSP
|
|
|
|
|
MOVW R0, 8(RSP) // argc
|
|
|
|
|
MOVD R1, 16(RSP) // argv
|
|
|
|
|
|
|
|
|
|
// create istack out of the given (operating system) stack.
|
|
|
|
|
// _cgo_init may update stackguard.
|
|
|
|
|
MOVD $runtime·g0(SB), g
|
|
|
|
|
MOVD RSP, R7
|
|
|
|
|
MOVD $(-64*1024)(R7), R0
|
|
|
|
|
MOVD R0, g_stackguard0(g)
|
|
|
|
|
MOVD R0, g_stackguard1(g)
|
|
|
|
|
MOVD R0, (g_stack+stack_lo)(g)
|
|
|
|
|
MOVD R7, (g_stack+stack_hi)(g)
|
|
|
|
|
|
|
|
|
|
// if there is a _cgo_init, call it using the gcc ABI.
|
|
|
|
|
MOVD _cgo_init(SB), R12
|
|
|
|
|
CMP $0, R12
|
|
|
|
|
BEQ nocgo
|
|
|
|
|
|
2015-04-03 02:37:22 -06:00
|
|
|
MRS_TPIDR_R0 // load TLS base pointer
|
|
|
|
|
MOVD R0, R3 // arg 3: TLS base pointer
|
2015-04-10 20:14:43 -06:00
|
|
|
#ifdef TLSG_IS_VARIABLE
|
|
|
|
|
MOVD $runtime·tls_g(SB), R2 // arg 2: tlsg
|
|
|
|
|
#else
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD $0x10, R2 // arg 2: tlsg TODO(minux): hardcoded for linux
|
2015-04-10 20:14:43 -06:00
|
|
|
#endif
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD $setg_gcc<>(SB), R1 // arg 1: setg
|
|
|
|
|
MOVD g, R0 // arg 0: G
|
|
|
|
|
BL (R12)
|
|
|
|
|
MOVD _cgo_init(SB), R12
|
|
|
|
|
CMP $0, R12
|
|
|
|
|
BEQ nocgo
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
nocgo:
|
|
|
|
|
// update stackguard after _cgo_init
|
|
|
|
|
MOVD (g_stack+stack_lo)(g), R0
|
|
|
|
|
ADD $const__StackGuard, R0
|
|
|
|
|
MOVD R0, g_stackguard0(g)
|
|
|
|
|
MOVD R0, g_stackguard1(g)
|
|
|
|
|
|
|
|
|
|
// set the per-goroutine and per-mach "registers"
|
|
|
|
|
MOVD $runtime·m0(SB), R0
|
|
|
|
|
|
|
|
|
|
// save m->g0 = g0
|
|
|
|
|
MOVD g, m_g0(R0)
|
|
|
|
|
// save m0 to g0->m
|
|
|
|
|
MOVD R0, g_m(g)
|
|
|
|
|
|
|
|
|
|
BL runtime·check(SB)
|
|
|
|
|
|
|
|
|
|
MOVW 8(RSP), R0 // copy argc
|
|
|
|
|
MOVW R0, -8(RSP)
|
|
|
|
|
MOVD 16(RSP), R0 // copy argv
|
|
|
|
|
MOVD R0, 0(RSP)
|
|
|
|
|
BL runtime·args(SB)
|
|
|
|
|
BL runtime·osinit(SB)
|
|
|
|
|
BL runtime·schedinit(SB)
|
|
|
|
|
|
|
|
|
|
// create a new goroutine to start program
|
2015-03-29 17:38:20 -06:00
|
|
|
MOVD $runtime·mainPC(SB), R0 // entry
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD RSP, R7
|
|
|
|
|
MOVD.W $0, -8(R7)
|
|
|
|
|
MOVD.W R0, -8(R7)
|
|
|
|
|
MOVD.W $0, -8(R7)
|
|
|
|
|
MOVD.W $0, -8(R7)
|
|
|
|
|
MOVD R7, RSP
|
|
|
|
|
BL runtime·newproc(SB)
|
|
|
|
|
ADD $32, RSP
|
|
|
|
|
|
|
|
|
|
// start this M
|
|
|
|
|
BL runtime·mstart(SB)
|
|
|
|
|
|
|
|
|
|
MOVD $0, R0
|
|
|
|
|
MOVD R0, (R0) // boom
|
|
|
|
|
UNDEF
|
|
|
|
|
|
2015-03-29 17:38:20 -06:00
|
|
|
DATA runtime·mainPC+0(SB)/8,$runtime·main(SB)
|
|
|
|
|
GLOBL runtime·mainPC(SB),RODATA,$8
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
TEXT runtime·breakpoint(SB),NOSPLIT,$-8-0
|
|
|
|
|
BRK
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·asminit(SB),NOSPLIT,$-8-0
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·reginit(SB),NOSPLIT,$-8-0
|
|
|
|
|
// initialize essential FP registers
|
|
|
|
|
FMOVD $4503601774854144.0, F27
|
|
|
|
|
FMOVD $0.5, F29
|
|
|
|
|
FSUBD F29, F29, F28
|
|
|
|
|
FADDD F29, F29, F30
|
|
|
|
|
FADDD F30, F30, F31
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* go-routine
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
// void gosave(Gobuf*)
|
|
|
|
|
// save state in Gobuf; setjmp
|
|
|
|
|
TEXT runtime·gosave(SB), NOSPLIT, $-8-8
|
|
|
|
|
MOVD buf+0(FP), R3
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, gobuf_sp(R3)
|
|
|
|
|
MOVD LR, gobuf_pc(R3)
|
|
|
|
|
MOVD g, gobuf_g(R3)
|
|
|
|
|
MOVD ZR, gobuf_lr(R3)
|
|
|
|
|
MOVD ZR, gobuf_ret(R3)
|
|
|
|
|
MOVD ZR, gobuf_ctxt(R3)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// void gogo(Gobuf*)
|
|
|
|
|
// restore state from Gobuf; longjmp
|
|
|
|
|
TEXT runtime·gogo(SB), NOSPLIT, $-8-8
|
|
|
|
|
MOVD buf+0(FP), R5
|
|
|
|
|
MOVD gobuf_g(R5), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
|
|
|
|
|
MOVD 0(g), R4 // make sure g is not nil
|
|
|
|
|
MOVD gobuf_sp(R5), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
MOVD gobuf_lr(R5), LR
|
|
|
|
|
MOVD gobuf_ret(R5), R0
|
|
|
|
|
MOVD gobuf_ctxt(R5), R26
|
|
|
|
|
MOVD $0, gobuf_sp(R5)
|
|
|
|
|
MOVD $0, gobuf_ret(R5)
|
|
|
|
|
MOVD $0, gobuf_lr(R5)
|
|
|
|
|
MOVD $0, gobuf_ctxt(R5)
|
|
|
|
|
CMP ZR, ZR // set condition codes for == test, needed by stack split
|
|
|
|
|
MOVD gobuf_pc(R5), R6
|
|
|
|
|
B (R6)
|
|
|
|
|
|
|
|
|
|
// void mcall(fn func(*g))
|
|
|
|
|
// Switch to m->g0's stack, call fn(g).
|
|
|
|
|
// Fn must never return. It should gogo(&g->sched)
|
|
|
|
|
// to keep running g.
|
|
|
|
|
TEXT runtime·mcall(SB), NOSPLIT, $-8-8
|
|
|
|
|
// Save caller state in g->sched
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(g)
|
|
|
|
|
MOVD LR, (g_sched+gobuf_pc)(g)
|
|
|
|
|
MOVD $0, (g_sched+gobuf_lr)(g)
|
|
|
|
|
MOVD g, (g_sched+gobuf_g)(g)
|
|
|
|
|
|
|
|
|
|
// Switch to m->g0 & its stack, call fn.
|
|
|
|
|
MOVD g, R3
|
|
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD m_g0(R8), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
CMP g, R3
|
|
|
|
|
BNE 2(PC)
|
|
|
|
|
B runtime·badmcall(SB)
|
|
|
|
|
MOVD fn+0(FP), R26 // context
|
|
|
|
|
MOVD 0(R26), R4 // code pointer
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R0
|
|
|
|
|
MOVD R0, RSP // sp = m->g0->sched.sp
|
|
|
|
|
MOVD R3, -8(RSP)
|
|
|
|
|
MOVD $0, -16(RSP)
|
|
|
|
|
SUB $16, RSP
|
|
|
|
|
BL (R4)
|
|
|
|
|
B runtime·badmcall2(SB)
|
|
|
|
|
|
|
|
|
|
// systemstack_switch is a dummy routine that systemstack leaves at the bottom
|
|
|
|
|
// of the G stack. We need to distinguish the routine that
|
|
|
|
|
// lives at the bottom of the G stack from the one that lives
|
|
|
|
|
// at the top of the system stack because the one at the top of
|
|
|
|
|
// the system stack terminates the stack walk (see topofstack()).
|
|
|
|
|
TEXT runtime·systemstack_switch(SB), NOSPLIT, $0-0
|
|
|
|
|
UNDEF
|
|
|
|
|
BL (LR) // make sure this function is not leaf
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// func systemstack(fn func())
|
|
|
|
|
TEXT runtime·systemstack(SB), NOSPLIT, $0-8
|
|
|
|
|
MOVD fn+0(FP), R3 // R3 = fn
|
|
|
|
|
MOVD R3, R26 // context
|
|
|
|
|
MOVD g_m(g), R4 // R4 = m
|
|
|
|
|
|
|
|
|
|
MOVD m_gsignal(R4), R5 // R5 = gsignal
|
|
|
|
|
CMP g, R5
|
|
|
|
|
BEQ noswitch
|
|
|
|
|
|
|
|
|
|
MOVD m_g0(R4), R5 // R5 = g0
|
|
|
|
|
CMP g, R5
|
|
|
|
|
BEQ noswitch
|
|
|
|
|
|
|
|
|
|
MOVD m_curg(R4), R6
|
|
|
|
|
CMP g, R6
|
|
|
|
|
BEQ switch
|
|
|
|
|
|
|
|
|
|
// Bad: g is not gsignal, not g0, not curg. What is it?
|
|
|
|
|
// Hide call from linker nosplit analysis.
|
|
|
|
|
MOVD $runtime·badsystemstack(SB), R3
|
|
|
|
|
BL (R3)
|
|
|
|
|
|
|
|
|
|
switch:
|
|
|
|
|
// save our state in g->sched. Pretend to
|
|
|
|
|
// be systemstack_switch if the G stack is scanned.
|
|
|
|
|
MOVD $runtime·systemstack_switch(SB), R6
|
|
|
|
|
ADD $8, R6 // get past prologue
|
|
|
|
|
MOVD R6, (g_sched+gobuf_pc)(g)
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(g)
|
|
|
|
|
MOVD $0, (g_sched+gobuf_lr)(g)
|
|
|
|
|
MOVD g, (g_sched+gobuf_g)(g)
|
|
|
|
|
|
|
|
|
|
// switch to g0
|
|
|
|
|
MOVD R5, g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R3
|
|
|
|
|
// make it look like mstart called systemstack on g0, to stop traceback
|
|
|
|
|
SUB $16, R3
|
|
|
|
|
AND $~15, R3
|
|
|
|
|
MOVD $runtime·mstart(SB), R4
|
|
|
|
|
MOVD R4, 0(R3)
|
|
|
|
|
MOVD R3, RSP
|
|
|
|
|
|
|
|
|
|
// call target function
|
|
|
|
|
MOVD 0(R26), R3 // code pointer
|
|
|
|
|
BL (R3)
|
|
|
|
|
|
|
|
|
|
// switch back to g
|
|
|
|
|
MOVD g_m(g), R3
|
|
|
|
|
MOVD m_curg(R3), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
MOVD $0, (g_sched+gobuf_sp)(g)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
noswitch:
|
|
|
|
|
// already on m stack, just call directly
|
|
|
|
|
MOVD 0(R26), R3 // code pointer
|
|
|
|
|
BL (R3)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* support for morestack
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
// Called during function prolog when more stack is needed.
|
|
|
|
|
// Caller has already loaded:
|
|
|
|
|
// R3 prolog's LR (R30)
|
|
|
|
|
//
|
|
|
|
|
// The traceback routines see morestack on a g0 as being
|
|
|
|
|
// the top of a stack (for example, morestack calling newstack
|
|
|
|
|
// calling the scheduler calling newm calling gc), so we must
|
|
|
|
|
// record an argument size. For that purpose, it has no arguments.
|
|
|
|
|
TEXT runtime·morestack(SB),NOSPLIT,$-8-0
|
|
|
|
|
// Cannot grow scheduler stack (m->g0).
|
|
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD m_g0(R8), R4
|
|
|
|
|
CMP g, R4
|
|
|
|
|
BNE 2(PC)
|
|
|
|
|
B runtime·abort(SB)
|
|
|
|
|
|
|
|
|
|
// Cannot grow signal stack (m->gsignal).
|
|
|
|
|
MOVD m_gsignal(R8), R4
|
|
|
|
|
CMP g, R4
|
|
|
|
|
BNE 2(PC)
|
|
|
|
|
B runtime·abort(SB)
|
|
|
|
|
|
|
|
|
|
// Called from f.
|
|
|
|
|
// Set g->sched to context in f
|
|
|
|
|
MOVD R26, (g_sched+gobuf_ctxt)(g)
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(g)
|
|
|
|
|
MOVD LR, (g_sched+gobuf_pc)(g)
|
|
|
|
|
MOVD R3, (g_sched+gobuf_lr)(g)
|
|
|
|
|
|
|
|
|
|
// Called from f.
|
|
|
|
|
// Set m->morebuf to f's callers.
|
|
|
|
|
MOVD R3, (m_morebuf+gobuf_pc)(R8) // f's caller's PC
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (m_morebuf+gobuf_sp)(R8) // f's caller's RSP
|
|
|
|
|
MOVD g, (m_morebuf+gobuf_g)(R8)
|
|
|
|
|
|
|
|
|
|
// Call newstack on m->g0's stack.
|
|
|
|
|
MOVD m_g0(R8), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
BL runtime·newstack(SB)
|
|
|
|
|
|
|
|
|
|
// Not reached, but make sure the return PC from the call to newstack
|
|
|
|
|
// is still in this function, and not the beginning of the next.
|
|
|
|
|
UNDEF
|
|
|
|
|
|
|
|
|
|
TEXT runtime·morestack_noctxt(SB),NOSPLIT,$-4-0
|
|
|
|
|
MOVW $0, R26
|
|
|
|
|
B runtime·morestack(SB)
|
|
|
|
|
|
runtime: implement GC stack barriers
This commit implements stack barriers to minimize the amount of
stack re-scanning that must be done during mark termination.
Currently the GC scans stacks of active goroutines twice during every
GC cycle: once at the beginning during root discovery and once at the
end during mark termination. The second scan happens while the world
is stopped and guarantees that we've seen all of the roots (since
there are no write barriers on writes to local stack
variables). However, this means pause time is proportional to stack
size. In particularly recursive programs, this can drive pause time up
past our 10ms goal (e.g., it takes about 150ms to scan a 50MB heap).
Re-scanning the entire stack is rarely necessary, especially for large
stacks, because usually most of the frames on the stack were not
active between the first and second scans and hence any changes to
these frames (via non-escaping pointers passed down the stack) were
tracked by write barriers.
To efficiently track how far a stack has been unwound since the first
scan (and, hence, how much needs to be re-scanned), this commit
introduces stack barriers. During the first scan, at exponentially
spaced points in each stack, the scan overwrites return PCs with the
PC of the stack barrier function. When "returned" to, the stack
barrier function records how far the stack has unwound and jumps to
the original return PC for that point in the stack. Then the second
scan only needs to proceed as far as the lowest barrier that hasn't
been hit.
For deeply recursive programs, this substantially reduces mark
termination time (and hence pause time). For the goscheme example
linked in issue #10898, prior to this change, mark termination times
were typically between 100 and 500ms; with this change, mark
termination times are typically between 10 and 20ms. As a result of
the reduced stack scanning work, this reduces overall execution time
of the goscheme example by 20%.
Fixes #10898.
The effect of this on programs that are not deeply recursive is
minimal:
name old time/op new time/op delta
BinaryTree17 3.16s ± 2% 3.26s ± 1% +3.31% (p=0.000 n=19+19)
Fannkuch11 2.42s ± 1% 2.48s ± 1% +2.24% (p=0.000 n=17+19)
FmtFprintfEmpty 50.0ns ± 3% 49.8ns ± 1% ~ (p=0.534 n=20+19)
FmtFprintfString 173ns ± 0% 175ns ± 0% +1.49% (p=0.000 n=16+19)
FmtFprintfInt 170ns ± 1% 175ns ± 1% +2.97% (p=0.000 n=20+19)
FmtFprintfIntInt 288ns ± 0% 295ns ± 0% +2.73% (p=0.000 n=16+19)
FmtFprintfPrefixedInt 242ns ± 1% 252ns ± 1% +4.13% (p=0.000 n=18+18)
FmtFprintfFloat 324ns ± 0% 323ns ± 0% -0.36% (p=0.000 n=20+19)
FmtManyArgs 1.14µs ± 0% 1.12µs ± 1% -1.01% (p=0.000 n=18+19)
GobDecode 8.88ms ± 1% 8.87ms ± 0% ~ (p=0.480 n=19+18)
GobEncode 6.80ms ± 1% 6.85ms ± 0% +0.82% (p=0.000 n=20+18)
Gzip 363ms ± 1% 363ms ± 1% ~ (p=0.077 n=18+20)
Gunzip 90.6ms ± 0% 90.0ms ± 1% -0.71% (p=0.000 n=17+18)
HTTPClientServer 51.5µs ± 1% 50.8µs ± 1% -1.32% (p=0.000 n=18+18)
JSONEncode 17.0ms ± 0% 17.1ms ± 0% +0.40% (p=0.000 n=18+17)
JSONDecode 61.8ms ± 0% 63.8ms ± 1% +3.11% (p=0.000 n=18+17)
Mandelbrot200 3.84ms ± 0% 3.84ms ± 1% ~ (p=0.583 n=19+19)
GoParse 3.71ms ± 1% 3.72ms ± 1% ~ (p=0.159 n=18+19)
RegexpMatchEasy0_32 100ns ± 0% 100ns ± 1% -0.19% (p=0.033 n=17+19)
RegexpMatchEasy0_1K 342ns ± 1% 331ns ± 0% -3.41% (p=0.000 n=19+19)
RegexpMatchEasy1_32 82.5ns ± 0% 81.7ns ± 0% -0.98% (p=0.000 n=18+18)
RegexpMatchEasy1_1K 505ns ± 0% 494ns ± 1% -2.16% (p=0.000 n=18+18)
RegexpMatchMedium_32 137ns ± 1% 137ns ± 1% -0.24% (p=0.048 n=20+18)
RegexpMatchMedium_1K 41.6µs ± 0% 41.3µs ± 1% -0.57% (p=0.004 n=18+20)
RegexpMatchHard_32 2.11µs ± 0% 2.11µs ± 1% +0.20% (p=0.037 n=17+19)
RegexpMatchHard_1K 63.9µs ± 2% 63.3µs ± 0% -0.99% (p=0.000 n=20+17)
Revcomp 560ms ± 1% 522ms ± 0% -6.87% (p=0.000 n=18+16)
Template 75.0ms ± 0% 75.1ms ± 1% +0.18% (p=0.013 n=18+19)
TimeParse 358ns ± 1% 364ns ± 0% +1.74% (p=0.000 n=20+15)
TimeFormat 360ns ± 0% 372ns ± 0% +3.55% (p=0.000 n=20+18)
Change-Id: If8a9bfae6c128d15a4f405e02bcfa50129df82a2
Reviewed-on: https://go-review.googlesource.com/10314
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
2015-05-20 14:30:49 -06:00
|
|
|
TEXT runtime·stackBarrier(SB),NOSPLIT,$0
|
|
|
|
|
// We came here via a RET to an overwritten LR.
|
|
|
|
|
// R0 may be live (see return0). Other registers are available.
|
|
|
|
|
|
|
|
|
|
// Get the original return PC, g.stkbar[g.stkbarPos].savedLRVal.
|
|
|
|
|
MOVD (g_stkbar+slice_array)(g), R4
|
|
|
|
|
MOVD g_stkbarPos(g), R5
|
|
|
|
|
MOVD $stkbar__size, R6
|
|
|
|
|
MUL R5, R6
|
|
|
|
|
ADD R4, R6
|
|
|
|
|
MOVD stkbar_savedLRVal(R6), R6
|
|
|
|
|
// Record that this stack barrier was hit.
|
|
|
|
|
ADD $1, R5
|
|
|
|
|
MOVD R5, g_stkbarPos(g)
|
|
|
|
|
// Jump to the original return PC.
|
|
|
|
|
B (R6)
|
|
|
|
|
|
2015-03-08 07:20:20 -06:00
|
|
|
// reflectcall: call a function with the given argument list
|
|
|
|
|
// func call(argtype *_type, f *FuncVal, arg *byte, argsize, retoffset uint32).
|
|
|
|
|
// we don't have variable-sized frames, so we use a small number
|
|
|
|
|
// of constant-sized-frame functions to encode a few bits of size in the pc.
|
|
|
|
|
// Caution: ugly multiline assembly macros in your future!
|
|
|
|
|
|
|
|
|
|
#define DISPATCH(NAME,MAXSIZE) \
|
|
|
|
|
MOVD $MAXSIZE, R27; \
|
|
|
|
|
CMP R27, R16; \
|
|
|
|
|
BGT 3(PC); \
|
|
|
|
|
MOVD $NAME(SB), R27; \
|
|
|
|
|
B (R27)
|
|
|
|
|
// Note: can't just "B NAME(SB)" - bad inlining results.
|
|
|
|
|
|
|
|
|
|
TEXT reflect·call(SB), NOSPLIT, $0-0
|
|
|
|
|
B ·reflectcall(SB)
|
|
|
|
|
|
|
|
|
|
TEXT ·reflectcall(SB), NOSPLIT, $-8-32
|
|
|
|
|
MOVWU argsize+24(FP), R16
|
|
|
|
|
// NOTE(rsc): No call16, because CALLFN needs four words
|
|
|
|
|
// of argument space to invoke callwritebarrier.
|
|
|
|
|
DISPATCH(runtime·call32, 32)
|
|
|
|
|
DISPATCH(runtime·call64, 64)
|
|
|
|
|
DISPATCH(runtime·call128, 128)
|
|
|
|
|
DISPATCH(runtime·call256, 256)
|
|
|
|
|
DISPATCH(runtime·call512, 512)
|
|
|
|
|
DISPATCH(runtime·call1024, 1024)
|
|
|
|
|
DISPATCH(runtime·call2048, 2048)
|
|
|
|
|
DISPATCH(runtime·call4096, 4096)
|
|
|
|
|
DISPATCH(runtime·call8192, 8192)
|
|
|
|
|
DISPATCH(runtime·call16384, 16384)
|
|
|
|
|
DISPATCH(runtime·call32768, 32768)
|
|
|
|
|
DISPATCH(runtime·call65536, 65536)
|
|
|
|
|
DISPATCH(runtime·call131072, 131072)
|
|
|
|
|
DISPATCH(runtime·call262144, 262144)
|
|
|
|
|
DISPATCH(runtime·call524288, 524288)
|
|
|
|
|
DISPATCH(runtime·call1048576, 1048576)
|
|
|
|
|
DISPATCH(runtime·call2097152, 2097152)
|
|
|
|
|
DISPATCH(runtime·call4194304, 4194304)
|
|
|
|
|
DISPATCH(runtime·call8388608, 8388608)
|
|
|
|
|
DISPATCH(runtime·call16777216, 16777216)
|
|
|
|
|
DISPATCH(runtime·call33554432, 33554432)
|
|
|
|
|
DISPATCH(runtime·call67108864, 67108864)
|
|
|
|
|
DISPATCH(runtime·call134217728, 134217728)
|
|
|
|
|
DISPATCH(runtime·call268435456, 268435456)
|
|
|
|
|
DISPATCH(runtime·call536870912, 536870912)
|
|
|
|
|
DISPATCH(runtime·call1073741824, 1073741824)
|
|
|
|
|
MOVD $runtime·badreflectcall(SB), R0
|
|
|
|
|
B (R0)
|
|
|
|
|
|
|
|
|
|
#define CALLFN(NAME,MAXSIZE) \
|
|
|
|
|
TEXT NAME(SB), WRAPPER, $MAXSIZE-24; \
|
|
|
|
|
NO_LOCAL_POINTERS; \
|
|
|
|
|
/* copy arguments to stack */ \
|
|
|
|
|
MOVD arg+16(FP), R3; \
|
|
|
|
|
MOVWU argsize+24(FP), R4; \
|
|
|
|
|
MOVD RSP, R5; \
|
|
|
|
|
ADD $(8-1), R5; \
|
|
|
|
|
SUB $1, R3; \
|
|
|
|
|
ADD R5, R4; \
|
|
|
|
|
CMP R5, R4; \
|
|
|
|
|
BEQ 4(PC); \
|
|
|
|
|
MOVBU.W 1(R3), R6; \
|
|
|
|
|
MOVBU.W R6, 1(R5); \
|
|
|
|
|
B -4(PC); \
|
|
|
|
|
/* call function */ \
|
|
|
|
|
MOVD f+8(FP), R26; \
|
|
|
|
|
MOVD (R26), R0; \
|
|
|
|
|
PCDATA $PCDATA_StackMapIndex, $0; \
|
|
|
|
|
BL (R0); \
|
|
|
|
|
/* copy return values back */ \
|
|
|
|
|
MOVD arg+16(FP), R3; \
|
|
|
|
|
MOVWU n+24(FP), R4; \
|
|
|
|
|
MOVWU retoffset+28(FP), R6; \
|
|
|
|
|
MOVD RSP, R5; \
|
|
|
|
|
ADD R6, R5; \
|
|
|
|
|
ADD R6, R3; \
|
|
|
|
|
SUB R6, R4; \
|
|
|
|
|
ADD $(8-1), R5; \
|
|
|
|
|
SUB $1, R3; \
|
|
|
|
|
ADD R5, R4; \
|
|
|
|
|
loop: \
|
|
|
|
|
CMP R5, R4; \
|
|
|
|
|
BEQ end; \
|
|
|
|
|
MOVBU.W 1(R5), R6; \
|
|
|
|
|
MOVBU.W R6, 1(R3); \
|
|
|
|
|
B loop; \
|
|
|
|
|
end: \
|
|
|
|
|
/* execute write barrier updates */ \
|
|
|
|
|
MOVD argtype+0(FP), R7; \
|
|
|
|
|
MOVD arg+16(FP), R3; \
|
|
|
|
|
MOVWU n+24(FP), R4; \
|
|
|
|
|
MOVWU retoffset+28(FP), R6; \
|
|
|
|
|
MOVD R7, 8(RSP); \
|
|
|
|
|
MOVD R3, 16(RSP); \
|
|
|
|
|
MOVD R4, 24(RSP); \
|
|
|
|
|
MOVD R6, 32(RSP); \
|
|
|
|
|
BL runtime·callwritebarrier(SB); \
|
|
|
|
|
RET
|
|
|
|
|
|
2015-07-29 13:04:30 -06:00
|
|
|
// These have 8 added to make the overall frame size a multiple of 16,
|
|
|
|
|
// as required by the ABI. (There is another +8 for the saved LR.)
|
|
|
|
|
CALLFN(·call16, 24 )
|
|
|
|
|
CALLFN(·call32, 40 )
|
|
|
|
|
CALLFN(·call64, 72 )
|
|
|
|
|
CALLFN(·call128, 136 )
|
|
|
|
|
CALLFN(·call256, 264 )
|
|
|
|
|
CALLFN(·call512, 520 )
|
|
|
|
|
CALLFN(·call1024, 1032 )
|
|
|
|
|
CALLFN(·call2048, 2056 )
|
|
|
|
|
CALLFN(·call4096, 4104 )
|
|
|
|
|
CALLFN(·call8192, 8200 )
|
|
|
|
|
CALLFN(·call16384, 16392 )
|
|
|
|
|
CALLFN(·call32768, 32776 )
|
|
|
|
|
CALLFN(·call65536, 65544 )
|
|
|
|
|
CALLFN(·call131072, 131080 )
|
|
|
|
|
CALLFN(·call262144, 262152 )
|
|
|
|
|
CALLFN(·call524288, 524296 )
|
|
|
|
|
CALLFN(·call1048576, 1048584 )
|
|
|
|
|
CALLFN(·call2097152, 2097160 )
|
|
|
|
|
CALLFN(·call4194304, 4194312 )
|
|
|
|
|
CALLFN(·call8388608, 8388616 )
|
|
|
|
|
CALLFN(·call16777216, 16777224 )
|
|
|
|
|
CALLFN(·call33554432, 33554440 )
|
|
|
|
|
CALLFN(·call67108864, 67108872 )
|
|
|
|
|
CALLFN(·call134217728, 134217736 )
|
|
|
|
|
CALLFN(·call268435456, 268435464 )
|
|
|
|
|
CALLFN(·call536870912, 536870920 )
|
|
|
|
|
CALLFN(·call1073741824, 1073741832 )
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
// bool cas(uint32 *ptr, uint32 old, uint32 new)
|
|
|
|
|
// Atomically:
|
|
|
|
|
// if(*val == old){
|
|
|
|
|
// *val = new;
|
|
|
|
|
// return 1;
|
|
|
|
|
// } else
|
|
|
|
|
// return 0;
|
|
|
|
|
TEXT runtime·cas(SB), NOSPLIT, $0-17
|
|
|
|
|
MOVD ptr+0(FP), R0
|
|
|
|
|
MOVW old+8(FP), R1
|
|
|
|
|
MOVW new+12(FP), R2
|
|
|
|
|
again:
|
|
|
|
|
LDAXRW (R0), R3
|
|
|
|
|
CMPW R1, R3
|
|
|
|
|
BNE ok
|
|
|
|
|
STLXRW R2, (R0), R3
|
|
|
|
|
CBNZ R3, again
|
|
|
|
|
ok:
|
|
|
|
|
CSET EQ, R0
|
|
|
|
|
MOVB R0, ret+16(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·casuintptr(SB), NOSPLIT, $0-25
|
|
|
|
|
B runtime·cas64(SB)
|
|
|
|
|
|
|
|
|
|
TEXT runtime·atomicloaduintptr(SB), NOSPLIT, $-8-16
|
|
|
|
|
B runtime·atomicload64(SB)
|
|
|
|
|
|
|
|
|
|
TEXT runtime·atomicloaduint(SB), NOSPLIT, $-8-16
|
|
|
|
|
B runtime·atomicload64(SB)
|
|
|
|
|
|
|
|
|
|
TEXT runtime·atomicstoreuintptr(SB), NOSPLIT, $0-16
|
|
|
|
|
B runtime·atomicstore64(SB)
|
|
|
|
|
|
2015-04-16 13:23:17 -06:00
|
|
|
// AES hashing not implemented for ARM64, issue #10109.
|
|
|
|
|
TEXT runtime·aeshash(SB),NOSPLIT,$-8-0
|
|
|
|
|
MOVW $0, R0
|
|
|
|
|
MOVW (R0), R1
|
|
|
|
|
TEXT runtime·aeshash32(SB),NOSPLIT,$-8-0
|
|
|
|
|
MOVW $0, R0
|
|
|
|
|
MOVW (R0), R1
|
|
|
|
|
TEXT runtime·aeshash64(SB),NOSPLIT,$-8-0
|
|
|
|
|
MOVW $0, R0
|
|
|
|
|
MOVW (R0), R1
|
|
|
|
|
TEXT runtime·aeshashstr(SB),NOSPLIT,$-8-0
|
|
|
|
|
MOVW $0, R0
|
|
|
|
|
MOVW (R0), R1
|
|
|
|
|
|
2015-03-08 07:20:20 -06:00
|
|
|
// bool casp(void **val, void *old, void *new)
|
|
|
|
|
// Atomically:
|
|
|
|
|
// if(*val == old){
|
|
|
|
|
// *val = new;
|
|
|
|
|
// return 1;
|
|
|
|
|
// } else
|
|
|
|
|
// return 0;
|
|
|
|
|
TEXT runtime·casp1(SB), NOSPLIT, $0-25
|
|
|
|
|
B runtime·cas64(SB)
|
|
|
|
|
|
|
|
|
|
TEXT runtime·procyield(SB),NOSPLIT,$0-0
|
|
|
|
|
MOVWU cycles+0(FP), R0
|
|
|
|
|
again:
|
|
|
|
|
YIELD
|
|
|
|
|
SUBW $1, R0
|
|
|
|
|
CBNZ R0, again
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// void jmpdefer(fv, sp);
|
|
|
|
|
// called from deferreturn.
|
|
|
|
|
// 1. grab stored LR for caller
|
|
|
|
|
// 2. sub 4 bytes to get back to BL deferreturn
|
|
|
|
|
// 3. BR to fn
|
|
|
|
|
TEXT runtime·jmpdefer(SB), NOSPLIT, $-8-16
|
|
|
|
|
MOVD 0(RSP), R0
|
|
|
|
|
SUB $4, R0
|
|
|
|
|
MOVD R0, LR
|
|
|
|
|
|
|
|
|
|
MOVD fv+0(FP), R26
|
|
|
|
|
MOVD argp+8(FP), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
SUB $8, RSP
|
|
|
|
|
MOVD 0(R26), R3
|
|
|
|
|
B (R3)
|
|
|
|
|
|
|
|
|
|
// Save state of caller into g->sched. Smashes R0.
|
|
|
|
|
TEXT gosave<>(SB),NOSPLIT,$-8
|
|
|
|
|
MOVD LR, (g_sched+gobuf_pc)(g)
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(g)
|
|
|
|
|
MOVD $0, (g_sched+gobuf_lr)(g)
|
|
|
|
|
MOVD $0, (g_sched+gobuf_ret)(g)
|
|
|
|
|
MOVD $0, (g_sched+gobuf_ctxt)(g)
|
|
|
|
|
RET
|
|
|
|
|
|
2015-04-27 01:32:23 -06:00
|
|
|
// func asmcgocall(fn, arg unsafe.Pointer) int32
|
2015-03-08 07:20:20 -06:00
|
|
|
// Call fn(arg) on the scheduler stack,
|
|
|
|
|
// aligned appropriately for the gcc ABI.
|
2015-04-03 02:37:22 -06:00
|
|
|
// See cgocall.go for more details.
|
2015-04-27 01:32:23 -06:00
|
|
|
TEXT ·asmcgocall(SB),NOSPLIT,$0-20
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD fn+0(FP), R1
|
|
|
|
|
MOVD arg+8(FP), R0
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
MOVD RSP, R2 // save original stack pointer
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD g, R4
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
// Figure out if we need to switch to m->g0 stack.
|
|
|
|
|
// We get called to create new OS threads too, and those
|
|
|
|
|
// come in on the m->g0 stack already.
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD m_g0(R8), R3
|
|
|
|
|
CMP R3, g
|
2015-03-08 07:20:20 -06:00
|
|
|
BEQ g0
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD R0, R9 // gosave<> and save_g might clobber R0
|
2015-03-08 07:20:20 -06:00
|
|
|
BL gosave<>(SB)
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD R3, g
|
2015-03-08 07:20:20 -06:00
|
|
|
BL runtime·save_g(SB)
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD (g_sched+gobuf_sp)(g), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
MOVD R9, R0
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
// Now on a scheduling stack (a pthread-created stack).
|
|
|
|
|
g0:
|
2015-04-03 02:37:22 -06:00
|
|
|
// Save room for two of our pointers /*, plus 32 bytes of callee
|
|
|
|
|
// save area that lives on the caller stack. */
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD RSP, R13
|
2015-04-03 02:37:22 -06:00
|
|
|
SUB $16, R13
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD R13, RSP
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD R4, 0(RSP) // save old g on stack
|
|
|
|
|
MOVD (g_stack+stack_hi)(R4), R4
|
|
|
|
|
SUB R2, R4
|
|
|
|
|
MOVD R4, 8(RSP) // save depth in old g stack (can't just save SP, as stack might be copied during a callback)
|
|
|
|
|
BL (R1)
|
|
|
|
|
MOVD R0, R9
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
// Restore g, stack pointer. R0 is errno, so don't touch it
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD 0(RSP), g
|
2015-03-08 07:20:20 -06:00
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_stack+stack_hi)(g), R5
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD 8(RSP), R6
|
2015-03-08 07:20:20 -06:00
|
|
|
SUB R6, R5
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD R9, R0
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD R5, RSP
|
2015-04-27 01:32:23 -06:00
|
|
|
|
|
|
|
|
MOVW R0, ret+16(FP)
|
2015-03-08 07:20:20 -06:00
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// cgocallback(void (*fn)(void*), void *frame, uintptr framesize)
|
|
|
|
|
// Turn the fn into a Go func (by taking its address) and call
|
|
|
|
|
// cgocallback_gofunc.
|
|
|
|
|
TEXT runtime·cgocallback(SB),NOSPLIT,$24-24
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD $fn+0(FP), R0
|
|
|
|
|
MOVD R0, 8(RSP)
|
|
|
|
|
MOVD frame+8(FP), R0
|
|
|
|
|
MOVD R0, 16(RSP)
|
|
|
|
|
MOVD framesize+16(FP), R0
|
|
|
|
|
MOVD R0, 24(RSP)
|
|
|
|
|
MOVD $runtime·cgocallback_gofunc(SB), R0
|
|
|
|
|
BL (R0)
|
2015-03-08 07:20:20 -06:00
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// cgocallback_gofunc(FuncVal*, void *frame, uintptr framesize)
|
2015-04-03 02:37:22 -06:00
|
|
|
// See cgocall.go for more details.
|
2015-07-29 13:04:30 -06:00
|
|
|
TEXT ·cgocallback_gofunc(SB),NOSPLIT,$24-24
|
2015-03-08 07:20:20 -06:00
|
|
|
NO_LOCAL_POINTERS
|
|
|
|
|
|
2015-04-03 02:37:22 -06:00
|
|
|
// Load g from thread-local storage.
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVB runtime·iscgo(SB), R3
|
2015-04-03 02:37:22 -06:00
|
|
|
CMP $0, R3
|
2015-03-08 07:20:20 -06:00
|
|
|
BEQ nocgo
|
2015-04-03 02:37:22 -06:00
|
|
|
BL runtime·load_g(SB)
|
2015-03-08 07:20:20 -06:00
|
|
|
nocgo:
|
|
|
|
|
|
|
|
|
|
// If g is nil, Go did not create the current thread.
|
|
|
|
|
// Call needm to obtain one for temporary use.
|
|
|
|
|
// In this case, we're running on the thread stack, so there's
|
|
|
|
|
// lots of space, but the linker doesn't know. Hide the call from
|
|
|
|
|
// the linker analysis by using an indirect call.
|
|
|
|
|
CMP $0, g
|
|
|
|
|
BNE havem
|
|
|
|
|
MOVD g, savedm-8(SP) // g is zero, so is m.
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD $runtime·needm(SB), R0
|
|
|
|
|
BL (R0)
|
2015-03-08 07:20:20 -06:00
|
|
|
|
|
|
|
|
// Set m->sched.sp = SP, so that if a panic happens
|
|
|
|
|
// during the function we are about to execute, it will
|
|
|
|
|
// have a valid SP to run on the g0 stack.
|
|
|
|
|
// The next few lines (after the havem label)
|
|
|
|
|
// will save this SP onto the stack and then write
|
|
|
|
|
// the same SP back to m->sched.sp. That seems redundant,
|
|
|
|
|
// but if an unrecovered panic happens, unwindm will
|
|
|
|
|
// restore the g->sched.sp from the stack location
|
|
|
|
|
// and then systemstack will try to use it. If we don't set it here,
|
|
|
|
|
// that restored SP will be uninitialized (typically 0) and
|
|
|
|
|
// will not be usable.
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD m_g0(R8), R3
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(R3)
|
|
|
|
|
|
|
|
|
|
havem:
|
|
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD R8, savedm-8(SP)
|
|
|
|
|
// Now there's a valid m, and we're running on its m->g0.
|
|
|
|
|
// Save current m->g0->sched.sp on stack and then set it to SP.
|
|
|
|
|
// Save current sp in m->g0->sched.sp in preparation for
|
|
|
|
|
// switch back to m->curg stack.
|
2015-04-03 02:37:22 -06:00
|
|
|
// NOTE: unwindm knows that the saved g->sched.sp is at 16(RSP) aka savedsp-16(SP).
|
|
|
|
|
// Beware that the frame size is actually 32.
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD m_g0(R8), R3
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(R3), R4
|
|
|
|
|
MOVD R4, savedsp-16(SP)
|
|
|
|
|
MOVD RSP, R0
|
|
|
|
|
MOVD R0, (g_sched+gobuf_sp)(R3)
|
|
|
|
|
|
|
|
|
|
// Switch to m->curg stack and call runtime.cgocallbackg.
|
|
|
|
|
// Because we are taking over the execution of m->curg
|
|
|
|
|
// but *not* resuming what had been running, we need to
|
|
|
|
|
// save that information (m->curg->sched) so we can restore it.
|
|
|
|
|
// We can restore m->curg->sched.sp easily, because calling
|
|
|
|
|
// runtime.cgocallbackg leaves SP unchanged upon return.
|
|
|
|
|
// To save m->curg->sched.pc, we push it onto the stack.
|
|
|
|
|
// This has the added benefit that it looks to the traceback
|
|
|
|
|
// routine like cgocallbackg is going to return to that
|
|
|
|
|
// PC (because the frame we allocate below has the same
|
|
|
|
|
// size as cgocallback_gofunc's frame declared above)
|
|
|
|
|
// so that the traceback will seamlessly trace back into
|
|
|
|
|
// the earlier calls.
|
|
|
|
|
//
|
|
|
|
|
// In the new goroutine, -16(SP) and -8(SP) are unused.
|
|
|
|
|
MOVD m_curg(R8), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R4 // prepare stack as R4
|
|
|
|
|
MOVD (g_sched+gobuf_pc)(g), R5
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD R5, -(24+8)(R4) // maintain 16-byte SP alignment
|
|
|
|
|
MOVD $-(24+8)(R4), R0
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD R0, RSP
|
|
|
|
|
BL runtime·cgocallbackg(SB)
|
|
|
|
|
|
|
|
|
|
// Restore g->sched (== m->curg->sched) from saved values.
|
|
|
|
|
MOVD 0(RSP), R5
|
|
|
|
|
MOVD R5, (g_sched+gobuf_pc)(g)
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD RSP, R4
|
|
|
|
|
ADD $(24+8), R4, R4
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD R4, (g_sched+gobuf_sp)(g)
|
|
|
|
|
|
|
|
|
|
// Switch back to m->g0's stack and restore m->g0->sched.sp.
|
|
|
|
|
// (Unlike m->curg, the g0 goroutine never uses sched.pc,
|
|
|
|
|
// so we do not have to restore it.)
|
|
|
|
|
MOVD g_m(g), R8
|
|
|
|
|
MOVD m_g0(R8), g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD (g_sched+gobuf_sp)(g), R0
|
|
|
|
|
MOVD R0, RSP
|
|
|
|
|
MOVD savedsp-16(SP), R4
|
|
|
|
|
MOVD R4, (g_sched+gobuf_sp)(g)
|
|
|
|
|
|
|
|
|
|
// If the m on entry was nil, we called needm above to borrow an m
|
|
|
|
|
// for the duration of the call. Since the call is over, return it with dropm.
|
|
|
|
|
MOVD savedm-8(SP), R6
|
|
|
|
|
CMP $0, R6
|
|
|
|
|
BNE droppedm
|
2015-04-03 02:37:22 -06:00
|
|
|
MOVD $runtime·dropm(SB), R0
|
|
|
|
|
BL (R0)
|
2015-03-08 07:20:20 -06:00
|
|
|
droppedm:
|
|
|
|
|
|
|
|
|
|
// Done!
|
|
|
|
|
RET
|
|
|
|
|
|
2015-04-03 02:37:22 -06:00
|
|
|
// Called from cgo wrappers, this function returns g->m->curg.stack.hi.
|
|
|
|
|
// Must obey the gcc calling convention.
|
2015-07-29 13:04:30 -06:00
|
|
|
TEXT _cgo_topofstack(SB),NOSPLIT,$24
|
2015-04-03 02:37:22 -06:00
|
|
|
// g (R28) and REGTMP (R27) might be clobbered by load_g. They
|
|
|
|
|
// are callee-save in the gcc calling convention, so save them.
|
|
|
|
|
MOVD R27, savedR27-8(SP)
|
|
|
|
|
MOVD g, saveG-16(SP)
|
|
|
|
|
|
|
|
|
|
BL runtime·load_g(SB)
|
|
|
|
|
MOVD g_m(g), R0
|
|
|
|
|
MOVD m_curg(R0), R0
|
|
|
|
|
MOVD (g_stack+stack_hi)(R0), R0
|
|
|
|
|
|
|
|
|
|
MOVD saveG-16(SP), g
|
|
|
|
|
MOVD savedR28-8(SP), R27
|
|
|
|
|
RET
|
|
|
|
|
|
2015-03-08 07:20:20 -06:00
|
|
|
// void setg(G*); set g. for use by needm.
|
|
|
|
|
TEXT runtime·setg(SB), NOSPLIT, $0-8
|
|
|
|
|
MOVD gg+0(FP), g
|
|
|
|
|
// This only happens if iscgo, so jump straight to save_g
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
RET
|
|
|
|
|
|
2015-04-03 02:37:22 -06:00
|
|
|
// void setg_gcc(G*); set g called from gcc
|
|
|
|
|
TEXT setg_gcc<>(SB),NOSPLIT,$8
|
|
|
|
|
MOVD R0, g
|
|
|
|
|
MOVD R27, savedR27-8(SP)
|
|
|
|
|
BL runtime·save_g(SB)
|
|
|
|
|
MOVD savedR27-8(SP), R27
|
2015-03-08 07:20:20 -06:00
|
|
|
RET
|
|
|
|
|
|
runtime: implement GC stack barriers
This commit implements stack barriers to minimize the amount of
stack re-scanning that must be done during mark termination.
Currently the GC scans stacks of active goroutines twice during every
GC cycle: once at the beginning during root discovery and once at the
end during mark termination. The second scan happens while the world
is stopped and guarantees that we've seen all of the roots (since
there are no write barriers on writes to local stack
variables). However, this means pause time is proportional to stack
size. In particularly recursive programs, this can drive pause time up
past our 10ms goal (e.g., it takes about 150ms to scan a 50MB heap).
Re-scanning the entire stack is rarely necessary, especially for large
stacks, because usually most of the frames on the stack were not
active between the first and second scans and hence any changes to
these frames (via non-escaping pointers passed down the stack) were
tracked by write barriers.
To efficiently track how far a stack has been unwound since the first
scan (and, hence, how much needs to be re-scanned), this commit
introduces stack barriers. During the first scan, at exponentially
spaced points in each stack, the scan overwrites return PCs with the
PC of the stack barrier function. When "returned" to, the stack
barrier function records how far the stack has unwound and jumps to
the original return PC for that point in the stack. Then the second
scan only needs to proceed as far as the lowest barrier that hasn't
been hit.
For deeply recursive programs, this substantially reduces mark
termination time (and hence pause time). For the goscheme example
linked in issue #10898, prior to this change, mark termination times
were typically between 100 and 500ms; with this change, mark
termination times are typically between 10 and 20ms. As a result of
the reduced stack scanning work, this reduces overall execution time
of the goscheme example by 20%.
Fixes #10898.
The effect of this on programs that are not deeply recursive is
minimal:
name old time/op new time/op delta
BinaryTree17 3.16s ± 2% 3.26s ± 1% +3.31% (p=0.000 n=19+19)
Fannkuch11 2.42s ± 1% 2.48s ± 1% +2.24% (p=0.000 n=17+19)
FmtFprintfEmpty 50.0ns ± 3% 49.8ns ± 1% ~ (p=0.534 n=20+19)
FmtFprintfString 173ns ± 0% 175ns ± 0% +1.49% (p=0.000 n=16+19)
FmtFprintfInt 170ns ± 1% 175ns ± 1% +2.97% (p=0.000 n=20+19)
FmtFprintfIntInt 288ns ± 0% 295ns ± 0% +2.73% (p=0.000 n=16+19)
FmtFprintfPrefixedInt 242ns ± 1% 252ns ± 1% +4.13% (p=0.000 n=18+18)
FmtFprintfFloat 324ns ± 0% 323ns ± 0% -0.36% (p=0.000 n=20+19)
FmtManyArgs 1.14µs ± 0% 1.12µs ± 1% -1.01% (p=0.000 n=18+19)
GobDecode 8.88ms ± 1% 8.87ms ± 0% ~ (p=0.480 n=19+18)
GobEncode 6.80ms ± 1% 6.85ms ± 0% +0.82% (p=0.000 n=20+18)
Gzip 363ms ± 1% 363ms ± 1% ~ (p=0.077 n=18+20)
Gunzip 90.6ms ± 0% 90.0ms ± 1% -0.71% (p=0.000 n=17+18)
HTTPClientServer 51.5µs ± 1% 50.8µs ± 1% -1.32% (p=0.000 n=18+18)
JSONEncode 17.0ms ± 0% 17.1ms ± 0% +0.40% (p=0.000 n=18+17)
JSONDecode 61.8ms ± 0% 63.8ms ± 1% +3.11% (p=0.000 n=18+17)
Mandelbrot200 3.84ms ± 0% 3.84ms ± 1% ~ (p=0.583 n=19+19)
GoParse 3.71ms ± 1% 3.72ms ± 1% ~ (p=0.159 n=18+19)
RegexpMatchEasy0_32 100ns ± 0% 100ns ± 1% -0.19% (p=0.033 n=17+19)
RegexpMatchEasy0_1K 342ns ± 1% 331ns ± 0% -3.41% (p=0.000 n=19+19)
RegexpMatchEasy1_32 82.5ns ± 0% 81.7ns ± 0% -0.98% (p=0.000 n=18+18)
RegexpMatchEasy1_1K 505ns ± 0% 494ns ± 1% -2.16% (p=0.000 n=18+18)
RegexpMatchMedium_32 137ns ± 1% 137ns ± 1% -0.24% (p=0.048 n=20+18)
RegexpMatchMedium_1K 41.6µs ± 0% 41.3µs ± 1% -0.57% (p=0.004 n=18+20)
RegexpMatchHard_32 2.11µs ± 0% 2.11µs ± 1% +0.20% (p=0.037 n=17+19)
RegexpMatchHard_1K 63.9µs ± 2% 63.3µs ± 0% -0.99% (p=0.000 n=20+17)
Revcomp 560ms ± 1% 522ms ± 0% -6.87% (p=0.000 n=18+16)
Template 75.0ms ± 0% 75.1ms ± 1% +0.18% (p=0.013 n=18+19)
TimeParse 358ns ± 1% 364ns ± 0% +1.74% (p=0.000 n=20+15)
TimeFormat 360ns ± 0% 372ns ± 0% +3.55% (p=0.000 n=20+18)
Change-Id: If8a9bfae6c128d15a4f405e02bcfa50129df82a2
Reviewed-on: https://go-review.googlesource.com/10314
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
2015-05-20 14:30:49 -06:00
|
|
|
TEXT runtime·getcallerpc(SB),NOSPLIT,$8-16
|
|
|
|
|
MOVD 16(RSP), R0 // LR saved by caller
|
|
|
|
|
MOVD runtime·stackBarrierPC(SB), R1
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
BNE nobar
|
|
|
|
|
// Get original return PC.
|
|
|
|
|
BL runtime·nextBarrierPC(SB)
|
|
|
|
|
MOVD 8(RSP), R0
|
|
|
|
|
nobar:
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD R0, ret+8(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
runtime: implement GC stack barriers
This commit implements stack barriers to minimize the amount of
stack re-scanning that must be done during mark termination.
Currently the GC scans stacks of active goroutines twice during every
GC cycle: once at the beginning during root discovery and once at the
end during mark termination. The second scan happens while the world
is stopped and guarantees that we've seen all of the roots (since
there are no write barriers on writes to local stack
variables). However, this means pause time is proportional to stack
size. In particularly recursive programs, this can drive pause time up
past our 10ms goal (e.g., it takes about 150ms to scan a 50MB heap).
Re-scanning the entire stack is rarely necessary, especially for large
stacks, because usually most of the frames on the stack were not
active between the first and second scans and hence any changes to
these frames (via non-escaping pointers passed down the stack) were
tracked by write barriers.
To efficiently track how far a stack has been unwound since the first
scan (and, hence, how much needs to be re-scanned), this commit
introduces stack barriers. During the first scan, at exponentially
spaced points in each stack, the scan overwrites return PCs with the
PC of the stack barrier function. When "returned" to, the stack
barrier function records how far the stack has unwound and jumps to
the original return PC for that point in the stack. Then the second
scan only needs to proceed as far as the lowest barrier that hasn't
been hit.
For deeply recursive programs, this substantially reduces mark
termination time (and hence pause time). For the goscheme example
linked in issue #10898, prior to this change, mark termination times
were typically between 100 and 500ms; with this change, mark
termination times are typically between 10 and 20ms. As a result of
the reduced stack scanning work, this reduces overall execution time
of the goscheme example by 20%.
Fixes #10898.
The effect of this on programs that are not deeply recursive is
minimal:
name old time/op new time/op delta
BinaryTree17 3.16s ± 2% 3.26s ± 1% +3.31% (p=0.000 n=19+19)
Fannkuch11 2.42s ± 1% 2.48s ± 1% +2.24% (p=0.000 n=17+19)
FmtFprintfEmpty 50.0ns ± 3% 49.8ns ± 1% ~ (p=0.534 n=20+19)
FmtFprintfString 173ns ± 0% 175ns ± 0% +1.49% (p=0.000 n=16+19)
FmtFprintfInt 170ns ± 1% 175ns ± 1% +2.97% (p=0.000 n=20+19)
FmtFprintfIntInt 288ns ± 0% 295ns ± 0% +2.73% (p=0.000 n=16+19)
FmtFprintfPrefixedInt 242ns ± 1% 252ns ± 1% +4.13% (p=0.000 n=18+18)
FmtFprintfFloat 324ns ± 0% 323ns ± 0% -0.36% (p=0.000 n=20+19)
FmtManyArgs 1.14µs ± 0% 1.12µs ± 1% -1.01% (p=0.000 n=18+19)
GobDecode 8.88ms ± 1% 8.87ms ± 0% ~ (p=0.480 n=19+18)
GobEncode 6.80ms ± 1% 6.85ms ± 0% +0.82% (p=0.000 n=20+18)
Gzip 363ms ± 1% 363ms ± 1% ~ (p=0.077 n=18+20)
Gunzip 90.6ms ± 0% 90.0ms ± 1% -0.71% (p=0.000 n=17+18)
HTTPClientServer 51.5µs ± 1% 50.8µs ± 1% -1.32% (p=0.000 n=18+18)
JSONEncode 17.0ms ± 0% 17.1ms ± 0% +0.40% (p=0.000 n=18+17)
JSONDecode 61.8ms ± 0% 63.8ms ± 1% +3.11% (p=0.000 n=18+17)
Mandelbrot200 3.84ms ± 0% 3.84ms ± 1% ~ (p=0.583 n=19+19)
GoParse 3.71ms ± 1% 3.72ms ± 1% ~ (p=0.159 n=18+19)
RegexpMatchEasy0_32 100ns ± 0% 100ns ± 1% -0.19% (p=0.033 n=17+19)
RegexpMatchEasy0_1K 342ns ± 1% 331ns ± 0% -3.41% (p=0.000 n=19+19)
RegexpMatchEasy1_32 82.5ns ± 0% 81.7ns ± 0% -0.98% (p=0.000 n=18+18)
RegexpMatchEasy1_1K 505ns ± 0% 494ns ± 1% -2.16% (p=0.000 n=18+18)
RegexpMatchMedium_32 137ns ± 1% 137ns ± 1% -0.24% (p=0.048 n=20+18)
RegexpMatchMedium_1K 41.6µs ± 0% 41.3µs ± 1% -0.57% (p=0.004 n=18+20)
RegexpMatchHard_32 2.11µs ± 0% 2.11µs ± 1% +0.20% (p=0.037 n=17+19)
RegexpMatchHard_1K 63.9µs ± 2% 63.3µs ± 0% -0.99% (p=0.000 n=20+17)
Revcomp 560ms ± 1% 522ms ± 0% -6.87% (p=0.000 n=18+16)
Template 75.0ms ± 0% 75.1ms ± 1% +0.18% (p=0.013 n=18+19)
TimeParse 358ns ± 1% 364ns ± 0% +1.74% (p=0.000 n=20+15)
TimeFormat 360ns ± 0% 372ns ± 0% +3.55% (p=0.000 n=20+18)
Change-Id: If8a9bfae6c128d15a4f405e02bcfa50129df82a2
Reviewed-on: https://go-review.googlesource.com/10314
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
2015-05-20 14:30:49 -06:00
|
|
|
TEXT runtime·setcallerpc(SB),NOSPLIT,$8-16
|
2015-03-08 07:20:20 -06:00
|
|
|
MOVD pc+8(FP), R0
|
runtime: implement GC stack barriers
This commit implements stack barriers to minimize the amount of
stack re-scanning that must be done during mark termination.
Currently the GC scans stacks of active goroutines twice during every
GC cycle: once at the beginning during root discovery and once at the
end during mark termination. The second scan happens while the world
is stopped and guarantees that we've seen all of the roots (since
there are no write barriers on writes to local stack
variables). However, this means pause time is proportional to stack
size. In particularly recursive programs, this can drive pause time up
past our 10ms goal (e.g., it takes about 150ms to scan a 50MB heap).
Re-scanning the entire stack is rarely necessary, especially for large
stacks, because usually most of the frames on the stack were not
active between the first and second scans and hence any changes to
these frames (via non-escaping pointers passed down the stack) were
tracked by write barriers.
To efficiently track how far a stack has been unwound since the first
scan (and, hence, how much needs to be re-scanned), this commit
introduces stack barriers. During the first scan, at exponentially
spaced points in each stack, the scan overwrites return PCs with the
PC of the stack barrier function. When "returned" to, the stack
barrier function records how far the stack has unwound and jumps to
the original return PC for that point in the stack. Then the second
scan only needs to proceed as far as the lowest barrier that hasn't
been hit.
For deeply recursive programs, this substantially reduces mark
termination time (and hence pause time). For the goscheme example
linked in issue #10898, prior to this change, mark termination times
were typically between 100 and 500ms; with this change, mark
termination times are typically between 10 and 20ms. As a result of
the reduced stack scanning work, this reduces overall execution time
of the goscheme example by 20%.
Fixes #10898.
The effect of this on programs that are not deeply recursive is
minimal:
name old time/op new time/op delta
BinaryTree17 3.16s ± 2% 3.26s ± 1% +3.31% (p=0.000 n=19+19)
Fannkuch11 2.42s ± 1% 2.48s ± 1% +2.24% (p=0.000 n=17+19)
FmtFprintfEmpty 50.0ns ± 3% 49.8ns ± 1% ~ (p=0.534 n=20+19)
FmtFprintfString 173ns ± 0% 175ns ± 0% +1.49% (p=0.000 n=16+19)
FmtFprintfInt 170ns ± 1% 175ns ± 1% +2.97% (p=0.000 n=20+19)
FmtFprintfIntInt 288ns ± 0% 295ns ± 0% +2.73% (p=0.000 n=16+19)
FmtFprintfPrefixedInt 242ns ± 1% 252ns ± 1% +4.13% (p=0.000 n=18+18)
FmtFprintfFloat 324ns ± 0% 323ns ± 0% -0.36% (p=0.000 n=20+19)
FmtManyArgs 1.14µs ± 0% 1.12µs ± 1% -1.01% (p=0.000 n=18+19)
GobDecode 8.88ms ± 1% 8.87ms ± 0% ~ (p=0.480 n=19+18)
GobEncode 6.80ms ± 1% 6.85ms ± 0% +0.82% (p=0.000 n=20+18)
Gzip 363ms ± 1% 363ms ± 1% ~ (p=0.077 n=18+20)
Gunzip 90.6ms ± 0% 90.0ms ± 1% -0.71% (p=0.000 n=17+18)
HTTPClientServer 51.5µs ± 1% 50.8µs ± 1% -1.32% (p=0.000 n=18+18)
JSONEncode 17.0ms ± 0% 17.1ms ± 0% +0.40% (p=0.000 n=18+17)
JSONDecode 61.8ms ± 0% 63.8ms ± 1% +3.11% (p=0.000 n=18+17)
Mandelbrot200 3.84ms ± 0% 3.84ms ± 1% ~ (p=0.583 n=19+19)
GoParse 3.71ms ± 1% 3.72ms ± 1% ~ (p=0.159 n=18+19)
RegexpMatchEasy0_32 100ns ± 0% 100ns ± 1% -0.19% (p=0.033 n=17+19)
RegexpMatchEasy0_1K 342ns ± 1% 331ns ± 0% -3.41% (p=0.000 n=19+19)
RegexpMatchEasy1_32 82.5ns ± 0% 81.7ns ± 0% -0.98% (p=0.000 n=18+18)
RegexpMatchEasy1_1K 505ns ± 0% 494ns ± 1% -2.16% (p=0.000 n=18+18)
RegexpMatchMedium_32 137ns ± 1% 137ns ± 1% -0.24% (p=0.048 n=20+18)
RegexpMatchMedium_1K 41.6µs ± 0% 41.3µs ± 1% -0.57% (p=0.004 n=18+20)
RegexpMatchHard_32 2.11µs ± 0% 2.11µs ± 1% +0.20% (p=0.037 n=17+19)
RegexpMatchHard_1K 63.9µs ± 2% 63.3µs ± 0% -0.99% (p=0.000 n=20+17)
Revcomp 560ms ± 1% 522ms ± 0% -6.87% (p=0.000 n=18+16)
Template 75.0ms ± 0% 75.1ms ± 1% +0.18% (p=0.013 n=18+19)
TimeParse 358ns ± 1% 364ns ± 0% +1.74% (p=0.000 n=20+15)
TimeFormat 360ns ± 0% 372ns ± 0% +3.55% (p=0.000 n=20+18)
Change-Id: If8a9bfae6c128d15a4f405e02bcfa50129df82a2
Reviewed-on: https://go-review.googlesource.com/10314
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
2015-05-20 14:30:49 -06:00
|
|
|
MOVD 16(RSP), R1
|
|
|
|
|
MOVD runtime·stackBarrierPC(SB), R2
|
|
|
|
|
CMP R1, R2
|
|
|
|
|
BEQ setbar
|
|
|
|
|
MOVD R0, 16(RSP) // set LR in caller
|
|
|
|
|
RET
|
|
|
|
|
setbar:
|
|
|
|
|
// Set the stack barrier return PC.
|
|
|
|
|
MOVD R0, 8(RSP)
|
|
|
|
|
BL runtime·setNextBarrierPC(SB)
|
2015-03-08 07:20:20 -06:00
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·getcallersp(SB),NOSPLIT,$0-16
|
|
|
|
|
MOVD argp+0(FP), R0
|
|
|
|
|
SUB $8, R0
|
|
|
|
|
MOVD R0, ret+8(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·abort(SB),NOSPLIT,$-8-0
|
|
|
|
|
B (ZR)
|
|
|
|
|
UNDEF
|
|
|
|
|
|
|
|
|
|
// memhash_varlen(p unsafe.Pointer, h seed) uintptr
|
|
|
|
|
// redirects to memhash(p, h, size) using the size
|
|
|
|
|
// stored in the closure.
|
|
|
|
|
TEXT runtime·memhash_varlen(SB),NOSPLIT,$40-24
|
|
|
|
|
GO_ARGS
|
|
|
|
|
NO_LOCAL_POINTERS
|
|
|
|
|
MOVD p+0(FP), R3
|
|
|
|
|
MOVD h+8(FP), R4
|
|
|
|
|
MOVD 8(R26), R5
|
|
|
|
|
MOVD R3, 8(RSP)
|
|
|
|
|
MOVD R4, 16(RSP)
|
|
|
|
|
MOVD R5, 24(RSP)
|
|
|
|
|
BL runtime·memhash(SB)
|
|
|
|
|
MOVD 32(RSP), R3
|
|
|
|
|
MOVD R3, ret+16(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·memeq(SB),NOSPLIT,$-8-25
|
|
|
|
|
MOVD a+0(FP), R1
|
|
|
|
|
MOVD b+8(FP), R2
|
|
|
|
|
MOVD size+16(FP), R3
|
|
|
|
|
ADD R1, R3, R6
|
|
|
|
|
MOVD $1, R0
|
|
|
|
|
MOVB R0, ret+24(FP)
|
|
|
|
|
loop:
|
|
|
|
|
CMP R1, R6
|
|
|
|
|
BEQ done
|
|
|
|
|
MOVBU.P 1(R1), R4
|
|
|
|
|
MOVBU.P 1(R2), R5
|
|
|
|
|
CMP R4, R5
|
|
|
|
|
BEQ loop
|
|
|
|
|
|
|
|
|
|
MOVB $0, ret+24(FP)
|
|
|
|
|
done:
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// memequal_varlen(a, b unsafe.Pointer) bool
|
|
|
|
|
TEXT runtime·memequal_varlen(SB),NOSPLIT,$40-17
|
|
|
|
|
MOVD a+0(FP), R3
|
|
|
|
|
MOVD b+8(FP), R4
|
|
|
|
|
CMP R3, R4
|
|
|
|
|
BEQ eq
|
|
|
|
|
MOVD 8(R26), R5 // compiler stores size at offset 8 in the closure
|
|
|
|
|
MOVD R3, 8(RSP)
|
|
|
|
|
MOVD R4, 16(RSP)
|
|
|
|
|
MOVD R5, 24(RSP)
|
|
|
|
|
BL runtime·memeq(SB)
|
|
|
|
|
MOVBU 32(RSP), R3
|
|
|
|
|
MOVB R3, ret+16(FP)
|
|
|
|
|
RET
|
|
|
|
|
eq:
|
|
|
|
|
MOVD $1, R3
|
|
|
|
|
MOVB R3, ret+16(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
2015-04-21 15:22:41 -06:00
|
|
|
TEXT runtime·cmpstring(SB),NOSPLIT,$-4-40
|
2015-04-07 05:38:02 -06:00
|
|
|
MOVD s1_base+0(FP), R2
|
|
|
|
|
MOVD s1_len+8(FP), R0
|
|
|
|
|
MOVD s2_base+16(FP), R3
|
|
|
|
|
MOVD s2_len+24(FP), R1
|
2015-04-21 15:22:41 -06:00
|
|
|
ADD $40, RSP, R7
|
|
|
|
|
B runtime·cmpbody<>(SB)
|
2015-04-07 05:38:02 -06:00
|
|
|
|
2015-04-21 15:22:41 -06:00
|
|
|
TEXT bytes·Compare(SB),NOSPLIT,$-4-56
|
2015-04-07 05:38:02 -06:00
|
|
|
MOVD s1+0(FP), R2
|
|
|
|
|
MOVD s1+8(FP), R0
|
|
|
|
|
MOVD s2+24(FP), R3
|
|
|
|
|
MOVD s2+32(FP), R1
|
2015-04-21 15:22:41 -06:00
|
|
|
ADD $56, RSP, R7
|
|
|
|
|
B runtime·cmpbody<>(SB)
|
2015-04-07 05:38:02 -06:00
|
|
|
|
|
|
|
|
// On entry:
|
|
|
|
|
// R0 is the length of s1
|
|
|
|
|
// R1 is the length of s2
|
|
|
|
|
// R2 points to the start of s1
|
|
|
|
|
// R3 points to the start of s2
|
2015-04-21 15:22:41 -06:00
|
|
|
// R7 points to return value (-1/0/1 will be written here)
|
2015-04-07 05:38:02 -06:00
|
|
|
//
|
|
|
|
|
// On exit:
|
2015-04-21 15:22:41 -06:00
|
|
|
// R4, R5, and R6 are clobbered
|
2015-04-07 05:38:02 -06:00
|
|
|
TEXT runtime·cmpbody<>(SB),NOSPLIT,$-4-0
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
CSEL LT, R1, R0, R6 // R6 is min(R0, R1)
|
|
|
|
|
|
|
|
|
|
ADD R2, R6 // R2 is current byte in s1, R6 is last byte in s1 to compare
|
|
|
|
|
loop:
|
|
|
|
|
CMP R2, R6
|
|
|
|
|
BEQ samebytes // all compared bytes were the same; compare lengths
|
|
|
|
|
MOVBU.P 1(R2), R4
|
|
|
|
|
MOVBU.P 1(R3), R5
|
|
|
|
|
CMP R4, R5
|
|
|
|
|
BEQ loop
|
|
|
|
|
// bytes differed
|
2015-04-21 15:22:41 -06:00
|
|
|
MOVD $1, R4
|
|
|
|
|
CSNEG LT, R4, R4, R4
|
|
|
|
|
MOVD R4, (R7)
|
2015-04-07 05:38:02 -06:00
|
|
|
RET
|
|
|
|
|
samebytes:
|
2015-04-21 15:22:41 -06:00
|
|
|
MOVD $1, R4
|
2015-04-07 05:38:02 -06:00
|
|
|
CMP R0, R1
|
2015-04-21 15:22:41 -06:00
|
|
|
CSNEG LT, R4, R4, R4
|
|
|
|
|
CSEL EQ, ZR, R4, R4
|
|
|
|
|
MOVD R4, (R7)
|
2015-04-07 05:38:02 -06:00
|
|
|
RET
|
|
|
|
|
|
2015-03-08 07:20:20 -06:00
|
|
|
// eqstring tests whether two strings are equal.
|
|
|
|
|
// The compiler guarantees that strings passed
|
|
|
|
|
// to eqstring have equal length.
|
|
|
|
|
// See runtime_test.go:eqstring_generic for
|
|
|
|
|
// equivalent Go code.
|
|
|
|
|
TEXT runtime·eqstring(SB),NOSPLIT,$0-33
|
|
|
|
|
MOVD s1str+0(FP), R0
|
|
|
|
|
MOVD s1len+8(FP), R1
|
|
|
|
|
MOVD s2str+16(FP), R2
|
|
|
|
|
ADD R0, R1 // end
|
|
|
|
|
loop:
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
BEQ equal // reaches the end
|
|
|
|
|
MOVBU.P 1(R0), R4
|
|
|
|
|
MOVBU.P 1(R2), R5
|
|
|
|
|
CMP R4, R5
|
|
|
|
|
BEQ loop
|
|
|
|
|
notequal:
|
|
|
|
|
MOVB ZR, ret+32(FP)
|
|
|
|
|
RET
|
|
|
|
|
equal:
|
|
|
|
|
MOVD $1, R0
|
|
|
|
|
MOVB R0, ret+32(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
//
|
|
|
|
|
// functions for other packages
|
|
|
|
|
//
|
|
|
|
|
TEXT bytes·IndexByte(SB),NOSPLIT,$0-40
|
|
|
|
|
MOVD b+0(FP), R0
|
|
|
|
|
MOVD b_len+8(FP), R1
|
|
|
|
|
MOVBU c+24(FP), R2 // byte to find
|
|
|
|
|
MOVD R0, R4 // store base for later
|
|
|
|
|
ADD R0, R1 // end
|
|
|
|
|
loop:
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
BEQ notfound
|
|
|
|
|
MOVBU.P 1(R0), R3
|
|
|
|
|
CMP R2, R3
|
|
|
|
|
BNE loop
|
|
|
|
|
|
|
|
|
|
SUB $1, R0 // R0 will be one beyond the position we want
|
|
|
|
|
SUB R4, R0 // remove base
|
|
|
|
|
MOVD R0, ret+32(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
notfound:
|
|
|
|
|
MOVD $-1, R0
|
|
|
|
|
MOVD R0, ret+32(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT strings·IndexByte(SB),NOSPLIT,$0-32
|
|
|
|
|
MOVD s+0(FP), R0
|
|
|
|
|
MOVD s_len+8(FP), R1
|
|
|
|
|
MOVBU c+16(FP), R2 // byte to find
|
|
|
|
|
MOVD R0, R4 // store base for later
|
|
|
|
|
ADD R0, R1 // end
|
|
|
|
|
loop:
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
BEQ notfound
|
|
|
|
|
MOVBU.P 1(R0), R3
|
|
|
|
|
CMP R2, R3
|
|
|
|
|
BNE loop
|
|
|
|
|
|
|
|
|
|
SUB $1, R0 // R0 will be one beyond the position we want
|
|
|
|
|
SUB R4, R0 // remove base
|
|
|
|
|
MOVD R0, ret+24(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
notfound:
|
|
|
|
|
MOVD $-1, R0
|
|
|
|
|
MOVD R0, ret+24(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// TODO: share code with memeq?
|
|
|
|
|
TEXT bytes·Equal(SB),NOSPLIT,$0-49
|
|
|
|
|
MOVD a_len+8(FP), R1
|
|
|
|
|
MOVD b_len+32(FP), R3
|
|
|
|
|
CMP R1, R3 // unequal lengths are not equal
|
|
|
|
|
BNE notequal
|
|
|
|
|
MOVD a+0(FP), R0
|
|
|
|
|
MOVD b+24(FP), R2
|
|
|
|
|
ADD R0, R1 // end
|
|
|
|
|
loop:
|
|
|
|
|
CMP R0, R1
|
|
|
|
|
BEQ equal // reaches the end
|
|
|
|
|
MOVBU.P 1(R0), R4
|
|
|
|
|
MOVBU.P 1(R2), R5
|
|
|
|
|
CMP R4, R5
|
|
|
|
|
BEQ loop
|
|
|
|
|
notequal:
|
|
|
|
|
MOVB ZR, ret+48(FP)
|
|
|
|
|
RET
|
|
|
|
|
equal:
|
|
|
|
|
MOVD $1, R0
|
|
|
|
|
MOVB R0, ret+48(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·fastrand1(SB),NOSPLIT,$-8-4
|
|
|
|
|
MOVD g_m(g), R1
|
|
|
|
|
MOVWU m_fastrand(R1), R0
|
|
|
|
|
ADD R0, R0
|
|
|
|
|
CMPW $0, R0
|
|
|
|
|
BGE notneg
|
|
|
|
|
EOR $0x88888eef, R0
|
|
|
|
|
notneg:
|
|
|
|
|
MOVW R0, m_fastrand(R1)
|
|
|
|
|
MOVW R0, ret+0(FP)
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·return0(SB), NOSPLIT, $0
|
|
|
|
|
MOVW $0, R0
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
// The top-most function running on a goroutine
|
|
|
|
|
// returns to goexit+PCQuantum.
|
|
|
|
|
TEXT runtime·goexit(SB),NOSPLIT,$-8-0
|
|
|
|
|
MOVD R0, R0 // NOP
|
|
|
|
|
BL runtime·goexit1(SB) // does not return
|
|
|
|
|
|
|
|
|
|
// TODO(aram): use PRFM here.
|
|
|
|
|
TEXT runtime·prefetcht0(SB),NOSPLIT,$0-8
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·prefetcht1(SB),NOSPLIT,$0-8
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·prefetcht2(SB),NOSPLIT,$0-8
|
|
|
|
|
RET
|
|
|
|
|
|
|
|
|
|
TEXT runtime·prefetchnta(SB),NOSPLIT,$0-8
|
|
|
|
|
RET
|
|
|
|
|
|
