llvm dev - Apr 2018 - Useless exit instruction in "main", replaceable with return inst

Hi,

LLVM optimizer seems to leave "call exit" instruction in
"main" function
but it could be replaced by a return instruction.
Any reason why leave it as is or this simple optimization could be
implemented e.g. in SimplifyLibCalls?

Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20180403/01d88af9/attachment.html>

> On Apr 3, 2018, at 12:24, Dávid Bolvanský via llvm-dev <llvm-dev at
lists.llvm.org> wrote:
> 
> LLVM optimizer seems to leave "call exit" instruction in
"main" function but it could be replaced by a return instruction.
> Any reason why leave it as is or this simple optimization could be
implemented e.g. in SimplifyLibCalls?
Hi Dávid,

The C standard forces the dynamic linker to basically invoke main like so
“exit(main(argc, argv))”, therefore returning from main isn’t any faster than
calling exit() directly. Also, for real world operating systems, the cost of
process startup and shutdown overhead is so high that the cycle or two
difference is deeply in the noise.

Dave

On 4/3/2018 9:24 AM, Dávid Bolvanský via llvm-dev wrote:
> Hi,
>
> LLVM optimizer seems to leave "call exit" instruction in
"main"
> function but it could be replaced by a return instruction.
> Any reason why leave it as is or this simple optimization could be 
> implemented e.g. in SimplifyLibCalls?
The usual answer to this sort of question is "nobody has implemented it 
yet".  And frankly, it's not really worth implementing; a minor size 
reduction on a cold path in main() has very little benefit. (And 
strictly speaking, C allows programs to recursively call main().)

-Eli

-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux
Foundation Collaborative Project