Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

7 messages, 2 authors, 2012-03-21 · open the first message on its own page

Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

From: Justin P. Mattock <hidden>
Date: 2012-02-05 22:59:20 

On Sun, 5 Feb 2012 23:21:41 +0100
Francois Romieu [off-list ref] wrote:

Larry Finger [off-list ref] :
quoted
On 02/05/2012 03:40 PM, Justin P. Mattock wrote:
quoted
this showed up as I woke the machine up from suspend. machine is connected
I just see networkmanager constantly spinning trying to connect then telling me
the wired network is not connected(even though its just wireless that I have).
[ 1273.447349] r8169 0000:06:00.0: Refused to change power state, currently in D3
[ 1273.571024] psmouse serio1: elantech: retrying ps2 command 0xf8 (2).
[ 1273.656709] r8169 0000:06:00.0: eth0: link up
                                         ^^^^^^^

Justin, you did not ever connect a cable, did you ?
nope! was using wireless the whole time. 

[...]
quoted
The error clearly comes from r8169, not from ath9k. If you unload
r8169 before suspending, does this still happen? My suspicion is
that NM is getting confused.
It could hide the problem but the D3 + link up messages are nonetheless
a bit unexpected.
quoted
P.S. Unless it happens with just the ath9k active, this ML is likely
not the correct one.
Yes, please Cc: netdev@vger.kernel.org + me instead.

Thanks.

-- 
Ueimor
removed linux-wireless and added netdev to the cc's
will monitor this and report anything that shows up.

-- 
Justin P. Mattock [off-list ref]

Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

From: Justin Mattock <hidden>
Date: 2012-03-19 04:01:05 

On Sun, Feb 5, 2012 at 2:59 PM, Justin P. Mattock
[off-list ref] wrote:
On Sun, 5 Feb 2012 23:21:41 +0100
Francois Romieu [off-list ref] wrote:
quoted
Larry Finger [off-list ref] :
quoted
On 02/05/2012 03:40 PM, Justin P. Mattock wrote:
quoted
this showed up as I woke the machine up from suspend. machine is connected
I just see networkmanager constantly spinning trying to connect then telling me
the wired network is not connected(even though its just wireless that I have).
[ 1273.447349] r8169 0000:06:00.0: Refused to change power state, currently in D3
[ 1273.571024] psmouse serio1: elantech: retrying ps2 command 0xf8 (2).
[ 1273.656709] r8169 0000:06:00.0: eth0: link up
                                         ^^^^^^^

Justin, you did not ever connect a cable, did you ?
nope! was using wireless the whole time.
quoted
[...]
quoted
The error clearly comes from r8169, not from ath9k. If you unload
r8169 before suspending, does this still happen? My suspicion is
that NM is getting confused.
It could hide the problem but the D3 + link up messages are nonetheless
a bit unexpected.
quoted
P.S. Unless it happens with just the ath9k active, this ML is likely
not the correct one.
Yes, please Cc: netdev@vger.kernel.org + me instead.

Thanks.

--
Ueimor
removed linux-wireless and added netdev to the cc's
will monitor this and report anything that shows up.

--
Justin P. Mattock [off-list ref]
seems I see this with the latest linux-next:

[21731.149774] wlan0: associate with 00:1e:2a:62:6b:1e (try 1/3)
[21731.152858] wlan0: RX AssocResp from 00:1e:2a:62:6b:1e (capab=0x31
status=0 aid=7)
[21731.152867] wlan0: associated
[21740.306348] ------------[ cut here ]------------
[21740.306363] WARNING: at net/sched/sch_generic.c:256
dev_watchdog+0x1f2/0x200()
[21740.306367] Hardware name: 0914
[21740.306370] NETDEV WATCHDOG: eth0 (r8169): transmit queue 0 timed out
[21740.306374] Modules linked in: dm_crypt xt_limit xt_tcpudp
xt_addrtype snd_hda_codec_hdmi xt_state joydev arc4
snd_hda_codec_conexant ath9k nf_conntrack_netbios_ns
nf_conntrack_broadcast snd_hda_intel snd_hda_codec mac80211 nf_nat_ftp
nf_nat snd_hwdep nf_conntrack_ipv4 nf_defrag_ipv4 snd_pcm
nf_conntrack_ftp snd_seq_midi ath9k_common ath9k_hw nf_conntrack
iptable_filter snd_rawmidi snd_seq_midi_event uvcvideo ath ip_tables
videobuf2_core videodev psmouse x_tables ideapad_laptop serio_raw
snd_seq cfg80211 lp snd_timer snd_seq_device snd videobuf2_vmalloc
videobuf2_memops soundcore parport sparse_keymap intel_ips
snd_page_alloc usb_storage uas i915 drm_kms_helper drm i2c_algo_bit
r8169 intel_agp intel_gtt agpgart video
[21740.306469] Pid: 0, comm: swapper/2 Not tainted 3.3.0-rc7-next-20120316 #1
[21740.306473] Call Trace:
[21740.306483]  [<c1049072>] warn_slowpath_common+0x72/0xa0
[21740.306489]  [<c14e00a2>] ? dev_watchdog+0x1f2/0x200
[21740.306493]  [<c14e00a2>] ? dev_watchdog+0x1f2/0x200
[21740.306498]  [<c1049143>] warn_slowpath_fmt+0x33/0x40
[21740.306503]  [<c14e00a2>] dev_watchdog+0x1f2/0x200
[21740.306520]  [<c1059c12>] run_timer_softirq+0x132/0x540
[21740.306525]  [<c1059b98>] ? run_timer_softirq+0xb8/0x540
[21740.306529]  [<c14dfeb0>] ? pfifo_fast_init+0x60/0x60
[21740.306534]  [<c1050e67>] __do_softirq+0x97/0x360
[21740.306539]  [<c1050dd0>] ? local_bh_enable+0xd0/0xd0
[21740.306542]  <IRQ>  [<c10513d6>] ? irq_exit+0x86/0xb0
[21740.306552]  [<c159f179>] ? smp_apic_timer_interrupt+0x59/0x88
[21740.306559]  [<c12cf1a8>] ? trace_hardirqs_off_thunk+0xc/0x14
[21740.306565]  [<c1597682>] ? apic_timer_interrupt+0x36/0x3c
[21740.306572]  [<c10a007b>] ? clocks_calc_mult_shift+0xfb/0x100
[21740.306580]  [<c1319fd9>] ? intel_idle+0xc9/0x130
[21740.306588]  [<c1495655>] ? cpuidle_enter+0x15/0x20
[21740.306592]  [<c1495b5a>] ? cpuidle_idle_call+0x8a/0x320
[21740.306600]  [<c10152c5>] ? cpu_idle+0xa5/0xf0
[21740.306607]  [<c1585d2d>] ? start_secondary+0x209/0x20f
[21740.306611] ---[ end trace 9a4d94a7e2dfb4d0 ]---
[21740.318685] r8169 0000:06:00.0: eth0: link up
[21752.292679] r8169 0000:06:00.0: eth0: link up
[21764.268569] r8169 0000:06:00.0: eth0: link up
[21776.254393] r8169 0000:06:00.0: eth0: link up
[21788.235797] r8169 0000:06:00.0: eth0: link up
[21800.196524] r8169 0000:06:00.0: eth0: link up
[21812.172497] r8169 0000:06:00.0: eth0: link up

one thing I have noticed these few weeks is once I wake the machine
up, wireless connects, but Networkmanager keeps trying to connect
ethernet(switching ethernet off in preferences or waiting a while
resolves this issue from the daemon), so I think there must be some
strange thing with NM
causing this(disabling "auto connect" gets rid of this issue)
the system is ubuntu(oneiric)


-- 
Justin P. Mattock

Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

From: Francois Romieu <romieu@fr.zoreil.com>
Date: 2012-03-20 09:39:12 

(Larry removed)

Justin Mattock [off-list ref] :
[...]
seems I see this with the latest linux-next:
Thanks for testing.

[...]
[21740.318685] r8169 0000:06:00.0: eth0: link up
[21752.292679] r8169 0000:06:00.0: eth0: link up
[21764.268569] r8169 0000:06:00.0: eth0: link up
[21776.254393] r8169 0000:06:00.0: eth0: link up
[21788.235797] r8169 0000:06:00.0: eth0: link up
[21800.196524] r8169 0000:06:00.0: eth0: link up
[21812.172497] r8169 0000:06:00.0: eth0: link up
This is completely broken. I could understand a few up/down link changes
until things settles but the driver should not claim periodically that
the link is up when there is no cable, at least not with a supported chipset.

Can you apply the debug helper below and report a complete dmesg from
boot with the same test (please remove l-k, netdev is good enough) ?

diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 61e6ab4..880264a 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -61,7 +61,8 @@
 #endif /* RTL8169_DEBUG */
 
 #define R8169_MSG_DEFAULT \
-	(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN)
+	(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN | \
+	 NETIF_MSG_LINK)
 
 #define TX_BUFFS_AVAIL(tp) \
 	(tp->dirty_tx + NUM_TX_DESC - tp->cur_tx - 1)
@@ -731,7 +732,7 @@ struct rtl8169_private {
 	void (*phy_reset_enable)(struct rtl8169_private *tp);
 	void (*hw_start)(struct net_device *);
 	unsigned int (*phy_reset_pending)(struct rtl8169_private *tp);
-	unsigned int (*link_ok)(void __iomem *);
+	unsigned int (*link_ok)(struct rtl8169_private *);
 	int (*do_ioctl)(struct rtl8169_private *tp, struct mii_ioctl_data *data, int cmd);
 
 	struct {
@@ -1260,14 +1261,28 @@ static unsigned int rtl8169_xmii_reset_pending(struct rtl8169_private *tp)
 	return rtl_readphy(tp, MII_BMCR) & BMCR_RESET;
 }
 
-static unsigned int rtl8169_tbi_link_ok(void __iomem *ioaddr)
+static unsigned int rtl8169_tbi_link_ok(struct rtl8169_private *tp)
 {
+	void __iomem *ioaddr = tp->mmio_addr;
+
 	return RTL_R32(TBICSR) & TBILinkOk;
 }
 
-static unsigned int rtl8169_xmii_link_ok(void __iomem *ioaddr)
+static unsigned int rtl8169_xmii_link_ok(struct rtl8169_private *tp)
 {
-	return RTL_R8(PHYstatus) & LinkStatus;
+	void __iomem *ioaddr = tp->mmio_addr;
+	struct net_device *dev = tp->dev;
+	u8 status;
+
+	status = RTL_R8(PHYstatus) & LinkStatus;
+	netif_info(tp, link, dev,
+		   "bmcr: %04x bmsr: %04x gbcr: %04x gbsr: %04x\n",
+		   rtl_readphy(tp, MII_BMCR),
+		   rtl_readphy(tp, MII_BMSR),
+		   rtl_readphy(tp, MII_CTRL1000),
+		   rtl_readphy(tp, MII_STAT1000));
+
+	return status;
 }
 
 static void rtl8169_tbi_reset_enable(struct rtl8169_private *tp)
@@ -1335,7 +1350,7 @@ static void __rtl8169_check_link_status(struct net_device *dev,
 					struct rtl8169_private *tp,
 					void __iomem *ioaddr, bool pm)
 {
-	if (tp->link_ok(ioaddr)) {
+	if (tp->link_ok(tp)) {
 		rtl_link_chg_patch(tp);
 		/* This is to cancel a scheduled suspend if there's one. */
 		if (pm)
@@ -3309,7 +3324,6 @@ static void rtl_hw_phy_config(struct net_device *dev)
 static void rtl_phy_work(struct rtl8169_private *tp)
 {
 	struct timer_list *timer = &tp->timer;
-	void __iomem *ioaddr = tp->mmio_addr;
 	unsigned long timeout = RTL8169_PHY_TIMEOUT;
 
 	assert(tp->mac_version > RTL_GIGA_MAC_VER_01);
@@ -3323,7 +3337,7 @@ static void rtl_phy_work(struct rtl8169_private *tp)
 		goto out_mod_timer;
 	}
 
-	if (tp->link_ok(ioaddr))
+	if (tp->link_ok(tp))
 		return;
 
 	netif_warn(tp, link, tp->dev, "PHY reset until link up\n");
-- 
1.7.7.6

Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

From: Justin P. Mattock <hidden>
Date: 2012-03-20 14:16:21 

On 03/20/2012 02:37 AM, Francois Romieu wrote:
(Larry removed)

Justin Mattock[off-list ref]  :
[...]
quoted
seems I see this with the latest linux-next:
Thanks for testing.

[...]
quoted
[21740.318685] r8169 0000:06:00.0: eth0: link up
[21752.292679] r8169 0000:06:00.0: eth0: link up
[21764.268569] r8169 0000:06:00.0: eth0: link up
[21776.254393] r8169 0000:06:00.0: eth0: link up
[21788.235797] r8169 0000:06:00.0: eth0: link up
[21800.196524] r8169 0000:06:00.0: eth0: link up
[21812.172497] r8169 0000:06:00.0: eth0: link up
This is completely broken. I could understand a few up/down link changes
until things settles but the driver should not claim periodically that
the link is up when there is no cable, at least not with a supported chipset.

Can you apply the debug helper below and report a complete dmesg from
boot with the same test (please remove l-k, netdev is good enough) ?
I removed linux-kernel from the Cc's and applied your patch.. system 
built fine, and suspend wakes up fine without the ethernet trying to 
connect like it was doing. will run this patch for a few Thanks for this!

dmesg is here: http://fpaste.org/jHSX/
(I have another patch added in, but it should not affect anything).
quoted hunk
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 61e6ab4..880264a 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -61,7 +61,8 @@
  #endif /* RTL8169_DEBUG */

  #define R8169_MSG_DEFAULT \
-	(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN)
+	(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN | \
+	 NETIF_MSG_LINK)

  #define TX_BUFFS_AVAIL(tp) \
  	(tp->dirty_tx + NUM_TX_DESC - tp->cur_tx - 1)
@@ -731,7 +732,7 @@ struct rtl8169_private {
  	void (*phy_reset_enable)(struct rtl8169_private *tp);
  	void (*hw_start)(struct net_device *);
  	unsigned int (*phy_reset_pending)(struct rtl8169_private *tp);
-	unsigned int (*link_ok)(void __iomem *);
+	unsigned int (*link_ok)(struct rtl8169_private *);
  	int (*do_ioctl)(struct rtl8169_private *tp, struct mii_ioctl_data *data, int cmd);

  	struct {
@@ -1260,14 +1261,28 @@ static unsigned int rtl8169_xmii_reset_pending(struct rtl8169_private *tp)
  	return rtl_readphy(tp, MII_BMCR)&  BMCR_RESET;
  }

-static unsigned int rtl8169_tbi_link_ok(void __iomem *ioaddr)
+static unsigned int rtl8169_tbi_link_ok(struct rtl8169_private *tp)
  {
+	void __iomem *ioaddr = tp->mmio_addr;
+
  	return RTL_R32(TBICSR)&  TBILinkOk;
  }

-static unsigned int rtl8169_xmii_link_ok(void __iomem *ioaddr)
+static unsigned int rtl8169_xmii_link_ok(struct rtl8169_private *tp)
  {
-	return RTL_R8(PHYstatus)&  LinkStatus;
+	void __iomem *ioaddr = tp->mmio_addr;
+	struct net_device *dev = tp->dev;
+	u8 status;
+
+	status = RTL_R8(PHYstatus)&  LinkStatus;
+	netif_info(tp, link, dev,
+		   "bmcr: %04x bmsr: %04x gbcr: %04x gbsr: %04x\n",
+		   rtl_readphy(tp, MII_BMCR),
+		   rtl_readphy(tp, MII_BMSR),
+		   rtl_readphy(tp, MII_CTRL1000),
+		   rtl_readphy(tp, MII_STAT1000));
+
+	return status;
  }

  static void rtl8169_tbi_reset_enable(struct rtl8169_private *tp)
@@ -1335,7 +1350,7 @@ static void __rtl8169_check_link_status(struct net_device *dev,
  					struct rtl8169_private *tp,
  					void __iomem *ioaddr, bool pm)
  {
-	if (tp->link_ok(ioaddr)) {
+	if (tp->link_ok(tp)) {
  		rtl_link_chg_patch(tp);
  		/* This is to cancel a scheduled suspend if there's one. */
  		if (pm)
@@ -3309,7 +3324,6 @@ static void rtl_hw_phy_config(struct net_device *dev)
  static void rtl_phy_work(struct rtl8169_private *tp)
  {
  	struct timer_list *timer =&tp->timer;
-	void __iomem *ioaddr = tp->mmio_addr;
  	unsigned long timeout = RTL8169_PHY_TIMEOUT;

  	assert(tp->mac_version>  RTL_GIGA_MAC_VER_01);
@@ -3323,7 +3337,7 @@ static void rtl_phy_work(struct rtl8169_private *tp)
  		goto out_mod_timer;
  	}

-	if (tp->link_ok(ioaddr))
+	if (tp->link_ok(tp))
  		return;

  	netif_warn(tp, link, tp->dev, "PHY reset until link up\n");





  

    
Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

    

      From: Justin P. Mattock <hidden>

      Date: 2012-03-20 21:31:26 
    

  

  
On 03/20/2012 07:16 AM, Justin P. Mattock wrote:
On 03/20/2012 02:37 AM, Francois Romieu wrote:
quoted
(Larry removed)

Justin Mattock[off-list ref] :
[...]
quoted
seems I see this with the latest linux-next:
Thanks for testing.

[...]
quoted
[21740.318685] r8169 0000:06:00.0: eth0: link up
[21752.292679] r8169 0000:06:00.0: eth0: link up
[21764.268569] r8169 0000:06:00.0: eth0: link up
[21776.254393] r8169 0000:06:00.0: eth0: link up
[21788.235797] r8169 0000:06:00.0: eth0: link up
[21800.196524] r8169 0000:06:00.0: eth0: link up
[21812.172497] r8169 0000:06:00.0: eth0: link up
This is completely broken. I could understand a few up/down link changes
until things settles but the driver should not claim periodically that
the link is up when there is no cable, at least not with a supported
chipset.

Can you apply the debug helper below and report a complete dmesg from
boot with the same test (please remove l-k, netdev is good enough) ?
I removed linux-kernel from the Cc's and applied your patch.. system
built fine, and suspend wakes up fine without the ethernet trying to
connect like it was doing. will run this patch for a few Thanks for this!

dmesg is here: http://fpaste.org/jHSX/
(I have another patch added in, but it should not affect anything).

well i dont know what happened.. I suspend throughout the day without 
any issues, but then on one wakeup everything went to crap ethernet 
started to try and connect without any wires in.. only thing I can think 
of is I was dongling with my phone(but the phone was disconnected before 
suspend),which uses a different module..

dmesg here:
http://fpaste.org/mcNC/

Justin P. Mattock





  

    
Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

    

      From: Francois Romieu <romieu@fr.zoreil.com>

      Date: 2012-03-20 23:29:02 
    

  

  
Justin P. Mattock [off-list ref] :
[...]
well i dont know what happened.. I suspend throughout the day
without any issues, but then on one wakeup everything went to crap
ethernet started to try and connect without any wires in.. only
thing I can think of is I was dongling with my phone(but the phone
was disconnected before suspend),which uses a different module..
The MII registers read as 0xff. I'll have to figure why the
suspend / resume cycle failed, or why it seemed to work. :o/

-- 
Ueimor





  

    
Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1f2/0x200()

    

      From: Justin P. Mattock <hidden>

      Date: 2012-03-21 05:16:02 
    

  

  
On 03/20/2012 04:27 PM, Francois Romieu wrote:
Justin P. Mattock[off-list ref]  :
[...]
quoted
well i dont know what happened.. I suspend throughout the day
without any issues, but then on one wakeup everything went to crap
ethernet started to try and connect without any wires in.. only
thing I can think of is I was dongling with my phone(but the phone
was disconnected before suspend),which uses a different module..
The MII registers read as 0xff. I'll have to figure why the
suspend / resume cycle failed, or why it seemed to work. :o/
not sure why this did this, I rebooted and have suspended numerous times 
without an issue with ethernet, even retraced my steps to try and 
re-created but nothing. seems solid to me beside that hickup.

I am trying to gather any info for you, if you have any ideas on what I 
can try let me know!

Thanks for the patch and info with this.

Justin P. Mattock




    
  

  
    

      

        Keyboard shortcuts
      

      

        
          
hback out one level

          
jnext message in thread

          
kprevious message in thread

          
ldrill in

          
Escclose help / fold thread tree

          
?toggle this help