This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: PATCH: Fix removed sections (Re: 2.12.90.0.12 Kernel Miscompile)
- From: Alan Modra <amodra at bigpond dot net dot au>
- To: "H. J. Lu" <hjl at lucon dot org>, "Leonard N. Zubkoff" <lnz at dandelion dot com>
- Cc: binutils at sources dot redhat dot com
- Date: Mon, 1 Jul 2002 21:10:46 +0930
- Subject: Re: PATCH: Fix removed sections (Re: 2.12.90.0.12 Kernel Miscompile)
- References: <20020625084948.A19332@lucon.org> <200206251703.g5PH3aFk001217@dandelion.com> <20020625115653.A22317@lucon.org> <20020625132405.A23560@lucon.org>
On Tue, Jun 25, 2002 at 01:24:05PM -0700, H. J. Lu wrote:
> On Tue, Jun 25, 2002 at 11:56:53AM -0700, H. J. Lu wrote:
> > http://sources.redhat.com/ml/binutils/2002-06/msg00289.html
> >
> > breaks the Linux kernel 2.2.21 on x86. Reverting
I'm curious to know exactly why. So far, I've only seen symptoms
reported, which could well be due to bugs in the kernel. (*) My
builds of 2.2.21 had no signifigant differences after applying your
patch, just the version string. I probably don't use the same
config as you do.
Can one of you provide a little more information about vmlinux file
differences? "readelf -S vmlinux" output before/after HJ's patch, if
different would be useful. My suspicion is that something in the
kernel is relying on a removed output section's alignment to affect
placement of the next output section.
*) Not that we want binutils to break kernel builds! However, if the
kernel is relying on what can be argued as ld _bugs_, then we need to
a) fix the kernel and b) consider whether we can live with a ld bug
in order to support old code.
--
Alan Modra
IBM OzLabs - Linux Technology Centre