On 4.7.2012 16:35, Martin Decky wrote: > the current observation is that revisions up to 1500 (and probably the > latest mainline, too -- untested yet) boot fine on PE 7150 if compiled > by the original toolchain from revision 1000 (i.e. GCC 4.6). > > I am now trying to bisect according to GCC versions and find out what is > the root cause of the regression. But it is likely related to the > compilation process, not to any particular change in the source code.
For the record, this is most likely the same thing as #465, caused by a GCC bug 53975. We may want to switch to the previous ia64 toolchain or try the suggested workaround, i.e. build with: -fno-fselective-scheduling -fno-fselective-scheduling2 Jakub _______________________________________________ HelenOS-devel mailing list [email protected] http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
