- 08 May, 2017 1 commit
-
-
Alex Brainman authored
Use closure parameter instead of external variable to remove 1 allocation. I tried to add test, but it is difficult to add something simple and not flake here. I did test this with: diff --git a/src/net/writev_test.go b/src/net/writev_test.go index 4c05be4..e417d68 100644 --- a/src/net/writev_test.go +++ b/src/net/writev_test.go @@ -99,6 +99,15 @@ func TestBuffers_WriteTo(t *testing.T) { } } +func TestBuffers_WriteToAllocs(t *testing.T) { + allocs := testing.AllocsPerRun(10, func() { + testBuffer_writeTo(t, 10, false) + }) + if allocs > 0 { + t.Fatalf("got %v; want 0", allocs) + } +} + func testBuffer_writeTo(t *testing.T, chunks int, useCopy bool) { oldHook := poll.TestHookDidWritev defer func() { poll.TestHookDidWritev = oldHook }() It makes allocation count go down by 1 after the fix. Before: C:\>u:\test -test.v -test.run=WriteToAllocs === RUN TestBuffers_WriteToAllocs --- FAIL: TestBuffers_WriteToAllocs (0.05s) writev_test.go:107: got 66; want 0 FAIL and after: C:\>u:\test -test.v -test.run=WriteToAllocs === RUN TestBuffers_WriteToAllocs --- FAIL: TestBuffers_WriteToAllocs (0.04s) writev_test.go:107: got 65; want 0 FAIL Thanks to @MichaelMonashev for report and the fix. Fixes #19222 Change-Id: I0f73cd9e2c8bbaa0653083f81f3ccb83b5ea84e1 Reviewed-on: https://go-review.googlesource.com/42893Reviewed-by:Brad Fitzpatrick <bradfitz@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
- 07 May, 2017 5 commits
-
-
Elias Naur authored
Ever since CL 33301 linking darwin/arm64 excutables has resulted in warnings like: ld: warning: -no_pie ignored for arm64 Remove -no_pie on darwin/arm64. Change-Id: I9f7685351fa8cce29795283e1a24fc7a6753d698 Reviewed-on: https://go-review.googlesource.com/42815 Run-TryBot: Elias Naur <elias.naur@gmail.com> Reviewed-by:
-
Kevin Burke authored
Also switch "stating" to "statting" to describe applying os.Stat to a resource; the former is more confusable than the latter. Change-Id: I9d8e3506bd383f8f1479c05948c03b8c633dc4af Reviewed-on: https://go-review.googlesource.com/42855Reviewed-by:
-
Marvin Stenger authored
In the common case, the grow method only needs to reslice the internal buffer. Making another function call to grow can be expensive when Write is called very often with small pieces of data (like a byte or rune). Thus, we add a tryGrowByReslice method that is inlineable so that we can avoid an extra call in most cases. name old time/op new time/op delta WriteByte-4 35.5µs ± 0% 17.4µs ± 1% -51.03% (p=0.000 n=19+20) WriteRune-4 55.7µs ± 1% 38.7µs ± 1% -30.56% (p=0.000 n=18+19) BufferNotEmptyWriteRead-4 304µs ± 5% 283µs ± 3% -6.86% (p=0.000 n=19+17) BufferFullSmallReads-4 87.0µs ± 5% 66.8µs ± 2% -23.26% (p=0.000 n=17+17) name old speed new speed delta WriteByte-4 115MB/s ± 0% 235MB/s ± 1% +104.19% (p=0.000 n=19+20) WriteRune-4 221MB/s ± 1% 318MB/s ± 1% +44.01% (p=0.000 n=18+19) Fixes #17857 Change-Id: I08dfb10a1c7e001817729dbfcc951bda12fe8814 Reviewed-on: https://go-review.googlesource.com/42813Reviewed-by:
-
Damien Lespiau authored
MOVSD is properly handled but its encoding test wasn't enabled. Enable it. For reference this was found with a little tool I wrote [1] to explore which instructions are missing or not tested in the go obj package and assembler: "which SSE2 instructions aren't tested? And don't list instructions which can take MMX operands" $ x86db-gogen list --extension SSE2 --not-tested --not-mmx CLFLUSH mem [m: np 0f ae /7] WILLAMETTE,SSE2 MOVSD xmmreg,xmmreg [rm: f2 0f 10 /r] WILLAMETTE,SSE2 MOVSD xmmreg,xmmreg [mr: f2 0f 11 /r] WILLAMETTE,SSE2 MOVSD mem64,xmmreg [mr: f2 0f 11 /r] WILLAMETTE,SSE2 MOVSD xmmreg,mem64 [rm: f2 0f 10 /r] WILLAMETTE,SSE2 (CLFLUSH was introduced with SSE2, but has its own CPUID bit) [1] https://github.com/dlespiau/x86db Change-Id: Ic3af3028cb8d4f02e53fdebb9b30fb311f4ee454 Reviewed-on: https://go-review.googlesource.com/42814Reviewed-by:
-
Alex Brainman authored
Currently windows Stat uses combination of Lstat and Readlink to walk symlinks until it reaches file or directory. Windows Readlink is implemented via Windows DeviceIoControl(FSCTL_GET_REPARSE_POINT, ...) call, but that call does not work on network shares or inside of Docker container (see issues #18555 ad #19922 for details). But Raymond Chen suggests different approach: https://blogs.msdn.microsoft.com/oldnewthing/20100212-00/?p=14963/ - he suggests to use Windows I/O manager to dereferences the symbolic link. This appears to work for all normal symlinks, but also for network shares and inside of Docker container. This CL implements described procedure. I also had to adjust TestStatSymlinkLoop, because the test is expecting Stat to return syscall.ELOOP for symlink with a loop. But new Stat returns Windows error of ERROR_CANT_RESOLVE_FILENAME = 1921 instead. I could map ERROR_CANT_RESOLVE_FILENAME into syscall.ELOOP, but I suspect the former is broader than later. And ERROR_CANT_RESOLVE_FILENAME message text of "The name of the file cannot be resolved by the system." sounds fine to me. Fixes #10935 Fixes #18555 Fixes #19922 Change-Id: I979636064cdbdb9c7c840cf8ae73fe2c24499879 Reviewed-on: https://go-review.googlesource.com/41834Reviewed-by:
Harshavardhana <hrshvardhana@gmail.com> Reviewed-by:
Ian Lance Taylor <iant@golang.org>
-
- 06 May, 2017 3 commits
-
-
Egon Elbre authored
gcc on windows writes invalid location for .debug_gdb_scripts which causes the executable loading to fail. Fixes #20183 Change-Id: I5134013bc926b44a55b528f66ab79555855d1f4d Reviewed-on: https://go-review.googlesource.com/42651Reviewed-by:
Alex Brainman <alex.brainman@gmail.com> Run-TryBot: Alex Brainman <alex.brainman@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Ben Shi authored
As discussion in issue #19141, the addend should be the third argument of MULA. This patch fixes it in both the front end and the back end of the assembler. And also tests are added to the encoding test. Fixes #19141 Change-Id: Idbc6f338b8fdfcad97a135f27a98c5b375b27d43 Reviewed-on: https://go-review.googlesource.com/42028 Run-TryBot: Cherry Zhang <cherryyz@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Cherry Zhang <cherryyz@google.com>
-
Nigel Tao authored
Fixes #20249 Change-Id: I5dceaef31de96345f8e6c155e12775dc4cc31bfb Reviewed-on: https://go-review.googlesource.com/42790Reviewed-by:
-
- 05 May, 2017 8 commits
-
-
Robert Griesemer authored
NewPackage required through documentation that the package name not be blank (which wasn't true since each time we check a new package we create one with a blank name (api.go:350). NewPackage also asserted that a package name not be "_". While it is invalid for a package name to be "_", one could conceivably create a package named "_" through export data manipulation. Furthermore, it is ok to import a package with package path "_" as long as the package itself is not named "_". - removed misleading documentation - removed unnecessary assertion - added safety checks when we actually do the import Fixes #20231. Change-Id: I1eb1ab7b5e3130283db715374770cf05d749d159 Reviewed-on: https://go-review.googlesource.com/42852 Run-TryBot: Robert Griesemer <gri@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Alan Donovan <adonovan@google.com>
-
Robert Griesemer authored
Fixes #20230. Change-Id: I2e9b9e9d2540eb66c8411ac7910962933bc2c0e9 Reviewed-on: https://go-review.googlesource.com/42870 Run-TryBot: Robert Griesemer <gri@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Josh Bleecher Snyder <josharian@gmail.com>
-
Josh Bleecher Snyder authored
Generated hash and eq routines don't need nil checks. Prior to this CL, this was accomplished by temporarily incrementing the global variable disable_checknil. However, that increment lasted only the lifetime of the call to funccompile. After CL 41503, funccompile may do nothing but enqueue the function for compilation, resulting in nil checks being generated. Fix this by adding an explicit flag to a function indicating whether nil checks should be disabled for that function. While we're here, allow concurrent compilation with the -w and -W flags, since that was needed to investigate this issue. Fixes #20242 Change-Id: Ib9140c22c49e9a09e62fa3cf350f5d3eff18e2bd Reviewed-on: https://go-review.googlesource.com/42591 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Marvin Stenger <marvin.stenger94@gmail.com> Reviewed-by:
Robert Griesemer <gri@golang.org>
-
Carlos Eduardo Seo authored
The Power processor manual states that "Branches not from the last instruction of an aligned quadword and not to the first instruction of an aligned quadword cause inefficiencies in the IBuffer". This changes the function alignment from 8 to 16 bytes to comply with that. Fixes #18963 Change-Id: Ibce9bf8302110a86c6ab05948569af9ffdfcf4bb Reviewed-on: https://go-review.googlesource.com/36390 TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
David Chase <drchase@google.com> Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
-
Samuel Tan authored
Allow the predefined escapers "html", "urlquery", and "js" to be used in pipelines when they have no potential to affect the correctness or safety of the escaped pipeline output. Specifically: - "urlquery" may be used if it is the last command in the pipeline. - "html" may be used if it is the last command in the pipeline, and the pipeline does not occur in an unquoted HTML attribute value context. - "js" may be used in any pipeline, since it does not affect the merging of contextual escapers. This change will loosens the restrictions on predefined escapers introduced in golang.org/cl/37880, which will hopefully ease the upgrade path for existing template users. This change brings back the escaper-merging logic, and associated unit tests, that were removed in golang.org/cl/37880. However, a few notable changes have been made: - "_html_template_nospaceescaper" is no longer considered equivalent to "html", since the former escapes spaces, while the latter does not (see #19345). This change should not silently break any templates, since pipelines where this substituion will happen will already trigger an explicit error. - An "_eval_args_" internal directive has been added to handle pipelines containing a single explicit call to a predefined escaper, e.g. {{html .X}} (see #19353). Also, the HTMLEscape function called by the predefined text/template "html" function now escapes the NULL character as well. This effectively makes it as secure as the internal html/template HTML escapers (see #19345). While this change is backward-incompatible, it will only affect illegitimate uses of this escaper, since the NULL character is always illegal in valid HTML. Fixes #19952 Change-Id: I9b5570a80a3ea284b53901e6a1f842fc59b33d3a Reviewed-on: https://go-review.googlesource.com/40936Reviewed-by:Russ Cox <rsc@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Dieter Plaetinck authored
Execute incurs separate writes for each "step", e.g. each variable that needs to be printed, and the final newline. While it is correct to state that templates can be executed concurrently, there is a more subtle nuance that is easily missed: when writing to the same writer, the writes from concurrent execute calls can be interleaved, leading to unexpected output. Change-Id: I0abbd7960d8a8d15e109a8a3eeff3b43b852bbbf Reviewed-on: https://go-review.googlesource.com/37444Reviewed-by:
Rob Pike <r@golang.org>
-
David Crawshaw authored
The external darwin linker has been printing: ld: warning: -read_only_relocs cannot be used with x86_64 for a long time. Now that it is printed by CL 33301, we may as well get rid of it. Fixes #20246 Change-Id: I1147cf1ff197fdfda228a1349f13627bcf9fc72f Reviewed-on: https://go-review.googlesource.com/42730 Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Todd Neal <todd@tneal.org> Reviewed-by:
-
Alex Brainman authored
For .bss section symbol ldelf does not set P (raw symbol data). Make ldpe do the same. Change-Id: Ib3d558456f505ee568d0972465fa9b08b5794a87 Reviewed-on: https://go-review.googlesource.com/42631Reviewed-by:
-
- 04 May, 2017 10 commits
-
-
Josh Bleecher Snyder authored
Fixes #20228 Change-Id: I1893ae3e192da01f9befe5469b2a32e534a691ba Reviewed-on: https://go-review.googlesource.com/42592Reviewed-by:
-
Josh Bleecher Snyder authored
If we've already complained about a type T, don't complain again about further expressions involving it. Fixes #20245 and hopefully all of its ilk. Change-Id: Ic0abe8235d52e8a7ac40e3615aea8f3a54fd7cec Reviewed-on: https://go-review.googlesource.com/42690 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Ian Lance Taylor authored
We already set it for mips32 objects. The native ELF linker warns when linking PIC objects with non-PIC objects. Our objects are PIC, but we were not marking them as such. Fixes #20243. Change-Id: Ifab131200b263e4c72cf81f7b131a65ac02a13a9 Reviewed-on: https://go-review.googlesource.com/42710 Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by:
-
David Crawshaw authored
This change passes runtime.Version from the go tool to the compiler. If the versions do not match, the compilation fails. The result is a go tool from one GOROOT will complain loudly if it is invoked with a different GOROOT value. Only release versions are checked, so that when developing Go you can still use "go install cmd/go" and "go install cmd/compile" separately. Fixes #19064 Change-Id: I17e184d07d3c1092b1d9af53ba55ed3ecf67791d Reviewed-on: https://go-review.googlesource.com/42595 Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Adam Bender authored
- remove <code> from <pre> - replace `` with <code></code> Change-Id: I46f0aec8b7645e2ac8cb53bca73aed55441acd65 Reviewed-on: https://go-review.googlesource.com/42612Reviewed-by:
-
Brad Fitzpatrick authored
(Found by making time.Time uncomparable and rerunning std tests locally.) Change-Id: I4fa6fb0ba7334965362387e2f6541c17a27ac3aa Reviewed-on: https://go-review.googlesource.com/42616 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> Reviewed-by:
Martin Möhrmann <moehrmann@google.com> Reviewed-by:
Damien Neil <dneil@google.com> Reviewed-by:
Bryan Mills <bcmills@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
David Crawshaw authored
Before this change, building a GOROOT using make.bash, and then moving the entire to a new path confused the go tool. Correct operation of the go tool under these conditions required either running make.bash again (not always possible if the new location was owned by a different system user) or setting the GOROOT environment variable. Setting GOROOT is unfortunate and discouraged, as it makes it too easy to use the go tool from one GOROOT and the compiler from another GOROOT. With this change, the go tool finds its GOROOT relative to its own location, using os.Executable. It checks it is in a GOROOT by searching for the GOROOT/pkg/tool directory, to avoid two plausible situations: ln -s $GOROOT/bin/go /usr/local/bin/go and PATH=$HOME/bin:$PATH GOPATH=$HOME ln -s $GOROOT/bin/go $HOME/bin/go Additionally, if the current executable path is not in a GOROOT, the tool will follow any symlinks for the executable and check to see if its original path is a GOROOT. Fixes #18678 Change-Id: I151d7d449d213164f98193cc176b616849e6332c Reviewed-on: https://go-review.googlesource.com/42533 Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Brad Fitzpatrick authored
Updates text from https://golang.org/cl/42511 Updates #14395 Change-Id: I711100525e074ab360e577520280c37645db1c95 Reviewed-on: https://go-review.googlesource.com/42614Reviewed-by:
Emmanuel Odeke <emm.odeke@gmail.com> Reviewed-by:
-
Brad Fitzpatrick authored
Change-Id: Ie4c92da0e3cbb97d3d7e03c7d15196c34f58a2cd Reviewed-on: https://go-review.googlesource.com/42613Reviewed-by:
-
Josh Bleecher Snyder authored
Compile: package p var f = func(...A) Before this CL: x.go:3:13: type %!v(PANIC=runtime error: invalid memory address or nil pointer dereference) is not an expression x.go:3:17: undefined: A After this CL: x.go:3:13: type func(...<T>) is not an expression x.go:3:17: undefined: A Found with go-fuzz. Fixes #20233 Change-Id: Ibb232b3954c4091071440eba48b44c4022a8083f Reviewed-on: https://go-review.googlesource.com/42610 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Matthew Dempsky <mdempsky@google.com>
-
- 03 May, 2017 8 commits
-
-
Andreas Auernhammer authored
This change explicitly documents that DES, MD5, RC4 and SHA-1 are insecure / broken - at all or at least within a commonly used scenario. Fixes #14395 Change-Id: Id1d543c85d67968ba64ed7495313501953c3ef3a Reviewed-on: https://go-review.googlesource.com/42511Reviewed-by:
-
Brad Fitzpatrick authored
Updates bundled http2 to x/net/http2 git rev feeb485 for: http2: add all bad ciphers, use package constants https://golang.org/cl/42510 Updates #20213 Change-Id: I851453e3785e6b126db7a5c5eec2ebbbf61358ae Reviewed-on: https://go-review.googlesource.com/42494 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:Dmitry Savintsev <dsavints@gmail.com> Reviewed-by:
-
Steven Hartland authored
Add the ability to override the default file and directory from which certificates are loaded by setting the OpenSSL compatible environment variables: SSL_CERT_FILE, SSL_CERT_DIR. If the variables are set the default locations are not checked. Added new default file "/usr/local/etc/ssl/cert.pem" for FreeBSD. Certificates in the first valid location found for both file and directory are added, instead of only the first file location if a valid one was found, which is consistent with OpenSSL. Fixes #3905 Fixes #14022 Fixes #14311 Fixes #16920 Fixes #18813 - If user sets SSL_CERT_FILE. Change-Id: Ia24fb7c1c2ffff4338b4cf214bd040326ce27bb0 Reviewed-on: https://go-review.googlesource.com/36093Reviewed-by:
-
Shenghou Ma authored
Fixes #17935. Change-Id: I49b0f6cee29ea76ed62b8faa5d6d1f51be41bf84 Reviewed-on: https://go-review.googlesource.com/33301 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Josh Bleecher Snyder authored
When compiling the program: package p func _(){ *;:= } Before: x.go:4:3: syntax error: unexpected semicolon, expecting expression x.go:4:4: non-name *%!v(PANIC=runtime error: invalid memory address or nil pointer dereference) on left side of := x.go:5:1: syntax error: unexpected }, expecting expression After: x.go:4:3: syntax error: unexpected semicolon, expecting expression x.go:4:4: non-name *<N> on left side of := x.go:5:1: syntax error: unexpected }, expecting expression No test because: (1) we don't have a good mechanism to check for the absence of the string "PANIC" in an error message (2) the string "*<N>", while better, is itself ugly enough that I don't want to actively check for it (3) the bug isn't very important, the kind of thing only fuzzers encounter (4) the fix is obvious and trivial Fixes #20220 Change-Id: I35faa986b60b671414ee999d6264b06937f250e3 Reviewed-on: https://go-review.googlesource.com/42498 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by:David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Martin Möhrmann authored
Brings in chacha20poly1305 directory from golang.org/x/crypto revision 12e9ca725de4806fbda1610fd95aacad15bd6810, adding: CL 41862: chacha20poly1305: add runtime internal independent cpu feature detection CL 39952: add import comment Change-Id: Ic46ff24b081bc1c66b6317334d33180e33bfd318 Reviewed-on: https://go-review.googlesource.com/42513 Run-TryBot: Martin Möhrmann <moehrmann@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
feilengcui008 authored
Change-Id: Ica7179d225c1fb79381f82f58ea5050ac6418b9c Reviewed-on: https://go-review.googlesource.com/42493Reviewed-by:
Daniel Martí <mvdan@mvdan.cc> Reviewed-by:
-
Filip Gruszczyński authored
This allows to pre-allocate the final size of the hashmap and avoid re-allocating as we insert entries. Furthermore for the current implementation of the hashmap it allows avoiding several rounds of evacuating hashmap entries after each re-allocation. DecodeComplex128Slice-8 51.9µs ± 1% 51.9µs ± 2% ~ (p=0.797 n=30+29) DecodeFloat64Slice-8 31.5µs ± 2% 31.6µs ± 2% ~ (p=0.050 n=28+28) DecodeInt32Slice-8 32.0µs ± 2% 31.9µs ± 3% ~ (p=0.666 n=29+28) DecodeStringSlice-8 57.7µs ± 2% 57.8µs ± 3% ~ (p=0.780 n=27+30) DecodeInterfaceSlice-8 498µs ± 2% 495µs ± 2% ~ (p=0.070 n=28+29) DecodeMap-8 300µs ± 2% 230µs ± 5% -23.31% (p=0.000 n=27+27) Updates #19525 Change-Id: Ia7233da49f05bae7a86c064d9ecebca966f5f2f7 Reviewed-on: https://go-review.googlesource.com/40113 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
- 02 May, 2017 5 commits
-
-
Filip Gruszczynski authored
Because the hint parameter is supposed to be treated purely as a hint, if it doesn't meet the requirements we disregard it and continue as if there was no hint at all. Fixes #19926 Change-Id: I86e7f99472fad6b99ba4e2fd33e4a9e55d55115e Reviewed-on: https://go-review.googlesource.com/40854 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Keith Randall <khr@golang.org>
-
Alberto Donizetti authored
There's no Settings->Agreement path for PolyGerrit users, but if we link directly to the page in the instructions, Gerrit will inform them that they can access the page by switching to the old UI. Fixes #20207 Change-Id: I0887ee854e4ac5975b5f305adb6259b81b41618f Reviewed-on: https://go-review.googlesource.com/42412Reviewed-by:
-
Brad Fitzpatrick authored
Change-Id: Ifbfd71a9c0d447e22c369c9d1209063b2a5c657b Reviewed-on: https://go-review.googlesource.com/42490 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Michael Munday authored
This CL modifies how MOV[DWHB] instructions that store a constant to memory are assembled to avoid them clobbering the condition code (flags). It also modifies zeroAuto to use MOVD instructions instead of CLEAR (which is assembled as XC). MOV[DWHB]storeconst ops also no longer clobbers flags. Note: this CL modifies the assembler so that it can no longer handle immediates outside the range of an int16 or offsets from SB, which reflects what the machine instructions support. The compiler doesn't need this capability any more and I don't think this affects any existing assembly, but it is easy to workaround if it does. Fixes #20187. Change-Id: Ie54947ff38367bd6a19962bf1a6d0296a4accffb Reviewed-on: https://go-review.googlesource.com/42179Reviewed-by:
-
Bryan C. Mills authored
Use Load instead of LoadOrStore in the fast path to save 1 alloc/op for existing keys. name old time/op new time/op delta IntAdd 6.39ns ± 7% 6.97ns ±19% ~ (p=0.105 n=8+8) IntAdd-6 12.3ns ± 1% 12.2ns ± 1% ~ (p=0.396 n=7+7) IntSet 6.41ns ± 6% 6.94ns ±21% ~ (p=0.168 n=8+8) IntSet-6 12.1ns ± 3% 11.7ns ± 9% ~ (p=0.496 n=7+8) FloatAdd 14.3ns ± 9% 14.7ns ± 4% ~ (p=0.088 n=8+8) FloatAdd-6 36.5ns ± 1% 36.6ns ± 0% ~ (p=0.709 n=7+6) FloatSet 6.59ns ± 7% 6.47ns ± 7% ~ (p=0.397 n=8+7) FloatSet-6 12.2ns ± 1% 12.2ns ± 2% ~ (p=0.748 n=7+7) StringSet 67.8ns ± 6% 68.7ns ± 6% ~ (p=0.342 n=8+8) StringSet-6 41.8ns ± 5% 41.7ns ± 5% ~ (p=0.979 n=8+8) MapSet 294ns ± 6% 234ns ± 4% -20.35% (p=0.000 n=8+8) MapSet-6 95.8ns ± 2% 89.4ns ± 3% -6.73% (p=0.000 n=8+8) MapSetDifferent 1.31µs ± 5% 1.07µs ± 4% -18.21% (p=0.000 n=8+8) MapSetDifferent-6 260ns ± 8% 210ns ± 9% -19.44% (p=0.000 n=8+8) MapSetString 294ns ± 6% 236ns ± 4% -19.92% (p=0.000 n=8+8) MapSetString-6 95.6ns ± 2% 89.9ns ± 2% -5.97% (p=0.000 n=7+8) MapAddSame 1.46µs ± 3% 1.46µs ± 5% ~ (p=0.721 n=8+8) MapAddSame-6 328ns ± 6% 330ns ± 4% ~ (p=0.776 n=8+8) MapAddDifferent 4.89µs ± 7% 4.98µs ± 6% ~ (p=0.505 n=8+8) MapAddDifferent-6 1.02µs ± 3% 1.01µs ± 4% ~ (p=0.352 n=7+8) MapAddSameSteadyState 62.1ns ± 7% 60.8ns ± 4% ~ (p=0.521 n=8+8) MapAddSameSteadyState-6 38.1ns ± 3% 37.7ns ± 0% ~ (p=0.185 n=7+6) MapAddDifferentSteadyState 290ns ± 5% 293ns ± 4% ~ (p=0.515 n=8+8) MapAddDifferentSteadyState-6 63.0ns ± 7% 63.7ns ±11% ~ (p=0.482 n=7+8) RealworldExpvarUsage 7.39µs ± 5% 7.51µs ± 5% ~ (p=0.382 n=8+8) RealworldExpvarUsage-6 3.07µs ±28% 3.04µs ±43% ~ (p=0.798 n=8+8) name old alloc/op new alloc/op delta IntAdd 0.00B 0.00B ~ (all equal) IntAdd-6 0.00B 0.00B ~ (all equal) IntSet 0.00B 0.00B ~ (all equal) IntSet-6 0.00B 0.00B ~ (all equal) FloatAdd 0.00B 0.00B ~ (all equal) FloatAdd-6 0.00B 0.00B ~ (all equal) FloatSet 0.00B 0.00B ~ (all equal) FloatSet-6 0.00B 0.00B ~ (all equal) StringSet 16.0B ± 0% 16.0B ± 0% ~ (all equal) StringSet-6 16.0B ± 0% 16.0B ± 0% ~ (all equal) MapSet 48.0B ± 0% 32.0B ± 0% -33.33% (p=0.000 n=8+8) MapSet-6 48.0B ± 0% 32.0B ± 0% -33.33% (p=0.000 n=8+8) MapSetDifferent 192B ± 0% 128B ± 0% -33.33% (p=0.000 n=8+8) MapSetDifferent-6 192B ± 0% 128B ± 0% -33.33% (p=0.000 n=8+8) MapSetString 48.0B ± 0% 32.0B ± 0% -33.33% (p=0.000 n=8+8) MapSetString-6 48.0B ± 0% 32.0B ± 0% -33.33% (p=0.000 n=8+8) MapAddSame 480B ± 0% 480B ± 0% ~ (all equal) MapAddSame-6 480B ± 0% 480B ± 0% ~ (all equal) MapAddDifferent 1.09kB ± 0% 1.09kB ± 0% ~ (all equal) MapAddDifferent-6 1.09kB ± 0% 1.09kB ± 0% ~ (all equal) MapAddSameSteadyState 0.00B 0.00B ~ (all equal) MapAddSameSteadyState-6 0.00B 0.00B ~ (all equal) MapAddDifferentSteadyState 0.00B 0.00B ~ (all equal) MapAddDifferentSteadyState-6 0.00B 0.00B ~ (all equal) RealworldExpvarUsage 0.00B 0.00B ~ (all equal) RealworldExpvarUsage-6 0.00B 0.00B ~ (all equal) name old allocs/op new allocs/op delta IntAdd 0.00 0.00 ~ (all equal) IntAdd-6 0.00 0.00 ~ (all equal) IntSet 0.00 0.00 ~ (all equal) IntSet-6 0.00 0.00 ~ (all equal) FloatAdd 0.00 0.00 ~ (all equal) FloatAdd-6 0.00 0.00 ~ (all equal) FloatSet 0.00 0.00 ~ (all equal) FloatSet-6 0.00 0.00 ~ (all equal) StringSet 1.00 ± 0% 1.00 ± 0% ~ (all equal) StringSet-6 1.00 ± 0% 1.00 ± 0% ~ (all equal) MapSet 3.00 ± 0% 2.00 ± 0% -33.33% (p=0.000 n=8+8) MapSet-6 3.00 ± 0% 2.00 ± 0% -33.33% (p=0.000 n=8+8) MapSetDifferent 12.0 ± 0% 8.0 ± 0% -33.33% (p=0.000 n=8+8) MapSetDifferent-6 12.0 ± 0% 8.0 ± 0% -33.33% (p=0.000 n=8+8) MapSetString 3.00 ± 0% 2.00 ± 0% -33.33% (p=0.000 n=8+8) MapSetString-6 3.00 ± 0% 2.00 ± 0% -33.33% (p=0.000 n=8+8) MapAddSame 11.0 ± 0% 11.0 ± 0% ~ (all equal) MapAddSame-6 11.0 ± 0% 11.0 ± 0% ~ (all equal) MapAddDifferent 31.0 ± 0% 31.0 ± 0% ~ (all equal) MapAddDifferent-6 31.0 ± 0% 31.0 ± 0% ~ (all equal) MapAddSameSteadyState 0.00 0.00 ~ (all equal) MapAddSameSteadyState-6 0.00 0.00 ~ (all equal) MapAddDifferentSteadyState 0.00 0.00 ~ (all equal) MapAddDifferentSteadyState-6 0.00 0.00 ~ (all equal) RealworldExpvarUsage 0.00 0.00 ~ (all equal) RealworldExpvarUsage-6 0.00 0.00 ~ (all equal) https://perf.golang.org/search?q=upload:20170501.1 Change-Id: I28fc3906473f2b7307f6d1ae05a8d9b01ef8a6f8 Reviewed-on: https://go-review.googlesource.com/42211 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
