On Thu, Oct 7, 2010 at 8:08 PM, Alexander Graf <ag...@suse.de> wrote: > > Am 07.10.2010 um 20:59 schrieb Blue Swirl <blauwir...@gmail.com>: > >> On Wed, Oct 6, 2010 at 9:39 PM, Alexander Graf <ag...@suse.de> wrote: >>> >>> On 06.10.2010, at 23:34, Blue Swirl wrote: >>> >>>> Compiling with GCC 4.6.0 20100925 produced warnings: >>>> /src/qemu/target-ppc/op_helper.c: In function 'helper_icbi': >>>> /src/qemu/target-ppc/op_helper.c:351:14: error: variable 'tmp' set but >>>> not used [-Werror=unused-but-set-variable] >>>> /src/qemu/target-ppc/op_helper.c: In function 'do_6xx_tlb': >>>> /src/qemu/target-ppc/op_helper.c:3805:28: error: variable 'EPN' set >>>> but not used [-Werror=unused-but-set-variable] >>>> /src/qemu/target-ppc/op_helper.c: In function 'do_74xx_tlb': >>>> /src/qemu/target-ppc/op_helper.c:3838:28: error: variable 'EPN' set >>>> but not used [-Werror=unused-but-set-variable] >>>> >>>> Fix by making the variable declarations and their uses also conditional >>>> to debug definition. Delete tmp. >>> >>> Maybe it would make more sense to get those LOG_* macros into static inline >>> functions. But for the issue at hand, the solution looks good to me. >> >> Perhaps all conditionally enabled debug printf stuff should be >> transformed into tracepoints? That would be more flexible than >> uncommenting DEBUG_foobar. Would it help with bitrot too? > > I haven't looked at the qemu tracepoints framework yet, but some of this > debug stuff is in hot paths and only very rarely used. So unless it is > overhead free, I'd prefer debug function stubs.
There is zero overhead with 'nop' backend, with other backends there is some overhead. I think only 'nop' backend should be used on production versions, so this should be comparable to current debug stuff.
