Bug 8091

Summary: sky2 losing speed in uplink
Product: Drivers Reporter: Peter Kerwien (peter)
Component: NetworkAssignee: Stephen Hemminger (stephen)
Status: REJECTED DUPLICATE    
Severity: normal CC: stephen
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.21-rc1 Subsystem:
Regression: --- Bisected commit-id:

Description Peter Kerwien 2007-02-26 12:44:33 UTC
Most recent kernel where this bug did *NOT* occur: N/A
Distribution: Gentoo amd64 2006.1
Hardware Environment: Asus A8R32-MVP Deluxe, Athlon64 3800+ x2, 2 x 512MB RAM
Software Environment: 
Problem Description:

I was trying the updated sky2 driver in 2.6.21-rc1 on my server running a
Gigabyte GA-965P-DS4 (FW: F8) motherboard with Marvell 8053 chip, Intel Core 2
Duo CPU. After approx. 48-72 hours, I lost almost all speed (going down to
approx. 15 kB/s) in uplink direction and sometimes lost connection completely
for a shorter while. I'm not sure what happened with my speed in downlink
direction. I then went back to Marvell own driver patched into a 2.6.20 kernel
on the server.

Meanwhile I started to try to reproduce it on my lab PC running a Asus A8R32-MVP
Deluxe motherboard (FW: 0602) which have two Gb LAN; one Marvell 8001 on the PCI
bus and one 8053 on the PCIe bus, Athlon64 3800+ x2 CPU. I installed 2.6.21-rc1
on this machine too, and started to use the 8053 with the sky2 driver. Made a
lot of FTP transfers in both direction, but mostly uploads to stress the uplink
direction. And after approx. 34 hours, I see the same symptom on this machine
too. A reboot recovers the bad speed.

Some system info (please tell me if you want more or other information):

dmesg:
======
...
Adding 996020k swap on /dev/hda2.  Priority:-1 extents:1 across:996020k
skge eth0: enabling interface
sky2 eth1: enabling interface
sky2 eth1: ram buffer 48K
sky2 eth1: Link is up at 100 Mbps, full duplex, flow control both
NETDEV WATCHDOG: eth1: transmit timed out
sky2 eth1: tx timeout
sky2 eth1: transmit ring 126 .. 85 report=126 done=126
sky2 eth1: disabling interface
sky2 eth1: enabling interface
sky2 eth1: ram buffer 48K
sky2 eth1: Link is up at 100 Mbps, full duplex, flow control both
NETDEV WATCHDOG: eth1: transmit timed out
sky2 eth1: tx timeout
sky2 eth1: transmit ring 227 .. 186 report=227 done=227
sky2 eth1: disabling interface
sky2 eth1: enabling interface
sky2 eth1: ram buffer 48K
sky2 eth1: Link is up at 100 Mbps, full duplex, flow control both
sky2 eth1: Link is down.
sky2 eth1: Link is up at 100 Mbps, full duplex, flow control both
usb 2-2: new low speed USB device using ohci_hcd and address 3
usb 2-2: configuration #1 chosen from 1 choice
input: Logitech HID compliant keyboard as /class/input/input3
input: USB HID v1.10 Keyboard [Logitech HID compliant keyboard] on
usb-0000:00:1c.0-2
input: Logitech HID compliant keyboard as /class/input/input4
input: USB HID v1.10 Device [Logitech HID compliant keyboard] on usb-0000:00:1c.0-2
...

lspci:
======
00:00.0 Host bridge: ATI Technologies Inc Unknown device 5952
00:02.0 PCI bridge: ATI Technologies Inc RS480 PCI-X Root Port
00:06.0 PCI bridge: ATI Technologies Inc RS480 PCI Bridge
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
HyperTransport Technology Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] DRAM
Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
Miscellaneous Control
00:1a.0 PCI bridge: ALi Corporation M5249 HTT to PCI Bridge
00:1c.0 USB Controller: ALi Corporation USB 1.1 Controller (rev 03)
00:1c.1 USB Controller: ALi Corporation USB 1.1 Controller (rev 03)
00:1c.2 USB Controller: ALi Corporation USB 1.1 Controller (rev 03)
00:1c.3 USB Controller: ALi Corporation USB 2.0 Controller (rev 01)
00:1d.0 Audio device: ALi Corporation High Definition Audio/AC'97 Host
Controller (rev 02)
00:1e.0 ISA bridge: ALi Corporation Unknown device 1575
00:1e.1 Bridge: ALi Corporation M7101 Power Management Controller [PMU]
00:1f.0 IDE interface: ALi Corporation M5229 IDE (rev c8)
00:1f.1 SATA controller: ALi Corporation ULi M5288 SATA (rev 10)
01:14.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit
Ethernet Controller (rev 14)
02:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8053 PCI-E Gigabit
Ethernet Controller (rev 20)
03:00.0 VGA compatible controller: nVidia Corporation G70 [GeForce 7600 GS] (rev a1)

cat /proc/interrupts:
=====================
           CPU0       CPU1       
  0:      39426   31020759   IO-APIC-edge      timer
  1:          0          2   IO-APIC-edge      i8042
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          0          4   IO-APIC-edge      i8042
 14:       1796     585677   IO-APIC-edge      ide0
 15:       1054    1112304   IO-APIC-edge      ide1
 17:          2       2993   IO-APIC-fasteoi   ohci_hcd:usb2
 18:       5288   30848696   IO-APIC-fasteoi   eth1, ohci_hcd:usb3
 19:          0          0   IO-APIC-fasteoi   libata, ohci_hcd:usb4
 21:         59     278355   IO-APIC-fasteoi   eth0
 23:          0          4   IO-APIC-fasteoi   ehci_hcd:usb1
NMI:       1102       1879 
LOC:   31061659   31061636 
ERR:          0

ethtool eth1:
=============
Settings for eth1:
	Supported ports: [ TP ]
	Supported link modes:   10baseT/Half 10baseT/Full 
	                        100baseT/Half 100baseT/Full 
	                        1000baseT/Half 1000baseT/Full 
	Supports auto-negotiation: Yes
	Advertised link modes:  10baseT/Half 10baseT/Full 
	                        100baseT/Half 100baseT/Full 
	                        1000baseT/Half 1000baseT/Full 
	Advertised auto-negotiation: Yes
	Speed: 100Mb/s
	Duplex: Full
	Port: Twisted Pair
	PHYAD: 0
	Transceiver: internal
	Auto-negotiation: on
	Supports Wake-on: pg
	Wake-on: d
	Current message level: 0x000000ff (255)
	Link detected: yes

Steps to reproduce:

Use the sky2 driver in 2.6.21-rc1 on a Marvell 8053 chip. Do a lot of transfers
and wait 48 hours or more. Hopefully something bad happens.
Comment 1 Stephen Hemminger 2007-03-14 14:57:35 UTC
Why is MSI disabled on this motherboard?
Is it because of the AMD chipset or did you explicitly turn it off.?
Comment 2 Peter Kerwien 2007-03-14 21:45:04 UTC
I haven't disabled MSI. BTW, I don't even know what it is.
Comment 3 Stephen Hemminger 2007-04-17 15:02:43 UTC
Several timeout fixes (related to flow control) went into later versions
of 2.6.21-rc4? and stable kernels 2.6.20.6.

Please try and reproduce with more recent kernel.
Comment 4 Stephen Hemminger 2007-04-18 13:19:55 UTC
This looks like same as previous bug (still not fixed).

*** This bug has been marked as a duplicate of 7546 ***