Re: why does r8375 change the ABI?
On Fri, Mar 23, 2007 at 09:51:02AM +0100, Bastian Blank wrote:
> On Thu, Mar 22, 2007 at 06:23:47PM -0600, dann frazier wrote:
> > The linux-2.6 build reports that this patch (fix for CVE-2006-5753)
> > breaks the ABI - but I can't seem to figure out why.
>
> Can you please provide the output of the abi check? Are only some
> functions affected or all?
dannf@pond:~/tmp/linux-2.6-2.6.18.dfsg.1$ python2.4 \
debian/bin/abicheck.py debian/build/build-i386-none-486 i386 none 486
ABI has changed! Refusing to continue.
Changed symbols:
is_bad_inode version: 0xa68da888-> 0xe0848c31
make_bad_inode version: 0xf14b2332-> 0x900a13d5
> > Troy Heber narrowed down the breakage to the addition of the
> > #include <linux/poll.h> line, but he couldn't identify the change in
> > the object files.
>
> He need to check the debugging output of the versions tool. It does a
> source-level check and may be confused by some construct.
Ok - what is the name of the versions tool?
--
dann frazier
Reply to: