Whoops. I'm surprised it even worked before. (Need two pipes,
not one.)
Also, remove the whole pipe registration business, since it
wasn't even required in the previous version. (I'd later fixed
it at the end of send100Response, but forgot to delete it)
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/8191044
This CL ensures we use the correct socket options for
passive and active open sockets.
For the passive open sockets created by Listen functions,
additional SO_REUSEADDR, SO_REUSEPORT options are required
for the quick service restart and/or multicasting.
For the active open sockets created by Dial functions, no
additional options are required.
R=golang-dev, dave, bradfitz, rsc
CC=golang-dev
https://golang.org/cl/7795050
"There are only two hard problems in computer science:
cache invalidation, naming things, and off-by-one errors."
The HTTP server code already strips Expect: 100-continue on
requests, so httputil.ReverseProxy should be unaffected, but
some servers send unsolicited HTTP/1.1 100 Continue responses,
so we need to skip over them if they're seen to avoid getting
off-by-one on Transport requests/responses.
This does change the behavior of people who were using Client
or Transport directly and explicitly setting "Expect: 100-continue"
themselves, but it didn't work before anyway. Now instead of the
user code seeing a 100 response and then things blowing up, now
it basically works, except the Transport will still blast away
the full request body immediately. That's the part that needs
to be finished to close this issue.
This is the safe quick fix.
Update #3665
R=golang-dev, dsymonds, dave, jgrahamc
CC=golang-dev
https://golang.org/cl/8166045
This benchmark verifies that CL #8173043 reduces time spent
sliding the Buffer's contents.
Results without and with CL #8173043 applied:
benchmark old ns/op new ns/op delta
BenchmarkBufferFullSmallReads 755336 175054 -76.82%
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/8174043
Also added a new benchmark from the same test:
benchmark old ns/op new ns/op delta
BenchmarkBufferNotEmptyWriteRead 2643698 709189 -73.17%
Fixes#5154
R=golang-dev, r, gri
CC=golang-dev
https://golang.org/cl/8164043
Adds the missing wildcard port assignment description to ListenUDP.
Also updates the wildcard port description on ListenTCP.
R=golang-dev, dave, r
CC=golang-dev
https://golang.org/cl/8063043
Also removes redundant tests that run Go 1.0 non-IPv6 support
Windows code on IPv6 enabled Windows kernels.
R=alex.brainman, golang-dev, bradfitz, dave
CC=golang-dev
https://golang.org/cl/7812052
With the faster strings package, the difference between
the specialized code and strings.Split is in the noise:
benchmark old ns/op new ns/op delta
BenchmarkPrint 16724291 16686729 -0.22%
(Measured on a Mac Pro, 2.8GHz Quad-core Intel Xeon,
4GB 800 MHz DDR2, Mac OS X 10.8.3)
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/8100044
Saves both the textproto.Reader allocation, and its internal
scratch buffer growing.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnWithKeepAliveLite 10324 10149 -1.70%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnWithKeepAliveLite 19 17 -10.53%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnWithKeepAliveLite 1559 1492 -4.30%
R=golang-dev, r, gri
CC=golang-dev
https://golang.org/cl/8094046
Removes another per-request allocation. Also makes the code more
readable, IMO. And more testable.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnWithKeepAliveLite 10539 10324 -2.04%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnWithKeepAliveLite 20 19 -5.00%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnWithKeepAliveLite 1609 1559 -3.11%
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/8118044
We already depend on strings in this file, so use it.
Plus strings.Index will be faster than a manual loop
once issue 3751 is finished.
R=golang-dev, khr
CC=golang-dev
https://golang.org/cl/8116043
A chunkWriter and a response are 1:1. Make them contiguous in
memory and save an allocation.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnWithKeepAliveLite 10715 10539 -1.64%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnWithKeepAliveLite 21 20 -4.76%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnWithKeepAliveLite 1626 1609 -1.05%
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/8114043
Add more tests around the various orders handlers can access
and flush response headers.
Also clarify the documentation on fields of response and
chunkWriter.
While there, remove an allocation (a header clone) for simple
handlers.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnWithKeepAliveLite 15245 14966 -1.83%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnWithKeepAliveLite 24 23 -4.17%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnWithKeepAliveLite 1717 1668 -2.85%
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/8101043
Motivated by garbage profiling in HTTP benchmarks. This
changes means new empty maps are just one small allocation
(the HMap) instead the HMap + the relatively larger h->buckets
allocation. This helps maps which remain empty throughout
their life.
benchmark old ns/op new ns/op delta
BenchmarkNewEmptyMap 196 107 -45.41%
benchmark old allocs new allocs delta
BenchmarkNewEmptyMap 2 1 -50.00%
benchmark old bytes new bytes delta
BenchmarkNewEmptyMap 195 50 -74.36%
R=khr, golang-dev, r
CC=golang-dev
https://golang.org/cl/7722046
There was another bufio.Writer not being reused, found with
GOGC=off and -test.memprofile.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnWithKeepAlive 18270 16046 -12.17%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnWithKeepAlive 38 36 -5.26%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnWithKeepAlive 4598 2488 -45.89%
Update #5100
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/8038047
A HMUL node appears in some constant divisions, but
to observe a false negative in race detector the divisor must be
suitably chosen to make sure the only memory access is
done for HMUL.
R=dvyukov
CC=golang-dev
https://golang.org/cl/7935045
For Go 1.1, stop checking the rlimit, because it broke now
that mheap is allocated using SysAlloc. See issue 5049.
R=r
CC=golang-dev
https://golang.org/cl/7741050
The arm gentraceback mishandled frame linkage values pointing
to the assembly return function. This function is special as
its frame size is zero and it contains only one instruction.
These conditions would preserve the frame pointer and result
in an off by one error when unwinding the caller.
Fixes#5124
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/8023043
Prevents storm of error messages if something goes wrong.
In the case of issue 5073 the epoll fd was closed by the test.
Update #5073.
R=golang-dev, r, rsc
CC=golang-dev
https://golang.org/cl/7966043
This CL avoids test data sharing in repetitive test runs;
e.g., go test net -cpu=1,1,1
R=golang-dev, fullung, bradfitz
CC=golang-dev
https://golang.org/cl/8011043
Handle interface comparison correctly,
add a few more tests, mark more nodes as impossible.
R=dvyukov, golang-dev
CC=golang-dev
https://golang.org/cl/7942045
Make the copy directly in the convert switch instead of an extra loop.
Also stops converting nil-[]byte to zero-[]byte when assigning to *interface
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/7962044
At some point in the past, I believe the GCD algorithm was setting d to
be negative. The RSA code has been correcting that ever since but, now,
it appears to have changed and the correction isn't needed.
Having d be too large is harmless, it's just a little odd and I
happened to notice.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7948044
This keeps the logic about how to set the thread-local variables
m and g in code compiled and linked by the gc toolchain,
an important property for upcoming cgo changes.
It's also just a nice cleanup: one less place to update when
these details change.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7560048
The right operand of a && and || is only executed conditionnally,
so the instrumentation must be more careful. In particular
it should not turn nodes assumed to be cheap after walk into
expensive ones.
Update #4228
R=dvyukov, golang-dev
CC=golang-dev
https://golang.org/cl/7986043
The ARM implementation of runtime.cgocallback_gofunc diverged
from the calling convention by leaving a word of garbage at
the top of the stack and storing the return PC above the
locals. This change stores the return PC at the top of the
stack and removes the save area above the locals.
Update #5124
This CL fixes first part of the ARM issues and added the unwind test.
R=golang-dev, bradfitz, minux.ma, cshapiro, rsc
CC=golang-dev
https://golang.org/cl/7728045
The edit makes Hypot's description match the form
used in the other routines in this package.
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/8003046
Adds the new debugging constant 'checkgc'. If its value is non-zero
all calls to mallocgc() from hashmap.c will start a garbage collection.
Fixes#5074.
R=golang-dev, khr
CC=golang-dev, rsc
https://golang.org/cl/7663051
Fixes performance of the current windows network poller
with the new scheduler.
Gives runtime a hint when GetQueuedCompletionStatus() will block.
Fixes#5068.
benchmark old ns/op new ns/op delta
BenchmarkTCP4Persistent 4004000 33906 -99.15%
BenchmarkTCP4Persistent-2 21790 17513 -19.63%
BenchmarkTCP4Persistent-4 44760 34270 -23.44%
BenchmarkTCP4Persistent-6 45280 43000 -5.04%
R=golang-dev, alex.brainman, coocood, rsc
CC=golang-dev
https://golang.org/cl/7612045
This CL updates CL 7511043;
- adds new test cases for both UDPConn and UnixConn,
- makes sure unnamed UnixAddr handling,
- replaces t.Errorf with t.Fatalf in sockname related test cases.
Fixes#3721 (again).
Fixes#3838 (again).
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7627048
I'm not sure how to write a test for this. The change in
behaviour is that if you somehow get a SIGBUS signal for an
address >= 0x1000, the program will now crash rather than
calling panic. As far as I know, on x86 GNU/Linux, the only
way to get a SIGBUS (rather than a SIGSEGV) is to set the
stack pointer to an invalid value.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7906045
E.g conversions from numeric types to RawBytes are missing, what makes RawBytes unusable in some cases.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/7783046
The GIF89a spec says: "Each image must fit within the
boundaries of the Logical Screen, as defined in the
Logical Screen Descriptor." Also, do not accept
GIFs which have too much data for the image size.
R=nigeltao, jra, r
CC=bradfitz, golang-dev
https://golang.org/cl/7602045
On Darwin and FreeBSD, the mmap syscall return value is returned
unmodified. This means that the return value will either be a
valid address or a positive error number.
Also check return value from mmap in SysReserve - the callers of
SysReserve expect nil to be returned if the allocation failed.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7871043
Rather than just checking for ENOMEM, check for a return value of less
than 4096, so that we catch other errors such as EACCES and EINVAL.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7942043
Saves over 8KB of allocations per new connection.
benchmark old ns/op new ns/op delta
BenchmarkServerFakeConnNoKeepAlive 28777 24927 -13.38%
benchmark old allocs new allocs delta
BenchmarkServerFakeConnNoKeepAlive 52 46 -11.54%
benchmark old bytes new bytes delta
BenchmarkServerFakeConnNoKeepAlive 13716 5286 -61.46%
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/7799047
After the revert of revision 9ea9e7e6e0c8
the related revision 76ff7da868c6 must be reverted too.
Fixes#5102.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7961044
If a handler didn't consume all its Request.Body, child.go was
closing the socket while the host was still writing to it,
causing the child to send a RST and the host (at least nginx)
to send an empty response body.
Now, we tell the host we're done with the request/response
first, and then close our input pipe after consuming a bit of
it. Consuming the body fixes the problem, and flushing to the
host first to tell it that we're done increases the chance
that the host cuts off further data to us, meaning we won't
have much to consume.
No new tests, because this package is lacking in tests.
Tested by hand with nginx. See issue for testing details.
Fixes#4183
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7939045
Also adjust the implementation of applyRelocationsAMD64
so that the test added in CL 6848044 still passes.
R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/7686049
XOR key into data 128 bits at a time instead of 64 bits
and pipeline half of state loads. Rotate loop to allow
single-register indexing for state[i].
On a MacBookPro10,2 (Core i5):
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 412 224 -45.63%
BenchmarkRC4_1K 3179 1613 -49.26%
BenchmarkRC4_8K 25223 12545 -50.26%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 310.51 570.42 1.84x
BenchmarkRC4_1K 322.09 634.48 1.97x
BenchmarkRC4_8K 320.97 645.32 2.01x
For comparison, on the same machine, openssl 0.9.8r reports
its rc4 speed as somewhat under 350 MB/s for both 1K and 8K
(it is operating 64 bits at a time).
On an Intel Xeon E5520:
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 418 259 -38.04%
BenchmarkRC4_1K 3200 1884 -41.12%
BenchmarkRC4_8K 25173 14529 -42.28%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 306.04 492.48 1.61x
BenchmarkRC4_1K 319.93 543.26 1.70x
BenchmarkRC4_8K 321.61 557.20 1.73x
For comparison, on the same machine, openssl 1.0.1
reports its rc4 speed as 587 MB/s for 1K and 601 MB/s for 8K.
R=agl
CC=golang-dev
https://golang.org/cl/7865046
-- amd64 --
On a MacBookPro10,2 (Core i5):
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 470 421 -10.43%
BenchmarkRC4_1K 3123 3275 +4.87%
BenchmarkRC4_8K 26351 25866 -1.84%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 272.22 303.40 1.11x
BenchmarkRC4_1K 327.80 312.58 0.95x
BenchmarkRC4_8K 307.24 313.00 1.02x
For comparison, on the same machine, openssl 0.9.8r reports
its rc4 speed as somewhat under 350 MB/s for both 1K and 8K.
The Core i5 performance can be boosted another 20%, but only
by making the Xeon performance significantly slower.
On an Intel Xeon E5520:
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 774 417 -46.12%
BenchmarkRC4_1K 6121 3200 -47.72%
BenchmarkRC4_8K 48394 25151 -48.03%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 165.18 306.84 1.86x
BenchmarkRC4_1K 167.28 319.92 1.91x
BenchmarkRC4_8K 167.29 321.89 1.92x
For comparison, on the same machine, openssl 1.0.1
(which uses a different implementation than 0.9.8r)
reports its rc4 speed as 587 MB/s for 1K and 601 MB/s for 8K.
It is using SIMD instructions to do more in parallel.
So there's still some improvement to be had, but even so,
this is almost 2x faster than what it replaced.
-- 386 --
On a MacBookPro10,2 (Core i5):
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 3491 421 -87.94%
BenchmarkRC4_1K 28063 3205 -88.58%
BenchmarkRC4_8K 220392 25228 -88.55%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 36.66 303.81 8.29x
BenchmarkRC4_1K 36.49 319.42 8.75x
BenchmarkRC4_8K 36.73 320.90 8.74x
On an Intel Xeon E5520:
benchmark old ns/op new ns/op delta
BenchmarkRC4_128 2268 524 -76.90%
BenchmarkRC4_1K 18161 4137 -77.22%
BenchmarkRC4_8K 142396 32350 -77.28%
benchmark old MB/s new MB/s speedup
BenchmarkRC4_128 56.42 244.13 4.33x
BenchmarkRC4_1K 56.38 247.46 4.39x
BenchmarkRC4_8K 56.86 250.26 4.40x
R=agl
CC=golang-dev
https://golang.org/cl/7547050
Fixes#5061.
Current code relies on the fact that fd's are automatically removed from epoll set when closed. However, it is not true. Underlying file description is removed from epoll set only when *all* fd's referring to it are closed.
There are 2 bad consequences:
1. Kernel delivers notifications on already closed fd's.
2. The following sequence of events leads to error:
- add fd1 to epoll
- dup fd1 = fd2
- close fd1 (not removed from epoll since we've dup'ed the fd)
- dup fd2 = fd1 (get the same fd as fd1)
- add fd1 to epoll = EEXIST
So, if fd can be potentially dup'ed of fork'ed, it's necessary to explicitly remove the fd from epoll set.
R=golang-dev, bradfitz, dave
CC=golang-dev
https://golang.org/cl/7870043
Add missing CLOSUREVAR in switch.
Mark MAKE, string conversion nodes as impossible.
Control statements do not need instrumentation.
Instrument COM and LROT nodes.
Instrument map length.
Update #4228
R=dvyukov, golang-dev
CC=golang-dev
https://golang.org/cl/7504047
Hashtable is arranged as an array of
8-entry buckets with chained overflow.
Each bucket has 8 extra hash bits
per key to provide quick lookup within
a bucket. Table is grown incrementally.
Update #3885
Go time drops from 0.51s to 0.34s.
R=r, rsc, m3b, dave, bradfitz, khr, ugorji, remyoudompheng
CC=golang-dev
https://golang.org/cl/7504044
Found this inconsistency from net/http's Server while
debugging Issue 4183
Unfortunately this package lacks testing around this,
or most of child.go. :/
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/7735046
No net package involved. And with ReportAllocs we can see
how much garbage is created per request.
R=adg, dave
CC=golang-dev
https://golang.org/cl/7913044
