On Thu, 30 Apr 2015 20:11:52 +0200 Volker Armin Hemmann wrote: > Am 30.04.2015 um 19:45 schrieb Andrew Savchenko: > > Hi, > > > > On Thu, 30 Apr 2015 18:26:22 +0200 Volker Armin Hemmann wrote: > >> That simple. The kernel is too important and the people programming it > >> know what they are doing. Don't set anything. It is retarded. > >>> - I don't like invoking 'CFLAGS="-O2 -march=foo"make' > >>> - I don't want to set CFLAGS as a persistent environment variable. > >>> - I don't want to modify the kernel Makefile > >>> > >>> Does it actually make sense to set an optimization level and -march? > >> no > > While I completely agree with you that kernel CFLAGS should not be > > randomly tampered with, I can't agree that -march itself is useless. > > Tests and results are available here: > > https://github.com/graysky2/kernel_gcc_patch > > > > Optimization is a very powerful tool if taken with care. Of course > > it may lead to a disastrous result if mindlessly used. > > > > Best regards, > > Andrew Savchenko > > if your mail client or browser is miscompiled, it is crashy, but worst > case, a bunch of emails or bookmarks are lost. > > If the kernel fucks up, it might write across partition boundaries and > destroy ALL your data. Or writes garbage instead of data. > > Don't f* with the kernel.
That's why we have tests. Follow the link above. As for a personal experience: we have kernels with this patch and gcc native optimization in production for several years. Results are fine (no kernel related issues). In order not to crash kernel, do not add -ffast-math there. You need to have some understanding before touching such stuff. Best regards, Andrew Savchenko
pgpoFIq0IXs1C.pgp
Description: PGP signature
