Message ID | 87k2cjuw6h.fsf@xmission.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On 11/04, Eric W. Biederman wrote: > > The following mostly correct patch modifies zap_other_threads in > the case of a de_thread to not wait for zombies to be reaped. The only > case that cares is ptrace (as threads are self reaping). So I don't > think this will cause any problems except removing the strace -f race. From my previous email: So the only plan I currently have is change de_thread() to wait until other threads pass exit_notify() or even exit_signals(), but I don't like this. And yes, I don't like this, but perhaps this is what we should do. The patch is incomplete and racy (afaics), and the SIGNAL_GROUP_EXIT checks doesn't look right, but off course technically this change should be simple enough. But not that simple. Just for example, the exiting sub-threads should not run with ->group_leader pointing to nowhere, in case it was reaped by de_thread. And we have another problem with PTRACE_EVENT_EXIT which can lead to the same deadlock. Unfortunately, the semantics of PTRACE_EVENT_EXIT was never defined. But this change will add the user-visible change. And if we add the user-visible changes, then perhaps we could simply untrace the traced sub-threads on exec. This change is simple, we do not even need to touch exec/de_thread, we could just change exit_notify() to ignore ->ptrace if exec is in progress. But I'm afraid we can't do this. Oleg. -- To unsubscribe from this list: send the line "unsubscribe linux-security-module" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 11/04, Oleg Nesterov wrote: > > On 11/04, Eric W. Biederman wrote: > > > > The following mostly correct patch modifies zap_other_threads in > > the case of a de_thread to not wait for zombies to be reaped. The only > > case that cares is ptrace (as threads are self reaping). So I don't > > think this will cause any problems except removing the strace -f race. > > From my previous email: > > So the only plan I currently have is change de_thread() to wait until > other threads pass exit_notify() or even exit_signals(), but I don't > like this. > > And yes, I don't like this, but perhaps this is what we should do. > > The patch is incomplete and racy (afaics), and the SIGNAL_GROUP_EXIT > checks doesn't look right, but off course technically this change should > be simple enough. > > But not that simple. Just for example, the exiting sub-threads should > not run with ->group_leader pointing to nowhere, in case it was reaped > by de_thread. Not to mention other potential problems outside of ptrace/exec. For example userns_install() can fail after mt-exec even without ptrace, simply because thread_group_empty() can be false. Sure, easy to fix, and probably _install() should use signal->live anyway, but still. And I didn't mention the fun with sighand unsharing. We simply can't do this until all sub-threads go away. IOW, your patch breaks the usage of ->siglock. The execing thread and the zombie threads will use different locks to, say, remove the task from thread-group. Again, this is fixable, but not that simple. > And we have another problem with PTRACE_EVENT_EXIT which can lead to the > same deadlock. Unfortunately, the semantics of PTRACE_EVENT_EXIT was never > defined. But this change will add the user-visible change. > > And if we add the user-visible changes, then perhaps we could simply untrace > the traced sub-threads on exec. This change is simple, we do not even need > to touch exec/de_thread, we could just change exit_notify() to ignore ->ptrace > if exec is in progress. But I'm afraid we can't do this. Eric, I hope you see my emails, I got the "Undelivered Mail Returned to Sender" ... This is the mail system at host mail.kernel.org. ... <ebiederm@xmission.com> (expanded from <security@kernel.org>): host mx.xmission.com[166.70.12.20] said: 550-XM-RJCT16: SPF Failure (ip=198.145.29.136, frm=oleg@redhat.com, 550 result=fail) (in reply to RCPT TO command) right now I have no idea what does this mean. Oleg. -- To unsubscribe from this list: send the line "unsubscribe linux-security-module" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Nov 4, 2016 at 11:45 AM, Oleg Nesterov <oleg@redhat.com> wrote: > Eric, I hope you see my emails, I got the "Undelivered Mail Returned to Sender" > ... > This is the mail system at host mail.kernel.org. > ... > <ebiederm@xmission.com> (expanded from <security@kernel.org>): host > mx.xmission.com[166.70.12.20] said: 550-XM-RJCT16: SPF Failure > (ip=198.145.29.136, frm=oleg@redhat.com, 550 result=fail) (in reply to RCPT > TO command) > > right now I have no idea what does this mean. This is a problem for Google folks too sometimes. This is saying that xmission.com is checking redhat.com's SPF records and refusing to let kernel.org deliver email as if it were redhat.com (due to security@kernel.org being an alias not a mailing list). There aren't good solutions for this, but best I've found is to have my security@kernel.org alias be a @kernel.org address instead of an @google.com address... -Kees
Kees Cook <keescook@chromium.org> writes: > On Fri, Nov 4, 2016 at 11:45 AM, Oleg Nesterov <oleg@redhat.com> wrote: >> Eric, I hope you see my emails, I got the "Undelivered Mail Returned to Sender" >> ... Oleg I can receive your messages directly and through vger.kernel.org lists, but I can't receive them through the email reflector at security@kernel.org. >> This is the mail system at host mail.kernel.org. >> ... >> <ebiederm@xmission.com> (expanded from <security@kernel.org>): host >> mx.xmission.com[166.70.12.20] said: 550-XM-RJCT16: SPF Failure >> (ip=198.145.29.136, frm=oleg@redhat.com, 550 result=fail) (in reply to RCPT >> TO command) >> >> right now I have no idea what does this mean. > > This is a problem for Google folks too sometimes. This is saying that > xmission.com is checking redhat.com's SPF records and refusing to let > kernel.org deliver email as if it were redhat.com (due to > security@kernel.org being an alias not a mailing list). There aren't > good solutions for this, but best I've found is to have my > security@kernel.org alias be a @kernel.org address instead of an > @google.com address... Ugh. Is even redhat configuring the redhat email to do that? I will have to look. Last I looked xmission.com was just enforcing the policy that the other mail domains were asking to be enforced on themselves. But those are policies that are incompatible with mailing lists in general. Although I do get confused about which part SPF and DKIM play in this mess. I just remember that the last several ``enhancements'' to email were busily breaking mailing lists and I thought they were completely insane. I can even find evidence that it is (or at least was) so bad that email standards comittee member's can't comminicate with each other via email lists. vger.kernel.org appears to rewrite the envelope sender to avoid problems. If xmission is doing any more than just performing what the domain of the senders of email asked them to do I will be happy to see if I can to sort it out. Eric -- To unsubscribe from this list: send the line "unsubscribe linux-security-module" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Nov 08, 2016 at 04:46:44PM -0600, Eric W. Biederman wrote: > Kees Cook <keescook@chromium.org> writes: ... > > This is a problem for Google folks too sometimes. This is saying that > > xmission.com is checking redhat.com's SPF records and refusing to let > > kernel.org deliver email as if it were redhat.com (due to > > security@kernel.org being an alias not a mailing list). There aren't > > good solutions for this, but best I've found is to have my > > security@kernel.org alias be a @kernel.org address instead of an > > @google.com address... > > Ugh. Is even redhat configuring the redhat email to do that? > I will have to look. > > Last I looked xmission.com was just enforcing the policy that the other > mail domains were asking to be enforced on themselves. But those are > policies that are incompatible with mailing lists in general. Although > I do get confused about which part SPF and DKIM play in this mess. > > I just remember that the last several ``enhancements'' to email were > busily breaking mailing lists and I thought they were completely insane. > I can even find evidence that it is (or at least was) so bad that email > standards comittee member's can't comminicate with each other via email > lists. > > vger.kernel.org appears to rewrite the envelope sender to avoid > problems. Envelope sender rewriting is insufficient, the From: lines need to be rewritten to be compliant. This is a pain in the ass for the @kvack.org mailing lists as well -- people with @google.com addresses don't see the mailing list postings of users from @google.com and other domains using "enhanced" email header "validation" techniques. -ben > If xmission is doing any more than just performing what the domain of > the senders of email asked them to do I will be happy to see if I can > to sort it out. > > Eric
Benjamin LaHaise <bcrl@kvack.org> writes: > On Tue, Nov 08, 2016 at 04:46:44PM -0600, Eric W. Biederman wrote: >> Kees Cook <keescook@chromium.org> writes: > ... >> > This is a problem for Google folks too sometimes. This is saying that >> > xmission.com is checking redhat.com's SPF records and refusing to let >> > kernel.org deliver email as if it were redhat.com (due to >> > security@kernel.org being an alias not a mailing list). There aren't >> > good solutions for this, but best I've found is to have my >> > security@kernel.org alias be a @kernel.org address instead of an >> > @google.com address... >> >> Ugh. Is even redhat configuring the redhat email to do that? >> I will have to look. >> >> Last I looked xmission.com was just enforcing the policy that the other >> mail domains were asking to be enforced on themselves. But those are >> policies that are incompatible with mailing lists in general. Although >> I do get confused about which part SPF and DKIM play in this mess. >> >> I just remember that the last several ``enhancements'' to email were >> busily breaking mailing lists and I thought they were completely insane. >> I can even find evidence that it is (or at least was) so bad that email >> standards comittee member's can't comminicate with each other via email >> lists. >> >> vger.kernel.org appears to rewrite the envelope sender to avoid >> problems. > > Envelope sender rewriting is insufficient, the From: lines need to be > rewritten to be compliant. This is a pain in the ass for the @kvack.org > mailing lists as well -- people with @google.com addresses don't see the > mailing list postings of users from @google.com and other domains using > "enhanced" email header "validation" techniques. That definitely happens in the worst case. At least for Oleg something less serious is happening because the from header does not get changed and the email gets to me through the vger.kernel.org lists. Eric -- To unsubscribe from this list: send the line "unsubscribe linux-security-module" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/kernel/exit.c b/kernel/exit.c index 9d68c45ebbe3..8c8556cab655 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -109,7 +109,8 @@ static void __exit_signal(struct task_struct *tsk) * If there is any task waiting for the group exit * then notify it: */ - if (sig->notify_count > 0 && !--sig->notify_count) + if ((sig->flags & SIGNAL_GROUP_EXIT) && + sig->notify_count > 0 && !--sig->notify_count) wake_up_process(sig->group_exit_task); if (tsk == sig->curr_target) @@ -690,6 +691,10 @@ static void exit_notify(struct task_struct *tsk, int group_dead) if (tsk->exit_state == EXIT_DEAD) list_add(&tsk->ptrace_entry, &dead); + if (!(tsk->signal->flags & SIGNAL_GROUP_EXIT) && + tsk->signal->notify_count > 0 && !--tsk->signal->notify_count) + wake_up_process(tsk->signal->group_exit_task); + /* mt-exec, de_thread() is waiting for group leader */ if (unlikely(tsk->signal->notify_count < 0)) wake_up_process(tsk->signal->group_exit_task); diff --git a/kernel/signal.c b/kernel/signal.c index 75761acc77cf..a3a5cd8dad0f 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1194,7 +1194,9 @@ int zap_other_threads(struct task_struct *p) while_each_thread(p, t) { task_clear_jobctl_pending(t, JOBCTL_PENDING_MASK); - count++; + if ((t->signal->flags & SIGNAL_GROUP_EXIT) || + !t->exit_state) + count++; /* Don't bother with already dead threads */ if (t->exit_state)