[llvm-bugs] [Bug 47588] New: Clang crashes when compiling __builtin_assume_aligned() with non-integer-constant alignment
via llvm-bugs
llvm-bugs at lists.llvm.org
Sat Sep 19 18:39:10 PDT 2020
https://bugs.llvm.org/show_bug.cgi?id=47588
Bug ID: 47588
Summary: Clang crashes when compiling
__builtin_assume_aligned() with non-integer-constant
alignment
Product: clang
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
Assignee: unassignedclangbugs at nondot.org
Reporter: bugsbugs at code.daltonmwoodard.com
CC: htmldeveloper at gmail.com, llvm-bugs at lists.llvm.org,
neeilans at live.com, richard-llvm at metafoo.co.uk
Created attachment 23985
--> https://bugs.llvm.org/attachment.cgi?id=23985&action=edit
Clang stacktrace
The following reliably leads to a crash within Clang (current trunk, 12.0
series as provided on compiler explorer). This is the most reduced from of the
crash I could come up with, but it was originally discovered in a larger
program source I'm not able to attach here.
// main.c
constexpr unsigned long long a() { return 4; }
int main()
{
int i = 0;
(void)__builtin_assume_aligned(&i, a());
}
I've backtested this and it seems to first reproduce in Clang 10.0. Note, the
following snippets containing alternative formulations do _not_ lead to a
crash.
// main.c (crash negative)
constexpr unsigned long long a = 4;
int main()
{
int i = 0;
(void)__builtin_assume_aligned(&i, a);
}
// main.c (crash negative)
struct a { static constexpr unsigned long long value = 4; };
int main()
{
int i = 0;
(void)__builtin_assume_aligned(&i, a::value);
}
Also note that when the `constexpr` is removed from the declaration of a(),
Clang does not crash and correctly diagnoses it as an error with:
<source>:6:11: error: argument to '__builtin_assume_aligned' must be a constant
integer
(void)__builtin_assume_aligned(&i, a());
Noting this just for completeness, but it's what I expected since the stack
trace is pointing to LLVM IR generation.
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20200920/6c04d80a/attachment-0001.html>
More information about the llvm-bugs
mailing list