More Linux Kernel Code Cleaned Up - Another Step Towards Building With Clang Or ICC

[phoronix]

Phoronix: More Linux Kernel Code Cleaned Up - Another Step Towards Building With Clang Or ICC

With the Linux 4.20~5.0, the kernel is now VLA-free as a step towards being able to compile the mainline code with the LLVM Clang compiler or other non-GCC compilers. Another step in this direction has been merged this cycle and that is cleaning up the compiler attributes code...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.20-Compiler-Attrs

[cen1]

Good to see this effort is still alive. The last time I checked the progress page it seemed dead.

[ms178]

Since Intel gives aways renewable free licences for private use for ICC nowadays, I am sure that this enabling Kernel work could stir up more interest. It also would provide Michael a new benchmark opportunity for compiler comparisons and performance testing. At least I wanted to try it out a couple of months ago, but unfortunately ICC turned out to lag behind in GLIBC compatibility which made it impossible to use on the latest (officially unsupported) distros.

I also wonder why Intel seems to lack a deeper collaboration between several internal teams. If I were Intel, I would use ClearLinux as a showcase for the greatness of my own compiler (next to still provide GCC/Clang as alternatives). They also work on GLIBC, the Kernel and other Linux projects and could fix these incompatibilities sooner ...

[Guest]

SJW conspiracy against GNU by George Soros something something something (foaming around the mouth)

/s

[michaelb1]

SJW conspiracy against GNU by George Soros something something something (foaming around the mouth)

/s

Funny quote, but George Soros is indeed a very bad person and a horrible example of late stage capitalism.

[ParticleBoard]

SJW conspiracy against GNU by George Soros something something something (foaming around the mouth)

/s

I come in here to read people's opinions on this and perhaps if anyone has tried it. Sure enough some "orange man bad" shitposter is in the first few comments talking to himself.

[ssokolow]

I also wonder why Intel seems to lack a deeper collaboration between several internal teams. If I were Intel, I would use ClearLinux as a showcase for the greatness of my own compiler (next to still provide GCC/Clang as alternatives). They also work on GLIBC, the Kernel and other Linux projects and could fix these incompatibilities sooner ...

According to Matt Pharr's The story of ispc, Intel has a very back-biting, competitive culture that management encourages in the belief that it'll weed out weak/bad ideas.

[ms178]

According to Matt Pharr's The story of ispc, Intel has a very back-biting, competitive culture that management encourages in the belief that it'll weed out weak/bad ideas.

Hehe, what a coincidence. I came across his blogpost some days ago and it was a great read! I hope Raja Koduri does not have to fight the same fights again with their compiler team.