svn commit: r216161 - in head/sys: amd64/amd64 i386/i386

Roman Divacky rdivacky at FreeBSD.org
Sat Dec 4 10:54:23 UTC 2010


On Fri, Dec 03, 2010 at 06:17:16PM -0500, Jung-uk Kim wrote:
> On Friday 03 December 2010 06:02 pm, Jung-uk Kim wrote:
> > On Friday 03 December 2010 05:43 pm, Jung-uk Kim wrote:
> > > On Friday 03 December 2010 05:08 pm, John Baldwin wrote:
> > > > On Friday, December 03, 2010 4:54:10 pm Jung-uk Kim wrote:
> > > > > Author: jkim
> > > > > Date: Fri Dec  3 21:54:10 2010
> > > > > New Revision: 216161
> > > > > URL: http://svn.freebsd.org/changeset/base/216161
> > > > >
> > > > > Log:
> > > > >   Explicitly initialize TSC frequency.  To calibrate TSC
> > > > > frequency, we use DELAY(9) and it may use TSC in turn if TSC
> > > > > frequency is non-zero.
> > > >
> > > > We zero the BSS, so these were already zero.  This just makes
> > > > the actual kernel file on disk larger by wasting space in .data
> > > > instead of .bss.
> > >
> > > Please note that I didn't touch other variables, e.g.,
> > > tsc_is_broken, because I knew that.  However, I just wanted to do
> > > that *explicitly*. Anyway, it is reverted now and SVN will
> > > remember what I wanted to do. ;-)
> > >
> > > BTW, if my memory serves, GCC (and all modern C compilers) put(s)
> > > zero-initialized variables back in .bss.
> >
> > I just tried it.  GCC generates identical binaries as I thought.
> > However, Clang doesn't do the optimization. :-/
> 
> Strangely, Clang increases .bss when a global variable is explicitly 
> initialized to zero.
> 
> -  2 .bss 00000004 0000000000000000 0000000000000000 00000540 2**2
> +  2 .bss 00000014 0000000000000000 0000000000000000 00000540 2**3

in my naive test gcc produces:

.globl foo
	.section	.bss
	.align 4
	.type	foo, @object
	.size	foo, 4
foo:
	.zero	4


and clang produces:

	.type	foo, at object             # @foo
	.bss
	.globl	foo
	.align	4
foo:
	.long	0                       # 0x0
	.size	foo, 4

ie. both put them into BSS


More information about the svn-src-head mailing list