Home » Mailing list archives » linux-netdev » 2008 » August » 5

Realtek 8111C transmit timed out

Previous thread: [PATCH 1/4] e1000e: convert ndev_ printks to something smaller by Jeff Kirsher on Monday, August 4, 2008 - 8:20 pm. (5 messages)

Next thread: [patch] tg3: Fix 'scheduling while atomic' errors by Matt Carlson on Tuesday, August 5, 2008 - 1:13 am. (2 messages)
[view original message]
From: John P Poet <jppoet@...>
To: <netdev@...>
Subject: Realtek 8111C transmit timed out
Date: Monday, August 4, 2008 - 9:12 pm

I recently bought a new motherboard with Realtek 8111C nics.  The nic

is detected as:


dmesg:

r8169 Gigabit Ethernet driver 2.2LK-NAPI loaded

ACPI: PCI Interrupt 0000:06:00.0[A] -> GSI 18 (level, low) -> IRQ 18

PCI: Setting latency timer of device 0000:06:00.0 to 64

eth0: RTL8168c/8111c at 0xffffc2000063e000, 00:1f:d0:20:01:57, XID

3c4000c0 IRQ 2292


lspci -v:

06:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd.

RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 02)

        Subsystem: Giga-byte Technology Unknown device e000

        Flags: bus master, fast devsel, latency 0, IRQ 2292

        I/O ports at b000 [size=256]

        Memory at de010000 (64-bit, prefetchable) [size=4K]

        Memory at de000000 (64-bit, prefetchable) [size=64K]

        [virtual] Expansion ROM at de020000 [disabled] [size=64K]

        Capabilities: [40] Power Management version 3

        Capabilities: [50] Message Signalled Interrupts: Mask- 64bit+

Queue=0/1 Enable+

        Capabilities: [70] Express Endpoint, MSI 01

        Capabilities: [b0] MSI-X: Enable- Mask- TabSize=2

        Capabilities: [d0] Vital Product Data <?>

        Capabilities: [100] Advanced Error Reporting <?>

        Capabilities: [140] Virtual Channel <?>

        Capabilities: [160] Device Serial Number 78-56-34-12-78-56-34-12

        Kernel driver in use: r8169

        Kernel modules: r8169


I have been having two problems with this nic:


1) On a cold boot, the nic is not brought up correctly.  It seems to

stay at 100mbit instead of going into gbit mode, and it fails to ping

the router even though it seems to get a valid IP address.  If I then

warm boot the computer, then then nic comes up in gbit just fine, and

generally works.


2) I am getting a lot of timeouts:


Aug  4 17:49:35 saphire kernel: NETDEV WATCHDOG: eth0: transmit timed out

Aug  4 17:49:35 saphire kernel: r8169: eth0: link up

Aug  4 17:50:05 saphire kernel: NETDEV WATCHDOG: eth0: transmit timed out

Aug  4 17:50:05 saph...
[ message continues ]
[view original message]
From: John P Poet <jppoet@...>
To: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Monday, August 4, 2008 - 10:00 pm

I probably should have mentioned that I am running the 2.6.26 kernel.


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
[view original message]
From: John P Poet <jppoet@...>
To: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Monday, August 4, 2008 - 11:42 pm

I just tried applying ALL the patches found here:


http://userweb.kernel.org/~romieu/r8169/2.6.27-rc1/20080803.tar.bz2


against 2.6.26.1.  Worked for a while, I then I got:


Aug  4 21:23:24 saphire kernel: NETDEV WATCHDOG: eth0: transmit timed out

Aug  4 21:23:24 saphire kernel: ------------[ cut here ]------------

Aug  4 21:23:24 saphire kernel: WARNING: at

net/sched/sch_generic.c:222 dev_watchdog+0xb6/0x116()

Aug  4 21:23:24 saphire kernel: Modules linked in: hdpvr videodev

v4l1_compat v4l2_common nfs nfsd lockd nfs_acl auth_rpcgss exportfs

autofs4 coretemp hwmon fuse sunrpc ipv6 acpi_cpufreq xfs dm_mirror

dm_log dm_multipath dm_mod raid456 async_xor async_memcpy async_tx xor

snd_hda_intel snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq

sr_mod snd_seq_device i2c_i801 serio_raw i2c_core snd_pcm_oss pcspkr

cdrom snd_mixer_oss snd_pcm snd_timer snd_page_alloc r8169 snd_hwdep

snd mii pl2303 usbserial soundcore sg button sata_mv sata_sil24 ahci

libata sd_mod scsi_mod ext3 jbd mbcache uhci_hcd ohci_hcd ehci_hcd

[last unloaded: microcode]

Aug  4 21:23:24 saphire kernel: Pid: 12472, comm: cc1plus Not tainted

2.6.26.1 #1

Aug  4 21:23:24 saphire kernel:

Aug  4 21:23:24 saphire kernel: Call Trace:

Aug  4 21:23:24 saphire kernel:  <IRQ>  [<ffffffff81031fa8>]

warn_on_slowpath+0x58/0x86

Aug  4 21:23:24 saphire kernel:  [<ffffffff8103a78e>] ? __mod_timer+0xc1/0xd3

Aug  4 21:23:24 saphire kernel:  [<ffffffff81041458>] ?

queue_delayed_work_on+0xc3/0xd6

Aug  4 21:23:24 saphire kernel:  [<ffffffff81216891>] ? dev_watchdog+0x0/0x116

Aug  4 21:23:24 saphire kernel:  [<ffffffff810414a7>] ?

queue_delayed_work+0x21/0x23

Aug  4 21:23:24 saphire kernel:  [<ffffffff810414c2>] ?

schedule_delayed_work+0x19/0x1b

Aug  4 21:23:24 saphire kernel:  [<ffffffff81216947>] dev_watchdog+0xb6/0x116

Aug  4 21:23:24 saphire kernel:  [<ffffffff8103a230>]

run_timer_softirq+0x16c/0x1e1

Aug  4 21:23:24 saphire kernel:  [<ffffffff81036a93>] __do_softirq+0...
[ message continues ]
[view original message]
From: Francois Romieu <romieu@...>
To: John P Poet <jppoet@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Tuesday, August 5, 2008 - 5:42 pm

(please Cc: the maintainer)


John P Poet <jppoet@gmail.com> :


Try to add 0002-r8169-avoid-thrashing-PCI-conf-space-above-RTL_GIGA.patch

from http://userweb.kernel.org/~romieu/r8169/2.6.26/20080717/


It is in 2.6.27-rc1 as 77332894c21165404496c56763d7df6c15c4bb09.


--

Ueimor

--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Wednesday, August 6, 2008 - 1:21 am

Thank you for the response.  That DOES seem to fix the cold-boot issue.


Unfortunately, it does not fix the timeout problem.  I am still

experiencing transmittion problems:


===================================================================

Aug  5 22:42:31 saphire kernel: NETDEV WATCHDOG: eth0: transmit timed out

Aug  5 22:42:31 saphire kernel: ------------[ cut here ]------------

Aug  5 22:42:31 saphire kernel: WARNING: at

net/sched/sch_generic.c:222 dev_watchdog+0xb6/0x116()

Aug  5 22:42:31 saphire kernel: Modules linked in: hdpvr videodev

v4l1_compat v4l2_common nfs nfsd lockd nfs_acl auth_rpcgss exportfs

autofs4 coretemp hwmon fuse sunrpc ipv6 acpi_cpufreq xfs dm_mirror

dm_log dm_multipath dm_mod raid456 async_xor async_memcpy async_tx xor

snd_hda_intel snd_seq_dummy snd_seq_oss snd_seq_midi_event i2c_i801

snd_seq serio_raw pcspkr sr_mod i2c_core cdrom snd_seq_device

snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd_page_alloc snd_hwdep

r8169 snd pl2303 usbserial soundcore sg button sata_mv sata_sil24 ahci

libata sd_mod scsi_mod ext3 jbd mbcache uhci_hcd ohci_hcd ehci_hcd

[last unloaded: microcode]

Aug  5 22:42:31 saphire kernel: Pid: 0, comm: swapper Not tainted 2.6.26.1b #1

Aug  5 22:42:31 saphire kernel:

Aug  5 22:42:31 saphire kernel: Call Trace:

Aug  5 22:42:31 saphire kernel:  <IRQ>  [<ffffffff81031fa8>]

warn_on_slowpath+0x58/0x86

Aug  5 22:42:31 saphire kernel:  [<ffffffff8103a78e>] ? __mod_timer+0xc1/0xd3

Aug  5 22:42:31 saphire kernel:  [<ffffffff81041458>] ?

queue_delayed_work_on+0xc3/0xd6

Aug  5 22:42:31 saphire kernel:  [<ffffffff81216891>] ? dev_watchdog+0x0/0x116

Aug  5 22:42:31 saphire kernel:  [<ffffffff810414a7>] ?

queue_delayed_work+0x21/0x23

Aug  5 22:42:31 saphire kernel:  [<ffffffff810414c2>] ?

schedule_delayed_work+0x19/0x1b

Aug  5 22:42:31 saphire kernel:  [<ffffffff81216947>] dev_watchdog+0xb6/0x116

Aug  5 22:42:31 saphire kernel:  [<ffffffff8103a230>]

run_timer_softirq+0x16c/0x1e1

Aug  5 22:42:3...
[ message continues ]
[view original message]
From: Francois Romieu <romieu@...>
To: John P Poet <jppoet@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Wednesday, August 6, 2008 - 3:03 pm

John P Poet <jppoet@gmail.com> :


Does the patch below (on top of the current pile) make a difference ?


diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c

index c421bf6..348846f 100644

--- a/drivers/net/r8169.c

+++ b/drivers/net/r8169.c

@@ -874,12 +874,9 @@ static int rtl8169_set_speed_xmii(struct net_device *dev,


 	auto_nego |= ADVERTISE_PAUSE_CAP | ADVERTISE_PAUSE_ASYM;


-	if ((tp->mac_version == RTL_GIGA_MAC_VER_12) ||

-	    (tp->mac_version == RTL_GIGA_MAC_VER_17)) {

-		/* Vendor specific (0x1f) and reserved (0x0e) MII registers. */

-		mdio_write(ioaddr, 0x1f, 0x0000);

-		mdio_write(ioaddr, 0x0e, 0x0000);

-	}

+	/* Vendor specific (0x1f) and reserved (0x0e) MII registers. */

+	mdio_write(ioaddr, 0x1f, 0x0000);

+	mdio_write(ioaddr, 0x0e, 0x0000);


 	tp->phy_auto_nego_reg = auto_nego;

 	tp->phy_1000_ctrl_reg = giga_ctrl;

--

Ueimor

--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Thursday, August 7, 2008 - 1:51 am

That does not make any difference.


I will say, that the problem seems to be reduced by this patch (or the

0002-r8169-avoid-thrashing-PCI-conf-space-above-RTL_GIGA.patch).


Before, I was getting failures two or three times an hour.  With

either of these patches applied, the failure rate has dropped to one

every two to four hours.


Thanks for trying.


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
[view original message]
From: Francois Romieu <romieu@...>
To: John P Poet <jppoet@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Thursday, August 7, 2008 - 6:31 pm

John P Poet <jppoet@gmail.com> :


I have updated the kit at http://userweb.kernel.org/~romieu/r8169/2.6.27-rc2


Can you give it a try ?


--

Ueimor

--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Thursday, August 7, 2008 - 11:28 pm

At first I *just* applied 20080807-r8169-test.patch, but the NIC did

not initialize during boot.  So, I also applied

0002-r8169-avoid-thrashing-PCI-conf-space-above-RTL_GIGA.patch.  After

that, it booted okay.


However, in general use it is much worse.  May just be bad luck (or

something in the 2.26.2 kernel?), but the timeout message are rampant:


Aug  7 21:14:58 saphire kernel: NETDEV WATCHDOG: eth0: transmit timed out

Aug  7 21:14:58 saphire kernel: r8169: eth0: link up


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Friday, August 22, 2008 - 3:55 pm

Any other ideas on this problem?  I am wondering if most people are

not complaining about it, because it is probably not noticeable for

web browsing (web browsing naturally has latencies involved).  It is a

killer for streaming stuff in "real time" between two machines,

though.


On my current machine with a 8111C NIC, I have disabled that NIC and

bought an Intel e1000 PCIe NIC.  That has solved the problem on that

computer, but I want to build another computer, and it seems like all

the new motherboards have the Realtek 8111C on them.


Should I try the "official" driver from Realtek?  How hard is it to integrate?


Thanks,


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
[view original message]
From: Francois Romieu <romieu@...>
To: John P Poet <jppoet@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Sunday, August 24, 2008 - 6:09 pm

John P Poet <jppoet@gmail.com> :


Yes, it would be interesting to know if it performs better.


Could you send the output of 'mii-tool -vv' ? I am curious to know how


It is supposed to build as an external module.


--

Ueimor

--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Thursday, August 7, 2008 - 11:51 pm

Hmmm.   Not just "much worse", but much, much, much, much, much worse!

 I am getting several timeouts every minute.


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
[view original message]
From: David Madsen <david.madsen@...>
To: John P Poet <jppoet@...>
Cc: Francois Romieu <romieu@...>, <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Tuesday, August 12, 2008 - 1:42 am

I also have a Realtek GigE card that was quite stable running on

2.6.24.  I recently updated my kernel briefly to 2.6.25.10 then

ultimately to 2.6.26.2 and started seeing similar timeouts in both

kernel versions.  My configuration didn't change much between the

kernels, but I do remember enabling MSI when I rebuit the kernel.  I

have not yet had a chance to disable MSI to see if that fixes the

timeouts but I thought I'd post what info I have in case that might

steer the debug in the right direction.  The frequency of the timeouts

has been quite low, and once the interface comes back up everything

seems to continue to function properly.  I haven't applied any of the

discussed patches yet, but I have set up the machine to disable MSI

the next time I am able to reboot it.  Prior to the kernel update from

2.6.24, MSI was diabled and I did not have any issues with timeouts on

the interface.  Let me know if I can provide any more information that

may help with this debug.


Here is the initialization:


Aug  6 17:02:13 [kernel] r8169 Gigabit Ethernet driver 2.2LK-NAPI loaded

Aug  6 17:02:13 [kernel] ACPI: PCI Interrupt 0000:04:00.0[A] -> GSI 16

(level, low) -> IRQ 16

Aug  6 17:02:13 [kernel] PCI: Setting latency timer of device 0000:04:00.0 to 64

Aug  6 17:02:13 [kernel] eth0: RTL8168b/8111b at 0xf881e000,

00:1a:4d:53:cd:0f, XID 38000000 IRQ 218


Several days later the first timeout:


NETDEV WATCHDOG: eth0: transmit timed out

------------[ cut here ]------------

WARNING: at net/sched/sch_generic.c:222 dev_watchdog+0xfd/0x110()

Modules linked in: xfs nfs coretemp it87 hwmon_vid eeprom nfsd lockd

sunrpc exportfs snd_pcm_oss snd_mixer_oss snd_seq_oss

snd_seq_midi_event snd_seq snd_seq_device fuse raid0 raid456 async_xor

async_memcpy async_tx xor md_mod raw1394 or51132 usbhid cx88_dvb

cx88_vp3054_i2c videobuf_dvb dvb_core tuner_simple tuner_types tda9887

tda8290 tuner tvaudio nvidia(P) msp3400 cx8800 cx88_alsa cx8802 cx88xx

bttv firmware_class compat_ioctl32 videodev v4l1_compat ir...
[ message continues ]
[view original message]
From: Francois Romieu <romieu@...>
To: David Madsen <david.madsen@...>
Cc: John P Poet <jppoet@...>, <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Tuesday, August 12, 2008 - 4:18 pm

David Madsen <david.madsen@gmail.com> :


Please note that the kernel complains more loudly about the watchdog

than it used to. Does it allow traffic to flow afterwards ?


--

Ueimor

--
[view original message]
From: David Madsen <david.madsen@...>
To: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Tuesday, August 12, 2008 - 8:39 pm

Likewise for me, the connection does come back up and continues to

function.  Prior to the upgrade and switch to MSI though I never had

any glitches like these that are occurring now.  Hopefully I'll be

able to disable MSI sometime this week to test out that theory.  Since

these timeouts are not easily forced (seems to be unaffected by

traffic load) It may take a while to confirm whether that fixes the

problem or not.


--David Madsen

--
[view original message]
From: John Patrick Poet <john@...>
To: Francois Romieu <romieu@...>
Cc: David Madsen <david.madsen@...>, John P Poet <jppoet@...>, <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Tuesday, August 12, 2008 - 5:44 pm

When one of these "hick-ups" happen while I am streaming video from my Myth

backend to my Myth frontend, it kills the playback, and I have to re-start

it.  It takes 5 to 10 seconds before I can re-start the playback, but it does

work without having to take any special action on my part.


So, to answer your question, yes traffic does flow afterwards.


John


--
[view original message]
From: John P Poet <jppoet@...>
To: Francois Romieu <romieu@...>
Cc: <netdev@...>
Subject: Re: Realtek 8111C transmit timed out
Date: Thursday, August 7, 2008 - 6:51 pm

*Instead* of the previous patches, right?


I will try applying 20080807-r8169-test.patch to the new 2.6.26.2 source.


Thanks,


John

--

A: Because it messes up the order in which people normally read text.

Q: Why is top-posting such a bad thing?

--
Previous thread: [PATCH 1/4] e1000e: convert ndev_ printks to something smaller by Jeff Kirsher on Monday, August 4, 2008 - 8:20 pm. (5 messages)

Next thread: [patch] tg3: Fix 'scheduling while atomic' errors by Matt Carlson on Tuesday, August 5, 2008 - 1:13 am. (2 messages)