MSVC: Implement runtime support for unwinding #26569
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This logic applies to all MSVC targets, so instead refactor it into platform.mk so it can one day apply to 32-bit MSVC.
|
(rust_highfive has picked a reviewer for you, use r? to override) |
|
r? @brson It may be somewhat premature to land this patch, but I'm curious what others' opinions are. This puts us into a state where unwinding is implemented (e.g. it won't abort the process), but it won't run any destructors as it unwinds a thread (and hence tests will fail). I've also been discovering some segfaults even with just |
alexcrichton
force-pushed
the
msvc-llvm-update
branch
from
b869318 to
6837bc5
Compare
Now that LLVM has been updated, the only remaining roadblock to implementing unwinding for MSVC is to fill out the runtime support in `std::rt::unwind::seh`. This commit does precisely that, fixing up some other bits and pieces along the way: * The `seh` unwinding module now uses `RaiseException` to initiate a panic. * The `rust_try.ll` file was rewritten for MSVC (as it's quite different) and is located at `rust_try_msvc_64.ll`, only included on MSVC builds for now. * The personality function for all landing pads generated by LLVM is hard-wired to `__C_specific_handler` instead of the standard `rust_eh_personality` lang item. This is required to get LLVM to emit SEH unwinding information instead of DWARF unwinding information. This also means that on MSVC the `rust_eh_personality` function is entirely unused (but is defined as it's a lang item). More details about how panicking works on SEH can be found in the `rust_try_msvc_64.ll` or `seh.rs` files, but I'm always open to adding more comments! A key aspect of this PR is missing, however, which is that **unwinding is still turned off by default for MSVC**. There is a [bug in llvm][llvm-bug] which causes optimizations to inline enough landing pads that LLVM chokes. If the compiler is optimized at `-O1` (where inlining isn't enabled) then it can bootstrap with unwinding enabled, but when optimized at `-O2` (inlining is enabled) then it hits a fatal LLVM error. [llvm-bug]: https://llvm.org/bugs/show_bug.cgi?id=23884
If a dylib doesn't actually export any symbols then link.exe won't emit a `foo.lib` file to link against (as one isn't necessary). Detect this case in the backend by omitting the `foo.lib` argument to the linker if it doesn't actually exist.
Makes this test case more robust by using standard libraries to ensure the binary can be built.
alexcrichton
force-pushed
the
msvc-llvm-update
branch
from
6837bc5 to
759a7f1
Compare
|
@bors r+ |
|
📌 Commit 759a7f1 has been approved by |
|
⌛ Testing commit 759a7f1 with merge d3c03d0... |
bors
added a commit
that referenced
this pull request
Jun 27, 2015
Now that LLVM has been updated, the only remaining roadblock to implementing unwinding for MSVC is to fill out the runtime support in `std::rt::unwind::seh`. This commit does precisely that, fixing up some other bits and pieces along the way: * The `seh` unwinding module now uses `RaiseException` to initiate a panic. * The `rust_try.ll` file was rewritten for MSVC (as it's quite different) and is located at `rust_try_msvc_64.ll`, only included on MSVC builds for now. * The personality function for all landing pads generated by LLVM is hard-wired to `__C_specific_handler` instead of the standard `rust_eh_personality` lang item. This is required to get LLVM to emit SEH unwinding information instead of DWARF unwinding information. This also means that on MSVC the `rust_eh_personality` function is entirely unused (but is defined as it's a lang item). More details about how panicking works on SEH can be found in the `rust_try_msvc_64.ll` or `seh.rs` files, but I'm always open to adding more comments! A key aspect of this PR is missing, however, which is that **unwinding is still turned off by default for MSVC**. There is a [bug in llvm][llvm-bug] which causes optimizations to inline enough landing pads that LLVM chokes. If the compiler is optimized at `-O1` (where inlining isn't enabled) then it can bootstrap with unwinding enabled, but when optimized at `-O2` (inlining is enabled) then it hits a fatal LLVM error. [llvm-bug]: https://llvm.org/bugs/show_bug.cgi?id=23884
bors
added a commit
that referenced
this pull request
Jun 28, 2015
This series of commits (currently rebased on #26569 to avoid conflicts) adds support for the standard library to run on Windows XP. The main motivation behind this PR is that to enable any Rust code in Firefox we need to support Windows XP. This PR doesn't yet intend to be a move to make Windows XP an officially supported platform, but instead simply get Rust code running on it. APIs like condition variables and RWLocks will immediately panic currently on XP, and it's unclear if that story wants to change much. Additionally, we may bind APIs like IOCP which aren't available on XP and would be *very* difficult to provide a fallback implementation. Essentially this PR enables running Rust on XP, but you still have to be careful to avoid non-XP portions of the standard library. The major components of this PR are: * Support for a new `i686-pc-windows-msvc` triple. This primarily involves a lot of build system hackery, but there are also a number of floating point functions which had to get switched up a bit. * All APIs not available on Windows are now accessed through our dynamic-detection mechanism * Mutexes on Windows were rewritten to use SRWLOCK as an optimization but can fall back to CRITICAL_SECTION.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Now that LLVM has been updated, the only remaining roadblock to implementing
unwinding for MSVC is to fill out the runtime support in
std::rt::unwind::seh.This commit does precisely that, fixing up some other bits and pieces along the
way:
sehunwinding module now usesRaiseExceptionto initiate a panic.rust_try.llfile was rewritten for MSVC (as it's quite different) and islocated at
rust_try_msvc_64.ll, only included on MSVC builds for now.to
__C_specific_handlerinstead of the standardrust_eh_personalitylangitem. This is required to get LLVM to emit SEH unwinding information instead
of DWARF unwinding information. This also means that on MSVC the
rust_eh_personalityfunction is entirely unused (but is defined as it's alang item).
More details about how panicking works on SEH can be found in the
rust_try_msvc_64.llorseh.rsfiles, but I'm always open to adding morecomments!
A key aspect of this PR is missing, however, which is that unwinding is still
turned off by default for MSVC. There is a bug in llvm which
causes optimizations to inline enough landing pads that LLVM chokes. If the
compiler is optimized at
-O1(where inlining isn't enabled) then it canbootstrap with unwinding enabled, but when optimized at
-O2(inlining isenabled) then it hits a fatal LLVM error.