Sync Fork from Upstream Repo #12
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
* Fix allunique(::StepRangeLen) * Add specialized unique(::AbstractRange) method
Defining
struct MyInt <: Integer
x::Int
end
(::Type{T})(x::MyInt) where T<:Integer = T(x.x)
triggers about 830 unique method invalidations. This fixes the majority
of them, but it's a lot of type-annotation.
When inlining declines to inline something, it instead turns them into :invoke statements. These are then turned into direct (non-inlined) calls by codegen or otherwise receive a fast path at runtime. While inlining has evolved quite a bit, this code has stayed much the same since it was introduced four years ago and doesn't seem to make much sense as is. In particular: 1. For the non-`invoke()` case we were doing an extra method look that seems entirely superfluous, because we already had to do the very same method lookup just to reach this point. The only thing this path was doing at that point was creating a "compilable" specialization (which might use a slightly different signature). We might as well do that directly. 2. For the invoke case, we were pro-actively adding the specialization to the `->invokes` dispatch cache. However, this doesn't make much sense a priori either, because the bail path does not go through the runtime `invoke()` code that uses that cache (it did many years ago when this code was introduced, but hasn't in a long time). There does not seem to be a good reason to believe that this signature will be any more likely than any other to be invoked using the runtime mechanism. This cleans up that path by getting rid of both the superfluous method lookup and the superfluous addition to the `->invokes` cache. There should be a slight performance improvement as well from avoiding this superfluous work, but the bail path is less common than one might expect (the vast majority of call sites are inlined) and in measurements the effect seems to be in the noise. Nevertheless, it seems like a nice simplification and is conceptually clearer.
Includes build fixes to ensure it is built thread-safe. Fixes #35796 regression.
We might print ANSI color codes around ERROR now, making this test unreliable.
Co-authored-by: tylebrow <tyler.brown@gartner.com>
There was some extra space
`nnz` counts the number of elements stored in the `SparseMatrixCSC`. "Structural nonzeros" is a tautology.
* Document returned LLVM type for a few functions * Require and assert `Vboxed` to be of type `T_prjlvalue` * Remove unused branch/specialization of `maybe_decay_tracked` * Add function to cast known-untracked value.
- enables building with TSAN for the runtime library as well as Julia code - updates the handling of `SANITIZE=1` in Make.inc - moves sanitizer to late in the pipeline, copies what Clang does - cleans up `options.h`, and `julia_internal.h` w.r.t sanitizers - update devdocs for sanitizer - adds a patch for TSAN to deal with Julia's usage of address spaces - don't use COPY_STACKS with TSAN - don't use DEEPBIND by default if a sanitizer is enabled
This provides a fallback `eltype` method specialized for imprecise NamedType types. Formerly we were calling the generic eltype fallback `eltype(::Type) = Any`, but relying on the generic fallback makes code vulnerable to invalidation when new `eltype` methods are added. Since this affects any poorly-inferred keyword-arg function, it's best to isolate this by defining a specialization.
Defining
struct MyInt <: Integer
x::Int
end
(::Type{T})(x::MyInt) where T<:Integer = T(x.x)
triggers about 830 unique method invalidations. This fixes the majority
of them, but it's a lot of type-annotation.
- it's been used as an alias to `Core.Const` and so there's been two ways to construct them, i.e. via `Const`'s constructor and `AbstractEvalCall` - I think this indirection is unnecessary, but rather just confusing so rename all the `AbstractEvalConstant` to `Const`
- use ThreadSynchronizer type alias in ReentrantLock - tweak the `timedwait` documentation; in particular, rename the `testcb` argument to `callback` - clarify an "Special note for Threads.Condition" to not reference a variable `c` that was never defined
Some cleanup and comment for addrspace tracking in codegen
Move the code that inlines results from `@pure` function earlier. It doesn't depend on any of the method analysis, so doing that work was just wasted time.
…6970) The metadata on the load may be striped by LLVM when moving code around
This allows much easier debugging of imaging mode codegen without hacks like specifying a dummy output and manually running module init functions. Also allows specializing codegen to give result more similar to code in actual sysimg in this debugging mode. Ref #36974.
…e vectors out of the loops in order to get a speedup in some cases where type inference otherwise takes tuple types. (#36975)
Co-authored-by: Ian <i.r.butterworth@gmail.com> Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com>
…_tfunc (#36926) * Avoid constructing an array only to splat it to make a tuple in tuple_tfunc * use ntuple
Removes ambiguity:
```
julia> Base.OrderStyle(Union{})
ERROR: MethodError: Base.OrderStyle(::Type{Union{}}) is ambiguous. Candidates:
```
This error is relevant as with current `unique!` definition that relies on `OrderStyle` one can have a problem in corner cases. E.g.:
```
julia> [i for i in ["1"] if i isa Int]
0-element Array{Union{},1}
```
Co-authored-by: Milan Bouchet-Valat <nalimilan@club.fr>
Also add a few missing llvm eltypes for load and GEPs.
Make alignment of tag more obvious for LLVM and add a few optimization passes after GC lowering to let LLVM delete some write barriers. This fixes a regression from LLVM < 5 where the pass ordering makes this happen automatically.
These were left over from when our IR wasn't quite linear yet and it was possible to have to evaluate another slot deep in the nesting context. Since that's not true anymore, these arguments are all unused.
This should make this previously non-atomic mutex threadsafe. Fixes #35117
This defines how much time we waste before letting another process run. Until we have performance measurements, this should probably be small. Closes #36952
The most common argument type for `to_indices` is a tuple of `Int`, yet we rely on the `Integer` fallback. This makes it vulnerable to invalidation by methods like https://github.com/JuliaMath/Interpolations.jl/blob/bcd05a3f0843661104589c31da8d257fecdbe265/src/Interpolations.jl#L273 which would rather convert non-`Int` integers. This prevents approximately 80 MethodInstance invalidations.
sthagen
pushed a commit
that referenced
this pull request
Apr 20, 2022
Follows up JuliaLang#44708 -- in that PR I missed the most obvious optimization opportunity, i.e. we can safely eliminate `isdefined` checks when all fields are defined at allocation site. This change allows us to eliminate capturing closure constructions when the body and callsite of capture closure is available within a optimized frame, e.g.: ```julia function abmult(r::Int, x0) if r < 0 r = -r end f = x -> x * r return @inline f(x0) end ``` ```diff diff --git a/_master.jl b/_pr.jl index ea06d865b75..c38f221090f 100644 --- a/_master.jl +++ b/_pr.jl @@ -1,24 +1,19 @@ julia> @code_typed abmult(-3, 3) CodeInfo( -1 ── %1 = Core.Box::Type{Core.Box} -│ %2 = %new(%1, r@_2)::Core.Box -│ %3 = Core.isdefined(%2, :contents)::Bool -└─── goto #3 if not %3 +1 ── goto #3 if not true 2 ── goto #4 3 ── $(Expr(:throw_undef_if_not, :r, false))::Any -4 ┄─ %7 = (r@_2 < 0)::Any -└─── goto #9 if not %7 -5 ── %9 = Core.isdefined(%2, :contents)::Bool -└─── goto #7 if not %9 +4 ┄─ %4 = (r@_2 < 0)::Any +└─── goto #9 if not %4 +5 ── goto #7 if not true 6 ── goto #8 7 ── $(Expr(:throw_undef_if_not, :r, false))::Any -8 ┄─ %13 = -r@_2::Any -9 ┄─ %14 = φ (#4 => r@_2, #8 => %13)::Any -│ %15 = Core.isdefined(%2, :contents)::Bool -└─── goto #11 if not %15 +8 ┄─ %9 = -r@_2::Any +9 ┄─ %10 = φ (#4 => r@_2, #8 => %9)::Any +└─── goto #11 if not true 10 ─ goto #12 11 ─ $(Expr(:throw_undef_if_not, :r, false))::Any -12 ┄ %19 = (x0 * %14)::Any +12 ┄ %14 = (x0 * %10)::Any └─── goto #13 -13 ─ return %19 +13 ─ return %14 ) => Any ```
sthagen
pushed a commit
that referenced
this pull request
Jun 29, 2022
When calling `jl_error()` or `jl_errorf()`, we must check to see if we
are so early in the bringup process that it is dangerous to attempt to
construct a backtrace because the data structures used to provide line
information are not properly setup.
This can be easily triggered by running:
```
julia -C invalid
```
On an `i686-linux-gnu` build, this will hit the "Invalid CPU Name"
branch in `jitlayers.cpp`, which calls `jl_errorf()`. This in turn
calls `jl_throw()`, which will eventually call `jl_DI_for_fptr` as part
of the backtrace printing process, which fails as the object maps are
not fully initialized. See the below `gdb` stacktrace for details:
```
$ gdb -batch -ex 'r' -ex 'bt' --args ./julia -C invalid
...
fatal: error thrown and no exception handler available.
ErrorException("Invalid CPU name "invalid".")
Thread 1 "julia" received signal SIGSEGV, Segmentation fault.
0xf75bd665 in std::_Rb_tree<unsigned int, std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo>, std::_Select1st<std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo> >, std::greater<unsigned int>, std::allocator<std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo> > >::lower_bound (__k=<optimized out>, this=0x248) at /usr/local/i686-linux-gnu/include/c++/9.1.0/bits/stl_tree.h:1277
1277 /usr/local/i686-linux-gnu/include/c++/9.1.0/bits/stl_tree.h: No such file or directory.
#0 0xf75bd665 in std::_Rb_tree<unsigned int, std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo>, std::_Select1st<std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo> >, std::greater<unsigned int>, std::allocator<std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo> > >::lower_bound (__k=<optimized out>, this=0x248) at /usr/local/i686-linux-gnu/include/c++/9.1.0/bits/stl_tree.h:1277
#1 std::map<unsigned int, JITDebugInfoRegistry::ObjectInfo, std::greater<unsigned int>, std::allocator<std::pair<unsigned int const, JITDebugInfoRegistry::ObjectInfo> > >::lower_bound (__x=<optimized out>, this=0x248) at /usr/local/i686-linux-gnu/include/c++/9.1.0/bits/stl_map.h:1258
#2 jl_DI_for_fptr (fptr=4155049385, symsize=symsize@entry=0xffffcfa8, slide=slide@entry=0xffffcfa0, Section=Section@entry=0xffffcfb8, context=context@entry=0xffffcf94) at /cache/build/default-amdci5-4/julialang/julia-master/src/debuginfo.cpp:1181
#3 0xf75c056a in jl_getFunctionInfo_impl (frames_out=0xffffd03c, pointer=4155049385, skipC=0, noInline=0) at /cache/build/default-amdci5-4/julialang/julia-master/src/debuginfo.cpp:1210
#4 0xf7a6ca98 in jl_print_native_codeloc (ip=4155049385) at /cache/build/default-amdci5-4/julialang/julia-master/src/stackwalk.c:636
#5 0xf7a6cd54 in jl_print_bt_entry_codeloc (bt_entry=0xf0798018) at /cache/build/default-amdci5-4/julialang/julia-master/src/stackwalk.c:657
#6 jlbacktrace () at /cache/build/default-amdci5-4/julialang/julia-master/src/stackwalk.c:1090
#7 0xf7a3cd2b in ijl_no_exc_handler (e=0xf0794010) at /cache/build/default-amdci5-4/julialang/julia-master/src/task.c:605
#8 0xf7a3d10a in throw_internal (ct=ct@entry=0xf070c010, exception=<optimized out>, exception@entry=0xf0794010) at /cache/build/default-amdci5-4/julialang/julia-master/src/task.c:638
#9 0xf7a3d330 in ijl_throw (e=0xf0794010) at /cache/build/default-amdci5-4/julialang/julia-master/src/task.c:654
#10 0xf7a905aa in ijl_errorf (fmt=fmt@entry=0xf7647cd4 "Invalid CPU name \"%s\".") at /cache/build/default-amdci5-4/julialang/julia-master/src/rtutils.c:77
#11 0xf75a4b22 in (anonymous namespace)::createTargetMachine () at /cache/build/default-amdci5-4/julialang/julia-master/src/jitlayers.cpp:823
#12 JuliaOJIT::JuliaOJIT (this=<optimized out>) at /cache/build/default-amdci5-4/julialang/julia-master/src/jitlayers.cpp:1044
#13 0xf7531793 in jl_init_llvm () at /cache/build/default-amdci5-4/julialang/julia-master/src/codegen.cpp:8585
#14 0xf75318a8 in jl_init_codegen_impl () at /cache/build/default-amdci5-4/julialang/julia-master/src/codegen.cpp:8648
#15 0xf7a51a52 in jl_restore_system_image_from_stream (f=<optimized out>) at /cache/build/default-amdci5-4/julialang/julia-master/src/staticdata.c:2131
#16 0xf7a55c03 in ijl_restore_system_image_data (buf=0xe859c1c0 <jl_system_image_data> "8'\031\003", len=125161105) at /cache/build/default-amdci5-4/julialang/julia-master/src/staticdata.c:2184
#17 0xf7a55cf9 in jl_load_sysimg_so () at /cache/build/default-amdci5-4/julialang/julia-master/src/staticdata.c:424
#18 ijl_restore_system_image (fname=0x80a0900 "/build/bk_download/julia-d78fdad601/lib/julia/sys.so") at /cache/build/default-amdci5-4/julialang/julia-master/src/staticdata.c:2157
#19 0xf7a3bdfc in _finish_julia_init (rel=rel@entry=JL_IMAGE_JULIA_HOME, ct=<optimized out>, ptls=<optimized out>) at /cache/build/default-amdci5-4/julialang/julia-master/src/init.c:741
#20 0xf7a3c8ac in julia_init (rel=<optimized out>) at /cache/build/default-amdci5-4/julialang/julia-master/src/init.c:728
#21 0xf7a7f61d in jl_repl_entrypoint (argc=<optimized out>, argv=0xffffddf4) at /cache/build/default-amdci5-4/julialang/julia-master/src/jlapi.c:705
#22 0x080490a7 in main (argc=3, argv=0xffffddf4) at /cache/build/default-amdci5-4/julialang/julia-master/cli/loader_exe.c:59
```
To prevent this, we simply avoid calling `jl_errorf` this early in the
process, punting the problem to a later PR that can update guard
conditions within `jl_error*`.
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.