|
From: Aaron S. <acs...@li...> - 2017-05-14 16:03:02
|
On Fri, 2017-05-12 at 10:55 +0200, Julian Seward wrote: > Yeah, to echo Ivo's point -- what you're looking at is a list of the > blocks > the first time they are executed -- because at that point they have > to be > instrumented. It's not an execution trace. All this really tells > you is > that the block at 0x6D04798 (the call target) has already been > instrumented. > You should be able to find evidence of that earlier in the debug > output. > > So to return to the original question -- do you have some other > reason to > believe that bits of code that should be executed, are being missed? > Julian, I tried it again, and this time set a breakpoint at the call target before trying to single step the bl at 0x6cf1b84. That breakpoint was triggered, so the problem seems to be just that stepi is allowing the subroutine call to execute, i.e. behaving like nexti instead. The segfault is probably due to issues relating to the new p9 instructions. Thanks for your assistance with this. Aaron -- Aaron Sawdey, Ph.D. acs...@li... 050-2/C113 (507) 253-7520 home: 507/263-0782 IBM Linux Technology Center - PPC Toolchain |