Re: the joy of scsi headers and glibc...

Andreas Jaeger (aj@suse.de)
07 Nov 1999 16:39:00 +0100


>>>>> Douglas Gilbert writes:

Douglas> Andreas,
Douglas> In my defence, in May I contacted Alan Cox about this
Douglas> thinking it was a RH6.0 distro problem. He told me it
Douglas> was a glibc problem and gave me an email address (in
Douglas> cygnus from memory) to report the problem to. I did
Douglas> and heard nothing.

The correct address for glibc problems is either bug-glibc@gnu.org or
libc-alpha@sourceware.cygnus.com. Sorry, that you heard nothing.

Douglas> Report it, document it and supply work arounds... at
Douglas> least the documentation finally got through.
;-)

Douglas> Further, if glibc documentation says that "`/usr/include/net' and `/usr/include/scsi'
Douglas> should *not* be symlinks" then why does
Douglas> the latter directory contain old versions of scsi.h and sg.h ?

Have a look at Andries' email and also at the glibc FAQ 3.5. Linus
asked us to reduce the usage of kernel headers to a minimum. The old
scsi.h and sg.h were not usable since they've got internal kernel data
and also external information. Therefore glibc comes with its own
headers.

Douglas> If they are Linux kernel headers (both are) then they should
Douglas> either be controlled by Linux kernel distribution methods or
Douglas> not be there at all.
Nope - you should strive to work with us so that this doesn't happen.
The information in scsi.h and sg.h should be independend of the kernel
version and therefore can live in glibc.

This leads to duplication but that's the way Linus likes to have it.

For most of the information that's duplicated, we (glibc developers)
try to get changes into the glibc sources asap. Sorry, that we've
missed to update the scsi headers so far.

I'm taking the scsi headers from 2.3.26 as reference and will try to
come up with a patch for glibc that I will send to for reviewing.

Andreas

P.S. Let's discuss this further privatly, I've set a reply-to.

-- 
 Andreas Jaeger   
  SuSE Labs aj@suse.de	
   private aj@arthur.rhein-neckar.de

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/