Since we already assert that the outgoing IR is in LCSSA, it is easy to
get misled into thinking that -indvars broke LCSSA if the incoming IR is
non-LCSSA. Checking this pre-condition will make such cases break in
more obvious ways.
Inspired by (but does _not_ fix) PR26682.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271196 91177308-0d34-0410-b5e6-96231b3b80d8
This chapter demonstrates lazily JITing from ASTs with the expressions being
executed on a remote machine via a TCP connection. It needs some polish, but is
substantially complete.
Currently x86-64 SysV ABI (Darwin and Linux) only, but other architectures
can be supported by changing the server code to use alternative ABI support
classes from llvm/include/llvm/ExecutionEngine/Orc/OrcABISupport.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271193 91177308-0d34-0410-b5e6-96231b3b80d8
This function failed to type-check as it was. No test case yet (we only have
regression tests for the remote-JIT code, and LLI don't use this function), but
an upcoming chapter of the Kaleidoscope Building A JIT tutorials will use
this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271189 91177308-0d34-0410-b5e6-96231b3b80d8
This will allow us to mock it out in a unit test to begin
writing unit tests for the various PDB and codeview classes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271168 91177308-0d34-0410-b5e6-96231b3b80d8
This will be used in a follow up commit to simplify code in clang that creates a ConstantDataVector and calls the other form.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271164 91177308-0d34-0410-b5e6-96231b3b80d8
Consolidate documentation by removing comments from the .cpp file where
the comments in the .cpp file were copy-pasted from the header.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271157 91177308-0d34-0410-b5e6-96231b3b80d8
Moreover, I've never seen globally scoped `static inline` functions
elsewhere in LLVM, so this pattern is unusual if not unidiomatic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271155 91177308-0d34-0410-b5e6-96231b3b80d8
- Use doxygen-style comments
- Don't repeat member names in comments
- Add newlines between declarations
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271154 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
If we can prove that an op.with.overflow intrinsic does not overflow, we
can get rid of the intrinsic, and replace it with non-wrapping
arithmetic.
This was first checked in at r265913 but reverted in r265950 because it
exposed some issues around how SCEV handled post-inc add recurrences.
Those issues have now been fixed.
Reviewers: atrick, regehr
Subscribers: sanjoy, mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D18685
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271153 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This change teaches SCEV to see reduce `(extractvalue
0 (op.with.overflow X Y))` into `op X Y` (with a no-wrap tag if
possible).
This was first checked in at r265912 but reverted in r265950 because it
exposed some issues around how SCEV handled post-inc add recurrences.
Those issues have now been fixed.
Reviewers: atrick, regehr
Subscribers: mcrosier, mzolotukhin, llvm-commits
Differential Revision: http://reviews.llvm.org/D18684
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271152 91177308-0d34-0410-b5e6-96231b3b80d8
Fixes PR27315.
The post-inc version of an add recurrence needs to "follow the same
rules" as a normal add or subtract expression. Otherwise we miscompile
programs like
```
int main() {
int a = 0;
unsigned a_u = 0;
volatile long last_value;
do {
a_u += 3;
last_value = (long) ((int) a_u);
if (will_add_overflow(a, 3)) {
// Leave, and don't actually do the increment, so no UB.
printf("last_value = %ld\n", last_value);
exit(0);
}
a += 3;
} while (a != 46);
return 0;
}
```
This patch changes SCEV to put no-wrap flags on post-inc add recurrences
only when the poison from a potential overflow will go ahead to cause
undefined behavior.
To avoid regressing performance too much, I've assumed infinite loops
without side effects is undefined behavior to prove poison<->UB
equivalence in more cases. This isn't ideal, but is not new to LLVM as
a whole, and far better than the situation I'm trying to fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271151 91177308-0d34-0410-b5e6-96231b3b80d8
This is a stripped down version of D19211, leaving out the questionable
"branching in poison is UB" bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271150 91177308-0d34-0410-b5e6-96231b3b80d8
If the begin iterator fails, we cannot dereference it's contents.
Instead, we must immediately stop processing symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271141 91177308-0d34-0410-b5e6-96231b3b80d8
This patch removes the llvm intrinsics VPMOVSX and (V)PMOVZX sign/zero extension intrinsics and auto-upgrades to SEXT/ZEXT calls instead. We already did this for SSE41 PMOVSX sometime ago so much of that implementation can be reused.
Reapplied now that the the companion patch (D20684) removes/auto-upgrade the clang intrinsics has been committed.
Differential Revision: http://reviews.llvm.org/D20686
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271131 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
When RF_NullMapMissingGlobalValues is set, mapValue can return null
for GlobalValue. When mapping the operands of a constant that is
referenced from metadata, we need to handle this case and actually
return null instead of mapping this constant.
Reviewers: dexonsmith, rafael
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D20713
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271129 91177308-0d34-0410-b5e6-96231b3b80d8
We were producing R_X86_64_GOTPCRELX for invalid instructions and
sometimes producing R_X86_64_GOTPCRELX instead of
R_X86_64_REX_GOTPCRELX.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@271118 91177308-0d34-0410-b5e6-96231b3b80d8
