On Jul 28, 2014, at 12:59 PM, Tomasz Paweł Gajc wrote: > You may have misread some lines. Having LLVM/clang and using it in OMV does > NOT have any influence on gcc and kernel compilation with gcc.
Not true: you generalized "LLVM is in de house" from a thread about a gcc -Os bug report in the kernel. I missed no lines. Think more carefully about the precision of you replies and the relevancy to topic please. 73 de Jeff > Wysłano z BlackBerry® w Orange > From: Jeffrey Johnson <[email protected]> > Date: Mon, 28 Jul 2014 12:44:23 -0400 > To: Tomasz Gajc<[email protected]> > Cc: Cooker OpenMandriva<[email protected]> > Subject: Re: [OM Cooker] gcc 4.9 heads up > > > On Jul 28, 2014, at 6:31 AM, Tomasz Gajc wrote: > >> We didn't switch to LLVM.clang because of some gcc <-> kernel issue. On that >> time when we decided on this, i think we were not even aware of that bug. >> >> IIRC kernel is not yet ready at least official branch to be compiled with >> clang. >> > > We agree on both of the above statements. > >> I'm not an kernel expert but CONFIG_CC_OPTIMIZE_FOR_SIZE set to y (not >> default value) may bring more bad rather good effects. >> > > I was surprised at your reply that having LLVM in OpenMandriva > might have anything to do with a gcc miscompilation of a kernel. > > Its a gcc bug, fixed, not the first or last bug, life goes on, no matter Linus > inflammatory rants widely disseminated. > > 73 de Jeff > >> >> >> 2014-07-28 11:23 GMT+02:00 Jeffrey Johnson <[email protected]>: >> >> > On Jul 28, 2014, at 4:54 AM, Tomasz Paweł Gajc <[email protected]> wrote: >> > >> > Yes we switched from gcc to clang for CC and CXX. Ofcourse software which >> > can't be build with clang are handled by gcc. >> >> Switching to LLVM for user land has little to >> do with a kernel miscompilation using -Os. >> >> Justifying a switch to LLVM because of a GCC kernel >> bug is illogical too. >> >> Or are you truly claiming that you are compiling >> kernels with LLVM instead of GCC? >> >> 73 de Jeff >> >
