Shenghou Ma [Fri, 26 Dec 2014 06:22:41 +0000 (01:22 -0500)]
cmd/dist, make.bash: darwin/arm support
cmd/dist: recognize darwin/arm as (host) goos/goarches. also hard
code GOARM=7 for darwin/arm.
make.bash: don't pass -mmacosx-version-min=10.6 when building for
darwin/arm.
Change-Id: If0ecd84a5179cd9bb61b801ac1899adc45f12f75
Reviewed-on: https://go-review.googlesource.com/2126 Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: David Crawshaw <crawshaw@golang.org>
Shenghou Ma [Fri, 26 Dec 2014 06:20:29 +0000 (01:20 -0500)]
misc/cgo/test: skip test7978 when using unsupported compilers
On Darwin/ARM, because libSystem doesn't provide functions for
__sync_fetch_and_add, and only clang can inline that function,
skip the test when building with GCC.
Change-Id: Id5e9d8f9bbe1e6bcb2f381f0f66cf68aa95277c7
Reviewed-on: https://go-review.googlesource.com/2125 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Shenghou Ma [Fri, 26 Dec 2014 06:09:07 +0000 (01:09 -0500)]
syscall: darwin/arm support
Change-Id: Id6f7fa12084204bc3a200f423c7966ce2a0b63a0
Reviewed-on: https://go-review.googlesource.com/2123 Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: David Crawshaw <crawshaw@golang.org>
Shenghou Ma [Fri, 26 Dec 2014 05:15:07 +0000 (00:15 -0500)]
liblink, cmd/ld, cmd/5l: darwin/arm support
liblink:
- set dummy value for ctxt->tlsoffset.
cmd/ld:
- always do external linking when using cgo on darwin/arm,
as our linker might not generate codesign-compatible binary.
cmd/5l:
- support generate ARM Mach-O binaries
- add machoreloc1() that translate our internal relocation to
macho relocations used by external linking.
Change-Id: Ic5454aeb87009aaf8f1453ec7fe33e6da55d5f06
Reviewed-on: https://go-review.googlesource.com/3273 Reviewed-by: David Crawshaw <crawshaw@golang.org>
Shenghou Ma [Fri, 26 Dec 2014 05:48:33 +0000 (00:48 -0500)]
liblink, cmd/ld, runtime: set the type of runtime.tlsg in runtime
In the old code, liblink, cmd/ld and runtime all have code determine
whether runtime.tlsg is an actual variable or a placeholder for TLS
relocation. This change consolidate them into one: the runtime/tls_arm.s
will ultimately determine the type of that variable.
Change-Id: I3b3f80791a1db4c2b7318f81a115972cd2237e43
Reviewed-on: https://go-review.googlesource.com/2118 Reviewed-by: David Crawshaw <crawshaw@golang.org> Reviewed-by: Dave Cheney <dave@cheney.net>
Shenghou Ma [Thu, 5 Feb 2015 19:16:13 +0000 (14:16 -0500)]
cmd/cgo: detect misuse of generated _cgo_export.h
Fixes #9742.
Change-Id: Ifedf7ff9465bc49534b708d414c8e435ee9ce6cd
Reviewed-on: https://go-review.googlesource.com/3970 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Rahul Chaudhry [Tue, 3 Feb 2015 18:52:18 +0000 (10:52 -0800)]
test: check for build constraints only upto the first blank line
The main issue is that the misc/cgo/{stdio,life} tests are silently
getting skipped when invoked from run.bash.
run.go should ignore any build tags after the first blank line in
source file. It already checks for test actions only upto the first
blank line. Build tags must be specified in the same block.
See http://golang.org/cl/3675 for background.
Change-Id: Id8abf000119e3335f7250d8ef34aac7811fc9dff
Reviewed-on: https://go-review.googlesource.com/3812 Reviewed-by: Minux Ma <minux@golang.org>
Rahul Chaudhry [Tue, 3 Feb 2015 20:41:52 +0000 (12:41 -0800)]
test: chdir before running go tool, cleanup afterwards.
issue9355 generated a file a.[568] in test/ directory and left it there.
For tests like these, it is best to chdir to a test specific directory
before generating any temporary files, since the tests are running
in parallel and might otherwise race with each other for the same files.
Change-Id: I58d96256d4d8ee3fda70d81077f19006064a7425
Reviewed-on: https://go-review.googlesource.com/3813 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Evan Phoenix [Fri, 30 Jan 2015 23:31:47 +0000 (15:31 -0800)]
expvar: Add benchmarks for perf sensitive operations
These benchmarks are only for functions commonly used in loops. The
other functions are typically used for inspection or setup and thus are
not performance sensitive.
Shenghou Ma [Mon, 2 Feb 2015 23:53:24 +0000 (18:53 -0500)]
cmd/ld: always set timestamp in PE header to 0
Fixes #9756.
Change-Id: If4ee6fe10f8f90294ff9c5e7480371494094b111 Signed-off-by: Shenghou Ma <minux@golang.org>
Reviewed-on: https://go-review.googlesource.com/3740 Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Rob Pike [Wed, 4 Feb 2015 21:50:08 +0000 (13:50 -0800)]
cmd/pack: simplify the testing logic slightly
Followup to https://go-review.googlesource.com/3910
We only need 1000 iteratinons.
Change-Id: Ib63ae53105176abec77bad9609d638aeda7bcd61
Reviewed-on: https://go-review.googlesource.com/3901 Reviewed-by: Minux Ma <minux@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Hyang-Ah Hana Kim [Sat, 24 Jan 2015 22:51:42 +0000 (17:51 -0500)]
runtime: support panic/print logging in android-L.
In android-L, logging is done through the logd daemon.
If logd daemon is available, send logging to logd.
Otherwise, fallback to the legacy mechanism (/dev/log files).
This change adds access/socket/connect calls to interact with the logd.
Fixes golang/go#9398.
Change-Id: I3c52b81b451f5862107d7c675f799fc85548486d
Reviewed-on: https://go-review.googlesource.com/3350 Reviewed-by: David Crawshaw <crawshaw@golang.org>
Russ Cox [Fri, 30 Jan 2015 04:35:14 +0000 (23:35 -0500)]
liblink: require use of TYPE_ADDR, not TYPE_CONST
Add Addr-checking for all Progs on input to liblink, in liblink/pass.c,
including requiring use of TYPE_ADDR, not TYPE_CONST.
Update compilers and assemblers to satisfy checks.
Chris Kastorff [Wed, 4 Feb 2015 12:43:00 +0000 (04:43 -0800)]
testing/quick: support generation of array types in Value
Generating array types like [4]int would fail even though the int type
is generatable. Allow generating values of array types when the inner
type is generatable.
Change-Id: I8d1bc0d2e471cd9357274204c9bc1fa67cbc272d
Reviewed-on: https://go-review.googlesource.com/3833 Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Andrew Gerrand <adg@golang.org>
Mikio Hara [Wed, 4 Feb 2015 02:27:08 +0000 (11:27 +0900)]
net: update TDDO
The issue #8432 has been marked as an issue for golang.org/x/net.
Change-Id: Ia39abd99b685c820ea6169ee6505b16028e7e77f
Reviewed-on: https://go-review.googlesource.com/3836 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Dmitry Vyukov [Tue, 3 Feb 2015 05:58:04 +0000 (08:58 +0300)]
runtime: bound defer pools
The unbounded list-based defer pool can grow infinitely.
This can happen if a goroutine routinely allocates a defer;
then blocks on one P; and then unblocked, scheduled and
frees the defer on another P.
The scenario was reported on golang-nuts list.
We've been here several times. Any unbounded local caches
are bad and grow to infinite size. This change introduces
central defer pool; local pools become fixed-size
with the only purpose of amortizing accesses to the
central pool.
Change-Id: Iadcfb113ccecf912e1b64afc07926f0de9de2248
Reviewed-on: https://go-review.googlesource.com/3741 Reviewed-by: Keith Randall <khr@golang.org>
Austin Clements [Wed, 4 Feb 2015 02:14:47 +0000 (21:14 -0500)]
runtime: fix RuntimeGogoBytes on windows/amd64
Before 3c0fee1, runtime.gogo was just long enough to align to 64 bytes
on OSs with short get_tls implementations and 80 bytes on OSs with
longer get_tls implementations (Windows, Solaris, and Plan 9). 3c0fee1 added a few instructions, which pushed it to 80 on most OSs,
including Windows and Plan 9, and 96 on Solaris.
Fixes #9770.
Change-Id: Ie84810657c14ab16dce9f0e0a932955251b0bf33
Reviewed-on: https://go-review.googlesource.com/3850 Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Jacob H. Haven [Wed, 4 Feb 2015 00:15:18 +0000 (16:15 -0800)]
crypto/tls: add support for AES_256_GCM_SHA384 cipher suites specified in RFC5289
Generalizes PRF calculation for TLS 1.2 to support arbitrary hashes (SHA-384 instead of SHA-256).
Testdata were all updated to correspond with the new cipher suites in the handshake.
Change-Id: I3d9fc48c19d1043899e38255a53c80dc952ee08f
Reviewed-on: https://go-review.googlesource.com/3265 Reviewed-by: Adam Langley <agl@golang.org>
Lynn Boger [Tue, 3 Feb 2015 18:13:31 +0000 (12:13 -0600)]
runtime: Change memprofrate to memprofilerate
Use memprofilerate in GODEBUG instead of memprofrate to be
consistent with other uses.
Change-Id: Iaf6bd3b378b1fc45d36ecde32f3ad4e63ca1e86b
Reviewed-on: https://go-review.googlesource.com/3800 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Russ Cox [Thu, 29 Jan 2015 21:16:24 +0000 (16:16 -0500)]
liblink: the zero Prog is now valid and ready for use
Use AXXX instead of AGOK (neither is a valid instruction but AXXX is zero)
for the initial setting of Prog.as, and now there are no non-zero default
field settings.
Remove the arch-specific zprog/zprg in favor of a single global zprog.
Remove the arch-specific prg constructor in favor of emallocz(sizeof(Prog)).
Russ Cox [Thu, 29 Jan 2015 20:35:56 +0000 (15:35 -0500)]
liblink: remove dead computation of p->back in span6/span8
Originally, when this code was part of 6l/8l, every new
Prog was constructed starting with zprg, which set back=2,
and then this code walked over the list setting back=1 for
backward branches, back=0 otherwise. The initial back=2
setting was used to identify forward branches (the branched-to
instruction had back == 2 since it hadn't yet been set to 0 or 1).
When the code was extracted into liblink and linked directly
with 6a/6g/8a/8g, those programs created the Prog struct
and did not set back=2, breaking this backward branch detection.
No one noticed, because the next loop recomputes the information.
The only requirement for the next loop is that p->back == 0 or 1 for
each of the Progs in the list.
The initialization of the zprg with back=2 would cause problems
in this second loop, for the few liblink-internally-generated instructions
that are created by copying zprg, except that the first loop was
making sure that back == 0 or 1.
The first loop's manipulation of p->back can thus be deleted,
provided we also delete the zprg.back = 2 initializations.
This is awful and my fault. I apologize.
While we're here, remove the .scale = 1 from the zprg init too.
Anything that sets up a scaled index should set the scale itself.
(And mostly those come from outside liblink anyway.)
Tested by checking that all generated code is bit-for-bit
identical to before this CL.
Russ Cox [Thu, 29 Jan 2015 19:58:54 +0000 (14:58 -0500)]
liblink: place TEXT/GLOBL flags in p->from3 always
Before, amd64 and 386 stored the flags in p->from.scale
and arm and ppc64 stored the flags in p->reg.
Both caused special cases in printing and in handling of the
addresses.
To avoid possible conflicts with the real meaning of p->from
and to avoid storing a non-register value in a reg field,
use from3 to hold a TYPE_CONST value giving the flags.
There is still a special case for printing, because the flags
are specified without a $, and normally a TYPE_CONST prints
with a $. But that's much less special than what came before.
This allows us to remove the textflag and settextflag methods
from LinkArch. They are no longer architecture-specific.
Russ Cox [Thu, 29 Jan 2015 18:50:19 +0000 (13:50 -0500)]
cmd/9a: fix GLOBL instruction
Because it was lumped in with the TEXT instruction,
the high 32 bits of the 64-bit constant holding the size
were always set to 0x80000000 (ArgsSizeUnknown).
This only worked because cmd/9l was reading the 64-bit
value into an int32.
While we're here, fix 5a.
It wasn't as much of a problem there because
the two values were being stored in two different fields.
But it was still wrong.
Dmitry Vyukov [Thu, 29 Jan 2015 09:47:30 +0000 (12:47 +0300)]
runtime: fix nosplit stack overflow
The overflow happens only with -gcflags="-N -l"
and can be reproduced with:
$ go test -gcflags="-N -l" -a -run=none net
runtime.cgocall: nosplit stack overflow
504 assumed on entry to runtime.cgocall
480 after runtime.cgocall uses 24
472 on entry to runtime.cgocall_errno
408 after runtime.cgocall_errno uses 64
400 on entry to runtime.exitsyscall
288 after runtime.exitsyscall uses 112
280 on entry to runtime.exitsyscallfast
152 after runtime.exitsyscallfast uses 128
144 on entry to runtime.writebarrierptr
88 after runtime.writebarrierptr uses 56
80 on entry to runtime.writebarrierptr_nostore1
24 after runtime.writebarrierptr_nostore1 uses 56
16 on entry to runtime.acquirem
-24 after runtime.acquirem uses 40
Move closure creation into separate function so that
frames of writebarrierptr_shadow and writebarrierptr_nostore1
are overlapped.
Trace writer goroutine synchronizes with StopTrace
using trace.shutdownSema runtime semaphore.
But race detector does not see that synchronization
and so produces false reports.
Teach race detector about the synchronization.
Austin Clements [Tue, 3 Feb 2015 14:18:15 +0000 (09:18 -0500)]
runtime: rearrange framepointer check condition
The test for the framepointer experiment flag is cheaper and more
branch-predictable than the other parts of this conditional, so move
it first. This is also more readable.
(Originally, the flag check required parsing the experiments string,
which is why it was done last. Now that flag is cached.)
Austin Clements [Tue, 3 Feb 2015 14:09:56 +0000 (09:09 -0500)]
runtime: use 2*regSize for saved frame pointer check
Previously, we checked for a saved frame pointer by looking for a
2*ptrSize gap between the argument pointer and the locals pointer.
The intent of this check was to look for a two stack slot gap (caller
IP and saved frame pointer), but stack slots are regSize, not ptrSize.
Correct this by checking instead for a 2*regSize gap.
On most platforms, this made no difference because ptrSize==regSize.
However, on amd64p32 (nacl), the saved frame pointer check incorrectly
fired when there was no saved frame pointer because the one stack slot
for the caller IP left an 8 byte gap, which is 2*ptrSize (but not
2*regSize) on amd64p32.
Mikio Hara [Mon, 2 Feb 2015 12:09:23 +0000 (21:09 +0900)]
net: case insensitivity of DNS labels in built-in stub resolver
This change adds support for case insensitivity of DNS labels to
built-in DNS stub resolver as described in RFC 4343.
Fixes #9215.
Change-Id: Ia752fe71866a3bfa3ea08371985b799d419ddea3
Reviewed-on: https://go-review.googlesource.com/3685 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Lynn Boger [Wed, 28 Jan 2015 18:28:59 +0000 (12:28 -0600)]
runtime: Add memprofrate value to GODEBUG
Add memprofrate as a value recognized in GODEBUG. The
value provided is used as the new setting for
runtime.MemProfileRate, allowing the user to
adjust memory profiling.
Change-Id: If129a247683263b11e2dd42473cf9b31280543d5
Reviewed-on: https://go-review.googlesource.com/3450 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Austin Clements [Wed, 14 Jan 2015 16:09:50 +0000 (11:09 -0500)]
cmd/6g, liblink, runtime: support saving base pointers
This adds a "framepointer" GOEXPERIMENT that that makes the amd64
toolchain maintain base pointer chains in the same way that gcc
-fno-omit-frame-pointer does. Go doesn't use these saved base
pointers, but this does enable external tools like Linux perf and
VTune to unwind Go stacks when collecting system-wide profiles.
This requires support in the compilers to not clobber BP, support in
liblink for generating the BP-saving function prologue and unwinding
epilogue, and support in the runtime to save BPs across preemption, to
skip saved BPs during stack unwinding and, and to adjust saved BPs
during stack moving.
As with other GOEXPERIMENTs, everything from the toolchain to the
runtime must be compiled with this experiment enabled. To do this,
run make.bash (or all.bash) with GOEXPERIMENT=framepointer.
Austin Clements [Tue, 27 Jan 2015 23:29:02 +0000 (18:29 -0500)]
runtime: eliminate uses of BP on amd64
Any place that clobbers BP in the runtime can potentially interfere
with frame pointer unwinding with GOEXPERIMENT=framepointer. This
change eliminates uses of BP in the runtime to address this problem.
We have spare registers everywhere this occurs, so there's no downside
to eliminating BP. Where possible, this uses the same new register as
the amd64p32 runtime, which doesn't use BP due to restrictions placed
on it by NaCL.
One nice side effect of this is that it will let perf/VTune unwind the
call stack even through a call to systemstack, which will let us get
really good call graphs from the garbage collector.
Austin Clements [Fri, 30 Jan 2015 20:30:41 +0000 (15:30 -0500)]
runtime: rename m.gcing to m.preemptoff and make it a string
m.gcing has become overloaded to mean "don't preempt this g" in
general. Once the garbage collector is preemptible, the one thing it
*won't* mean is that we're in the garbage collector.
So, rename gcing to "preemptoff" and make it a string giving a reason
that preemption is disabled. gcing was never set to anything but 0 or
1, so we don't have to worry about there being a stack of reasons.
Austin Clements [Fri, 30 Jan 2015 16:18:24 +0000 (11:18 -0500)]
cmd/9g: note suboptimal copy code
9g generates needlessly complex code for small copies. There are a
few other things that need to be improved about the copy code, so for
now just note the problem.
Dmitri Shuralyov [Mon, 2 Feb 2015 09:28:10 +0000 (01:28 -0800)]
go/printer: set prefix correctly when all comment lines blank
In stripCommonPrefix, the prefix was correctly calculated in all cases,
except one. That unhandled case is when there are more than 2 lines,
but all lines are blank (other than the first and last lines,
which contain /* and */ respectively).
This change detects that case and correctly sets the prefix calculated
from the last line. This is consistent with the (correct) behavior
that happens when there's at least one non-blank line.
That fixes issue #9751 that occurs for problematic input,
where cmd/gofmt and go/source would insert extra indentation on
every format operation. It also allows go/printer itself to print
such parsed files in an expected way.
Fixes #9751.
Change-Id: Id3dfb945beb59ffad3705085a3c285fca30a5f87
Reviewed-on: https://go-review.googlesource.com/3684 Reviewed-by: Robert Griesemer <gri@golang.org>
Rahul Chaudhry [Fri, 30 Jan 2015 23:07:11 +0000 (15:07 -0800)]
os: allow EROFS in TestMkdirAllAtSlash
On some systems (e.g. ChromeOS), / is mounted read-only.
This results in error code syscall.EROFS, which I guess
is just as valid as syscall.EACCES for this test.
Change-Id: I9188d5437a1b5ac1daa9c68b95b8dcb447666ca3
Reviewed-on: https://go-review.googlesource.com/3676 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Dmitry Vyukov [Fri, 30 Jan 2015 11:36:12 +0000 (14:36 +0300)]
runtime: fix system memory allocator on plan9
The following line in sysFree:
n += (n + memRound) &^ memRound
doubles value of n (n += n).
Which is wrong and can lead to memory corruption.
Fixes #9712
Change-Id: I3c141b71da11e38837c09408cf4f1d22e8f7f36e
Reviewed-on: https://go-review.googlesource.com/3602 Reviewed-by: David du Colombier <0intro@gmail.com>