Messages in this thread | | | From | Pedro Alves <> | Subject | Re: [PATCH v3] Make PTRACE_SEIZE set ptrace options specified in 'data' | Date | Sat, 10 Sep 2011 12:19:30 +0100 |
| |
On Friday 09 September 2011 21:03:10, Denys Vlasenko wrote: > On Friday 09 September 2011 19:09, Pedro Alves wrote:
> No need to shout.
Sorry.
> execve is such a rare syscall the one extra stop on it is not > going to be a problem. > > > And about not needing to handle the magic unadorned SIGTRAP. > > The magic unadorned post-exec SIGTRAP does not have `status & 0xff00' > > set, it is not a ptrace event! > > What SIGTRAP? With PTRACE_O_TRACEEXEC, there is no SIGTRAP.
But _without_ PTRACE_O_TRACEEXEC there is. You've raised its existence as justification for needing to be able to set options directly on PTRACE_SEIZE. Point is, if we don't get rid of the SIGTRAP when PTRACE_O_TRACEEXEC is _not_ in effect, then _everyone_ will always pass PTRACE_O_TRACEEXEC to SEIZE. If that is true, you might as well make it default... But I'm claiming that a tracer may not want to see exec events at all, so making it so that when you don't specify PTRACE_O_TRACEEXEC, then you also don't get the magic SIGTRAP, is more useful, and eliminates your justification too. Oleg already showed it's a super trivial patch too. If you want to be able to specify options directly on SEIZE, fine, I can see that it is useful (of course gdb does the same "I still need to set options on this child" dance as strace does).
> > If we don't disable the magic SIGTRAP, there's no way for a > > tracer to do a very non-invasive SEIZE, say, a GDB mode that > > only cares to let the tracer run free to catch SIGSEGVs > > in some child, while later on during the run, the user remembers > > to set a breakpoint. At that point the tracer needs to catch > > exec events, so it'd enable TRACE_O_EVENTEXEC. Getting rid of > > the SIGTRAP gets rid of the spurious stops when TRACE_O_EVENTEXEC > > is not enabled. > > This part I don't understand.
Say, you run the whole of gcc's testsuite under gdb, and let it run until one of the children SIGSEGVs. You do "gdb make; run". Currently, all the children stop momentarily for fork/vfork/exec, which slows down the run significantly (there are thousands of forks/execs). We should be able to only SEIZE the shell that runs "make" (gdb runs the child through the shell, like "sh -c make"), and let all its children run free, the least invasive way possible. When a SIGSEGV happens, gdb can sync up about the process that crashed from /proc.
We can't get rid of the magic SIGTRAP on PTRACE_ATTACH/PTRACE_TRACEME for backwards compatibility reasons, but SEIZE is new.
> (btw, PTRACE_O_TRACEEXEC, not TRACE_O_EVENTEXEC).
Thanks.
-- Pedro Alves
| |