- From: Frank Mori Hess <fmhess_at_users.sourceforge.net>
- Date: Tue, 9 Sep 2003 15:28:43 -0500
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Tuesday 09 September 2003 02:32 pm, bvds_at_bvds.geneva.edu wrote: > >On Tue, Sep 09, 2003 at 01:49:25PM -0400, bvds_at_bvds.geneva.edu wrote: > >> Generally, when I have written such code, I used "size_t" instead > >> of "int" to handle this issue. Would that be an acceptable fix here? > > > >No, because it isn't size_t. It's unsigned long, iirc. > > I wasn't sure if the code actually *depended* on signed vs. > unsigned here. (It seemed to me that unsigned would be more > appropriate.) So you would prefer something like: > > #ifdef __x86_64__ > long iobase; > #else > int iobase; > #endif > > in include/linux/comedidev.h > I'm not sure why you want to #ifdef this. I've already changed the iobase parameter that was producing the warnings to unsigned long. I think the official approved C type for holding pointer values would be intptr_t or uintptr_t, but they haven't even made their way into the Linux kernel include files yet. - -- Frank -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE/Xjf+5vihyNWuA4URAuxbAKCw5jQio3T1+Q7iHw06Sw3W7f3LMACeLp22 tSd32aMnSW7XD/zx0qFOB5I= =/a16 -----END PGP SIGNATURE-----
Received on 2003-09-09Z19:28:43