Implement global_asm!() (RFC 1548) by mrhota · Pull Request #40702 · rust-lang/rust

mrhota · 2017-03-21T15:11:03Z

This is a first attempt. One (potential) problem I haven't solved is how to handle multiple usages of global_asm! in a module/crate. It looks like LLVMSetModuleInlineAsm overwrites module asm, and LLVMAppendModuleInlineAsm is not provided in LLVM C headers 😦

I can provide more detail as needed, but honestly, there's not a lot going on here.

r? @eddyb

CC @Amanieu @jackpot51

Tracking issue: #35119

rust-highfive · 2017-03-21T15:11:18Z

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @pnkfelix (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

mrhota · 2017-03-21T15:13:05Z

I will rebase this evening.

eddyb · 2017-03-21T15:58:30Z

cc @nagisa @michaelwoerister

mattico · 2017-03-21T17:51:04Z

If an LLVM API is not provided by the C API, you can add a C++ wrapper in rustllvm and the corresponding Rust types in librustc_llvm. Look at the rest of the file to see the conventions that are used (to_rust, from_rust, etc.)

mrhota · 2017-03-21T19:24:30Z

Thank you @mattico!

@steveklabnik I updated the unstable-book, so you might want to look it over.

parched · 2017-03-21T23:24:00Z

+my_asm_func:
+ ret
+__other_cancel:
+ jmp __cancel


This would need to be mangled, wouldn't it?

ah, yes, we need a #[no_mangle]. Thanks!

mrhota · 2017-03-22T05:22:07Z

We have a feature-gate test and a couple codegen tests! We also have a new unstable-book page. My initial problems/questions have been resolved.

Amanieu

I think a test should be added which uses include_str! with global_asm!.

Amanieu · 2017-03-22T22:50:42Z

This should be include_str! (missing !).

mrhota · 2017-03-23T00:17:28Z

r? @pnkfelix

mrhota · 2017-03-23T00:33:30Z

@Amanieu we have a new test using include_str!. Hopefully it's what you had in mind.

Amanieu · 2017-03-23T00:35:52Z

@mrhota Yes, that's what I had in mind.

mrhota · 2017-03-24T18:15:08Z

@eddyb @nagisa @michaelwoerister any thoughts here? not sure who should review

nagisa

Seems good to me overall.

I’d like, however, more tests. Namely to ensure stuff like these compile. Compilation here is the important distinction, as codegen tests do not actually go through the assembler, and only generate LLVM-IR, which does not insure the actual codegen works.

// cfg-gating works properly
#[cfg(target_arch="x86")] global_asm! { ... }
#[cfg(target_arch="x86_64")] global_asm! { ... }

// codegen tests for all borderline T-1 architectures
global_asm!("foo blt foo") // arm
global_asm!("foo: j foo") // mips
global_asm!("foo: b foo") // ppc

nagisa · 2017-03-24T18:31:42Z

Since assembly is for x86, this list should be significantly longer.
Namely we also support: MIPS, PowerPC, SystemZ, wasm, asmjs, nacl, nvptx, avr, msp430 and probably more stuff I forgot to mention.

A good approach would be to git grep '// ignore-' src/tests and add every single suffix that's seems unrelated to x86.

nagisa · 2017-03-24T18:32:02Z

Since assembly is for x86, this list should be significantly longer.

I copied this from another codegen test, I think. I'll fill out the list.

nagisa · 2017-03-24T18:36:32Z

Seems to be unused/unnecessary?

mrhota · 2017-03-25T00:54:35Z

@nagisa I might be missing something, but what's the point of testing full-compile in addition to codegen? The real meat of this feature enables insertion of module asm into llvm-ir by using global_asm!, which the codegen tests ensure.

If we test full-compile, then we're just testing that LLVM does what it says it'll do with module asms. If the compile fails at that point, rustc is off the hook, since it's either an LLVM problem, or the programmer provided bad assembly.

global_asm!("put on your time travel trousers!");

should in fact go all the way through codegen, and should produce

... module asm "put on your time travel trousers!"

which will obviously fail to assemble, but that's not our problem. The assembly template provided to global_asm! is (semantically) completely inaccessible to rustc and LLVM.

But, again, the codegen tests ensure rustc faithfully has handed the provided assembly template to LLVM, which itself promises faithfully to deliver it to the assembler.

nagisa · 2017-03-25T09:07:03Z

If the compile fails at that point, rustc is off the hook, since it's either an LLVM problem, or the programmer provided bad assembly.

This is irrelevant from the perspective of the end user – they use rustc, not LLVM. You can assign blame to some dependency, but the fact of dependencies not working propagates to the end products.

Now, I don’t feel strongly about this, since this is will be unstable for a foreseeable future, but such tests will have to materialise before this gets stabilised. Might as well materialise them right now, then, right?

EDIT: It might as well point out the fact of us having LLVM misconfigured in a way that it is unable to work with the module level assembly :)

mrhota · 2017-03-26T22:12:51Z

@nagisa I'm at a loss; I have no idea why the builder keeps failing codegen tests since I added more // ignore-s. It passes locally.

mrhota · 2017-03-27T20:42:06Z

r? @nikomatsakis you seemed to be somewhat interested in how this feature turns out

bors · 2017-03-29T18:03:19Z

☔ The latest upstream changes (presumably #40899) made this pull request unmergeable. Please resolve the merge conflicts.

nagisa · 2017-03-29T20:06:40Z

I'm at a loss; I have no idea why the builder keeps failing codegen tests since I added more // ignore-s. It passes locally.

I’ve got no idea why it fails either. I’ll make a mental note to try and build this locally, but am likely to forget.

mrhota · 2017-03-30T01:51:27Z

@nagisa thanks. I ran the failing docker container's build on my mac and it passed. I ran the tests on my mac, too, and they pass.

mrhota · 2017-03-30T14:30:48Z

@nagisa OK, resolved build failure. Was that the last thing on your checklist?

@Amanieu github indicates that you requested changes. I think I resolved the issues you mentioned, but I don't know if you manually have to check the box...

Amanieu · 2017-03-30T15:22:50Z

I don't see a box to check. Anyways, I'm happy with the changes.

bors · 2017-04-13T15:26:58Z

💔 Test failed - status-appveyor

TimNN · 2017-04-13T15:28:28Z

@bors retry

fatal: unable to access 'https://github.com/rust-lang/rust.git/': Unknown SSL protocol error in connection to github.com:443 ?!

bors · 2017-04-13T15:30:46Z

⌛ Testing commit 63a0747 with merge a2e97b2...

bors · 2017-04-13T15:33:44Z

💔 Test failed - status-appveyor

mrhota · 2017-04-13T15:38:01Z

same SSL error on Windows

alexcrichton · 2017-04-13T15:43:08Z

@bors: retry

appveyor git failure - Tracking issue for spurious network failures on bots #40474

bors · 2017-04-13T15:45:01Z

⌛ Testing commit 63a0747 with merge a5d5852...

@eddyb

Implement global_asm!() (RFC 1548) This is a first attempt. ~~One (potential) problem I haven't solved is how to handle multiple usages of `global_asm!` in a module/crate. It looks like `LLVMSetModuleInlineAsm` overwrites module asm, and `LLVMAppendModuleInlineAsm` is not provided in LLVM C headers 😦~~ I can provide more detail as needed, but honestly, there's not a lot going on here. r? @eddyb CC @Amanieu @jackpot51 Tracking issue: #35119

bors · 2017-04-13T17:32:19Z

💔 Test failed - status-appveyor

frewsxcv · 2017-04-13T17:56:30Z

That appveyor failure looks legit.

mrhota · 2017-04-13T17:56:45Z

failures:
---- [run-pass] run-pass\simple_global_asm.rs stdout ----
error: compilation failed!
status: exit code: 101
command: PATH="C:\projects\rust\build\i686-pc-windows-gnu\stage2\bin;C:\projects\rust\build\i686-pc-windows-gnu\stage2-tools\i686-pc-windows-gnu\release\deps;C:\projects\rust\build\i686-pc-windows-gnu\stage2\lib\rustlib\i686-pc-windows-gnu\lib;C:\Program Files (x86)\Inno Setup 5;C:\Python27;C:\projects\rust\mingw32\bin;C:\msys64\usr\bin;C:\Perl\site\bin;C:\Perl\bin;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Program Files\7-Zip;C:\Program Files\Microsoft\Web Platform Installer;C:\Tools\GitVersion;C:\Tools\PsTools;C:\Program Files\Git LFS;C:\Program Files (x86)\Subversion\bin;C:\Program Files\Microsoft SQL Server\120\Tools\Binn;C:\Program Files\Microsoft SQL Server\Client SDK\ODBC\110\Tools\Binn;C:\Program Files (x86)\Microsoft SQL Server\120\Tools\Binn;C:\Program Files\Microsoft SQL Server\120\DTS\Binn;C:\Program Files (x86)\Microsoft SQL Server\120\Tools\Binn\ManagementStudio;C:\Tools\WebDriver;C:\Program Files (x86)\Microsoft SDKs\TypeScript\1.4;C:\Program Files (x86)\Microsoft Visual Studio 12.0\Common7\IDE\PrivateAssemblies;C:\Program Files (x86)\Microsoft SDKs\Azure\CLI\wbin;C:\Ruby193\bin;C:\Tools\NUnit\bin;C:\Tools\xUnit;C:\Tools\MSpec;C:\Tools\Coverity\bin;C:\Program Files (x86)\CMake\bin;C:\go\bin;C:\Program Files\Java\jdk1.8.0\bin;C:\Python27;C:\Program Files\nodejs;C:\Program Files (x86)\iojs;C:\Program Files\iojs;C:\Users\appveyor\AppData\Roaming\npm;C:\Program Files\Microsoft SQL Server\130\Tools\Binn;C:\Program Files (x86)\MSBuild\14.0\Bin;C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\IDE\Extensions\Microsoft\SQLDB\DAC\120;C:\Tools\NuGet;C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\IDE\CommonExtensions\Microsoft\TestWindow;C:\Program Files\Microsoft DNX\Dnvm;C:\Program Files\Microsoft SQL Server\Client SDK\ODBC\130\Tools\Binn;C:\Program Files (x86)\Microsoft SQL Server\130\Tools\Binn;C:\Program Files (x86)\Microsoft SQL Server\130\DTS\Binn;C:\Program Files\Microsoft SQL Server\130\DTS\Binn;C:\Program Files (x86)\Microsoft SQL Server\110\DTS\Binn;C:\Program Files (x86)\Microsoft SQL Server\120\DTS\Binn;C:\Program Files (x86)\Apache\Maven\bin;C:\Python27\Scripts;C:\Tools\NUnit3;C:\ProgramData\chocolatey\bin;C:\Program Files\Mercurial;C:\Program Files (x86)\Yarn\bin;C:\Program Files\LLVM\bin;C:\Program Files\dotnet;C:\Program Files\erl8.3\bin;C:\Tools\curl\bin;C:\Program Files\Microsoft Service Fabric\bin\Fabric\Fabric.Code;C:\Program Files\Microsoft SDKs\Service Fabric\Tools\ServiceFabricLocalClusterManager;C:\Program Files (x86)\nodejs;C:\Program Files\Git\cmd;C:\Program Files\Git\usr\bin;C:\Program Files\Amazon\AWSCLI;C:\Users\appveyor\AppData\Local\Yarn\.bin;C:\Users\appveyor\AppData\Roaming\npm;C:\Program Files\AppVeyor\BuildAgent;C:\projects\rust;C:\projects\rust\handle" C:\projects\rust\build\i686-pc-windows-gnu\stage2\bin\rustc.exe C:\projects\rust\src/test\run-pass\simple_global_asm.rs -L C:\projects\rust\build\i686-pc-windows-gnu\test\run-pass --target=i686-pc-windows-gnu --error-format json -L C:\projects\rust\build\i686-pc-windows-gnu\test\run-pass\simple_global_asm.stage2-i686-pc-windows-gnu.run-pass.libaux -C prefer-dynamic -o C:\projects\rust\build\i686-pc-windows-gnu\test\run-pass\simple_global_asm.stage2-i686-pc-windows-gnu.exe -Crpath -O -Lnative=C:\projects\rust\build\i686-pc-windows-gnu\native\rust-test-helpers
stdout:
------------------------------------------
------------------------------------------
stderr:
------------------------------------------
{"message":"linking with `gcc` failed: exit code: 1","code":null,"level":"error","spans":[],"children":[{"message":"\"gcc\" \"-Wl,--enable-long-section-names\" \"-fno-use-linker-plugin\" \"-Wl,--nxcompat\" \"-nostdlib\" \"-Wl,--large-address-aware\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\\\\crt2.o\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\\\\rsbegin.o\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\test\\\\run-pass\\\\simple_global_asm.0.o\" \"-o\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\test\\\\run-pass\\\\simple_global_asm.stage2-i686-pc-windows-gnu.exe\" \"-Wl,--gc-sections\" \"-nodefaultlibs\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\test\\\\run-pass\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\test\\\\run-pass\\\\simple_global_asm.stage2-i686-pc-windows-gnu.run-pass.libaux\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\native\\\\rust-test-helpers\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\" \"-L\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\" \"-l\" \"std-4d6881ec6132b951\" \"-Wl,-Bstatic\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\\\\libcompiler_builtins-7ac5a34e9b48514f.rlib\" \"-Wl,-Bdynamic\" \"-l\" \"advapi32\" \"-l\" \"ws2_32\" \"-l\" \"userenv\" \"-l\" \"shell32\" \"-lmingwex\" \"-lmingw32\" \"-lgcc\" \"-lmsvcrt\" \"-luser32\" \"-lkernel32\" \"C:\\\\projects\\\\rust\\\\build\\\\i686-pc-windows-gnu\\\\stage2\\\\lib\\\\rustlib\\\\i686-pc-windows-gnu\\\\lib\\\\rsend.o\"","code":null,"level":"note","spans":[],"children":[],"rendered":null},{"message":"C:\\projects\\rust\\build\\i686-pc-windows-gnu\\test\\run-pass\\simple_global_asm.0.o:(.text+0x15): undefined reference to `foo'\r\ncollect2.exe: error: ld returned 1 exit status\r\n","code":null,"level":"note","spans":[],"children":[],"rendered":null}],"rendered":null}
{"message":"aborting due to previous error","code":null,"level":"error","spans":[],"children":[],"rendered":null}
------------------------------------------
thread '[run-pass] run-pass\simple_global_asm.rs' panicked at 'explicit panic', src\tools\compiletest\src\runtest.rs:2627
note: Run with `RUST_BACKTRACE=1` for a backtrace.
failures:
[run-pass] run-pass\simple_global_asm.rs
test result: FAILED. 2632 passed; 1 failed; 17 ignored; 0 measured

@alexcrichton @nagisa any pointers? I don't have access to a MSYS_BITS=32, RUST_CONFIGURE_ARGS=--build=i686-pc-windows-gnu environment. Maybe there's something obvious I missed, but I don't see how foo should be undefined in src/test/run-pass/simple_global_asm.rs on that environment.

Also, note that codegen tests pass on that platform, which is to say global_asm's argument is faithfully passed to LLVM and written out in llvm-ir, which is the whole point of this feature. A test like this was my worry when, as I indicated to @nagisa before, this test is now only testing that platform's assembler and linker, not rustc. I'm now caught in an endless loop of fiddling with a test until it passes. That is, it's a pointless test.

alexcrichton · 2017-04-13T18:02:29Z

The failure here is:

undefined reference to 'foo'

That's probably because of the way symbols are named across systems. I believe on some platforms (like OSX and Windows) you need a leading underscore (e.g. _foo in the asm) whereas others like Linux you can omit it.

mrhota · 2017-04-13T19:06:28Z

@alexcrichton that seems to be it. Thanks!

mrhota · 2017-04-14T14:16:33Z

@alexcrichton ready again...

I declared both foo and _foo in the assembly. Hopefully this will satisfy windows' linker:

 .global foo
.global _foo
foo:
_foo:
ret

nikomatsakis · 2017-04-14T19:46:42Z

@bors r=nagisa

bors · 2017-04-14T19:46:43Z

📌 Commit a35c4e3 has been approved by nagisa

@eddyb

Implement global_asm!() (RFC 1548) This is a first attempt. ~~One (potential) problem I haven't solved is how to handle multiple usages of `global_asm!` in a module/crate. It looks like `LLVMSetModuleInlineAsm` overwrites module asm, and `LLVMAppendModuleInlineAsm` is not provided in LLVM C headers 😦~~ I can provide more detail as needed, but honestly, there's not a lot going on here. r? @eddyb CC @Amanieu @jackpot51 Tracking issue: rust-lang#35119

bors · 2017-04-14T22:30:24Z

⌛ Testing commit a35c4e3 with merge ea2c4f5...

frewsxcv · 2017-04-14T23:15:29Z

@bors retry reprioritizing

Rollup of 4 pull requests - Successful merges: #40702, #41172, #41249, #41303 - Failed merges:

bors · 2017-04-15T01:53:29Z

⌛ Testing commit a35c4e3 with merge f0ca5d4...

rust-highfive assigned pnkfelix Mar 21, 2017

mrhota changed the title ~~[WIP] global_asm!()~~[WIP] Implement global_asm!() (RFC 1548)Mar 21, 2017

parched reviewed Mar 21, 2017
View reviewed changes

mrhota force-pushed the global_asm branch from f40985b to 42cd7ccCompare March 22, 2017 00:54

mrhota changed the title ~~[WIP] Implement global_asm!() (RFC 1548)~~Implement global_asm!() (RFC 1548)Mar 22, 2017

Amanieu requested changes Mar 22, 2017
View reviewed changes

nagisa reviewed Mar 24, 2017
View reviewed changes

rust-highfive assigned nikomatsakis and unassigned pnkfelix Mar 27, 2017

mrhota force-pushed the global_asm branch from 0a56f79 to 25655acCompare March 29, 2017 22:19

mrhota force-pushed the global_asm branch from d01d22c to 907ba1bCompare March 30, 2017 05:30

Mark-Simulacrum added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Apr 14, 2017

Comply with windows symbol name conventions
a35c4e3

frewsxcv mentioned this pull request Apr 14, 2017
Rollup of 4 pull requests #41305
Merged

bors added a commit that referenced this pull request Apr 14, 2017
Auto merge of#41305- frewsxcv:rollup, r=frewsxcv
f0ca5d4
Rollup of 4 pull requests - Successful merges: #40702, #41172, #41249, #41303 - Failed merges:

bors merged commit a35c4e3 into rust-lang:masterApr 15, 2017

mrhota deleted the global_asm branch April 15, 2017 02:35

Mrmaxmeier mentioned this pull request Apr 16, 2017
Rustup to 1.18.0-nightly (d5cf1cb64 2017-04-15) rust-lang/rust-clippy#1682
Merged

Uh oh!

Uh oh!

Conversation

mrhota commented Mar 21, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rust-highfive commented Mar 21, 2017

Uh oh!

mrhota commented Mar 21, 2017

Uh oh!

eddyb commented Mar 21, 2017

Uh oh!

mattico commented Mar 21, 2017

Uh oh!

mrhota commented Mar 21, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mrhota commented Mar 22, 2017

Uh oh!

Amanieu left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mrhota commented Mar 23, 2017

Uh oh!

mrhota commented Mar 23, 2017

Uh oh!

Amanieu commented Mar 23, 2017

Uh oh!

mrhota commented Mar 24, 2017

Uh oh!

nagisa left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nagisaMar 24, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mrhota commented Mar 25, 2017

Uh oh!

nagisa commented Mar 25, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mrhota commented Mar 26, 2017

Uh oh!

mrhota commented Mar 27, 2017

Uh oh!

bors commented Mar 29, 2017

Uh oh!

nagisa commented Mar 29, 2017

Uh oh!

mrhota commented Mar 30, 2017

Uh oh!

mrhota commented Mar 30, 2017

Uh oh!

Amanieu commented Mar 30, 2017

Uh oh!

bors commented Apr 13, 2017

Uh oh!

TimNN commented Apr 13, 2017

Uh oh!

bors commented Apr 13, 2017

Uh oh!

bors commented Apr 13, 2017

mrhota commented Mar 21, 2017•
edited
Loading

nagisaMar 24, 2017•
edited
Loading

nagisa commented Mar 25, 2017•
edited
Loading