gc does not report this as an error, but go/types does.
(I suspect that constructing a closure counts as a reference
to &all in gc's implementation).
This is not a tool bug, since the spec doesn't require
implementations to implement this check, but it does
illustrate that dialect variations are always a nuisance.
LGTM=rsc, bradfitz
R=bradfitz
CC=golang-codereviews, gri, rsc
https://golang.org/cl/73850043
Previously, passing a long duration to ParseDuration could result in
random, even negative, values.
LGTM=r
R=golang-codereviews, bradfitz, r
CC=golang-codereviews
https://golang.org/cl/72120043
CL 69340044 requires that syscall.SO_ERROR be defined on all unix like platforms. Add SO_ERROR to the list of dummy constants in sycall/net_nacl.go.
LGTM=bradfitz
R=iant, rsc, mikioh.mikioh, bradfitz
CC=golang-codereviews
https://golang.org/cl/73100043
If we report a leak, make sure we've waited long enough to be sure.
The new sleep regimen waits 1.05 seconds before failing; the old
one waited 0.005 seconds.
(The single linux/amd64 failure in this test feels more like a
timing problem than a leak. I don't want to spend time on it unless
we're sure.)
LGTM=bradfitz
R=bradfitz
CC=golang-codereviews
https://golang.org/cl/72630043
We provide amd64p32 implementations for md5 and sha1 so we need to exclude amd64p32 from the generic implementations in those packages.
Fixes build once CL 72360044 lands.
LGTM=agl, remyoudompheng
R=rsc, bradfitz, agl, remyoudompheng
CC=golang-codereviews
https://golang.org/cl/72460043
This code being buggy is the only explanation I can come up
with for issue 7325. It's probably not, but the only alternative
is a Windows kernel bug. Comment this out to see what breaks
or gets fixed.
Update #7325
LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/72590044
From the trace it appears that stackalloc is being
called with 0x1800 which is 6k = 4k + (StackSystem=2k).
Make StackSystem 4k too, to make stackalloc happy.
It's already 4k on windows/amd64.
TBR=khr
CC=golang-codereviews
https://golang.org/cl/72600043
There are at least 3 bugs:
1. g->stacksize accounting is broken during copystack/shrinkstack
2. stktop->free is not properly maintained during copystack/shrinkstack
3. stktop->free logic is broken:
we can have stktop->free==FixedStack,
and we will free it into stack cache,
but it actually comes from heap as the result of non-copying segment shrink
This shows as at least spurious races on race builders (maybe something else as well I don't know).
The idea behind the refactoring is to consolidate stacksize and
segment origin logic in stackalloc/stackfree.
Fixes#7490.
LGTM=rsc, khr
R=golang-codereviews, rsc, khr
CC=golang-codereviews
https://golang.org/cl/72440043
Recursive panics leave dangling Panic structs in g->panic stack.
At best it leads to a Defer leak and incorrect output on a subsequent panic.
At worst it arbitrary corrupts heap.
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/72480043
One reason the sync.Pool finalizer test can fail is that
this function's ef1 contains uninitialized data that just
happens to point at some of the old pool. I've seen this cause
retention of a single pool cache line (32 elements) on arm.
Really we need liveness information for C functions, but
for now we can be more careful about data in long-lived
C functions that block.
LGTM=bradfitz, dvyukov
R=golang-codereviews, bradfitz, dvyukov
CC=golang-codereviews, iant, khr
https://golang.org/cl/72490043
Instead, split the underlying storage in half and
free just half of it.
Shrinking without copying lets us reclaim storage used
by a previously profligate Go routine that has now blocked
inside some C code.
To shrink in place, we need all stacks to be a power of 2 in size.
LGTM=rsc
R=golang-codereviews, rsc
CC=golang-codereviews
https://golang.org/cl/69580044
Two memory allocator bug fixes.
- efence is not maintaining the proper heap metadata
to make eventual memory reuse safe, so use SysFault.
- now that our heap PageSize is 8k but most hardware
uses 4k pages, SysAlloc and SysReserve results must be
explicitly aligned. Do that in a few more call sites and
document this fact in malloc.h.
Fixes#7448.
LGTM=iant
R=golang-codereviews, josharian, iant
CC=dvyukov, golang-codereviews
https://golang.org/cl/71750048
I've just needed the G status on fault to debug runtime bug.
For some reason we print everything except header here.
Make it more informative and consistent.
R=rsc
CC=golang-codereviews
https://golang.org/cl/67870056
Implement custom assembly thunks for hot race calls (memory accesses and function entry/exit).
The thunks extract caller pc, verify that the address is in heap or global and switch to g0 stack.
Before:
ok regexp 3.692s
ok compress/bzip2 9.461s
ok encoding/json 6.380s
After:
ok regexp 2.229s (-40%)
ok compress/bzip2 4.703s (-50%)
ok encoding/json 3.629s (-43%)
For comparison, normal non-race build:
ok regexp 0.348s
ok compress/bzip2 0.304s
ok encoding/json 0.661s
Race build:
ok regexp 2.229s (+540%)
ok compress/bzip2 4.703s (+1447%)
ok encoding/json 3.629s (+449%)
Also removes some race-related special cases from cgocall and scheduler.
In long-term it will allow to remove cyclic runtime/race dependency on cmd/cgo.
Fixes#4249.
Fixes#7460.
Update #6508
Update #6688
R=iant, rsc, bradfitz
CC=golang-codereviews
https://golang.org/cl/55100044
During the glob decoding process interface values are set to concrete
values after a test for assignability. If the assignability test fails
a slightly vague error message is produced. While technically accurate
the error message does not clearly describe the problem.
Rewrite the error message to include the usage of the word assignable,
which makes it clear the concrete value type is not assignable to the
interface value type.
Fixes#6467.
LGTM=r
R=golang-codereviews, rsc, r
CC=golang-codereviews
https://golang.org/cl/71590043
32-bit Windows uses "structured exception handling" (SEH) to
handle hardware faults: that there is a per-thread linked list
of fault handlers maintained in user space instead of
something like Unix's signal handlers. The structures in the
linked list are required to live on the OS stack, and the
usual discipline is that the function that pushes a record
(allocated from the current stack frame) onto the list pops
that record before returning. Not to pop the entry before
returning creates a dangling pointer error: the list head
points to a stack frame that no longer exists.
Go pushes an SEH record in the top frame of every OS thread,
and that record suffices for all Go execution on that thread,
at least until cgo gets involved.
If we call into C using cgo, that called C code may push its
own SEH records, but by the convention it must pop them before
returning back to the Go code. We assume it does, and that's
fine.
If the C code calls back into Go, we want the Go SEH handler
to become active again, not whatever C has set up. So
runtime.callbackasm1, which handles a call from C back into
Go, pushes a new SEH record before calling the Go code and
pops it when the Go code returns. That's also fine.
It can happen that when Go calls C calls Go like this, the
inner Go code panics. We allow a defer in the outer Go to
recover the panic, effectively wiping not only the inner Go
frames but also the C calls. This sequence was not popping the
SEH stack up to what it was before the cgo calls, so it was
creating the dangling pointer warned about above. When
eventually the m stack was used enough to overwrite the
dangling SEH records, the SEH chain was lost, and any future
panic would not end up in Go's handler.
The bug in TestCallbackPanic and friends was thus creating a
situation where TestSetPanicOnFault - which causes a hardware
fault - would not find the Go fault handler and instead crash
the binary.
Add checks to TestCallbackPanicLocked to diagnose the mistake
in that test instead of leaving a bad state for another test
case to stumble over.
Fix bug by restoring SEH chain during deferred "endcgo"
cleanup.
This bug is likely present in Go 1.2.1, but since it depends
on Go calling C calling Go, with the inner Go panicking and
the outer Go recovering the panic, it seems not important
enough to bother fixing before Go 1.3. Certainly no one has
complained.
Fixes#7470.
LGTM=alex.brainman
R=golang-codereviews, alex.brainman
CC=golang-codereviews, iant, khr
https://golang.org/cl/71440043
Regenerate z-files for DragonFly BSD 3.6.
F_DUP_FD_CLOEXEC is now supported, so remove the zero value constant
from types_dragonfly.go so that we use the generated value from the
z-files.
LGTM=mikioh.mikioh
R=golang-codereviews, mikioh.mikioh
CC=golang-codereviews
https://golang.org/cl/70080047
The format of the DragonFly BSD syscalls.master file has changed
slightly - update mksysnum_dragonfly.pl to match.
LGTM=mikioh.mikioh
R=golang-codereviews, mikioh.mikioh
CC=golang-codereviews
https://golang.org/cl/71460044
Disable the "udp" to IPv6 unicast address on the loopback interface
test under DragonFly BSD. This currently returns a local address of
0.0.0.1, rather than an IPv6 address with zone identifier.
Update #7473
LGTM=mikioh.mikioh
R=golang-codereviews, mikioh.mikioh
CC=golang-codereviews
https://golang.org/cl/71500044
Performing multiple connect system calls on a non-blocking socket
under DragonFly BSD does not necessarily result in errors from earlier
connect calls being returned, particularly if we are connecting to
localhost. Instead, once netpoll tells us that the socket is ready,
get the SO_ERROR socket option to see if the connection succeeded
or failed.
Fixes#7474
LGTM=mikioh.mikioh
R=mikioh.mikioh
CC=golang-codereviews
https://golang.org/cl/69340044
Added test cases and expanded test harness to handle token end
positions.
Also: Make sure token end positions are never outside the valid
position range, as was possible in case of parse errors.
Fixes#7458.
LGTM=bradfitz
R=bradfitz
CC=golang-codereviews
https://golang.org/cl/70190046
the crypto/tls revision d3d43f270632 (CL 67010043, requiring ServerName or InsecureSkipVerify) breaks net/smtp,
since it seems impossible to do SMTP via TLS anymore. i've tried to fix this by simply using a tls.Config with
ServerName, instead of a nil *tls.Config. without this fix, doing SMTP with TLS results in error "tls: either
ServerName or InsecureSkipVerify must be specified in the tls.Config".
testing: the new method TestTlsClient(...) sets up a skeletal smtp server with tls capability, and test client
injects a "fake" certificate allowing tls to work on localhost; thus, the modification to SendMail(...) enabling
this.
Fixes#7437.
LGTM=bradfitz
R=golang-codereviews, josharian, bradfitz
CC=golang-codereviews
https://golang.org/cl/70380043
The network connection dies differently from the server's
perspective on (some) Windows when the client goes away. Match
on the common prefix (common to Unix and Windows) instead of
the network error part.
Fixes#7456
LGTM=josharian
R=golang-codereviews, josharian
CC=alex.brainman, golang-codereviews, iant
https://golang.org/cl/70010050
For non-closure functions, the context register is uninitialized
on entry and will not be used, but morestack saves it and then the
garbage collector treats it as live. This can be a source of memory
leaks if the context register points at otherwise dead memory.
Avoid this by introducing a parallel set of morestack functions
that clear the context register, and use those for the non-closure functions.
I hope this will help with some of the finalizer flakiness, but it probably won't.
Fixes#7244.
LGTM=dvyukov
R=khr, dvyukov
CC=golang-codereviews
https://golang.org/cl/71030044
I have one machine where this 25 test run is flaky
and fails ("21 >= 21"), but 50 works everywhere.
LGTM=josharian
R=josharian
CC=golang-codereviews
https://golang.org/cl/67870053
It mentioned true and false for error values. Instead, just
don't mention the error semantics, as they match normal Go
conventions (if error is non-nil, the other value is
meaningless). We generally only document error values when
they're interesting (where non-nil, non-nil is valid, or the
error value can be certain known values or types).
Fixes#7464
LGTM=agl
R=agl
CC=golang-codereviews
https://golang.org/cl/68440044
The flakiness appears to be just in tests, not in the actual code.
Specifically, the many tests call runtime.GC once and expect that
the finalizers will be running in the background when GC returns.
Now that the sweep phase is concurrent with execution, however,
the finalizers will not be run until sweep finishes, which might
be quite a bit later. To force sweep to finish, implement runtime.GC
by calling the actual collection twice. The second will complete the
sweep from the first.
This was reliably broken after a few runs before the CL and now
passes tens of runs:
while GOMAXPROCS=2 ./runtime.test -test.run=Finalizer -test.short \
-test.timeout=300s -test.cpu=$(perl -e 'print ("1,2,4," x 100) . "1"')
do true; done
Fixes#7328.
LGTM=dvyukov
R=dvyukov, dave
CC=golang-codereviews
https://golang.org/cl/71080043
The exception handler runs on the ordinary g stack,
and the stack copier is now trying to do a traceback
across it. That's never been needed before, so it was
unimplemented. Implement it, in all its ugliness.
Fixes windows/amd64 build.
TBR=khr
CC=golang-codereviews
https://golang.org/cl/71030043
1) Move StateHijacked callback earlier, to make it
panic-proof. A Hijack followed by a panic didn't previously
result in ConnState getting fired for StateHijacked. Move it
earlier, to the time of hijack.
2) Don't fire StateActive unless any bytes were read off the
wire while waiting for a request. This means we don't
transition from New or Idle to Active if the client
disconnects or times out. This was documented before, but not
implemented properly.
This CL is required for an pending fix for Issue 7264
LGTM=josharian
R=josharian
CC=golang-codereviews
https://golang.org/cl/69860049
Unfortunately FreeBSD 10 has changed its syscall arguments for
some reasons but as per request at golang-dev this CL does not
generate some structures, syscall numbers and constants as
compatible to FreeBSD 10 as follows:
Structure: Stat_t, IfData, IfMsghdr are based on 8-STABLE
Syscall number: Capsicum API is based on 9-STABLE
Constant: IFT_CARP, SIOCAIFADDR, SIOCSIFPHYADDR are based on 9-STABLE
Update #7193
FreeBSD 10 breaking changes:
r205792: Rename st_*timespec fields to st_*tim for POSIX 2008
compliance.
http://svnweb.freebsd.org/base?view=revision&revision=205792
r254804: Restructure the mbuf pkthdr to make it fit for upcoming
capabilities and features.
http://svnweb.freebsd.org/base?view=revision&revision=254804
r255219: Change the cap_rights_t type from uint64_t to a structure
that we can extend in the future in a backward compatible (API and
ABI) way.
http://svnweb.freebsd.org/base?view=revision&revision=255219
LGTM=iant
R=golang-codereviews, rsc, minux.ma, gobot, iant
CC=golang-codereviews
https://golang.org/cl/56770044
In Go 1.2, closing a request body without reading to EOF
causes the underlying TCP connection to not be reused. This
client code following redirects was never updated when that
happened.
This was part of a previous CL but moved to its own CL at
Josh's request. Now with test too.
LGTM=josharian
R=josharian
CC=golang-codereviews
https://golang.org/cl/70800043
Currently a write error will cause future reads to return that same error.
However, there may have been extra information from a peer pending on
the read direction that is now unavailable.
This change splits the single connErr into errors for the read, write and
handshake. (Splitting off the handshake error is needed because both read
and write paths check the handshake error.)
Fixes#7414.
LGTM=bradfitz, r
R=golang-codereviews, r, bradfitz
CC=golang-codereviews
https://golang.org/cl/69090044
The addition of Server.ConnState provides all the necessary
hooks to stop a Server gracefully, but StateNew previously
could fire concurrently with Serve exiting (as it does when
its net.Listener is closed). This previously meant one
couldn't use a WaitGroup incremented in the StateNew hook
along with calling Wait after Serve. Now you can.
Update #4674
LGTM=bradfitz
R=bradfitz
CC=golang-codereviews
https://golang.org/cl/70410044
Update #7338
The nil deref tests are currently failing on the *bsd/arm platforms. In an effort to avoid the build deteriorating further I would like to skip these tests on freebsd/arm and netbsd/arm.
LGTM=bradfitz, minux.ma
R=golang-codereviews, bradfitz, minux.ma
CC=golang-codereviews
https://golang.org/cl/69870045
cgocall.c: define the CBARGS macro for GOARCH_amd64p32. I don't think the value of this macro will ever be used under nacl/amd64p32 but it is required to compile even if cgo is not used.
hashmap.goc: amd64p32 uses 32bit words.
LGTM=iant
R=rsc, iant
CC=golang-codereviews
https://golang.org/cl/69960044
This is a user error, but killing -1 kills everything, which
is a pretty bad failure mode.
Fixes#7434
LGTM=iant
R=iant
CC=golang-codereviews
https://golang.org/cl/70140043
While reviewing uses of the lower-level Client API in code, I found
that in many cases, code was using Client only because it needed a
timeout on the connection. DialWithDialer allows a timeout (and
other values) to be specified without resorting to the low-level API.
LGTM=r
R=golang-codereviews, r, bradfitz
CC=golang-codereviews
https://golang.org/cl/68920045
Update #4674
This allows for all sorts of graceful shutdown policies,
without picking a policy (e.g. lameduck period) and without
adding lots of locking to the server core. That policy and
locking can be implemented outside of net/http now.
LGTM=adg
R=golang-codereviews, josharian, r, adg, dvyukov
CC=golang-codereviews
https://golang.org/cl/69260044
This CL replays the following one CL from the rsc-go13nacl repo.
This is the last replay CL: after this CL the main repo will have
everything the rsc-go13nacl repo did. Changes made to the main
repo after the rsc-go13nacl repo branched off probably mean that
NaCl doesn't actually work after this CL, but all the code is now moved
over and just needs to be redebugged.
---
cmd/6l, cmd/8l, cmd/ld: support for Native Client
See golang.org/s/go13nacl for design overview.
This CL is publicly visible but not CC'ed to golang-dev,
to avoid distracting from the preparation of the Go 1.2
release.
This CL and the others will be checked into my rsc-go13nacl
clone repo for now, and I will send CLs against the main
repo early in the Go 1.3 development.
R≡khr
https://golang.org/cl/15750044
---
LGTM=bradfitz, dave, iant
R=dave, bradfitz, iant
CC=golang-codereviews
https://golang.org/cl/69040044
over multiple scans. Previously, the Go code assumed that DC was
synonymous with interleaved and AC with non-interleaved.
Fixes#6767.
The test files were generated with libjpeg's cjpeg program, version 9a,
with the following patch, since cjpeg is hard-coded to output
interleaved DC.
$ diff -u jpeg-9a*/jcparam.c
--- jpeg-9a-clean/jcparam.c 2013-07-01 21:13:28.000000000 +1000
+++ jpeg-9a/jcparam.c 2014-02-27 11:40:41.236889852 +1100
@@ -572,7 +572,7 @@
{
int ci;
- if (ncomps <= MAX_COMPS_IN_SCAN) {
+ if (0) {
/* Single interleaved DC scan */
scanptr->comps_in_scan = ncomps;
for (ci = 0; ci < ncomps; ci++)
@@ -610,7 +610,7 @@
(cinfo->jpeg_color_space == JCS_YCbCr ||
cinfo->jpeg_color_space == JCS_BG_YCC)) {
/* Custom script for YCC color images. */
- nscans = 10;
+ nscans = 14;
} else {
/* All-purpose script for other color spaces. */
if (ncomps > MAX_COMPS_IN_SCAN)
LGTM=r
R=r
CC=golang-codereviews
https://golang.org/cl/69000046
Before GC, we flush all the per-P allocation caches. Doing
stack shrinking mid-GC causes these caches to fill up. At the
end of gc, the sweepgen is incremented which causes all of the
data in these caches to be in a bad state (cached but not yet
swept).
Move the stack shrinking until after sweepgen is incremented,
so any caching that happens as part of shrinking is done with
already-swept data.
Reenable stack copying.
LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/69620043
No change to $GOROOT/src, misc formatting.
Nice side-effect: almost 3% faster runs because it's much faster to compute
line number differences in the generated output than the incoming source.
Benchmark run, best of 5 runs, before and after:
BenchmarkPrint 200 12347587 ns/op
BenchmarkPrint 200 11999061 ns/op
Fixes#4504.
LGTM=adonovan
R=golang-codereviews, adonovan
CC=golang-codereviews
https://golang.org/cl/69260045
This is the simple half of https://golang.org/cl/53560043/ with
a new benchmark. pongad is in the C+A files already.
benchmark old ns/op new ns/op delta
BenchmarkReaderWriteToOptimal 2054 825 -59.83%
Update #6373
LGTM=iant, gri
R=golang-codereviews, iant, gri
CC=golang-codereviews
https://golang.org/cl/69220046
This test currently deadlocks on dragonfly/386.
Update #7421
LGTM=minux.ma
R=golang-codereviews, minux.ma
CC=golang-codereviews
https://golang.org/cl/69380043
warning: src/pkg/runtime/mem_plan9.c:72 param declared and not used: n
src/pkg/runtime/mem_plan9.c:73 name not declared: nbytes
src/pkg/runtime/mem_plan9.c:73 bad in naddr: NAME nbytes<>+0(SB)
LGTM=minux.ma, bradfitz
R=khr, minux.ma, bradfitz
CC=golang-codereviews
https://golang.org/cl/69360043
On stack overflow, if all frames on the stack are
copyable, we copy the frames to a new stack twice
as large as the old one. During GC, if a G is using
less than 1/4 of its stack, copy the stack to a stack
half its size.
TODO
- Do something about C frames. When a C frame is in the
stack segment, it isn't copyable. We allocate a new segment
in this case.
- For idempotent C code, we can abort it, copy the stack,
then retry. I'm working on a separate CL for this.
- For other C code, we can raise the stackguard
to the lowest Go frame so the next call that Go frame
makes triggers a copy, which will then succeed.
- Pick a starting stack size?
The plan is that eventually we reach a point where the
stack contains only copyable frames.
LGTM=rsc
R=dvyukov, rsc
CC=golang-codereviews
https://golang.org/cl/54650044
MCaches now hold a MSpan for each sizeclass which they have
exclusive access to allocate from, so no lock is needed.
Modifying the heap bitmaps also no longer requires a cas.
runtime.free gets more expensive. But we don't use it
much any more.
It's not much faster on 1 processor, but it's a lot
faster on multiple processors.
benchmark old ns/op new ns/op delta
BenchmarkSetTypeNoPtr1 24 23 -0.42%
BenchmarkSetTypeNoPtr2 33 34 +0.89%
BenchmarkSetTypePtr1 51 49 -3.72%
BenchmarkSetTypePtr2 55 54 -1.98%
benchmark old ns/op new ns/op delta
BenchmarkAllocation 52739 50770 -3.73%
BenchmarkAllocation-2 33957 34141 +0.54%
BenchmarkAllocation-3 33326 29015 -12.94%
BenchmarkAllocation-4 38105 25795 -32.31%
BenchmarkAllocation-5 68055 24409 -64.13%
BenchmarkAllocation-6 71544 23488 -67.17%
BenchmarkAllocation-7 68374 23041 -66.30%
BenchmarkAllocation-8 70117 20758 -70.40%
LGTM=rsc, dvyukov
R=dvyukov, bradfitz, khr, rsc
CC=golang-codereviews
https://golang.org/cl/46810043
Functions that "fit" on one line and were on one
line in the original source are not broken up into
two lines anymore simply because they contain a comment.
- Fine-tuned use of separating blanks after /*-style comments, so:
( /* extra blank after this comment */ )
(a int /* no extra blank after this comment*/)
- Factored out comment state (from printer state) into commentInfo.
- No impact on $GOROOT/src, misc formatting.
Fixes#5543.
LGTM=r
R=golang-codereviews, r
CC=golang-codereviews
https://golang.org/cl/68630043
This partly addresses issue 6099 where a gofmt rewrite is behaving
unexpectedly because the provided rewrite term is not a valid expression
but is silently consumed anyway.
LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/68920044