public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Badness in local_bh_enable at kernel/softirq.c:122
@ 2004-05-27 17:45 Jurjen Oskam
  2004-05-27 18:19 ` Jurjen Oskam
                   ` (2 more replies)
  0 siblings, 3 replies; 24+ messages in thread
From: Jurjen Oskam @ 2004-05-27 17:45 UTC (permalink / raw)
  To: linux-kernel

Hi there,

I'm using pptp-linux to connect to my ADSL-provider. I've been doing this
for several years now, without significant problems. A week ago, I
upgraded

May 27 11:35:41 calvin kernel: Badness in local_bh_enable at kernel/softirq.c:122
May 27 11:35:41 calvin kernel: Call Trace:
May 27 11:35:41 calvin kernel:  [<c011ac00>] local_bh_enable+0x60/0x80
May 27 11:35:41 calvin kernel:  [<c4cc2b54>] ppp_sync_push+0x54/0x140 [ppp_synctty]
May 27 11:35:41 calvin kernel:  [<c4cc25a5>] ppp_sync_wakeup+0x25/0x60 [ppp_synctty]
May 27 11:35:41 calvin kernel:  [<c01f20e7>] pty_unthrottle+0x47/0x60
May 27 11:35:41 calvin kernel:  [<c01eefb1>] check_unthrottle+0x31/0x40
May 27 11:35:41 calvin kernel:  [<c01ef02b>] n_tty_flush_buffer+0xb/0x60
May 27 11:35:41 calvin kernel:  [<c01f2488>] pty_flush_buffer+0x48/0x60
May 27 11:35:41 calvin kernel:  [<c01ec116>] do_tty_hangup+0x2d6/0x340
May 27 11:35:41 calvin kernel:  [<c01ed387>] release_dev+0x547/0x580
May 27 11:35:41 calvin kernel:  [<c010c038>] timer_interrupt+0xb8/0xc0
May 27 11:35:41 calvin kernel:  [<c012544d>] rcu_check_quiescent_state+0x4d/0x60
May 27 11:35:41 calvin kernel:  [<c012551e>] rcu_process_callbacks+0xbe/0xe0
May 27 11:35:41 calvin kernel:  [<c011ad7a>] tasklet_action+0x3a/0x60
May 27 11:35:41 calvin kernel:  [<c0108805>] do_IRQ+0xc5/0xe0
May 27 11:35:41 calvin kernel:  [<c0106ed8>] common_interrupt+0x18/0x20
May 27 11:35:41 calvin kernel:  [<c01ed709>] tty_release+0x9/0x20
May 27 11:35:41 calvin kernel:  [<c0145532>] __fput+0xd2/0x100
May 27 11:35:41 calvin kernel:  [<c0143ee3>] filp_close+0x43/0x80
May 27 11:35:41 calvin kernel:  [<c0118795>] put_files_struct+0x55/0xc0
May 27 11:35:41 calvin kernel:  [<c01192eb>] do_exit+0x14b/0x2e0
May 27 11:35:41 calvin kernel:  [<c0119548>] do_group_exit+0x48/0x80
May 27 11:35:41 calvin kernel:  [<c0120bf9>] get_signal_to_deliver+0x199/0x2c0
May 27 11:35:41 calvin kernel:  [<c0106a65>] do_signal+0x45/0xc0
May 27 11:35:41 calvin kernel:  [<c01134ea>] recalc_task_prio+0x8a/0x1a0
May 27 11:35:41 calvin kernel:  [<c014463d>] vfs_read+0xbd/0xe0
May 27 11:35:41 calvin kernel:  [<c014482b>] sys_read+0x2b/0x60
May 27 11:35:41 calvin kernel:  [<c0106b29>] do_notify_resume+0x49/0x50
May 27 11:35:41 calvin kernel:  [<c0106d0e>] work_notifysig+0x13/0x15
May 27 11:35:41 calvin kernel:


Each time the pppd link was retried, the "Badness in..." message appeared.

This is on a 450 MHz K6-II.
-- 
Jurjen Oskam

"Avoid putting a paging file on a fault-tolerant drive, such as a mirrored
volume or a RAID-5 volume. Paging files do not need fault-tolerance."-MS Q308417

^ permalink raw reply	[flat|nested] 24+ messages in thread
* Re: [PATCH][RFC] 2.6.6 tty_io.c hangup locking
@ 2004-06-13 15:05 Paul Fulghum
  0 siblings, 0 replies; 24+ messages in thread
From: Paul Fulghum @ 2004-06-13 15:05 UTC (permalink / raw)
  To: Jurjen Oskam; +Cc: linux-kernel@vger.kernel.org

[-- Attachment #1: Type: text/plain, Size: 300 bytes --]

And Evolution mangled it again.

I can't seem to prevent Evolution from destroying
the message contents, so let's try forwarding the
original patch as an attachment.

If that does not work, this will have to wait
until I can get into work tomorrow, where I have a
working mail client.

Thanks,
Paul


[-- Attachment #2: Forwarded message - [PATCH][RFC] 2.6.6 ppp_synctty.c --]
[-- Type: message/rfc822, Size: 8677 bytes --]

From: Paul Fulghum <paulkf@microgate.com>
To: linux-kernel <linux-kernel@vger.kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Subject: [PATCH][RFC] 2.6.6 ppp_synctty.c
Date: 02 Jun 2004 15:49:12 -0500
Message-ID: <1086209351.2163.7.camel@deimos.microgate.com>

This patch adapts the changes made by Paul Mackerras
to ppp_async.c which allow for the fact that
a line discipline receive and wakeup callbacks
can be called at hard interrupt context and/or
with interrupts disabled.

I have tested with a synchronous serial PPP connection.

Comments and other testing requested.

Thanks,
Paul

--
Paul Fulghum
paulkf@microgate.com

--- linux-2.6.6/drivers/net/ppp_synctty.c	2004-04-03 21:36:57.000000000 -0600
+++ linux-2.6.6-mg1/drivers/net/ppp_synctty.c	2004-06-02 15:01:35.177725315 -0500
@@ -65,7 +65,9 @@
 	struct sk_buff	*tpkt;
 	unsigned long	last_xmit;
 
-	struct sk_buff	*rpkt;
+	struct sk_buff_head rqueue;
+
+	struct tasklet_struct tsk;
 
 	atomic_t	refcnt;
 	struct semaphore dead_sem;
@@ -88,6 +90,7 @@
 static int ppp_sync_send(struct ppp_channel *chan, struct sk_buff *skb);
 static int ppp_sync_ioctl(struct ppp_channel *chan, unsigned int cmd,
 			  unsigned long arg);
+static void ppp_sync_process(unsigned long arg);
 static int ppp_sync_push(struct syncppp *ap);
 static void ppp_sync_flush_output(struct syncppp *ap);
 static void ppp_sync_input(struct syncppp *ap, const unsigned char *buf,
@@ -217,6 +220,9 @@
 	ap->xaccm[3] = 0x60000000U;
 	ap->raccm = ~0U;
 
+	skb_queue_head_init(&ap->rqueue);
+	tasklet_init(&ap->tsk, ppp_sync_process, (unsigned long) ap);
+
 	atomic_set(&ap->refcnt, 1);
 	init_MUTEX_LOCKED(&ap->dead_sem);
 
@@ -267,10 +273,10 @@
 	 */
 	if (!atomic_dec_and_test(&ap->refcnt))
 		down(&ap->dead_sem);
+	tasklet_kill(&ap->tsk);
 
 	ppp_unregister_channel(&ap->chan);
-	if (ap->rpkt != 0)
-		kfree_skb(ap->rpkt);
+	skb_queue_purge(&ap->rqueue);
 	if (ap->tpkt != 0)
 		kfree_skb(ap->tpkt);
 	kfree(ap);
@@ -369,17 +375,24 @@
 	return 65535;
 }
 
+/*
+ * This can now be called from hard interrupt level as well
+ * as soft interrupt level or mainline.
+ */
 static void
 ppp_sync_receive(struct tty_struct *tty, const unsigned char *buf,
-		  char *flags, int count)
+		  char *cflags, int count)
 {
 	struct syncppp *ap = sp_get(tty);
+	unsigned long flags;
 
 	if (ap == 0)
 		return;
-	spin_lock_bh(&ap->recv_lock);
-	ppp_sync_input(ap, buf, flags, count);
-	spin_unlock_bh(&ap->recv_lock);
+	spin_lock_irqsave(&ap->recv_lock, flags);
+	ppp_sync_input(ap, buf, cflags, count);
+	spin_unlock_irqrestore(&ap->recv_lock, flags);
+	if (skb_queue_len(&ap->rqueue))
+		tasklet_schedule(&ap->tsk);
 	sp_put(ap);
 	if (test_and_clear_bit(TTY_THROTTLED, &tty->flags)
 	    && tty->driver->unthrottle)
@@ -394,8 +407,8 @@
 	clear_bit(TTY_DO_WRITE_WAKEUP, &tty->flags);
 	if (ap == 0)
 		return;
-	if (ppp_sync_push(ap))
-		ppp_output_wakeup(&ap->chan);
+	set_bit(XMIT_WAKEUP, &ap->xmit_flags);
+	tasklet_schedule(&ap->tsk);
 	sp_put(ap);
 }
 
@@ -449,9 +462,9 @@
 		if (get_user(val, (int *) arg))
 			break;
 		ap->flags = val & ~SC_RCV_BITS;
-		spin_lock_bh(&ap->recv_lock);
+		spin_lock_irq(&ap->recv_lock);
 		ap->rbits = val & SC_RCV_BITS;
-		spin_unlock_bh(&ap->recv_lock);
+		spin_unlock_irq(&ap->recv_lock);
 		err = 0;
 		break;
 
@@ -512,6 +525,32 @@
 }
 
 /*
+ * This is called at softirq level to deliver received packets
+ * to the ppp_generic code, and to tell the ppp_generic code
+ * if we can accept more output now.
+ */
+static void ppp_sync_process(unsigned long arg)
+{
+	struct syncppp *ap = (struct syncppp *) arg;
+	struct sk_buff *skb;
+
+	/* process received packets */
+	while ((skb = skb_dequeue(&ap->rqueue)) != NULL) {
+		if (skb->len == 0) {
+			/* zero length buffers indicate error */
+			ppp_input_error(&ap->chan, 0);
+			kfree_skb(skb);
+		}
+		else
+			ppp_input(&ap->chan, skb);
+	}
+
+	/* try to push more stuff out */
+	if (test_bit(XMIT_WAKEUP, &ap->xmit_flags) && ppp_sync_push(ap))
+		ppp_output_wakeup(&ap->chan);
+}
+
+/*
  * Procedures for encapsulation and framing.
  */
 
@@ -600,7 +639,6 @@
 	struct tty_struct *tty = ap->tty;
 	int tty_stuffed = 0;
 
-	set_bit(XMIT_WAKEUP, &ap->xmit_flags);
 	if (!spin_trylock_bh(&ap->xmit_lock))
 		return 0;
 	for (;;) {
@@ -667,15 +705,44 @@
  * Receive-side routines.
  */
 
-static inline void
-process_input_packet(struct syncppp *ap)
+/* called when the tty driver has data for us. 
+ *
+ * Data is frame oriented: each call to ppp_sync_input is considered
+ * a whole frame. If the 1st flag byte is non-zero then the whole
+ * frame is considered to be in error and is tossed.
+ */
+static void
+ppp_sync_input(struct syncppp *ap, const unsigned char *buf,
+		char *flags, int count)
 {
 	struct sk_buff *skb;
 	unsigned char *p;
-	int code = 0;
 
-	skb = ap->rpkt;
-	ap->rpkt = 0;
+	if (count == 0)
+		return;
+
+	if (ap->flags & SC_LOG_INPKT)
+		ppp_print_buffer ("receive buffer", buf, count);
+
+	/* stuff the chars in the skb */
+	if ((skb = dev_alloc_skb(ap->mru + PPP_HDRLEN + 2)) == 0) {
+		printk(KERN_ERR "PPPsync: no memory (input pkt)\n");
+		goto err;
+	}
+	/* Try to get the payload 4-byte aligned */
+	if (buf[0] != PPP_ALLSTATIONS)
+		skb_reserve(skb, 2 + (buf[0] & 1));
+
+	if (flags != 0 && *flags) {
+		/* error flag set, ignore frame */
+		goto err;
+	} else if (count > skb_tailroom(skb)) {
+		/* packet overflowed MRU */
+		goto err;
+	}
+
+	p = skb_put(skb, count);
+	memcpy(p, buf, count);
 
 	/* strip address/control field if present */
 	p = skb->data;
@@ -693,59 +760,15 @@
 	} else if (skb->len < 2)
 		goto err;
 
-	/* pass to generic layer */
-	ppp_input(&ap->chan, skb);
+	/* queue the frame to be processed */
+	skb_queue_tail(&ap->rqueue, skb);
 	return;
 
- err:
-	kfree_skb(skb);
-	ppp_input_error(&ap->chan, code);
-}
-
-/* called when the tty driver has data for us. 
- *
- * Data is frame oriented: each call to ppp_sync_input is considered
- * a whole frame. If the 1st flag byte is non-zero then the whole
- * frame is considered to be in error and is tossed.
- */
-static void
-ppp_sync_input(struct syncppp *ap, const unsigned char *buf,
-		char *flags, int count)
-{
-	struct sk_buff *skb;
-	unsigned char *sp;
-
-	if (count == 0)
-		return;
-
-	/* if flag set, then error, ignore frame */
-	if (flags != 0 && *flags) {
-		ppp_input_error(&ap->chan, *flags);
-		return;
-	}
-
-	if (ap->flags & SC_LOG_INPKT)
-		ppp_print_buffer ("receive buffer", buf, count);
-
-	/* stuff the chars in the skb */
-	if ((skb = ap->rpkt) == 0) {
-		if ((skb = dev_alloc_skb(ap->mru + PPP_HDRLEN + 2)) == 0) {
-			printk(KERN_ERR "PPPsync: no memory (input pkt)\n");
-			ppp_input_error(&ap->chan, 0);
-			return;
-		}
-		/* Try to get the payload 4-byte aligned */
-		if (buf[0] != PPP_ALLSTATIONS)
-			skb_reserve(skb, 2 + (buf[0] & 1));
-		ap->rpkt = skb;
-	}
-	if (count > skb_tailroom(skb)) {
-		/* packet overflowed MRU */
-		ppp_input_error(&ap->chan, 1);
-	} else {
-		sp = skb_put(skb, count);
-		memcpy(sp, buf, count);
-		process_input_packet(ap);
+err:
+	/* queue zero length packet as error indication */
+	if (skb || (skb = dev_alloc_skb(0))) {
+		skb_trim(skb, 0);
+		skb_queue_tail(&ap->rqueue, skb);
 	}
 }
 



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

^ permalink raw reply	[flat|nested] 24+ messages in thread
end of thread, other threads:[~2004-06-13 15:05 UTC | newest]

Thread overview: 24+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-05-27 17:45 Badness in local_bh_enable at kernel/softirq.c:122 Jurjen Oskam
2004-05-27 18:19 ` Jurjen Oskam
2004-05-27 20:41 ` Paul Fulghum
2004-05-28 18:42 ` [PATCH][RFC] 2.6.6 tty_io.c hangup locking Paul Fulghum
2004-05-28 20:11   ` Jurjen Oskam
2004-05-28 20:33     ` Paul Fulghum
2004-05-28 23:06   ` Andrew Morton
2004-05-29 17:45     ` Paul Fulghum
2004-06-01 20:51     ` [PATCH] 2.6.6 synclinkmp.c Paul Fulghum
2004-06-01 20:57       ` Russell King
2004-06-01 21:25         ` Paul Fulghum
2004-06-02 21:22           ` Russell King
2004-06-02 22:04             ` Paul Fulghum
2004-06-02 14:13         ` Paul Fulghum
2004-06-01 20:51     ` [PATCH] 2.6.6 synclink.c Paul Fulghum
2004-06-02 14:15       ` Paul Fulghum
2004-06-01 20:53     ` [PATCH] 2.6.6 synclink_cs.c Paul Fulghum
2004-06-01 21:00       ` Russell King
2004-06-01 23:04         ` Paul Fulghum
2004-06-13  9:05   ` [PATCH][RFC] 2.6.6 tty_io.c hangup locking Jurjen Oskam
2004-06-13 13:29     ` Paul Fulghum
2004-06-13 14:24       ` Jurjen Oskam
2004-06-13 14:39         ` Paul Fulghum
  -- strict thread matches above, loose matches on Subject: below --
2004-06-13 15:05 Paul Fulghum

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox