[llvm-bugs] [Bug 41568] New: [LTO] -Os and -Oz -flto results in error message from LTO code generator.
via llvm-bugs
llvm-bugs at lists.llvm.org
Tue Apr 23 07:31:23 PDT 2019
https://bugs.llvm.org/show_bug.cgi?id=41568
Bug ID: 41568
Summary: [LTO] -Os and -Oz -flto results in error message from
LTO code generator.
Product: tools
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: lto
Assignee: unassignedbugs at nondot.org
Reporter: peter.smith at linaro.org
CC: llvm-bugs at lists.llvm.org
clang t.c -Os -flto -o optlto -fuse-ld=lld
ld.lld: error: -plugin-opt=Os: number expected, but got 's'
clang-9: error: linker command failed with exit code 1 (use -v to see
invocation)
Anything that will compile will do for the t.c file, I used an empty main.
This is the same for -Oz. While the error is reported by LLD it is coming from
the LTO Code Generator.
This has been raised before as https://bugs.llvm.org/show_bug.cgi?id=31665
however it was closed by the raiser as invalid as the optsize and minsize
attributes are added to the functions when they were compiled.
However I think that there still could be some work to do here:
- It isn't convenient to change the optimization flags just for the link steps
in some build systems. It also isn't discoverable what the equivalent
optimization setting should be given as a replacement. I believe -O2 comes
closest.
- When using the new pass manager there are some passes that check the
PassBuilder optimisation setting using isOptimizingForSize(). I think the
impact on code-generation will be low, but this may change in the future.
At this stage I don't think that supporting -Os and -Oz need be much more than
translating to one of the other optimization levels. Of course there could be a
custom size focused -Os, -Oz LTO added later.
--
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/20190423/b68b85fc/attachment.html>
More information about the llvm-bugs
mailing list