lkml.org 
[lkml]   [2015]   [Apr]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] x86: Align jump targets to 1 byte boundaries

* Markus Trippelsdorf <markus@trippelsdorf.de> wrote:

> > I'm not so sure about that one, our data access patterns are
> > usually a lot more well thought out than our code alignment (which
> > is really mostly compiler controlled). It also gives limited
> > savings:
> >
> > 9202488 vmlinux gcc-5
> > 9186105 vmlinux gcc-5 (-malign-data=abi)
> >
> > Which is 0.1%. I've got a handful of options in that size range:
> >
> > + # Reduces vmlinux size by 0.25%:
> > + KBUILD_CFLAGS += -fno-caller-saves
> > +
> > + # Reduces vmlinux size by 1.10%:
> > + KBUILD_CFLAGS += -fno-inline-small-functions
> > +
> > + # Reduces vmlinux size by about 0.95%:
> > + KBUILD_CFLAGS += -fno-tree-ch
> >
> > but obviously they are more obscure and thus riskier. Find below
> > an updated "Ingo's combo patch". It gives more than 10% savings
> > here on x86 defconfig using gcc 4.9, without LTO.
>
> Well obviously, if you do not care about performance you can reduce
> the text size further and further. [...]

Yes, but I picked GCC options that I don't think impact performance
negatively and offer a sizable debloating effect. Especially with
inlining if code size increases it's probably a net loss.

> [...] But what is interesting is to keep the performance up (or even
> increase it) and still reduce the text size.

By my (admittedly quick) review I think those 3 extra options I added
still generate pretty OK code in the end. I.e. they are not like -Os
that generates utter crap to save a byte.

Thanks,

Ingo


\
 
 \ /
  Last update: 2015-04-14 14:41    [W:0.120 / U:1.048 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site