[8346] in linux-scsi channel archive
Re: udivdi3 and umoddi3
daemon@ATHENA.MIT.EDU (Matti Aarnio)
Fri Mar 10 03:21:42 2000
Date: Fri, 10 Mar 2000 10:16:19 +0200
From: Matti Aarnio <matti.aarnio@sonera.fi>
To: Philip Blundell <Philip.Blundell@pobox.com>
Cc: Eric Youngdale <eric@andante.org>,
Stanley Wu <wudragon2@yahoo.com>, linux-scsi@vger.rutgers.edu
Message-ID: <20000310101619.Z13396@mea.tmt.tele.fi>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
In-Reply-To: <E12TKF1-0001r5-00@kings-cross.london.uk.eu.org>; from Philip Blundell on Fri, Mar 10, 2000 at 07:53:46AM +0000
On Fri, Mar 10, 2000 at 07:53:46AM +0000, Philip Blundell wrote:
> >> unresolved symbol __udivdi3
> >> unresolved symbol __umoddi3
...
> Those functions are only invoked for `long long' arithmetic. One option is
> simply to avoid this in the kernel; the other option is to link the kernel
> with libgcc.a and export its symbols. The latter already happens on some
> architectures.
Yes indeed. Kernel is NOT supposed to include libgcc.a,
because such will allow hiding EXTREMELY slow operations
like these into the system.
You are supposed to figure out why a "long long" value is
divided by some other, and if that can be turned e.g. into
a SHIFT statement ?
> p.
/Matti Aarnio
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
