On Tue, 24 Apr 2007, Arlin Davis wrote:
> Roland Dreier wrote: > > > > ./dapl/udapl/linux/dapl_osd.h:82:24: error: asm/atomic.h: No such file or > > directory > > > In file included from ./dapl/include/dapl.h:50, > > > from dapl/udapl/dapl_init.c:39: > > > ./dapl/udapl/linux/dapl_osd.h: In function 'dapl_os_atomic_inc': > > > ./dapl/udapl/linux/dapl_osd.h:163: warning: implicit declaration of > > function 'IA64_FETCHADD' > > > > I seem to recall udapl does some very bogus things with <asm/atomic.h> > > and atomic operations in general. > > > > The built-in atomics in gcc had some ia64 issues until gcc 4.1.1 so we had no > choice but to get down and do some bogus things. :^) > > Turns out Redhat 5 removed asm/atomic.h to force users back to the built-in > atomics which is fine given that Redhat 5 provides gcc 4.1.1. We will go back > and revisit all the hacks. In the short term, for OFED 1.2, we may have to > simply add another conditional in the build (configure.in) to handle the > Redhat 5 case. > > James, do you have any input here? No. That is a good summary. _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
