Re: [RFC] Proposal for ptrace improvements

[Denys Vlasenko]

On Friday 04 March 2011 18:12, Linus Torvalds wrote:
> On Fri, Mar 4, 2011 at 9:05 AM, Denys Vlasenko <vda.linux@xxxxxxxxxxxxxx> wrote:
> >
> > Here's a typical sequence of operations strace is doing:
> >
> > wait4(-1, [{WIFSTOPPED(s) && WSTOPSIG(s) == SIGTRAP}], __WALL, NULL) = 10445
> > ptrace(PTRACE_GETSIGINFO, 10445, 0, {si_signo=SIGTRAP, si_code=0x5,
> > si_pid=10445, si_uid=0, si_value={int=0, ptr=0}}) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*ORIG_RAX, [0x36]) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*CS, [0x23]) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*RAX, [0xffffffffffffffda]) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*RBX, [0x1]) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*RCX, [0x5401]) = 0
> > ptrace(PTRACE_PEEKUSER, 10445, 8*RDX, [0xffb83f10]) = 0
> > ptrace(PTRACE_SYSCALL, 10445, 0x1, SIG_0) = 0
> >
> > IOW: strace wastes a lot of time just transiting to the kernel and back
> > with simple syscalls to get the information.
> >
> > What would help here is a "vectorized" waitpid operation
> > which retrieves much more information in one go:
> >
> > waitpid_vec(int max_results, int status_vector[], siginfo_t
> > si_vector[], struct pt_regs reg_vector[])
> 
> No. Start off with what we already have: make ptrace use
> "PTRACE_GETREGSET" instead of PTRACE_PEEKUSER.
> 
> That should get the whole "user_regs_struct" in one single system call.

Aye. That's the idea: get more info in one go.

I bet a part of the rationale for PTRACE_GETREGSET
was to get not one register per syscall, but all of them.

We have more of these:

readv/writev

Get not one signal, but many of them (signalfd).

Get not one task's wait notification, but many of them,
    and also get accompanying siginfo etc (waitpidv?
    or better, waitfd?)

-- 
vda
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/