[OpenAFS-devel] not sure what to do about rlim --- also on other 64 bit machines?

Jeffrey Hutzelman jhutz@cmu.edu
Thu, 29 Jun 2006 14:32:45 -0400


On Thursday, June 29, 2006 12:16:55 PM -0400 chas williams - CONTRACTOR 
<chas@cmf.nrl.navy.mil> wrote:

> In message <4B581E744A3961C820B47CE8@mariner.pc.cs.cmu.edu>,Jeffrey
> Hutzelman w rites:
>>> way KBUILD_BASENAME is handled and requires a new macro KBUILD_STR.
>>> i rewrote the linux kernel tests (see openafs bug 34561) to understand
>>> this.  however, it would probably be a good idea to make the tests a bit
>>> more kernel friendly under 2.6 by using the existing kernel build
>>> system. this would always test with the 'right' flags.
>>
>> I agree, as do several others.  Find us a tuit.  Any shape will do; I'll
>> round it off myself if necessary (hm, though that might require a round
>> one  to use as a template...).
>
> i rewrote all the kernel test macros to use a single template so that
> is a pretty good start.  the problem with sles 10 apparently is the same
> as 2.6.17.  some headers require you atleast include autoconf.h now.
>
> using the kernel test environment should be fairly simple, but guessing
> which you need (2.2/2.4 or 2.6 tests) is annoying.

Basically, we need AC_TRY_KBUILD or some such thing.  It's a little messy, 
but doable.