| View previous topic :: View next topic |
| Author |
Message |
platojones
Veteran
Joined: 23 Oct 2002
Posts: 1602
Location: Just over the horizon
|
 Posted: Mon Mar 02, 2009 5:52 pm Post subject: |
 |
|
| Likvid wrote: | The sky2 driver is really bad.
I have the Asus P5B-Deluxe loaded with latest v.1238 firmware now and running kernel 2.6.28
I can't get it to work with my Marvell 88E8056 card at all.
Sometimes it works and now i get Destination Hos Unreachable all the time.
Any solution to this as there are still major problems with the code in this driver? |
That's weird...I'm using the same kernel, same board, and have never had a problem with this driver. But I'm not using any firmware either. |
|
| Back to top |
|
 |
c0nv1ct
n00b
Joined: 29 May 2007
Posts: 8
|
| Posted: Tue Mar 10, 2009 10:16 am Post subject: |
|
|
I'm having the same issue as the OP with the onbard marvell NIC on my Asus P6T Deluxe. This thread is quite old, please tell me this has been resolved by now or has the problem come back? Kernels 2.6.28 and 2.6.29-rc7 both result in network connectivity loss whenever I sync to a local rsync server, copy a large file from CIFS or NFS mounts, or anything more stressful on the NIC than streaming a simple MP3 song.
Ethernet controller: Marvell Technology Group Ltd. 88E8056 PCI-E Gigabit Ethernet Controller (rev 12) |
|
| Back to top |
|
|
svrmarty
n00b
Joined: 24 Sep 2007
Posts: 9
|
| Posted: Thu Jun 25, 2009 8:11 pm Post subject: |
|
|
| i've the same issue with 2.6.30 and sky2 |
|
| Back to top |
|
|
forsaken1
n00b
Joined: 30 May 2008
Posts: 27
Location: Lovely Skåne, Sweden
|
| Posted: Sun Jun 28, 2009 6:49 pm Post subject: |
|
|
| I'm also getting this problem with 2.6.29.4 and now 2.6.30 on a gigabyte 965p-ds3. |
|
| Back to top |
|
|
Mark Gray
n00b
Joined: 14 Feb 2007
Posts: 14
Location: Berkshire, UK
|
| Posted: Sat Jul 04, 2009 4:48 pm Post subject: |
|
|
And another - 965P-DS4 - Marvell 88E8053
I don't mind the receiver hang/down/up sequence. My linux box and network are configured to recover from it automatically so the worst that happens is a restarted download, usually while I'm away from the system.
However, when I upgraded to 2.6.27, it caused a kernel panic as well. Same with 2.6.28, and I'm now trying 2.6.29. A network drop-out is one thing, but I take exception to a kernel panic (and data loss).
Looking around the mailing lists I found another report where this had happened when the interface had been brought down manually. I'd assume the two are related but I can't find evidence that a fix has made it into mainline. There's a patch that masks the bug though so I'll either try that or go back to 2.6.24 if 2.6.29 does the same.
PS - I've got a capture of the panic if anyone wants it. |
|
| Back to top |
|
|
forsaken1
n00b
Joined: 30 May 2008
Posts: 27
Location: Lovely Skåne, Sweden
|
| Posted: Sat Jul 04, 2009 5:45 pm Post subject: |
|
|
| I took the sky2 driver from 2.6.31-rc1 into 2.6.29.5 and so far so good. |
|
| Back to top |
|
|
Mark Gray
n00b
Joined: 14 Feb 2007
Posts: 14
Location: Berkshire, UK
|
| Posted: Sat Jul 04, 2009 7:01 pm Post subject: |
|
|
Can you give me a quick guide to how to do that please. I'm happy hacking software to pieces but draw the line at doing that in the kernel without a safety net in the form of a few instructions. 
Thankyou.
Edit - I just swapped in the sky2.c from 2.6.31-rc2, but that paniced instantly. Dropped back to 2.6.24 for now. Reading the changelog suggests it does include the bug fix I think I might need (along with quite a few more), so here's hoping for 2.6.31 (or an easy backport). |
|
| Back to top |
|
|
forsaken1
n00b
Joined: 30 May 2008
Posts: 27
Location: Lovely Skåne, Sweden
|
| Posted: Sat Jul 04, 2009 10:19 pm Post subject: |
|
|
| Mark Gray wrote: | Can you give me a quick guide to how to do that please. I'm happy hacking software to pieces but draw the line at doing that in the kernel without a safety net in the form of a few instructions.
Thankyou.
Edit - I just swapped in the sky2.c from 2.6.31-rc2, but that paniced instantly. Dropped back to 2.6.24 for now. Reading the changelog suggests it does include the bug fix I think I might need (along with quite a few more), so here's hoping for 2.6.31 (or an easy backport). |
I took sky2.c and sky2.h from 2.6.31-rc1 and put them in 2.6.29.5 worked without any problems for me. |
|
| Back to top |
|
|
forsaken1
n00b
Joined: 30 May 2008
Posts: 27
Location: Lovely Skåne, Sweden
|
| Posted: Sun Jul 05, 2009 11:52 am Post subject: |
|
|
| I got a timeout from the netdev watchdog after about a day using the .31rc1 driver so it seems it didn't solve the problems. |
|
| Back to top |
|
|
gringo
Advocate
Joined: 27 Apr 2003
Posts: 3793
|
| Posted: Tue Jul 21, 2009 9:36 am Post subject: |
|
|
im also a "happy" owner of a crappy marvell 88E8056 based ethernet card and apparently lowering MTU rate to 1492 does help a lot here.
Just in case this helps.
cheers
_________________
Error: Failing not supported by current locale |
|
| Back to top |
|
|
svrmarty
n00b
Joined: 24 Sep 2007
Posts: 9
|
| Posted: Sun Aug 09, 2009 6:55 pm Post subject: |
|
|
| gringo wrote: | im also a "happy" owner of a crappy marvell 88E8056 based ethernet card and apparently lowering MTU rate to 1492 does help a lot here.
Just in case this helps.
cheers |
hey great
seems to work here too
the only error sometimes happens now is
Aug 9 20:54:28 dvb kernel: sky2 eth1: rx length error: status 0x5ea0100 length 1510 |
|
| Back to top |
|
|
svrmarty
n00b
Joined: 24 Sep 2007
Posts: 9
|
| Posted: Wed Aug 12, 2009 7:32 pm Post subject: |
|
|
seems the major problem is now away
but i now get errors like
Aug 12 21:23:42 dvb kernel: sky2 eth1: hung mac 123:6 fifo 195 (64:59)
Aug 12 21:23:42 dvb kernel: sky2 eth1: receiver hang detected
Aug 12 21:23:42 dvb kernel: sky2 eth1: disabling interface
Aug 12 21:23:42 dvb kernel: sky2 eth1: enabling interface
Aug 12 21:23:44 dvb kernel: sky2 eth1: Link is up at 100 Mbps, full duplex, flow control both |
|
| Back to top |
|
|
gringo
Advocate
Joined: 27 Apr 2003
Posts: 3793
|
| Posted: Thu Aug 13, 2009 8:23 am Post subject: |
|
|
| Quote: | | seems to work here too |
well, after playing around a bit, it turns out it wasn´t my hardware´s fault, it was dhcpcd´s ( im using openrc from git ); in a default setup it requests the mtu rate from the server and apparently gets an wrong value ( at least here). Overriding this solved it for me and i have no need to set the mtu rate manually, default rate works without problems now.
cheers
_________________
Error: Failing not supported by current locale |
|
| Back to top |
|
|
cach0rr0
Bodhisattva
Joined: 13 Nov 2008
Posts: 4123
Location: Houston, Republic of Texas
|
| Posted: Thu Aug 13, 2009 4:26 pm Post subject: |
|
|
FWIW, no issues here whatsoever with:
| Code: |
Ethernet controller: Marvell Technology Group Ltd. 88E8056 PCI-E Gigabit Ethernet Controller (rev 12)
|
and
| Code: |
*-core
description: Motherboard
product: M3A78-T
vendor: ASUSTeK Computer INC.
physical id: 0
version: Rev 1.xx
serial: MS1C91BJ5100589
slot: To Be Filled By O.E.M.
|
and
| Code: |
Linux gentoob0x 2.6.28-hardened-r7 #8 SMP Tue Jul 28 09:03:33 CDT 2009 x86_64 AMD Phenom(tm) 9950 Quad-Core Processor AuthenticAMD GNU/Linux
|
driver is working flawlessly, even under heavy load
_________________
Lost configuring your system?
dump lspci -n here | see Pappy's guide | Link Stash |
|
| Back to top |
|
|
HelAu
n00b
Joined: 09 Feb 2004
Posts: 59
|
| Posted: Tue Dec 08, 2009 1:39 pm Post subject: |
|
|
Hi
I have a XFX Geforce 8300 MoBo with marvel 88E8056 onboard. When I use the sky2 module (kernel 2.6.31 and 2.6.32) everything works but horribly slow ( about 10 MB per MINUTE ).
Then I tried the sk98lin from the marvell homepage ( Version Linux_10.81.6.3 ), but this one results in a kernel oops.
Any hints what I can try (without using another network card) ? |
|
| Back to top |
|
|
HeXiLeD
Veteran
Joined: 20 Aug 2005
Posts: 1159
Location: Online
|
| Posted: Tue Dec 08, 2009 3:27 pm Post subject: |
|
|
Flawlessly working here with an ASUS P5Q Premium
| Code: | | Ethernet controller: Marvell Technology Group Ltd. 88E8056 PCI-E Gigabit Ethernet Controller (rev 12) |
| Code: | | Curr: 14473.90 kBytes (using sky2 loaded as module and the builtin NICs - which i use all 4 as a router FLAWLESSLY) |
You might want to check this topic
_________________
Do you hear the sound of inevitability?
With age, comes great grumpiness and that, was 20 years ago...
CertFP: becbbd161d5a5c31de3c45171b77bf710911db29 / d985d21f89fe2977b593c4d381a1a86802e62990d9328d893db76d59f9935244 |
|
| Back to top |
|
|
HelAu
n00b
Joined: 09 Feb 2004
Posts: 59
|
| Posted: Tue Dec 08, 2009 4:02 pm Post subject: |
|
|
Hi,
I've already read that topic, but didn't find a hint how to solve my problem. |
|
| Back to top |
|
|
HelAu
n00b
Joined: 09 Feb 2004
Posts: 59
|
| Posted: Tue Dec 08, 2009 10:39 pm Post subject: |
|
|
Hi,
My problem is solved. The reason for the slow network was not the sky2 driver, a broken switch caused this ! Sorry for bothering you ... |
|
| Back to top |
|
|
-neX-
n00b
Joined: 29 Apr 2005
Posts: 45
|
| Posted: Wed Apr 07, 2010 3:14 pm Post subject: |
|
|
hello fellow sky2 users...
using the sky2 driver on a Asus P5GD1 with gentoo-sources 2.6.31-r6 and 2.6.31-r10. the driver had been a problem in the past, and went away and i forgot about it. a couple of days ago, the module took its typical dive and the kernel was not able to recover. I had one of the other admins of this box, who is far better at understanding debug output than i am, take a look at the output dumped to the kern.log. transcript of IRC below.
the errors mentioned were generated under 2.6.31-gentoo-r6.
| freenode.net wrote: | [17:45:46] <xxx> yeah im quite certain that it was the nic
[17:45:52] <xxx> Apr 5 23:55:22 omega kernel: sky2 eth0: hung mac 123:120 fifo 194 (67:61)
[17:45:53] <xxx> Apr 5 23:55:22 omega kernel: sky2 eth0: receiver hang detected
[17:45:53] <xxx> Apr 5 23:55:22 omega kernel: sky2 eth0: disabling interface
[17:46:05] <xxx> Apr 5 23:55:22 omega kernel: [<c1375183>] ? sky2_up+0x80d/0x998
[17:46:14] <xxx> Apr 5 23:55:22 omega kernel: [<c1373ab9>] ? sky2_reset+0x83/0x548
[17:46:14] <xxx> Apr 5 23:55:22 omega kernel: [<c1376128>] ? sky2_restart+0xc4/0xe5
[17:46:15] <xxx> Apr 5 23:55:22 omega kernel: [<c1376064>] ? sky2_restart+0x0/0xe5
[17:46:24] <xxx> looks like the interface reset failed
[17:46:51] <xxx> yeah see how it called the restart method twice?
[17:46:59] <xxx> and the restart method was called to two memory locations
[17:47:15] <xxx> i think that's the problem
|
typically, i will see the quoted text below show up in my logs. this also causes a SNMPTrap to be sent out by the switch this interface is directly connected to. this indicates that the port is indeed going down. now that ive given it some thought, and poked through some old logs, this is related to the card/module and not the switching equipment, at least in my case. its been connected to three separate switches at two different physical locations.
these errors generated under my current kernel, 2.6.31-gentoo-r10.
| Code: | Apr 6 19:45:24 omega kernel: sky2 eth0: hung mac 123:73 fifo 194 (44:38)
Apr 6 19:45:24 omega kernel: sky2 eth0: receiver hang detected
Apr 6 19:45:24 omega kernel: sky2 eth0: disabling interface
Apr 6 19:45:24 omega kernel: sky2 eth0: enabling interface
Apr 6 19:45:25 omega kernel: sky2 eth0: Link is up at 100 Mbps, full duplex, flow control rx
|
some interface hardware info from lshw.
| Code: | # lshw -class network
*-network
description: Ethernet interface
product: 88E8053 PCI-E Gigabit Ethernet Controller
vendor: Marvell Technology Group Ltd.
physical id: 0
bus info: pci@0000:02:00.0
logical name: eth0
version: 15
serial: 00:11:d8:27:75:46
size: 100MB/s
capacity: 1GB/s
width: 64 bits
clock: 33MHz
capabilities: pm vpd msi pciexpress bus_master cap_list rom ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
configuration: autonegotiation=on broadcast=yes driver=sky2 driverversion=1.23 duplex=full firmware=N/A ip=1.2.3.4 latency=0 link=yes multicast=yes port=twisted pair speed=100MB/s
resources: irq:17 memory:d7ffc000-d7ffffff ioport:d800(size=256) memory:d7fc0000-d7fdffff(prefetchable) |
at this point, im basically just planning to put in a different NIC. the marvell one has been forever unreliable.
edit: removed the ip from lshw out. oops! |
|
| Back to top |
|
|
paluszak
Apprentice
Joined: 28 Jun 2004
Posts: 266
Location: Warsaw, Poland
|
| Posted: Mon Apr 12, 2010 10:07 am Post subject: |
|
|
Hi,
I've been using sky2 driver from stable 2.6.30-2.6.32 kernels on my Toshiba Tecra:
| Code: | | 02:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8053 PCI-E Gigabit Ethernet Controller (rev 15) |
since 7-8 month ago and the driver's working without a glitch.
J. |
|
| Back to top |
|
|
Cyker
Veteran
Joined: 15 Jun 2006
Posts: 1746
|
| Posted: Mon Apr 12, 2010 7:28 pm Post subject: |
|
|
I still get the hung mac thing but nowhere near as much as I used to and usually the kernel can re-init itself. It cuts off all connections which is a pain when I'm transferring loads of stuff but I think it's about as good as it's likely to get  |
|
| Back to top |
|
|
Cyker
Veteran
Joined: 15 Jun 2006
Posts: 1746
|
| Posted: Wed Apr 20, 2011 11:45 am Post subject: |
|
|
Kernel 2.6.38-r1, still got this problem! 
It's a lot less problematic as the kernel detects when it happens and can restart the driver, but it's annoying as it cuts off things in mid-transfer.
| Quote: | Apr 20 12:27:36 CyServ sky2 0000:02:00.0: eth0: hung mac 0:69 fifo 96 (109:5)
Apr 20 12:27:36 sky2 0000:02:00.0: eth0: receiver hang detected
Apr 20 12:27:36 br0: port 1(eth0) entering forwarding state
Apr 20 12:27:39 sky2 0000:02:00.0: eth0: Link is up at 1000 Mbps, full duplex, flow control both
Apr 20 12:27:39 br0: port 1(eth0) entering listening state
Apr 20 12:27:39 br0: port 1(eth0) entering listening state
Apr 20 12:27:40 br0: port 1(eth0) entering learning state
Apr 20 12:27:41 br0: topology change detected, propagating
Apr 20 12:27:41 br0: port 1(eth0) entering forwarding state
|
As always, can only duplicate it when the interface is suddenly placed under a very heavy/saturating load... |
|
| Back to top |
|
|
warer
n00b
Joined: 26 Oct 2002
Posts: 38
|
| Posted: Wed May 04, 2011 10:59 pm Post subject: |
|
|
| I got this problem back in kernel 2.6.38 crap. It have been fine for a few years and the bug is back. |
|
| Back to top |
|
|
ktsaou
Tux's lil' helper
Joined: 10 Jul 2003
Posts: 80
|
| Posted: Sun Jun 12, 2011 10:55 pm Post subject: |
|
|
I have a few logs too, on 2.6.38-gentoo-r6:
| Code: | [ 1.225672] sky2: driver version 1.28
[ 1.225865] sky2 0000:04:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 1.226050] sky2 0000:04:00.0: setting latency timer to 64
[ 1.226090] sky2 0000:04:00.0: Yukon-2 EC Ultra chip revision 5
[ 1.226339] sky2 0000:04:00.0: irq 45 for MSI/MSI-X
[ 1.226824] sky2 0000:04:00.0: eth0: addr 00:50:8d:b7:e9:c9
[ 25.725969] sky2 0000:04:00.0: eth0: enabling interface
[ 28.158983] sky2 0000:04:00.0: eth0: Link is up at 1000 Mbps, full duplex, flow control rx
[81289.789869] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1212
[81289.789873] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1324
[81289.789884] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1020
[89336.994567] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1084
[89336.994575] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1510
[89336.994586] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 372
[89336.994634] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1084
[89336.994717] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1020
[89336.994741] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 684
[89336.994766] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1084
[89336.994811] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 236
[89336.994839] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 572
[89336.994910] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 428
[89348.668514] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1468
[89348.668557] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1468
[89348.668587] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 316
[89348.669741] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 828
[89348.669748] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 620
[89348.669764] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1510
[89348.669833] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1028
[89348.669861] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 804
[89348.669865] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 684
[89348.669892] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 836
[89357.673845] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1084
[89357.673893] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 108
[89357.675308] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 828
[89368.200131] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 956
[89368.200262] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 700
[89368.200327] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 300
[89368.200351] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1468
[89368.200381] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1020
[90067.363692] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 700
[90067.363696] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1388
[90067.363700] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 748
[90067.363703] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1004
[90067.363712] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 748
[90067.363715] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1084
[90067.363786] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 492
[90067.363792] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 172
[90067.363797] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 1452
[90067.363800] sky2 0000:04:00.0: eth0: rx error, status 0x7ffc0001 length 108
|
The rx errors appear under load.
Any way to fix it? |
|
| Back to top |
|
|
warer
n00b
Joined: 26 Oct 2002
Posts: 38
|
| Posted: Tue Jun 14, 2011 7:36 am Post subject: |
|
|
| I found a solution that works i bought a Intel Pro/1000CT network adapter. It is automatically deteceted and no driver flaws. I am fed up with SKY2. |
|
| Back to top |
|
|
|
Kernel & Hardware
