1. In functions with heap-allocated result variables or with
defer statements, the return sequence requires more than
just a single RET instruction. There is an optimization that
arranges for all returns to jump to a single copy of the return
epilogue in this case. Unfortunately, that optimization is
fundamentally incompatible with PC-based liveness information:
it takes PCs at many different points in the function and makes
them all land at one PC, making the combined liveness information
at that target PC a mess. Disable this optimization, so that each
return site gets its own copy of the 'call deferreturn' and the
copying of result variables back from the heap.
This removes quite a few spurious 'ambiguously live' variables.
2. Let orderexpr allocate temporaries that are passed by address
to a function call and then die on return, so that we can arrange
an appropriate VARKILL.
2a. Do this for ... slices.
2b. Do this for closure structs.
2c. Do this for runtime.concatstring, which is the implementation
of large string additions. Change representation of OADDSTR to
an explicit list in typecheck to avoid reconstructing list in both
walk and order.
3. Let orderexpr allocate the temporary variable copies used for
range loops, so that they can be killed when the loop is over.
Similarly, let it allocate the temporary holding the map iterator.
CL 81940043 reduced the number of ambiguously live temps
in the godoc binary from 860 to 711.
This CL reduces the number to 121. Still more to do, but another
good checkpoint.
Update #7345
LGTM=khr
R=khr
CC=golang-codereviews
https://golang.org/cl/83090046
REP MOVSQ and REP STOSQ have a really high startup overhead.
Use a Duff's device to do the repetition instead.
benchmark old ns/op new ns/op delta
BenchmarkClearFat32 7.20 1.60 -77.78%
BenchmarkCopyFat32 6.88 2.38 -65.41%
BenchmarkClearFat64 7.15 3.20 -55.24%
BenchmarkCopyFat64 6.88 3.44 -50.00%
BenchmarkClearFat128 9.53 5.34 -43.97%
BenchmarkCopyFat128 9.27 5.56 -40.02%
BenchmarkClearFat256 13.8 9.53 -30.94%
BenchmarkCopyFat256 13.5 10.3 -23.70%
BenchmarkClearFat512 22.3 18.0 -19.28%
BenchmarkCopyFat512 22.0 19.7 -10.45%
BenchmarkCopyFat1024 36.5 38.4 +5.21%
BenchmarkClearFat1024 35.1 35.0 -0.28%
TODO: use for stack frame zeroing
TODO: REP prefixes are still used for "reverse" copying when src/dst
regions overlap. Might be worth fixing.
LGTM=rsc
R=golang-codereviews, rsc
CC=golang-codereviews, r
https://golang.org/cl/81370046
The new channel and map runtime routines take pointers
to values, typically temporaries. Without help, the compiler
cannot tell when those temporaries stop being needed,
because it isn't sure what happened to the pointer.
Arrange to insert explicit VARKILL instructions for these
temporaries so that the liveness analysis can avoid seeing
them as "ambiguously live".
The change is made in order.c, which was already in charge of
introducing temporaries to preserve the order-of-evaluation
guarantees. Now its job has expanded to include introducing
temporaries as needed by runtime routines, and then also
inserting the VARKILL annotations for all these temporaries,
so that their lifetimes can be shortened.
In order to do its job for the map runtime routines, order.c arranges
that all map lookups or map assignments have the form:
x = m[k]
x, y = m[k]
m[k] = x
where x, y, and k are simple variables (often temporaries).
Likewise, receiving from a channel is now always:
x = <-c
In order to provide the map guarantee, order.c is responsible for
rewriting x op= y into x = x op y, so that m[k] += z becomes
t = m[k]
t2 = t + z
m[k] = t2
While here, fix a few bugs in order.c's traversal: it was failing to
walk into select and switch case bodies, so order of evaluation
guarantees were not preserved in those situations.
Added tests to test/reorder2.go.
Fixes#7671.
In gc/popt's temporary-merging optimization, allow merging
of temporaries with their address taken as long as the liveness
ranges do not intersect. (There is a good chance of that now
that we have VARKILL annotations to limit the liveness range.)
Explicitly killing temporaries cuts the number of ambiguously
live temporaries that must be zeroed in the godoc binary from
860 to 711, or -17%. There is more work to be done, but this
is a good checkpoint.
Update #7345
LGTM=khr
R=khr
CC=golang-codereviews
https://golang.org/cl/81940043
e.g., don't delete /dev/null. this fix inspired by gnu libiberty,
unlink-if-ordinary.c.
Fixes#7563
LGTM=iant
R=golang-codereviews, iant, 0intro
CC=golang-codereviews, r
https://golang.org/cl/76810045
chanrecv now expects a pointer to the data to be filled in.
mapiterinit expects a pointer to the hash iterator to be filled in.
In both cases, the temporary being pointed at changes from
dead to alive during the call. In order to make sure it is
preserved if a garbage collection happens after that transition
but before the call returns, the temp must be marked as live
during the entire call.
But if it is live during the entire call, it needs to be safe for
the garbage collector to scan at the beginning of the call,
before the new data has been filled in. Therefore, it must be
zeroed by the caller, before the call. Do that.
My previous attempt waited to mark it live until after the
call returned, but that's unsafe (see first paragraph);
undo that change in plive.c.
This makes powser2 pass again reliably.
I looked at every call to temp in the compiler.
The vast majority are followed immediately by an
initialization of temp, so those are fine.
The only ones that needed changing were the ones
where the next operation is to pass the address of
the temp to a function call, and there aren't too many.
Maps are exempted from this because mapaccess
returns a pointer to the data and lets the caller make
the copy.
Fixes many builds.
TBR=khr
CC=golang-codereviews
https://golang.org/cl/80700046
cgo represents all 0-sized and unsized types internally as [0]byte. This means that pointers to incomplete types would be interchangable, even if given a name by typedef.
Fixes#7409.
LGTM=iant
R=golang-codereviews, bradfitz, iant
CC=golang-codereviews
https://golang.org/cl/76450043
1. On entry to a function, only zero the ambiguously live stack variables.
Before, we were zeroing all stack variables containing pointers.
The zeroing is pretty inefficient right now (issue 7624), but there are also
too many stack variables detected as ambiguously live (issue 7345),
and that must be addressed before deciding how to improve the zeroing code.
(Changes in 5g/ggen.c, 6g/ggen.c, 8g/ggen.c, gc/pgen.c)
Fixes#7647.
2. Make the regopt word-based liveness analysis preserve the
whole-variable liveness property expected by the garbage collection
bitmap liveness analysis. That is, if the regopt liveness decides that
one word in a struct needs to be preserved, make sure it preserves
the entire struct. This is particularly important for multiword values
such as strings, slices, and interfaces, in which all the words need
to be present in order to understand the meaning.
(Changes in 5g/reg.c, 6g/reg.c, 8g/reg.c.)
Fixes#7591.
3. Make the regopt word-based liveness analysis treat a variable
as having its address taken - which makes it preserved across
all future calls - whenever n->addrtaken is set, for consistency
with the gc bitmap liveness analysis, even if there is no machine
instruction actually taking the address. In this case n->addrtaken
is incorrect (a nicer way to put it is overconservative), and ideally
there would be no such cases, but they can happen and the two
analyses need to agree.
(Changes in 5g/reg.c, 6g/reg.c, 8g/reg.c; test in bug484.go.)
Fixes crashes found by turning off "zero everything" in step 1.
4. Remove spurious VARDEF annotations. As the comment in
gc/pgen.c explains, the VARDEF must immediately precede
the initialization. It cannot be too early, and it cannot be too late.
In particular, if a function call sits between the VARDEF and the
actual machine instructions doing the initialization, the variable
will be treated as live during that function call even though it is
uninitialized, leading to problems.
(Changes in gc/gen.c; test in live.go.)
Fixes crashes found by turning off "zero everything" in step 1.
5. Do not treat loading the address of a wide value as a signal
that the value must be initialized. Instead depend on the existence
of a VARDEF or the first actual read/write of a word in the value.
If the load is in order to pass the address to a function that does
the actual initialization, treating the load as an implicit VARDEF
causes the same problems as described in step 4.
The alternative is to arrange to zero every such value before
passing it to the real initialization function, but this is a much
easier and more efficient change.
(Changes in gc/plive.c.)
Fixes crashes found by turning off "zero everything" in step 1.
6. Treat wide input parameters with their address taken as
initialized on entry to the function. Otherwise they look
"ambiguously live" and we will try to emit code to zero them.
(Changes in gc/plive.c.)
Fixes crashes found by turning off "zero everything" in step 1.
7. An array of length 0 has no pointers, even if the element type does.
Without this change, the zeroing code complains when asked to
clear a 0-length array.
(Changes in gc/reflect.c.)
LGTM=khr
R=khr
CC=golang-codereviews
https://golang.org/cl/80160044
Zeroing the outputs makes sure that during function calls
in those functions we do not let the garbage collector
treat uninitialized values as pointers.
The garbage collector may still see uninitialized values
if a preemption occurs during the function prologue,
before the zeroing has had a chance to run.
This reduces the number of 'bad pointer' messages when
that runtime check is enabled, but it doesn't fix all of them,
so the check is still disabled.
It will also avoid leaks, although I doubt any of these were
particularly serious.
LGTM=iant, khr
R=iant, khr
CC=golang-codereviews
https://golang.org/cl/80850044
If you compile a program that has cgo LDFLAGS directives, those are exported to an environment variable to be used by subsequent compiler tool invocations. The linking phase when using the gccgo toolchain did not consider the envvar CGO_LDFLAGS's linking directives resulting in undefined references when using cgo+gccgo.
Fixes#7573
LGTM=iant
R=golang-codereviews, iant
CC=golang-codereviews
https://golang.org/cl/80780043
GCC on OS X 10.6 doesn't support -Wuninitialized without -O.
Fixes#7492.
LGTM=iant
R=golang-codereviews, dave, iant
CC=golang-codereviews
https://golang.org/cl/72360045
Fixes#7627.
CL 61970044 changed the order in which .a files are passed to gccgo's link phase. However by reversing the order it caused gccgo to complain if both internal (liba.a) and external (liba_test.a) versions of a package were presented as the former would not contain all the necessary symbols, and the latter would duplicate symbols already defined.
This change ensures that all 'fake' targets remain at the top of the final link order which should be fine as a package compiled as an external test is a superset of its internal sibling.
Looking at how gcToolchain links tests I think this change now accurately mirrors those actions which present $WORK/_test before $WORK in the link order.
LGTM=iant
R=rsc, iant
CC=golang-codereviews, michael.hudson
https://golang.org/cl/80300043
Change two-bit stack map entries to encode:
0 = dead
1 = scalar
2 = pointer
3 = multiword
If multiword, the two-bit entry for the following word encodes:
0 = string
1 = slice
2 = iface
3 = eface
That way, during stack scanning we can check if a string
is zero length or a slice has zero capacity. We can avoid
following the contained pointer in those cases. It is safe
to do so because it can never be dereferenced, and it is
desirable to do so because it may cause false retention
of the following block in memory.
Slice feature turned off until issue 7564 is fixed.
Update #7549
LGTM=rsc
R=golang-codereviews, bradfitz, rsc
CC=golang-codereviews
https://golang.org/cl/76380043
Revision 3ae4607a43ff introduced CONVNOP layers
to fix type checking issues arising from comparisons.
The added complexity made 8g run out of registers
when compiling an equality function in go.net/ipv6.
A similar issue occurred in test/sizeof.go on
amd64p32 with 6g.
Fixes#7405.
LGTM=khr
R=rsc, dave, iant, khr
CC=golang-codereviews
https://golang.org/cl/78100044
Removes most uses of the REP prefix, which has a high startup cost.
LGTM=iant
R=golang-codereviews, iant, khr
CC=golang-codereviews
https://golang.org/cl/77920043
A too large float constant is an error.
A too small float constant is rounded to zero.
Fixes#7419
Update #6902
LGTM=iant
R=golang-codereviews, iant
CC=golang-codereviews
https://golang.org/cl/76730046
Old versions of DTrace (as those shipped in OS X and FreeBSD)
don't support unicode characters in symbol names. Replace '·'
to '.' to make DTrace happy.
Fixes#7493
LGTM=aram, rsc
R=aram, rsc, gobot, iant
CC=golang-codereviews
https://golang.org/cl/72280043
The lowering to runtime calls introduces hidden pointers to the
arguments of select clauses. When implicit conversions were
involved it could end up with incompatible pointers. Since the
pointed-to types have the same representation, we can introduce a
forced conversion.
Fixes#6847.
LGTM=rsc
R=rsc, iant, khr
CC=golang-codereviews
https://golang.org/cl/72380043
It was using a REP STOSQ but putting in CX the number of 32-bit
words to clear.
LGTM=dave
R=rsc, dave
CC=golang-codereviews
https://golang.org/cl/75240043
Under some circumstances, gccgoToolchain's ld can pass the path of
build outputs that have been deleted to the link command.
Fixes#7303.
LGTM=rsc
R=golang-codereviews, dave, michael.hudson, rsc
CC=golang-codereviews
https://golang.org/cl/61970044
The compilers expect to not be interrupted by floating
point exceptions. On Plan 9, every process starts with
interrupts enabled for invalid operation, stack overflow,
and divide by zero exceptions.
LGTM=rsc
R=rsc, 0intro
CC=golang-codereviews
https://golang.org/cl/72750043
Acid can't produce a stack trace without .frame symbols.
Of course, it can only unwind through linear stacks but
this is still better than nothing. (I wrote an acid func
to do the full unwind a long time ago but lost it and
haven't worked up the courage to write it again).
Note that these will only be present in the native symbol
table for Plan 9 binaries.
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/72450045
debuglive >= 1 is not the condition under which we
start recording messages (we avoid printing for
init functions even if debuglive is set).
LGTM=bradfitz, iant
R=golang-codereviews, bradfitz
CC=golang-codereviews, iant, khr
https://golang.org/cl/74390043
This fixes the following amd64p32 issue:
pkg/time/format.go:724: internal compiler error: twobitwalktype1: invalid initial alignment, Time
caused by the pointer zone ending on a 32-bit-aligned boundary.
LGTM=rsc
R=rsc, dave
CC=golang-codereviews
https://golang.org/cl/72270046
Replaces CL 70000043.
Switch to the amd64p32 linker model if we are building under nacl/amd64p32.
No need to introduce linkarchinit() as 6a contains its own main() function.
LGTM=rsc
R=rsc, minux.ma
CC=golang-codereviews
https://golang.org/cl/72020043
Replaces CL 70000043.
Introduce linkarchinit() from cmd/ld.
For cmd/6g, switch to the amd64p32 linker model if we are building under nacl/amd64p32.
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/71330045
This is a test case for CL 34680044.
Fixes#6333.
LGTM=bradfitz
R=golang-codereviews, bradfitz, minux.ma
CC=golang-codereviews
https://golang.org/cl/71230049
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
Tested GOARM=6 on Raspberry Pi, and I found only a few tests that
use sub-normal numbers fails. I have a patch to NetBSD kernel pending
that fixes this issue (NetBSD kernel doesn't allow us to disable the
Flush-to-Zero feature).
LGTM=jsing
R=golang-codereviews, jsing
CC=golang-codereviews
https://golang.org/cl/70730043
Essentialy for running tests without a working cmd/go.
While we're at it, also fix a typo.
LGTM=rsc
R=golang-codereviews, rsc
CC=golang-codereviews
https://golang.org/cl/70640043
