On Wed, Aug 09, 2017 at 02:48:40PM +0200, Juerg Haefliger wrote:
> On Wed, Aug 9, 2017 at 2:14 PM, Josh Poimboeuf <jpoim...@redhat.com> wrote:
> > On Wed, Aug 09, 2017 at 09:28:12AM +0200, Juerg Haefliger wrote:
> >> Hi,
> >>
> >> Is there a specific reason why we're using
> >>
> >> unset KPATCH_GCC_TEMPDIR
> >> CROSS_COMPILE="$TOOLSDIR/kpatch-gcc " make ...
> >>
> >> to build the original kernel? This doesn't seem to do anything special
> >> unless I'm missing something.
> >
> > Good question.  If I remember right, it's for compile speed.  Otherwise,
> > when building the patched kernel, make will notice the compiler binary
> > changed and it will rebuild everything.
> 
> Hmm. OK. But isn't everything rebuilt anyways for the patched kernel?
> I'll run some tests to better understand this.

If you're only patching one .c file, then make is smart enough to only
rebuild that one object file.

> Also, why not build two separate trees for the original and patched
> kernels and then diff the trees instead of using the kpatch-gcc magic?
> Speed?

Yes, kpatch-build is already slow as molasses, so any speedups are
greatly appreciated by the developers :-)

-- 
Josh

_______________________________________________
kpatch mailing list
kpatch@redhat.com
https://www.redhat.com/mailman/listinfo/kpatch

Reply via email to