> John David Anglin wrote: > >> It will currently only work with 64bit kernels, because the gcc options -pg > >> and -ffunction-sections can't be enabled at the same time and -ffunction-sections > >> is still needed to be able to link 32bit kernels. > > > > I have done some testing. As far as I can tell, -pg works with > > -ffunction-sections on hppa-unknown-linux-gnu and hppa64-hp-hpux11. > > That's great! > Which gcc version did you tested? 4.4.0. > For me, (cross-compiler) hppa-linux-gcc 3.3.4 didn't linked without -ffunction-sections. There's been quite a few change since 3.3.4. In principle, -ffunction-sections shouldn't be needed to compile and link the kernel although it may produce better code. Long calls should be generated automatically if a pc-relative branch can't reach the stub table. As I said, ld -r is not supposed to merge .text sections. If I remember correctly from some previous discussions, 3.3.4 is too old. There was a bug in the handling of weak functions affecting kernel builds. Currently, the only supported GCC branches are 4.3 and 4.4. The 4.2 branch will be closed soon and no further releases will be done with this branch. 4.4 will be forked when the number of P1 bugs reaches zero. From my standpoint, it would be best to test with 4.3 or 4.4. I think the major linux vendors are using 4.3 for their current releases. > Did you tested modules (on 32bit) as well, e.g. the xfs module which had the > relocation problems? No, I just did a build and check of gcc, and tested a testsuite program to see that it ran correctly with -pg and -ffunctions-sections (i.e., I ran gprof using the gmon.out). I also visually inspected the generated assembly code. Dave -- J. David Anglin dave.anglin@xxxxxxxxxxxxxx National Research Council of Canada (613) 990-0752 (FAX: 952-6602) -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
