Uclibc fixes #1
Open
Uclibc fixes #1
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
* Remove backtrace header and bump version * re-trigger CI
rust-lang#298) Fixes rust-lang#289 `mmap` should be available on all platforms we support (except Windows), so we can use `libbacktrace`'s more efficient mmap-based code.
If dwarf debug info isn't available we fall back to searching the symbol table. Typically this is done with `dladdr` on most platforms but with gimli we typically have all the infrastructure already in place to do the search ourselves. This functionality was already filled out on Linux and Windows, but it wasn't implemented on macOS yet because it wasn't necessary. Implementing a pretty simple version, however, shows substantial speedups for the various benchmarks. Presumably `dladdr` isn't exactly the fastest thing in the world and our sorted list search which is cached must be much faster here! The current comparison of before/after this change looks like: ``` name before ns/iter after ns/iter diff ns/iter diff % speedup new 81,472 9,047 -72,425 -88.90% x 9.01 new_unresolved 2,126 2,009 -117 -5.50% x 1.06 new_unresolved_and_resolve_separate 82,252 9,134 -73,118 -88.90% x 9.01 trace 1,273 1,185 -88 -6.91% x 1.07 trace_and_resolve_callback 67,403 2,123 -65,280 -96.85% x 31.75 trace_and_resolve_separate 76,452 2,822 -73,630 -96.31% x 27.09 ```
* Add regression test for SIGPIPE on macOS * Disable `mmap` on macOS See rust-lang/rust#45866 for more details * Run `cargo fmt` * Remove unused variable name * s/macos/darwin/ * Move macOS test to its own file * Move macOS dSYM test to its own directory * Remove 'darwin' check to verify new test * Fix rustfmt * Re-add darwin check * Move macOS test to a non-workspace crate
* Add test case for line-tables-only
* Support line-tables-only when using libbacktrace
Previously when `backtrace_pcinfo` succeeded, but failed to obtain a
filename or a function name, the line number would be ignored. Instead,
when successful combine all available information.
For example, when using clang `-g1` or `-gline-tables-only` before:
```
1: baz
2: bar
3: foo
```
and after:
```
1: baz
at src/callback.c:5
2: bar
at src/callback.c:9
3: foo
at src/callback.c:13
```
Co-authored-by: Tomasz Miąsko <[email protected]>
* Update libbacktrace to latest master * Temporarily switch to t6/libbacktrace which has 5d4a5ff0d17040b5dd5dfaf2c3694b9a7b67b85d * Define HAVE_MACH_O_DYLD_H on macOS
…ng#294) * Define HAVE_KERN_PROC on FreeBSD to fix rust-lang/rust#54434 For the same reason define HAVE_KERN_PROC_ARGS on NetBSD. This depends on having [1] available in libbacktrace. [1] ianlancetaylor/libbacktrace@0f06cda * Add dragonfly too
Try to slim down the dependencies of `gimli-symbolize`
This is a pretty small dependency for the `gimli-symbolize` feature, and in an effort to reduce the number of external dependencies this commit inlines the small amount of code we used from that crate here directly. This should hopefully make this a bit easier to port to libstd eventually.
* Remove coresymbolication implementation This isn't used anywhere by default any more and I don't think there's any reason to enable it. This older implementation uses private APIs which aren't supposed to be used. Additionally libbacktrace nowadays should be workable enough and with gimli soon on the horizon that should be a suitable default for macOS and iOS. * Disable android emulator in CI
* Remove dladdr fallback and implementation This has been present for a very long time but I believe this has never actually been that necessary. Non-MSVC platforms all use libbacktrace by default, and libbacktrace will consult symbol tables of object files to do what `dladdr` does, just inside of libbacktrace. Additionally gimli implements the same logic. I believe that this means that `dladdr` isn't necessary for resolving any symbols since our other strategies should already be doing everything for us. This commit makes the feature defunkt and otherwise removes the various forms of fallback to dladdr. * Switch to the `object` crate for object parsing This commit switches the gimli feature from the `goblin` crate to the `object` crate for parsing object files. The main motivation here is trimming the dependencies of the `gimli-symbolize` feature to a bare minimum. The `object` crate itself has no dependencies now and should be a relatively easy drop-in replacement for the `goblin` crate. * Parse fat libraries on macOS This commit updates the object parsing code for macOS to support fat libraries. This enables gimli to symbolize addresses coming from system libraries which are currently installed frequently as fat libraries. Closes rust-lang#319 * Fix macOS symbolication of system libraries This commit fixes an issue where symbolication of system libraries didn't work on macOS. Symbolication through the symbol table was always off by a slide amount for the library. It's not entirely clear why this kept happening or what was going on, but some poking in LLDB's source revealed a way we can differentiate and figure out what addresses need to be looked up in the symbol table. Some more information is contained in the comments of the commit itself. Closes rust-lang#318 Co-authored-by: Philip Craig <[email protected]>
Prunes out the `gimli-symbolize` dependency tree to just 3 crates!
It's not entirely clear to me if this'll fix the issue, or why this issue is coming up, but maybe some biases are negative in some systems?
Make sure we keep the code up-to-date!
Try to stay on the update train for QEMU fixes
This commit uses the information and APIs learned from rust-lang/rust#71060 to implement support for dynamic libraries on MinGW. Previously symbols only worked if they came from the main executable, but now the process's list of dynamic libraries are also searched so we can symbolicate, for example, symbols in the compiler which come from loaded libraries rather than the main executable.
ELF files allow debug info sections to be compressed. The libbacktrace backed supported these compressed sections, but the Gimli backend did not. This commit adds that support to the Gimli backend. In my tests (with the BFD linker, lld, and gold) these debug info sections do not obey the alignment requirements that the object crate expects for the gABI compression header, so this commit additionally enables the "unaligned" feature in the upcoming version of the object crate. There is a bit of unsafe to ensure the lifetime of the decompressed sections matches the lifetime of the mmap'd file. I don't think there is a way around this unsafe code, unless we are willing to ditch Gimli's EndianSlice for an (apparently slower) EndianReader backed by a Cow<[u8]>. Fix rust-lang#342.
* Prepare for this crate to go into libstd This commit is a preparation of this crate to be included as a submodule into the standard library. I'm not 100% sold on this yet but I'm somewhat convinced that this is going to happen this way. This is progress on rust-lang#328 and a preview of what it might look like to implement this strategy. Currently I don't plan to merge this to the `master` branch unless it's decided to move forward with this integration strategy of the gimli feature of the backtrace crate. * Update as-if-std integration
…ust-lang#361) * Disable libbacktrace on DevkitA64 * Force noop symbolizer on DevkitA64 (for now) * Add support for linkle's DWARF format to gimli symbolizer * fmt * Address review * Subtract bias from len, revert back to wrapping_add * Use Vec::with_capacity in mmap_fake * fmt
Special-case platforms with `mmap` support, falling back to the `std::fs` implementation for platforms that don't have it.
* Provide access to source column numbers of symbol locations, in addition to filename and line numbers (where available - currently only gimli). * Print column numbers in backtraces, if available. * cargo fmt * Fixed breaking-change to public API, by delegating to new print_raw_with_column (as suggested by @alexcrichton in review rust-lang#367 (comment)) * Added tests for column numbers * Reflect actual conditional compilation logic in smoke test guards
* Module base address added to the Frame * Test added for module base address of the Frame * Process handle cached
Duplication introduced in 4b8f2be.
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.
No description provided.