problem with special keys in lfs-cvs (home, end, F1...)

Greg Schafer gschafer at zip.com.au
Fri Aug 23 03:20:02 PDT 2002


On Fri, Aug 23, 2002 at 02:18:45AM -0500, DJ Lucas wrote:
> Greg Schafer wrote:
> >
> > Possibly unrelated but you should be aware of a bad bug recently
> > discovered in loadkeys. If you patched your gcc-3.x to change the
> > default arch or you compiled kbd with -march=i686 or greater you
> > will probably hit the bug.
> >
> > If this is indeed the case, compile kbd with -march=i586.
> >
> > Greg
> 
> Greg, and everybody actually,
> 
> Any good possibility that this IS almost definately related to the buggy
> kbd?  I was the first (that I know of) to have this issue with the six
> function keys, right after...maybe just before 3.1 was released.  I
> didn't even really bother to trouble shoot it.  I did a complete system
> rebuild from the crippled system and it fixed itself.  Three people have
> reported that a kernel re-compile ONLY from a freshly unpacked source,
> fixes the issue.  Problem is that very few people have seen the 'issue'.
> Does kbd have anything to do with console input if you are not using the
> loadkeys script?  And what software packages are going to know where the
> are, and alter our kernel sources?

gcc-3.1 and above is known to miscompile the loadkeys binary at -march=i686
and above. If you don't use loadkeys then you cannot be affected :-) In
other words, loadkeys does not come into play if you don't run the loadkeys
script.

But having said that, if gcc has been proven to miscompile some particular
piece of code, who knows what other code it is silently miscompiling? So I
suppose the answer to your initial question is "maybe". It is quite possible
that some keyboard handling piece of kernel code gets miscompiled.

Greg
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe blfs-support' in the subject header of the message



More information about the blfs-support mailing list