|
Bugzilla – Full Text Bug Listing |
| Summary: | Realtek 8139 does not work with acpi on laptop Mitac 8050QDA (odd workaround found) | ||
|---|---|---|---|
| Product: | [openSUSE] SUSE LINUX 10.0 | Reporter: | k. t. <cooling.crystals> |
| Component: | Kernel | Assignee: | Thomas Renninger <trenn> |
| Status: | RESOLVED DUPLICATE | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Normal | ||
| Priority: | P3 - Medium | CC: | cooling.crystals, tiwai |
| Version: | Final | ||
| Target Milestone: | --- | ||
| Hardware: | i686 | ||
| OS: | Other | ||
| Whiteboard: | |||
| Found By: | Other | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Attachments: |
dmesg
acpidmp new file hwinfo --sound |
||
|
Description
k. t.
2005-09-11 16:13:22 UTC
Please, at the very first, tell us what does not work? 'hwinfo --netcard'? Is the driver for your NIC loaded automatically? Do you get an network interface? Is the interface set up properly? Without further information this bug report is just unusable (--> invalid). 8050QDA:/home/kristjan # hwinfo --netcard
26: PCI 204.0: 0200 Ethernet controller
[Created at pci.277]
UDI: /org/freedesktop/Hal/devices/pci_10ec_8139
Unique ID: rBUF.fpFizmEv0jF
Parent ID: 6NW+.pbi2aV6ikBD
SysFS ID:
/devices/pci0000:00/0000:00:1e.0/0000:02:04.0
SysFS BusID: 0000:02:04.0
Hardware Class: network
Model: "Mitac RTL-8139/8139C/8139C+"
Vendor: pci 0x10ec "Realtek Semiconductor Co., Ltd."
Device: pci 0x8139 "RTL-8139/8139C/8139C+"
SubVendor: pci 0x1071 "Mitac"
SubDevice: pci 0x8054
Revision: 0x10
Driver: "8139too"
Device File: eth0
I/O Ports: 0xa000-0xbfff (rw)
Memory Range: 0xb0001000-0xb00010ff
(rw,non-prefetchable)
IRQ: 5 (3727 events)
HW Address: 00:40:d0:76:49:a3
Link detected: yes
Module Alias:
"pci:v000010ECd00008139sv00001071sd00008054bc02sc00i00"
Driver Info #0:
Driver Status: 8139too is active
Driver Activation Cmd: "modprobe 8139too"
Driver Info #1:
Driver Status: 8139cp is not active
Driver Activation Cmd: "modprobe 8139cp"
Config Status: cfg=no, avail=yes, need=no,
active=unknown
Attached to: #15 (PCI bridge)
27: PCI 205.0: 0282 WLAN controller
[Created at pci.277]
UDI: /org/freedesktop/Hal/devices/pci_8086_4220
Unique ID: ZcKW.gy1ZasUC0k5
Parent ID: 6NW+.pbi2aV6ikBD
SysFS ID:
/devices/pci0000:00/0000:00:1e.0/0000:02:05.0
SysFS BusID: 0000:02:05.0
Hardware Class: network
Model: "Intel PRO/Wireless 2200BG"
Vendor: pci 0x8086 "Intel Corporation"
Device: pci 0x4220 "PRO/Wireless 2200BG"
SubVendor: pci 0x8086 "Intel Corporation"
SubDevice: pci 0x2702
Revision: 0x05
Features: WLAN
Memory Range: 0xb0000000-0xb0000fff
(rw,non-prefetchable)
IRQ: 11 (no events)
Requires: ipw-firmware
Module Alias:
"pci:v00008086d00004220sv00008086sd00002702bc02sc80i00"
Driver Info #0:
Driver Status: ipw2200 is active
Driver Activation Cmd: "modprobe ipw2200"
Config Status: cfg=no, avail=yes, need=no,
active=unknown
Attached to: #15 (PCI bridge)
Did I really explain my problem in too complicated
way?
Ok here is another example:
*run knights*
kristjan@8050QDA:~> ping -c 3 www.yahoo.com
PING www.yahoo.akadns.net (68.142.226.45) 56(84) bytes
of data.
64 bytes from p14.www.re2.yahoo.com (68.142.226.45):
icmp_seq=1 ttl=54 time=148 ms
64 bytes from p14.www.re2.yahoo.com (68.142.226.45):
icmp_seq=2 ttl=54 time=146 ms
64 bytes from p14.www.re2.yahoo.com (68.142.226.45):
icmp_seq=3 ttl=54 time=145 ms
--- www.yahoo.akadns.net ping statistics ---
3 packets transmitted, 3 received, 0% packet loss,
time 2008ms
rtt min/avg/max/mdev = 145.111/146.893/148.921/1.565
ms
*now I shut down knights and ping again*
kristjan@8050QDA:~> ping -c 3 www.yahoo.com
ping: unknown host www.yahoo.com
*dmesg*
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a05a. (queue head)
eth0: Tx descriptor 1 is 9008a24e.
eth0: Tx descriptor 2 is 9008a04e.
eth0: Tx descriptor 3 is 9008a046.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a24e. (queue head)
eth0: Tx descriptor 1 is 9008a046.
eth0: Tx descriptor 2 is 9008a046.
eth0: Tx descriptor 3 is 9008a24e.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a24e. (queue head)
eth0: Tx descriptor 1 is 9008a24e.
eth0: Tx descriptor 2 is 9008a24e.
eth0: Tx descriptor 3 is 9008a24e.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 407 dirty entry 403.
eth0: Tx descriptor 0 is 9008a048.
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a05b.
eth0: Tx descriptor 3 is 9008a048. (queue head)
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timeout, status 0c 0005 c07f media 18.
eth0: Tx queue start entry 4 dirty entry 0.
eth0: Tx descriptor 0 is 9008a03c. (queue head)
eth0: Tx descriptor 1 is 9008a03c.
eth0: Tx descriptor 2 is 9008a03c.
eth0: Tx descriptor 3 is 9008a03c.
eth0: link up, 10Mbps, half-duplex, lpa 0x0000
> Did I really explain my problem in too complicated way?
No. You didn't explain it at all.
OK. I have laptop Mitac 8050QDA. I am unable to use onboard eth0 Realtek RTL-8139/8139C+ (rev 10), UNLESS I run CPU intensive? tasks/programs (such as game Knights!) I need to know why I can't connect to internet if I don't have amaroK/Knights running (suse 9.3/suse10RC1). I know by now that if I boot with option acpi=off I will not have any problem with internet connection. Network interface is set up properly. Module 8139too loads automatically. In conclusion if I want to be connected to the internet (without boot option acpi=off) I have to launch game called 'Knights' (this includes network card configuration using YaST). Final conclusion after some research: This has something to do with arts or esd running. If these soundservers aren't used then realtek isn't working (responding?). So it isn't about cpu usage, it's more about esd or arts running in backround (are being used/accesed by some program) as prerequisite. This is acpi (5: 459879 XT-PIC Intel ICH6, eth0) problem and looks like I found very original workaround! Could you try with pci=noacpi - if this works it should be your preferred workaround as you should have full ACPI support (beside interrupt routing). We could still have a look what goes wrong without any acpi boot options... Important for me is dmesg output when booting without any acpi options and possibly the output of acpidmp. Created attachment 54795 [details]
dmesg
Created attachment 54796 [details]
acpidmp
Internet works with boot option pci=noacpi ! Sadly X fails to start: Oct 19 17:42:52 6lu kernel: nvidia: module not supported by Novell, setting U taint flag. Oct 19 17:42:52 6lu kernel: nvidia: module license 'NVIDIA' taints kernel. Oct 19 17:42:52 6lu kernel: PCI: No IRQ known for interrupt pin A of device 0000:01:00.0. Please try using pci=biosirq. Oct 19 17:42:52 6lu kernel: NVRM: Can't find an IRQ for your NVIDIA card! Oct 19 17:42:52 6lu kernel: NVRM: Please check your BIOS settings. Oct 19 17:42:52 6lu kernel: NVRM: [Plug & Play OS] should be set to NO Oct 19 17:42:52 6lu kernel: NVRM: [Assign IRQ to VGA] should be set to YES Oct 19 17:42:52 6lu kernel: nvidia: probe of 0000:01:00.0 failed with error -1 Oct 19 17:42:52 6lu kernel: NVRM: The NVIDIA probe routine failed for 1 device(s). Oct 19 17:42:52 6lu kernel: NVRM: None of the NVIDIA graphics adapters were initialized! Oct 19 17:42:53 6lu kdm: :0[5366]: IO Error in XOpenDisplay Oct 19 17:42:53 6lu kdm[5330]: X server for display :0 terminated unexpectedly Oct 19 17:42:53 6lu kdm[5330]: Display :0 cannot be opened Oct 19 17:42:53 6lu kdm[5330]: Unable to fire up local display :0; disabling. I know that I could try with "nv" driver, but image quality isn't the same (it is more like 16bit even with 24bit). About Plug and Play OS and Assign IRQ to vga - not available in my bios settings... dmesg and acpidmp have been created then system was booted with following line : selinux=0 resume=/dev/sda1 splash=off showopts The acpidmp output is cut, the beginning is missing. Could you repost, please. So with pci=noacpi your NVidia card doesn't get a valid IRQ and without your network card does not get one (except your soundcard is running) ... This is really strange... Takashi, do you have any idea? (How do you know that network card doesn't get one ?) Created attachment 54867 [details]
new file
I don't know how to edit old one so I post new file.
Hmmm, I don't see anything obvious wrong... Takashi, do you have any ideas on how to get there further or why the network card works after sound initialisation? Which sound driver and model? Show /proc/asound/cards. The problem is likely the interrupt sharing and handling... Also, do you have any special module option? Check /etc/modprobe.d/sound. Sorry, /proc/asound/cards is empty file (opened with text editor kate) /etc/modprobe.d/sound: options snd-intel8x0 enable=1 index=0 # r0Vg.CiSFcYKExl1:82801FB/FBM/FR/FW/FRW (ICH6 Family) AC'97 Audio Controller alias snd-card-0 snd-intel8x0 I will add hwinfo --sound Created attachment 54928 [details]
hwinfo --sound
ah... you mean cat /proc/asound/cards ?
kristjan@kohvi:~> cat /proc/asound/cards
0 [ICH6 ]: ICH4 - Intel ICH6
Intel ICH6 with ALC655 at 0xd0000000, irq 5
At the moment I use pci=noacpi and "nv" driver (Option "FPDither" enables better image quality, almost normal color gradient now)
First you should ensure, that you have the latest BIOS installed. As this seem to be a machine specific problem, even with specific devices, I like to close this one. If it still happens with the most recent BIOS (also try pci=noacpi, then noapic if that does not help try both). I think further tracking is only worth if this still happens on newest 10.1 Alpha installations. As said the problem seems very machine specific and is likely to be a BIOS bug... Please reopen if you still encounter that problem on a new 10.1 Alpha installation with new BIOS. irqpoll fixes the problem (BIOS not updated); dhcp works; network is set up at boot; no more "transmit timeout" errors; I will test 10.1 as soon as possible. My notebook is 8050QDA. I have done BIOS update (EC and SYSTEM are latest now). I have 10.1 beta3 installed and ethernet card still has same strange behaviour that can be fixed/avoided with boot option 'irqpoll'. I have tried to free IRQ for 'nvidia' by removing unused modem card and it gets independent irq but still complains with 'pci=noacpi' and does not startx (this is with 10.0 ;'nvidia' not installed in beta3 ; modem card is still removed since I don't need it) What's next ? Is this a bug in BIOS (still not fixed in newest version) or can we say it's acpi in linux or 8139too module ? Very interesting. In 10.1 beta3 network fails with pci=noacpi or acpi=off or normal boot. I can get network working with irqpoll or doing "knights trick" (the game I start and realtek will magically work). does it work with "apic" and with/without pci=noacpi? Not sure what changed there, please follow up #149407 *** This bug has been marked as a duplicate of 149407 *** Great! I got it (realtek) running with 'apic'. But isn't this dangerous?
kbeta@kohvi:~> cat /proc/interrupts
CPU0
0: 58357 IO-APIC-edge timer
1: 330 IO-APIC-edge i8042
8: 2 IO-APIC-edge rtc
9: 196 IO-APIC-level acpi
12: 24420 IO-APIC-edge i8042
14: 11850 IO-APIC-edge libata
15: 1592 IO-APIC-edge libata
177: 1 IO-APIC-level uhci_hcd:usb4, yenta
209: 2 IO-APIC-level uhci_hcd:usb3, ohci1394
217: 0 IO-APIC-level ehci_hcd:usb1, uhci_hcd:usb2
225: 0 IO-APIC-level Intel ICH6
233: 1235 IO-APIC-level eth1
NMI: 0
LOC: 257999
ERR: 0
MIS: 0
I'm confused. What should I do with bug 149407 ?
Using "apic" boot option instead of "irqpoll" is not good option for this hardware; I got complete touchpad freeze (although it happened with 10.0 I think it's possible that it will happen with 10.1 also) luckily keyboard was still working and I rebooted. After reboot to 10.0 X failed to start so I rebooted to 10.1 beta 3 and same happened there. Next I did shut down and turned system on again after 10 seconds and booted to 10.0 this time everything worked again. Ofcourse- bug 148094 could play a role here but I have had no more crashes (touchpad+keboard) since using boot option "psmouse.rate=40" (with "irqpoll"). There is the same problem in bug #149407. Best is you copy and paste the last two comments as they are there. The info is very interesting and it should not split over several bugs if there is the same problem. On bug #149407 they are heavily working on. They even think about setting apic as default AFAIK. You might also add vojtech@suse.de to CC in this bug if he's not already. He is in charge of kernel keyboard/touchpad stuff. Currently only 10.1 is interesting for us. If this one is out we could think about backporting *safe* things (no already working machines are allowed to break, this is often very difficult to ensure). I think I'll do nothing now unless I'm asked to do so at bug 149407. -this is my last comment here at bug 116397- Why do you need psmouse.rate=40? Have you also tried this to add/delete this with apic option? I'm scared to boot without "psmouse.rate=40" (see the bug 148094). I don't believe that "apic" would fix 2 bugs at the same time (realtek irq + touchpad/keyboard hang) but nevertheless I'll try it explicitly (rebooting to beta3 10.1 and giving it a serious test run with "apic")... OK. "apic" only boot without "psmouse.rate=40" resulted in keyboard+touchpad freeze. So "psmouse.rate=40" is must have for me (there is already patch; see bug 148094 for more details). Want me to test "apic" + "psmouse.rate=40" in beta3 suse10.1 (I don't think this is necessary; because I had touchpad freeze in 10.0 with with the same boot line already)? |