Re: [CRIU] [PATCH 3/3] signalfd: add ability to read siginfo-s without dequeuing signals (v2)

On 02/10, Andrew Vagin wrote:

On Sat, Feb 09, 2013 at 11:53:04PM +0100, Michael Kerrisk (man-pages) wrote:

quoted

This looks promising, but I am not sure I understand the user-space
API. Could you explain how it would look to (say) pull all per-thread
signals from user space?

	siginfo_t *signals = malloc(...);
	...

	ret = ptrace(PTRACE_ATTACH, pid, NULL, NULL);

Yes. I assume this should not be a problem since c/r needs to ptrace
the task anyway?

For shared (per-process) signals one line should be changed:
 ret = ptrace(PTRACE_PEEKSIGNAL, pid, PTRACE_PEEK_SHARED + i, signals + i);

Yes. Except I'd say "PTRACE_PEEK_SHARED | i" to clarify that
PTRACE_PEEK_SHARED is "flag". Note that we can more flags, say,
PTRACE_PEEK_DEQUEUE which obviously means "dequeue signal" (but in
this case we should filter out SIGKILL/SIGSTOP).

And note that either way (I mean ptrace or signalfd) this can not
actually peek a signal which doesn't have the sigqueue (say,
SEND_SIG_FORCED). I guess this is not the problem, c/r can look
at /proc/pid/status#*Pnd aditionally.

OK, you seem to agree with this approach, I'll try to make the
working patch tomorrow. At least ptrace_peek_signal() needs another
is_compat argument for compat_ptrace_request().

Oleg.

quoted

quoted

--- x/kernel/ptrace.c
+++ x/kernel/ptrace.c

@@ -618,6 +618,35 @@ static int ptrace_setsiginfo(struct task
        return error;
 }

+static int ptrace_peek_signal(struct task_struct *child,
+                               unsigned long addr, siginfo_t __user *uinfo)
+{
+       siginfo_t info;
+       struct sigpending *pending;
+       int ret = -ESOMETHING;
+
+       pending = &child->pending;
+       if (addr & PTRACE_PEEK_SHARED) {
+               addr &= ~PTRACE_PEEK_SHARED;
+               pending = &child->signal->shared_pending;
+       }
+
+       spin_lock_irq(&child->sighand->siglock);
+       list_for_each_entry(q, &pending->list, list) {
+               if (!addr--) {
+                       copy_siginfo(info, &q->info);
+                       ret = 0;
+                       break;
+               }
+       }
+       spin_lock_irq(&child->sighand->siglock);
+
+       if (!ret)
+               ret = copy_siginfo_to_user(uinfo, info);
+       if (!ret)
+               ret = __put_user(info, si_code);
+       return ret;
+}

 #ifdef PTRACE_SINGLESTEP
 #define is_singlestep(request)         ((request) == PTRACE_SINGLESTEP)

@@ -742,6 +771,10 @@ int ptrace_request(struct task_struct *c
                ret = put_user(child->ptrace_message, datalp);
                break;

+       case PTRACE_PEEKSIGNAL:
+               ret = ptrace_peek_signal(child, addr, datavp);
+               break;
+
        case PTRACE_GETSIGINFO:
                ret = ptrace_getsiginfo(child, &siginfo);
                if (!ret)

--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Author of "The Linux Programming Interface"; http://man7.org/tlpi/
_______________________________________________
CRIU mailing list
CRIU-GEFAQzZX7r8dnm+yROfE0A@public.gmane.org
https://lists.openvz.org/mailman/listinfo/criu