- 24 Mar, 2015 18 commits
-
-
Shenghou Ma authored
REGRT1 and REGRT2 are also reserved on arm64 for runtime (duffzero and duffcopy). Change-Id: If098527a7f29d16f94bdcec05fd55950b9076e35 Signed-off-by:
Shenghou Ma <minux@golang.org> Reviewed-on: https://go-review.googlesource.com/7977Reviewed-by:
Aram Hăvărneanu <aram@mgk.ro> Reviewed-by:
Dave Cheney <dave@cheney.net> Reviewed-by:
Rob Pike <r@golang.org>
-
Shenghou Ma authored
Fixes newly introduced test on linux/arm64 because linux/arm64 doesn't have the getpgrp syscall. Getpgid(0) is documented to be equivalent to Getpgrp. Change-Id: I8f30f4f8de8c32fe04a29c9c4a9330d4e4e6b46d Signed-off-by:
-
David Crawshaw authored
Previously the extra m needed for cgo callbacks was created on the first callback. This works for cgo, however the cgocallback mechanism is also borrowed by badsignal which can run before any cgo calls are made. Now we initialize the extra M at runtime startup before any signal handlers are registered, so badsignal cannot be called until the extra M is ready. Updates #10207. Change-Id: Iddda2c80db6dc52d8b60e2b269670fbaa704c7b3 Reviewed-on: https://go-review.googlesource.com/7978Reviewed-by:
Ian Lance Taylor <iant@golang.org> Run-TryBot: David Crawshaw <crawshaw@golang.org>
-
Shenghou Ma authored
Change-Id: I5bfeeef53665b1ef79ec0e4a6ae0eb4aa3e95603 Signed-off-by:
Brad Fitzpatrick <bradfitz@golang.org>
-
Rick Hudson authored
There are calls to stdcall when the GC thinks the world is stopped and stdcall write a *g for the CPU profiler. This produces a write barrier but the GC is not prepared to deal with write barriers when it thinks the world is stopped. Since the g is on allg it does not need a write barrier to keep it alive so eliminate the write barrier. Change-Id: I937633409a66553d7d292d87d7d58caba1fad0b6 Reviewed-on: https://go-review.googlesource.com/7979Reviewed-by:
Austin Clements <austin@google.com> Run-TryBot: Rick Hudson <rlh@golang.org>
-
Alex Brainman authored
The test is a simple reproduction of issue 9356. Update #8948. Update #9356. Change-Id: Ia77bc36d12ed0c3c4a8b1214cade8be181c9ad55 Reviewed-on: https://go-review.googlesource.com/7618Reviewed-by:
-
Shenghou Ma authored
Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e33 Signed-off-by:
-
Shenghou Ma authored
Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e32 Signed-off-by:
Alex Brainman <alex.brainman@gmail.com> Reviewed-by:
-
Shenghou Ma authored
Fixes #4069. Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e31 Signed-off-by:
-
Shenghou Ma authored
Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e30 Signed-off-by:
-
Shenghou Ma authored
Signed-off-by:
-
Shenghou Ma authored
Update #4069: this CL fixes the issue on windows/386. Signed-off-by:
-
Shenghou Ma authored
windows/386 also wants underscore prefix for external names. This CL is in preparation of external linking support. Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e2d Signed-off-by:
-
Shenghou Ma authored
When external linking, we must link to implib provided by mingw, so we must use properly decorated names for stdcalls. Because the feature is only used in the runtime, I've designed a new decoration scheme so that we can use the same decorated name for both 386 and amd64. A stdcall function named FooEx from bar16.dll which takes 3 parameters will be imported like this: //go:cgo_import_dynamic runtime._FooEx FooEx%3 "bar16.dll" Depending on the size of uintptr, the linker will later transform it to _FooEx@12 or _FooEx@24. This is in prepration for the next CL that adds external linking support for windows/386. Change-Id: I2d2ea233f976aab3f356f9b508cdd246d5013e2c Signed-off-by:
-
Shenghou Ma authored
Update #10215. Change-Id: Ib588f90279a4ef5461492553d50ad77c742b3560 Signed-off-by:
-
Nigel Tao authored
This is in preparation for inlining the color.YCbCrToRGB calls in a follow-up change. Change-Id: I30750ace11a8ef6016b3c1e0b4bfdbcc8151f9a5 Reviewed-on: https://go-review.googlesource.com/7951Reviewed-by:
-
Matthew Dempsky authored
The debug/dwarf and encoding/asn1 examples were added in 2009, a few months before Go added implicit semicolons, and never updated. The go/ast node types have always been named just "Expr", "Stmt", and "Decl", so the comments about "ExprNode", "StmtNode", and "DeclNode" were likely just mistaken because the interface tag methods are "exprNode", "stmtNode", and "declNode", respectively. Change-Id: I9d138cc3a16c1a51453da1406914d7b320bf6270 Reviewed-on: https://go-review.googlesource.com/7980Reviewed-by:
-
Matthew Dempsky authored
Instead of reimplementing chained hash tables, just use maps. Use bool instead of uint8 for variables only set to 0 or 1. Fix parsing of `import foo "foo" // indirect` lines. Previously, this was treated as an import of package path `"foo" // indirect`, which could result in the cycle-detection code failing to detect a cycle because it would be treated as a separate package from `"foo"`. Also, since there are theoretically multiple quoted forms for a package path, use strconv.Unquote to normalize them. Side benefit: Unquote will complain if any trailing comments sneak back in. Aside: For most Go archives, Go package data is only present in the __.PKGDEF member, but unless -u is used, ldpkg is only called on the _go_.6 member. Consequently, importcycles is a no-op when -u isn't used as it has no package data to inspect. Change-Id: I7076cf91a66726a8d9c5676adfea13c5532001fa Reviewed-on: https://go-review.googlesource.com/7002Reviewed-by:
-
- 23 Mar, 2015 13 commits
-
-
Nigel Tao authored
The imageutil.DrawYCbCr function lives in an internal package because it is needed by both the image/draw and image/jpeg packages, but it doesn't seem right for one of those two to depend on the other. It could eventually go into the image package, but that would require committing to an API for the rest of Go 1.x. Change-Id: I7b12555c970d86409365e99eef9360702aaffa30 Reviewed-on: https://go-review.googlesource.com/7925Reviewed-by:
-
Sebastien Binet authored
Previously, running 'go get' with a local path would correctly download the package but fail to install it. This is because a sticky error - resulting from discovering that the package needed to be downloaded - was still around. Theoretically, such sticky errors would be cleared but they weren't because the map tracking these errors were indexed with the correct canonical import path of the package (e.g. "ex.com/x/pkg") whereas the clearing was done with the local path (e.g. "./pkg".) Always use the canonical import path. Fixes #9767 Change-Id: Ia0e8a51ac591d4c833d11285da5b767ef7ed8ad2 Reviewed-on: https://go-review.googlesource.com/6266Reviewed-by:
-
Brad Fitzpatrick authored
Fixes #10224 Change-Id: I21037379b4667575e51ab0b6b683138c505c3f68 Reviewed-on: https://go-review.googlesource.com/7960Reviewed-by:
David Crawshaw <crawshaw@golang.org>
-
Brad Fitzpatrick authored
Change-Id: I4d35a96ace71b68a91561c08fd7331859d9dd19d Reviewed-on: https://go-review.googlesource.com/5593Reviewed-by:
Dmitry Vyukov <dvyukov@google.com>
-
Brad Fitzpatrick authored
Change-Id: Ie45d45649ea978becf84eb911cf4ace1dd237d9e Reviewed-on: https://go-review.googlesource.com/7918Reviewed-by:
-
Robert Griesemer authored
Change-Id: Ib8df771957f9a5bf00bee7eb8cb361d4acd5813f Reviewed-on: https://go-review.googlesource.com/7911Reviewed-by:
-
David Crawshaw authored
This test was introduced in cl/5130 and broke the darwin/arm builder. Also check some errors, which was making the failure hard to decipher. Change-Id: Ifb1d60b9971782cf8d2e979d83f8a81249d7ee9b Reviewed-on: https://go-review.googlesource.com/7932Reviewed-by:
-
Joe Harrison authored
Since there is internal usage of svg files in cmd/trace served via HTTP, add it to the (small) map of mime types. MIME type as documented at http://www.w3.org/TR/SVG11/mimereg.html Change-Id: If14f07cbe2e55d8d68af663f30920a530e2e7974 Reviewed-on: https://go-review.googlesource.com/7942Reviewed-by:
-
Michael Matloob authored
This is a port of RE2's bitstate backtracker, which triggers under the same conditions that the RE2 backtracker triggers. However I wasn't sure how to port over some of the optimizations in the RE2 backtracker, and there is a ~2% penalty on benchmarks that don't trigger the backtracker. benchmark old ns/op new ns/op delta BenchmarkLiteral 312 189 -39.42% BenchmarkNotLiteral 4435 3001 -32.33% BenchmarkMatchClass 5758 4378 -23.97% BenchmarkMatchClass_InRange 5385 4084 -24.16% BenchmarkReplaceAll 5291 3505 -33.76% BenchmarkAnchoredLiteralShortNonMatch 190 200 +5.26% BenchmarkAnchoredLiteralLongNonMatch 189 194 +2.65% BenchmarkAnchoredShortMatch 479 304 -36.53% BenchmarkAnchoredLongMatch 478 499 +4.39% BenchmarkOnePassShortA 791 798 +0.88% BenchmarkNotOnePassShortA 3202 1571 -50.94% BenchmarkOnePassShortB 614 633 +3.09% BenchmarkNotOnePassShortB 2685 881 -67.19% BenchmarkOnePassLongPrefix 152 154 +1.32% BenchmarkOnePassLongNotPrefix 505 533 +5.54% BenchmarkMatchEasy0_32 139 171 +23.02% BenchmarkMatchEasy0_1K 653 1797 +175.19% BenchmarkMatchEasy0_32K 12032 13346 +10.92% BenchmarkMatchEasy0_1M 462882 461272 -0.35% BenchmarkMatchEasy0_32M 15015339 15365238 +2.33% BenchmarkMatchEasy1_32 122 168 +37.70% BenchmarkMatchEasy1_1K 3339 2612 -21.77% BenchmarkMatchEasy1_32K 72330 71721 -0.84% BenchmarkMatchEasy1_1M 2545410 2652284 +4.20% BenchmarkMatchEasy1_32M 80072063 82609750 +3.17% BenchmarkMatchMedium_32 2359 1980 -16.07% BenchmarkMatchMedium_1K 75939 58593 -22.84% BenchmarkMatchMedium_32K 2450907 2501106 +2.05% BenchmarkMatchMedium_1M 78707697 80174418 +1.86% BenchmarkMatchMedium_32M 2535146010 2570896441 +1.41% BenchmarkMatchHard_32 4297 2960 -31.11% BenchmarkMatchHard_1K 133592 88997 -33.38% BenchmarkMatchHard_32K 4240445 4336907 +2.27% BenchmarkMatchHard_1M 136187006 139350238 +2.32% BenchmarkMatchHard_32M 4350855890 4478537306 +2.93% benchmark old MB/s new MB/s speedup BenchmarkMatchEasy0_32 228.74 186.11 0.81x BenchmarkMatchEasy0_1K 1565.91 569.64 0.36x BenchmarkMatchEasy0_32K 2723.31 2455.10 0.90x BenchmarkMatchEasy0_1M 2265.32 2273.22 1.00x BenchmarkMatchEasy0_32M 2234.68 2183.79 0.98x BenchmarkMatchEasy1_32 261.08 190.22 0.73x BenchmarkMatchEasy1_1K 306.59 391.91 1.28x BenchmarkMatchEasy1_32K 453.03 456.88 1.01x BenchmarkMatchEasy1_1M 411.95 395.35 0.96x BenchmarkMatchEasy1_32M 419.05 406.18 0.97x BenchmarkMatchMedium_32 13.56 16.16 1.19x BenchmarkMatchMedium_1K 13.48 17.48 1.30x BenchmarkMatchMedium_32K 13.37 13.10 0.98x BenchmarkMatchMedium_1M 13.32 13.08 0.98x BenchmarkMatchMedium_32M 13.24 13.05 0.99x BenchmarkMatchHard_32 7.45 10.81 1.45x BenchmarkMatchHard_1K 7.67 11.51 1.50x BenchmarkMatchHard_32K 7.73 7.56 0.98x BenchmarkMatchHard_1M 7.70 7.52 0.98x BenchmarkMatchHard_32M 7.71 7.49 0.97x Fixes #4154 Change-Id: Iff7fb9507f0872b320d08afc08679751ed1b28bc Reviewed-on: https://go-review.googlesource.com/2153Reviewed-by:
Russ Cox <rsc@golang.org>
-
Michael MacInnis authored
On Unix, when placing a child in a new process group, allow that group to become the foreground process group. Also, allow a child process to join a specific process group. When setting the foreground process group, Ctty is used as the file descriptor of the controlling terminal. Ctty has been added to the BSD and Solaris SysProcAttr structures and the handling of Setctty changed to match Linux. Change-Id: I18d169a6c5ab8a6a90708c4ff52eb4aded50bc8c Reviewed-on: https://go-review.googlesource.com/5130 Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by:
-
David Crawshaw authored
Issue #10043 was resolved by https://golang.org/cl/7072. This test now passes. Change-Id: I7bdef63d9ee4edcede23613a14a2ddce14018f34 Reviewed-on: https://go-review.googlesource.com/7931Reviewed-by:
-
Mikio Hara authored
Fixes #10214. Change-Id: If29d26c590eb53d4976e0a6ace0ed33aaf0c3e80 Reviewed-on: https://go-review.googlesource.com/7924Reviewed-by:
-
Brad Fitzpatrick authored
ServeContent ignored zero time.Time{} values when generating Last-Modified response headers and checking If-Modified-Since request headers. Do the same for a time.Time representing the Unix epoch zero value, as this is a common bogus value. Callers who really want to send that value (incredibly unlikely) can add a nanosecond to it and it will be truncated to second granularity anyway. Fixes #9842 Change-Id: I69f697bfc4017404a92a34e3fe57e2711c1e299d Reviewed-on: https://go-review.googlesource.com/7915Reviewed-by:David Symonds <dsymonds@golang.org>
-
- 21 Mar, 2015 3 commits
-
-
Nigel Tao authored
Change-Id: I9968f53a8286a0e5ccc197a9b5fae499e2f95326 Reviewed-on: https://go-review.googlesource.com/7790Reviewed-by:
-
Joel Sing authored
Change-Id: I5b057d16eed1b364e608ff0fd74de323da6492bc Reviewed-on: https://go-review.googlesource.com/7679Reviewed-by:
-
Dave Cheney authored
Implement runtime.atomicand8 for amd64p32 which was overlooked in CL 7861. Change-Id: Ic7eccddc6fd6c4682cac1761294893928f5428a2 Reviewed-on: https://go-review.googlesource.com/7920Reviewed-by:
-
- 20 Mar, 2015 6 commits
-
-
Robert Griesemer authored
to map element keys Composite literals containing element values that are themselves composite literals may leave away the element's literal types if they are identical to the enclosing composite literal's element type. (http://golang.org/ref/spec#Composite_literals) When we made this change, we forgot to apply the analogous rule to map literal keys. This change generalizes that rule. Added more examples, including one showing the recursive application of the elision rules. This is a fully backward-compatible language change. It was discussed some time back. Fixes #8589. To be submitted once all compilers accept the extension. Change-Id: I4d45b64b5970f0d5501572945d5a097e64a9458b Reviewed-on: https://go-review.googlesource.com/2591Reviewed-by:
-
Jeremy Schlatter authored
Change-Id: I5e11a76ebb20284618144be2ef5449d3202b6222 Reviewed-on: https://go-review.googlesource.com/7900Reviewed-by:
-
Robert Griesemer authored
Per pending https://go-review.googlesource.com/2591 . Change-Id: I1ce9d1c629e9fc43dbd862b3433aa5840f46656c Reviewed-on: https://go-review.googlesource.com/2621Reviewed-by:
Alan Donovan <adonovan@google.com>
-
Robert Griesemer authored
To use a pure Go implementation of the low-level arithmetic functions (when no platform-specific assembly implementations are available), set the build tag math_big_pure_go. This will make it easy to vendor the math/big package where no assembly is available (for instance for use with gc which relies on 1.4 functionality for now). Change-Id: I91e17c0fdc568a20ec1512d7c64621241dc60c17 Reviewed-on: https://go-review.googlesource.com/7856Reviewed-by:
-
Russ Cox authored
These can be implemented with just a compare and a move instruction. Do so, avoiding the overhead of a call into the runtime. These assertions are a significant cost in Go code that uses interface{} as a safe alternative to C's void* (or unsafe.Pointer), such as the current version of the Go compiler. *T here includes pointer to T but also any Go type represented as a single pointer (chan, func, map). It does not include [1]*T or struct{*int}. That requires more work in other parts of the compiler; there is a TODO. Change-Id: I7ff681c20d2c3eb6ad11dd7b3a37b1f3dda23965 Reviewed-on: https://go-review.googlesource.com/7862Reviewed-by: -
Russ Cox authored
This CL moves the bulk of the code that has been copy-and-pasted since the initial 386 port back into a shared place, cutting 5 copies to 1. The motivation here is not cleanup per se but instead to reduce the cost of introducing changes in shared concepts like regalloc or general expression evaluation. For example, a change after this one will implement x.(*T) without a call into the runtime. This CL makes that followup work 5x easier. The single copy still has more special cases for architecture details than I'd like, but having them called out explicitly like this at least opens the door to generalizing the conditions and smoothing out the distinctions in the future. This is a LARGE CL. I started by trying to pull in one function at a time in a sequence of CLs and it became clear that everything was so interrelated that it had to be moved as a whole. Apologies for the size. It is not clear how many more releases this code will matter for; eventually it will be replaced by Keith's SSA work. But as noted above, the deduplication was necessary to reduce the cost of working on the current code while we have it. Passes tests on amd64, 386, arm, and ppc64le. Can build arm64 binaries but not tested there. Being able to build binaries means it is probably very close. Change-Id: I735977f04c0614f80215fb12966dfe9bbd1f5861 Reviewed-on: https://go-review.googlesource.com/7853Reviewed-by:
-
