/* net/atm/signaling.c - ATM signaling */

/* Written 1995-2000 by Werner Almesberger, EPFL LRC/ICA */

#define pr_fmt(fmt) KBUILD_MODNAME ":%s: " fmt, __func__

#include <linux/errno.h>	/* error codes */
#include <linux/kernel.h>	/* printk */
#include <linux/skbuff.h>
#include <linux/wait.h>
#include <linux/sched.h>	/* jiffies and HZ */
#include <linux/atm.h>		/* ATM stuff */
#include <linux/atmsap.h>
#include <linux/atmsvc.h>
#include <linux/atmdev.h>
#include <linux/bitops.h>
#include <linux/slab.h>

#include "resources.h"
#include "signaling.h"

struct atm_vcc *sigd = NULL;

static void sigd_put_skb(struct sk_buff *skb)
{
	if (!sigd) {
		pr_debug("atmsvc: no signaling daemon\n");
		kfree_skb(skb);
		return;
	}
	atm_force_charge(sigd, skb->truesize);
	skb_queue_tail(&sk_atm(sigd)->sk_receive_queue, skb);
	sk_atm(sigd)->sk_data_ready(sk_atm(sigd));
}

static void modify_qos(struct atm_vcc *vcc, struct atmsvc_msg *msg)
{
	struct sk_buff *skb;

	if (test_bit(ATM_VF_RELEASED, &vcc->flags) ||
	    !test_bit(ATM_VF_READY, &vcc->flags))
		return;
	msg->type = as_error;
	if (!vcc->dev->ops->change_qos)
		msg->reply = -EOPNOTSUPP;
	else {
		/* should lock VCC */
		msg->reply = vcc->dev->ops->change_qos(vcc, &msg->qos,
						       msg->reply);
		if (!msg->reply)
			msg->type = as_okay;
	}
	/*
	 * Should probably just turn around the old skb. But the, the buffer
	 * space accounting needs to follow the change too. Maybe later.
	 */
	while (!(skb = alloc_skb(sizeof(struct atmsvc_msg), GFP_KERNEL)))
		schedule();
	*(struct atmsvc_msg *)skb_put(skb, sizeof(struct atmsvc_msg)) = *msg;
	sigd_put_skb(skb);
}

static int sigd_send(struct atm_vcc *vcc, struct sk_buff *skb)
{
	struct atmsvc_msg *msg;
	struct atm_vcc *session_vcc;
	struct sock *sk;

	msg = (struct atmsvc_msg *) skb->data;
	atomic_sub(skb->truesize, &sk_atm(vcc)->sk_wmem_alloc);
	vcc = *(struct atm_vcc **) &msg->vcc;
	pr_debug("%d (0x%lx)\n", (int)msg->type, (unsigned long)vcc);
	sk = sk_atm(vcc);

	switch (msg->type) {
	case as_okay:
		sk->sk_err = -msg->reply;
		clear_bit(ATM_VF_WAITING, &vcc->flags);
		if (!*vcc->local.sas_addr.prv && !*vcc->local.sas_addr.pub) {
			vcc->local.sas_family = AF_ATMSVC;
			memcpy(vcc->local.sas_addr.prv,
			       msg->local.sas_addr.prv, ATM_ESA_LEN);
			memcpy(vcc->local.sas_addr.pub,
			       msg->local.sas_addr.pub, ATM_E164_LEN + 1);
		}
		session_vcc = vcc->session ? vcc->session : vcc;
		if (session_vcc->vpi || session_vcc->vci)
			break;
		session_vcc->itf = msg->pvc.sap_addr.itf;
		session_vcc->vpi = msg->pvc.sap_addr.vpi;
		session_vcc->vci = msg->pvc.sap_addr.vci;
		if (session_vcc->vpi || session_vcc->vci)
			session_vcc->qos = msg->qos;
		break;
	case as_error:
		clear_bit(ATM_VF_REGIS, &vcc->flags);
		clear_bit(ATM_VF_READY, &vcc->flags);
		sk->sk_err = -msg->reply;
		clear_bit(ATM_VF_WAITING, &vcc->flags);
		break;
	case as_indicate:
		vcc = *(struct atm_vcc **)&msg->listen_vcc;
		sk = sk_atm(vcc);
		pr_debug("as_indicate!!!\n");
		lock_sock(sk);
		if (sk_acceptq_is_full(sk)) {
			sigd_enq(NULL, as_reject, vcc, NULL, NULL);
			dev_kfree_skb(skb);
			goto as_indicate_complete;
		}
		sk->sk_ack_backlog++;
		skb_queue_tail(&sk->sk_receive_queue, skb);
		pr_debug("waking sk_sleep(sk) 0x%p\n", sk_sleep(sk));
		sk->sk_state_change(sk);
as_indicate_complete:
		release_sock(sk);
		return 0;
	case as_close:
		set_bit(ATM_VF_RELEASED, &vcc->flags);
		vcc_release_async(vcc, msg->reply);
		goto out;
	case as_modify:
		modify_qos(vcc, msg);
		break;
	case as_addparty:
	case as_dropparty:
		sk->sk_err_soft = -msg->reply;
					/* < 0 failure, otherwise ep_ref */
		clear_bit(ATM_VF_WAITING, &vcc->flags);
		break;
	default:
		pr_alert("bad message type %d\n", (int)msg->type);
		return -EINVAL;
	}
	sk->sk_state_change(sk);
out:
	dev_kfree_skb(skb);
	return 0;
}

void sigd_enq2(struct atm_vcc *vcc, enum atmsvc_msg_type type,
	       struct atm_vcc *listen_vcc, const struct sockaddr_atmpvc *pvc,
	       const struct sockaddr_atmsvc *svc, const struct atm_qos *qos,
	       int reply)
{
	struct sk_buff *skb;
	struct atmsvc_msg *msg;
	static unsigned int session = 0;

	pr_debug("%d (0x%p)\n", (int)type, vcc);
	while (!(skb = alloc_skb(sizeof(struct atmsvc_msg), GFP_KERNEL)))
		schedule();
	msg = (struct atmsvc_msg *)skb_put(skb, sizeof(struct atmsvc_msg));
	memset(msg, 0, sizeof(*msg));
	msg->type = type;
	*(struct atm_vcc **) &msg->vcc = vcc;
	*(struct atm_vcc **) &msg->listen_vcc = listen_vcc;
	msg->reply = reply;
	if (qos)
		msg->qos = *qos;
	if (vcc)
		msg->sap = vcc->sap;
	if (svc)
		msg->svc = *svc;
	if (vcc)
		msg->local = vcc->local;
	if (pvc)
		msg->pvc = *pvc;
	if (vcc) {
		if (type == as_connect && test_bit(ATM_VF_SESSION, &vcc->flags))
			msg->session = ++session;
			/* every new pmp connect gets the next session number */
	}
	sigd_put_skb(skb);
	if (vcc)
		set_bit(ATM_VF_REGIS, &vcc->flags);
}

void sigd_enq(struct atm_vcc *vcc, enum atmsvc_msg_type type,
	      struct atm_vcc *listen_vcc, const struct sockaddr_atmpvc *pvc,
	      const struct sockaddr_atmsvc *svc)
{
	sigd_enq2(vcc, type, listen_vcc, pvc, svc, vcc ? &vcc->qos : NULL, 0);
	/* other ISP applications may use "reply" */
}

static void purge_vcc(struct atm_vcc *vcc)
{
	if (sk_atm(vcc)->sk_family == PF_ATMSVC &&
	    !test_bit(ATM_VF_META, &vcc->flags)) {
		set_bit(ATM_VF_RELEASED, &vcc->flags);
		clear_bit(ATM_VF_REGIS, &vcc->flags);
		vcc_release_async(vcc, -EUNATCH);
	}
}

static void sigd_close(struct atm_vcc *vcc)
{
	struct sock *s;
	int i;

	pr_debug("\n");
	sigd = NULL;
	if (skb_peek(&sk_atm(vcc)->sk_receive_queue))
		pr_err("closing with requests pending\n");
	skb_queue_purge(&sk_atm(vcc)->sk_receive_queue);

	read_lock(&vcc_sklist_lock);
	for (i = 0; i < VCC_HTABLE_SIZE; ++i) {
		struct hlist_head *head = &vcc_hash[i];

		sk_for_each(s, head) {
			vcc = atm_sk(s);

			purge_vcc(vcc);
		}
	}
	read_unlock(&vcc_sklist_lock);
}

static struct atmdev_ops sigd_dev_ops = {
	.close = sigd_close,
	.send =	sigd_send
};

static struct atm_dev sigd_dev = {
	.ops =		&sigd_dev_ops,
	.type =		"sig",
	.number =	999,
	.lock =		__SPIN_LOCK_UNLOCKED(sigd_dev.lock)
};

int sigd_attach(struct atm_vcc *vcc)
{
	if (sigd)
		return -EADDRINUSE;
	pr_debug("\n");
	sigd = vcc;
	vcc->dev = &sigd_dev;
	vcc_insert_socket(sk_atm(vcc));
	set_bit(ATM_VF_META, &vcc->flags);
	set_bit(ATM_VF_READY, &vcc->flags);
	return 0;
}
the event will have been detached from the task, but the
grouping will have been retained, such that group operations might
still work properly while there are live child events etc.

This does however mean that we can miss a perf_group_detach() call
when the group decomposes, this in turn can then lead to
use-after-free.

Fix it by explicitly doing the group detach if its still required.

Reported-by: Dmitry Vyukov &lt;dvyukov@google.com&gt;
Tested-by: Dmitry Vyukov &lt;dvyukov@google.com&gt;
Signed-off-by: Peter Zijlstra (Intel) &lt;peterz@infradead.org&gt;
Cc: Alexander Shishkin &lt;alexander.shishkin@linux.intel.com&gt;
Cc: Arnaldo Carvalho de Melo &lt;acme@kernel.org&gt;
Cc: Arnaldo Carvalho de Melo &lt;acme@redhat.com&gt;
Cc: Jiri Olsa &lt;jolsa@redhat.com&gt;
Cc: Linus Torvalds &lt;torvalds@linux-foundation.org&gt;
Cc: Mathieu Desnoyers &lt;mathieu.desnoyers@efficios.com&gt;
Cc: Peter Zijlstra &lt;peterz@infradead.org&gt;
Cc: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Cc: stable@vger.kernel.org # v4.5+
Cc: syzkaller &lt;syzkaller@googlegroups.com&gt;
Fixes: 63b6da39bb38 ("perf: Fix perf_event_exit_task() race")
Link: http://lkml.kernel.org/r/20170126153955.GD6515@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar &lt;mingo@kernel.org&gt;
</div><div class='diffstat-header'><a href='/cgit.cgi/linux/net-next.git/diff/?id=a76a82a3e38c8d3fb6499e3dfaeb0949241ab588'>Diffstat</a> (limited to 'include/trace/events/skb.h')</div><table summary='diffstat' class='diffstat'>