This allows implementing address-of-global
as a pc-relative address instead of as a
32-bit integer constant.
LGTM=rminnich, iant
R=golang-codereviews, rminnich, iant
CC=golang-codereviews
https://golang.org/cl/128070045
The name linkwriteobj is misleading because it implies that
the function has something to do with the linker, which it
does not. The name is historical: the function performs an
operation that was previously performed by the linker, but no
longer is.
LGTM=rsc
R=rsc, minux.ma
CC=golang-codereviews
https://golang.org/cl/88210045
When I did the original 386 ports on Linux and OS X, I chose to
define GS-relative expressions like 4(GS) as relative to the actual
thread-local storage base, which was usually GS but might not be
(it might be FS, or it might be a different constant offset from GS or FS).
The original scope was limited but since then the rewrites have
gotten out of control. Sometimes GS is rewritten, sometimes FS.
Some ports do other rewrites to enable shared libraries and
other linking. At no point in the code is it clear whether you are
looking at the real GS/FS or some synthesized thing that will be
rewritten. The code manipulating all these is duplicated in many
places.
The first step to fixing issue 7719 is to make the code intelligible
again.
This CL adds an explicit TLS pseudo-register to the 386 and amd64.
As a register, TLS refers to the thread-local storage base, and it
can only be loaded into another register:
MOVQ TLS, AX
An offset from the thread-local storage base is written off(reg)(TLS*1).
Semantically it is off(reg), but the (TLS*1) annotation marks this as
indexing from the loaded TLS base. This emits a relocation so that
if the linker needs to adjust the offset, it can. For example:
MOVQ TLS, AX
MOVQ 8(AX)(TLS*1), CX // load m into CX
On systems that support direct access to the TLS memory, this
pair of instructions can be reduced to a direct TLS memory reference:
MOVQ 8(TLS), CX // load m into CX
The 2-instruction and 1-instruction forms correspond roughly to
ELF TLS initial exec mode and ELF TLS local exec mode, respectively.
Liblink applies this rewrite on systems that support the 1-instruction form.
The decision is made using only the operating system (and probably
the -shared flag, eventually), not the link mode. If some link modes
on a particular operating system require the 2-instruction form,
then all builds for that operating system will use the 2-instruction
form, so that the link mode decision can be delayed to link time.
Obviously it is late to be making changes like this, but I despair
of correcting issue 7719 and issue 7164 without it. To make sure
I am not changing existing behavior, I built a "hello world" program
for every GOOS/GOARCH combination we have and then worked
to make sure that the rewrite generates exactly the same binaries,
byte for byte. There are a handful of TODOs in the code marking
kludges to get the byte-for-byte property, but at least now I can
explain exactly how each binary is handled.
The targets I tested this way are:
darwin-386
darwin-amd64
dragonfly-386
dragonfly-amd64
freebsd-386
freebsd-amd64
freebsd-arm
linux-386
linux-amd64
linux-arm
nacl-386
nacl-amd64p32
netbsd-386
netbsd-amd64
openbsd-386
openbsd-amd64
plan9-386
plan9-amd64
solaris-amd64
windows-386
windows-amd64
There were four exceptions to the byte-for-byte goal:
windows-386 and windows-amd64 have a time stamp
at bytes 137 and 138 of the header.
darwin-386 and plan9-386 have five or six modified
bytes in the middle of the Go symbol table, caused by
editing comments in runtime/sys_{darwin,plan9}_386.s.
Fixes#7164.
LGTM=iant
R=iant, aram, minux.ma, dave
CC=golang-codereviews
https://golang.org/cl/87920043
Replaces CL 70000043.
Introduce linkarchinit() from cmd/ld.
For cmd/6c, switch to the amd64p32 linker model if we are building under nacl/amd64p32.
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/72010043
From the original description in CL 15770043
The main change here is to consult $GOARCH.
In 6c, when GOOS=nacl, some of the more complex addressing modes must be disabled, and the BP and R15 registers must not be used.
See golang.org/s/go13nacl for design overview.
LGTM=rsc
R=golang-codereviews, gobot, rsc
CC=golang-codereviews
https://golang.org/cl/69020044
We now use the %A, %D, %P, and %R routines from liblink
across the board.
Fixes#7178.
Fixes#7055.
LGTM=iant
R=golang-codereviews, gobot, rsc, dave, iant, remyoudompheng
CC=golang-codereviews
https://golang.org/cl/49170043
- add buffered stdout to all tools and provide to link ctxt.
- avoid extra \n before ! in .6 files written by assemblers
(makes them match the C compilers).
- use linkwriteobj instead of linkouthist+linkwritefuncs.
- in assemblers and C compilers, record pc explicitly in Prog,
for use by liblink.
- in C compilers, preserve jump target links.
- in Go compilers (gsubr.c) attach gotype directly to
corresponding LSym* instead of rederiving from instruction stream.
- in Go compilers, emit just one definition for runtime.zerovalue
from each compilation.
This CL consists entirely of small adjustments.
The heavy lifting is in CL 39680043.
Each depends on the other.
R=golang-dev, dave, iant
CC=golang-dev
https://golang.org/cl/37030045
Preparation for golang.org/s/go13linker work.
This CL does not build by itself. It depends on 35740044
and 35790044 and will be submitted at the same time.
R=iant
CC=golang-dev
https://golang.org/cl/34580044
Also introduce BGET2/4, BPUT2/4 as they are widely used.
Slightly improve BGETC/BPUTC implementation.
This gives ~5% CPU time improvement on go install -a -p1 std.
Before:
real user sys
0m23.561s 0m16.625s 0m5.848s
0m23.766s 0m16.624s 0m5.846s
0m23.742s 0m16.621s 0m5.868s
after:
0m22.999s 0m15.841s 0m5.889s
0m22.845s 0m15.808s 0m5.850s
0m22.889s 0m15.832s 0m5.848s
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12745047
This CL introduces a FUNCDATA number for runtime-specific
garbage collection metadata, changes the C and Go compilers
to emit that metadata, and changes the runtime to expect it.
The old pseudo-instructions that carried this information
are gone, as is the linker code to process them.
R=golang-dev, dvyukov, cshapiro
CC=golang-dev
https://golang.org/cl/11406044
whose argument size is unknown (C vararg functions, and
assembly code without an explicit specification).
We used to use 0 to mean "unknown" and 1 to mean "zero".
Now we use ArgsSizeUnknown (0x80000000) to mean "unknown".
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/11590043
Fixes#5085.
{6,8}c/swt.c allocates a third Biobuf in automatic memory which is not terminated at the end of the function. This causes the buffer to be 'in use' when the batexit handler fires, confusing valgrind.
Huge thanks to DMorsing for the diagnosis.
R=golang-dev, daniel.morsing, rsc
CC=golang-dev
https://golang.org/cl/7844044
Switch to new pragma names, but leave old ones available for now.
Merge the three cgo-related sections in the .6 files into a single
cgo section.
R=golang-dev, iant, ality
CC=golang-dev
https://golang.org/cl/7424048
Also:
- faster code for example extraction
- simplify handling of command documentation:
all "main" packages are treated as commands
- various minor cleanups along the way
For commands written in Go, any doc.go file containing
documentation must now be part of package main (rather
then package documentation), otherwise the documentation
won't show up in godoc (it will still build, though).
For commands written in C, documentation may still be
in doc.go files defining package documentation, but the
recommended way is to explicitly ignore those files with
a +build ignore constraint to define package main.
Fixes#4806.
R=adg, rsc, dave, bradfitz
CC=golang-dev
https://golang.org/cl/7333046
The routine that adds an automatic to the stack was
adding ptrsize-1 to the size before rounding up.
That addition would only make sense to turn a round down
into a round up. Before a round up, it just wastes a word.
The effect was that a 6c function with one local and
one two-word function call used (8+8)+(16+8) = 40 bytes
instead of 8+16 = 24 bytes.
The wasted space mostly didn't matter, but one place where
it does matter is when trying to stay within the 128-byte
total frame constraint for #pragma textflag 7 functions.
This only affects the C compilers, not the Go compilers.
5c already had correct code, which is now copied to 6c and 8c.
R=ken2
CC=golang-dev
https://golang.org/cl/7303099
The FmtLong flag should only be used with the %D verb
when printing an ATEXT Prog. It was erroneously used
for every Prog except ADATA. This caused a preponderance
of exclamation points, "!!", in the assembly listings.
I also cleaned up the code so that the list.c files look
very similar. Now the real differences are easily spotted
with a simple diff.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7128045
For simplicity, only recognizes expressions of the exact form
"(x << a) | (x >> b)" where x is a variable and a and b are
integer constant expressions that add to x's bit width.
Fixes#4629.
$ cat rotate.c
unsigned int
rotate(unsigned int x)
{
x = (x << 3) | (x >> (sizeof(x) * 8 - 3));
return x;
}
## BEFORE
$ go tool 6c -S rotate.c
(rotate.c:2) TEXT rotate+0(SB),$0-8
(rotate.c:2) MOVL x+0(FP),!!DX
(rotate.c:4) MOVL DX,!!AX
(rotate.c:4) SALL $3,!!AX
(rotate.c:4) MOVL DX,!!CX
(rotate.c:4) SHRL $29,!!CX
(rotate.c:4) ORL CX,!!AX
(rotate.c:5) RET ,!!
(rotate.c:5) RET ,!!
(rotate.c:5) END ,!!
## AFTER
$ go tool 6c -S rotate.c
(rotate.c:2) TEXT rotate+0(SB),$0-8
(rotate.c:4) MOVL x+0(FP),!!AX
(rotate.c:4) ROLL $3,!!AX
(rotate.c:5) RET ,!!
(rotate.c:5) RET ,!!
(rotate.c:5) END ,!!
R=rsc, minux.ma
CC=golang-dev
https://golang.org/cl/7069056
Reference the 80386 compiler documentation now that the
documentation for the 68020 is offline.
R=golang-dev, minux.ma, rsc
CC=golang-dev
https://golang.org/cl/7127053
The fixjmp step eliminates redundant chains of JMP
instructions that are produced by the compiler during
code generation.
It is already implemented in gc, and can be adapted to 6c/8c with
the exception that JMPs refer to destination by pc instead of by
pointer. The algorithm is modified to operate on Regs instead of Progs
for this reason. The pcs are already restored later by regopt.
R=goalng-dev, rsc
CC=golang-dev
https://golang.org/cl/6865046
On ARM, char is unsigned, and the code generation for
multiplication gets totally broken.
Fixes#4354.
R=golang-dev, dave, minux.ma, rsc
CC=golang-dev
https://golang.org/cl/6826079
Some newer Linux distributions (Ubuntu ARM at least) use a new multiarch
directory organization, where dynamic linker is no longer in the hardcoded
path in our linker.
For example, Ubuntu 12.04 ARM hardfloat places its dynamic linker at
/lib/arm-linux-gnueabihf/ld-linux.so.3
Ref: http://lackof.org/taggart/hacking/multiarch/
Also, to support Debian GNU/kFreeBSD as a FreeBSD variant, we need this capability, so it's part of issue 3533.
This CL add a new pragma (#pragma dynlinker "path") to cc.
R=iant, rsc
CC=golang-dev
https://golang.org/cl/6086043
This makes it possible to inline the prefetch of upcoming
memory addresses during garbage collection, instead of
needing to flush registers, make a function call, and
reload registers. On garbage collection-heavy workloads,
this results in a 5% speedup.
Fixes#3493.
R=dvyukov, ken, r, dave
CC=golang-dev
https://golang.org/cl/5990066
Without an explicit signal for a truncation, copy propagation
will sometimes propagate a 32-bit truncation and end up
overwriting uses of the original 64-bit value.
The case that arose in practice is in C but I believe
that the same could plausibly happen in Go.
The main reason we didn't run into the same in Go
is that I (perhaps incorrectly?) drop MOVL AX, AX
during gins, so the truncation was never generated, so
it didn't confuse the optimizer.
Fixes#1315.
Fixes#3488.
R=ken2
CC=golang-dev
https://golang.org/cl/6002043
cc: add #pragma textflag to set it
runtime: mark mheap to go into noptr-bss.
remove special case in garbage collector
Remove the ARM from.flag field created by CL 5687044.
The DUPOK flag was already in p->reg, so keep using that.
Otherwise test/nilptr.go creates a very large binary.
Should fix the arm build.
Diagnosed by minux.ma; replacement for CL 5690044.
R=golang-dev, minux.ma, r
CC=golang-dev
https://golang.org/cl/5686060
As a convenience to people working on the tools,
leave Makefiles that invoke the go dist tool appropriately.
They are not used during the build.
R=golang-dev, bradfitz, n13m3y3r, gustavo
CC=golang-dev
https://golang.org/cl/5636050
This can happen on Plan 9 if we we're building
with the 32-bit and 64-bit host compilers, one
after the other.
R=rsc
CC=golang-dev
https://golang.org/cl/5599053
Also delete gotest, since it's messy to fix and slated for deletion anyway.
A couple of things outside src can't be tested any more. "go test" will be
fixed and these tests will be re-enabled. They're noisy for now.
Fixes#284.
R=rsc
CC=golang-dev
https://golang.org/cl/5598049
Change the name of cas() in cc to newcase() to avoid a NIX conflict.
cas() is used in cc to create a new Case struct. There is a name
conflict in that cas() is a commonly-used
name for compare and swap. Since cas() is only used internally
in the compiler in 3 places, change the name to avoid a wider
conflict with the NIX runtime. This issue might well come up on
other OSes in the future anyway, as the name is fairly common.
R=rsc
CC=golang-dev
https://golang.org/cl/5294071
src/cmd/5c/reg.c:
. Added USED() attribute.
src/cmd/6c/cgen.c:
. Revised code around "REGARG" to resemble use in "8c" and
consequently remove a warning.
src/cmd/6l/asm.c:
. Added USED() attributes.
. Removed an unnecessary assignment.
R=golang-dev
CC=golang-dev, rsc
https://golang.org/cl/4836045