[intrepid] PCMCIA network-card working in hardy does not work

[Mika Pflüger]

exakt version of the kernel: 2.6.27-4-generic (Ubuntu 2.6.27-4.6-generic)

I am testing Intrepid on a fairly old Laptop without build-in network card, thus I use a Netgear FA411 Fast Ethernet PCMCIA network card. Under hardy it works like a charm (with every version of kernel that was ever present in hardy), no problems, works after suspend/resume and hibernate/resume, everything's fine. But under Intrepid it doesn't work. To be more specific: the driver refuses to drive the card claiming somebody else is responsible, but no other driver shows up, so the cards stays undriven.

The symptoms: the card's leds are on, ifconfig reports only loopback device, the gnome tool in the upper right-hand corner does not see a device.

I have the output of lspcmcia -vv under hardy and intrepid as well as the output of dmesg under both attached, the most interesting lines I cite:

lspcmcia -vv hardy:
Socket 1 Device 0: [pcnet_cs] (bus ID: 1.0)

lspcmcia -vv intrepid:
Socket 1 Device 0: [-- no driver --] (bus ID: 1.0)

dmesg hardy:
[ 90.051734] pccard: PCMCIA card inserted into slot 1
[ 102.844908] cs: IO port probe 0x100-0x3af: clean.
[ 102.847719] cs: IO port probe 0x3e0-0x4ff: clean.
[ 102.849158] cs: IO port probe 0x820-0x8ff: clean.
[ 102.850492] cs: IO port probe 0xc00-0xcf7: clean.
[ 102.853025] cs: IO port probe 0xa00-0xaff: clean.
[ 103.286116] cs: memory probe 0xa0000000-0xa0ffffff: clean.
[ 103.294603] pcmcia: registering new device pcmcia1.0
[ 103.308750] cs: IO port probe 0x100-0x3af: clean.
[ 103.311598] cs: IO port probe 0x3e0-0x4ff: clean.
[ 103.313037] cs: IO port probe 0x820-0x8ff: clean.
[ 103.319870] cs: IO port probe 0xc00-0xcf7: clean.
[ 103.322501] cs: IO port probe 0xa00-0xaff: clean.
[ 104.536688] eth0: NE2000 Compatible: io 0x320, irq 3, hw_addr 00:40:f4:24:70:ec

dmesg intrepid:
[ 50.771501] pccard: PCMCIA card inserted into slot 1
[ 52.287586] input: PS2++ Logitech TouchPad 3 as /class/input/input2
[ 52.374044] cs: IO port probe 0x100-0x3af: excluding 0x200-0x207 0x220-0x22f 0x300-0x307 0x378-0x37f 0x388-0x38f
[ 52.376950] cs: IO port probe 0x3e0-0x4ff: clean.
[ 52.378348] cs: IO port probe 0x820-0x8ff: clean.
[ 52.388953] cs: IO port probe 0x100-0x3af: excluding 0x200-0x207 0x220-0x22f 0x300-0x307 0x378-0x37f 0x388-0x38f
[ 52.391827] cs: IO port probe 0x3e0-0x4ff: clean.
[ 52.393223] cs: IO port probe 0x820-0x8ff: clean.
[ 52.394451] cs: IO port probe 0xc00-0xcf7: clean.
[ 52.396832] cs: IO port probe 0xa00-0xaff: clean.
[ 52.398230] cs: memory probe 0xa0000000-0xa0ffffff: clean.
[ 52.407451] cs: IO port probe 0xc00-0xcf7: clean.
[ 52.409796] cs: IO port probe 0xa00-0xaff: clean.
[ 52.416103] pcmcia: registering new device pcmcia1.0
[ 54.255498] pcnet_cs: this is an AX88190 card!
[ 54.255529] pcnet_cs: use axnet_cs instead.

attached the complete outputs. If you need further information - ask.

[Mika Pflüger (mikapflueger)]

I looked quickly what the driver axnet_cs (that is proposed by pnet_cs) is for: it is for spcial AXIS-based cards. I did not find any hints that the Netgear FA411 card is based on this chipset. I would rather say, it ist not. So maybe the error is, that pcnet_cs falsely detects an AXIS card.

[Mika Pflüger (mikapflueger)]

Hi,

I don't think the bug is a duplicate of bug #256054, because:
1. I tried the workaround described at bug #256054, that means I disabled the NetworkManager completely. It isn't loaded anymore (does not appear in boot log, or whatever the name of the output ist, that you get when you give nosplash at boot prompt) and 'ps aux | grep net' does not show any running instances (of course after a reboot after disabling the NetworkManager). Although NetworkManager can't do anything anymore, the dmesg-output is the same (not exactly, NetworkManager does not appear in dmesg, but it didn't appear around the PCMCIA-probes before, anyway, as I point out later), eth0 does not appear in ifconfig-output, and lspcmcia-vv gives exactly the same output. Or in short: It didn't change anything.

2. The problem arises far earlier than NetworkManager is being loaded. The problem arises during the stage in booting process when the hardware drivers are being loaded. That is before any runlevel. After that, when entering runlevel 2, roughly, ifup is excuted (without any special impact, as /dev/eth0 does not exist) and way after that (when entering runlevel 5), NetworkManager is executed and /could/ indeed interfere with any settings, if there would even be a network device. But there is none, and there are no special settings, it would just work fine with DHCP, if NM would set any device to that, but there is no networking device (virtually, physically it is sitting there blinking). Furthermore, before removing NetworkManager I could access NetworkManager and it was reporting "No ethernet device" or something like this. That means I can confirm that NetworkManager did not find any devices and can therefore not mess the configuration of these.

So I would say this is not a duplicate of bug #256054.

[Chris Coulson (chrisccoulson)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. Unfortunately we can't fix it without more information.

Please include the information requested at https://wiki.ubuntu.com/DebuggingHardwareDetection as separate attachments.

[Mika Pflüger (mikapflueger)]

It just came to my mind that this might matter as well (I almost forgot it because it is so everyday for me...):
I am booting the kernel with the kerneloption pci=irqmask=0x0e98. That is because neither pci=biosirq nor without any pci-options wrok. Both fail to detect the harddrive and the cd-rom drive, although they are both standard IDE (or at least claim to be). I read about the pci=irqmask=0x0e98 option that it is required for this notebook, and actually you won't get far without (it won't boot), this is known for this laptop at least since 2001.
I don't use other kernel boot parameters, only ro, pci=irqmask=0x0e98 and nosplash.

[Chris Coulson (chrisccoulson)]

Mika, what happens if you manually load the pcnet_cs driver:

sudo modprobe pcnet_cs

[Mika Pflüger (mikapflueger)]

Hi,

shortly: it doesn't happen anything.

long version:
I describe what I do, just in case I misunderstood you.
I boot intrepid, log in and run
sudo modprobe pcnet_cs
No output at all, dmseg|tail does not reveal any messages since boottime.
No other log in /var/log that I checked has any messages.
Thus I reboot and try this time
sudo modprobe -v --first-time pcnet_cs
to get at least some output. I get
FATAL: Module pcnet_cs already in kernel.
To check that, I reboot and say
lsmod | grep pcnet_cs (output see attachment, four lines)
and then I sudo modprobe.
After loading the driver manually lsmod looks the same (lsmod|grep pcnet_cs gives the same four lines, I diffed them, it is exactly the same, as in the attachment)

In ifconfig the network interface does not show up.

I don't know if there is a way to manually bind a driver to a pcmcia-card (at least pccardctl does not have this function), if there is a way, I'd give it a try.

[spiderbatdad (spiderbatdad)]

I am having the same problem on my friends machine...the card worked wel in gutsy and hardy. The following occurs during boot now...intrepid 2.6.27-5

[ 271.664079] wlan0: issue_cmd(): timed out waiting for CMD_COMPLETE. irq bits:0x0000 irq_status:0x0000 timeout:49ms cmd_status:0 (Idle)
[ 271.664104] wlan0: issue_cmd(cmd:0x0002) FAILED
[ 271.664118] Pid: 4903, comm: NetworkManager Not tainted 2.6.27-5-generic #1
[ 271.664130] [<c037d686>] ? printk+0x1d/0x1f
[ 271.664177] [<cfb85749>] acxpci_s_issue_cmd_timeo+0x269/0x3b0 [acx]
[ 271.664268] [<cfb7f599>] acx_s_configure+0x99/0xc0 [acx]
[ 271.664293] [<cfb81217>] acx_s_update_card_settings+0x5f7/0x1110 [acx]
[ 271.664318] [<c012a571>] ? hrtick_start_fair+0x181/0x1a0
[ 271.664350] [<cfb7bfd6>] acx_ioctl_commit+0x26/0x40 [acx]
[ 271.664370] [<cfb7bfb0>] ? acx_ioctl_commit+0x0/0x40 [acx]
[ 271.664391] [<c036b12c>] ioctl_standard_call+0xbc/0x110
[ 271.664409] [<c02f3b85>] ? __dev_get_by_name+0x85/0xb0
[ 271.664425] [<c036a99e>] wireless_process_ioctl+0xbe/0x150
[ 271.664438] [<cfb7ba50>] ? acx_ioctl_set_encode+0x0/0x1a0 [acx]
[ 271.664459] [<c037e6b0>] ? mutex_lock+0x10/0x20
[ 271.664470] [<c036aaa0>] wext_handle_ioctl+0x70/0xe0
[ 271.664482] [<c036b070>] ? ioctl_standard_call+0x0/0x110
[ 271.664494] [<c036ac60>] ? ioctl_private_call+0x0/0x180
[ 271.664507] [<c02f5d2f>] dev_ioctl+0x45f/0x520
[ 271.664518] [<c0342d39>] ? inet_create+0x229/0x300
[ 271.664531] [<c037f8ad>] ? _spin_lock+0xd/0x10
[ 271.664543] [<c01e36f5>] ? inotify_d_instantiate+0x45/0x60
[ 271.664559] [<c0215548>] ? security_d_instantiate+0x18/0x30
[ 271.664578] [<c01c7477>] ? d_instantiate+0x57/0x60
[ 271.664592] [<c02e62f0>] ? sock_ioctl+0x0/0x250
[ 271.664605] [<c02e63e5>] sock_ioctl+0xf5/0x250
[ 271.664614] [<c02e62f0>] ? sock_ioctl+0x0/0x250
[ 271.664624] [<c01c067d>] vfs_ioctl+0x2d/0x90
[ 271.664638] [<c01c0866>] do_vfs_ioctl+0x66/0x1f0
[ 271.664650] [<c01c0a5b>] sys_ioctl+0x6b/0x70
[ 271.664661] [<c0103f7b>] sysenter_do_call+0x12/0x2f
[ 271.664677] =======================
[ 271.664684] wlan0: configure(type:0x1010) FAILED
[ 271.724128] wlan0: issue_cmd(): timed out waiting for CMD_COMPLETE. irq bits:0x0000 irq_status:0x0000 timeout:49ms cmd_status:0 (Idle)
[ 271.724160] wlan0: issue_cmd(cmd:0x0016) FAILED
[ 271.724173] Pid: 4903, comm: NetworkManager Not tainted 2.6.27-5-generic #1
[ 271.724185] [<c037d686>] ? printk+0x1d/0x1f
[ 271.724230] [<cfb85749>] acxpci_s_issue_cmd_timeo+0x269/0x3b0 [acx]
[ 271.724322] [<cfb80a8a>] acx_s_set_probe_request_template+0x14a/0x160 [acx]
[ 271.724354] [<cfb818c3>] acx_s_update_card_settings+0xca3/0x1110 [acx]
[ 271.724378] [<c014a72d>] ? enqueue_hrtimer+0x7d/0x120
[ 271.724398] [<c014b304>] ? hrtimer_start+0xc4/0x1c0
[ 271.724413] [<c012a571>] ? hrtick_start_fair+0x181/0x1a0
[ 271.724431] [<c012ad2d>] ? check_preempt_wakeup+0x14d/0x1b0
[ 271.724444] [<c012b4ee>] ? try_to_wake_up+0xde/0x290
[ 271.724457] [<c031038d>] ? __nla_put+0x1d/0x50
[ 271.724477] [<c036a69f>] ? wireless_send_event+0x15f/0x230
[ 271.724497] [<cfb7bfd6>] acx_ioctl_commit+0x26/0x40 [acx]
[ 271.724517] [<cfb7bfb0>] ? acx_ioctl_commit+0x0...

[Chris Coulson (chrisccoulson)]

spiderbatdad - yours is a different problem. Please open a new bug report.

Mika - Are you saying that the pcnet_cs module is already loaded automatically when you boot?

Thanks

[Mika Pflüger (mikapflueger)]

Yes, it is loaded automatically on boot time and it is used by the modules pcmcia_core and pcmcia (as lsmod shows). I really don't know much about the linux driver system, but to me it looks as if pcnet_cs fails to be somehow bound to the pcmcia card.
In dmesg you can see, that pcnet_cs (or somebody else with the name 'cs') even probes IO ports on boot time. So it is loaded in the kernel.

Thanks for your efforts so far, Chris.

[dunbrokin (pj-king)]

I have a similar problem with my PCMCIA card....it worked perfectly under 8.04 but now cannot see the router.

Here is a printout:

fitz@fitz-laptop:~$ sudo lshw -C Network
*-network
description: Ethernet interface
product: 3c556B CardBus [Tornado]
vendor: 3Com Corporation
physical id: 3
bus info: pci@0000:00:03.0
logical name: eth0
version: 20
serial: 00:04:76:60:6f:87
size: 10MB/s
capacity: 100MB/s
width: 32 bits
clock: 33MHz
capabilities: pm bus_master cap_list ethernet physical tp mii 10bt 10bt-fd 100bt 100bt-fd autonegotiation
configuration: autonegotiation=on broadcast=yes driver=3c59x duplex=half latency=64 link=no maxlatency=10 mingnt=10 module=3c59x multicast=yes port=MII speed=10MB/s
*-network DISABLED
description: Wireless interface
product: ACX 111 54Mbps Wireless Interface
vendor: Texas Instruments
physical id: 1
bus info: pci@0000:06:00.0
logical name: wlan0
version: 00
serial: 00:0f:3d:57:76:56
width: 32 bits
clock: 33MHz
capabilities: pm bus_master cap_list ethernet physical wireless
configuration: broadcast=yes driver=acx_pci latency=64 module=acx multicast=yes wireless=IEEE 802.11b+/g+
*-network DISABLED
description: Ethernet interface
physical id: 1
logical name: pan0
serial: d2:f8:de:93:ae:85
capabilities: ethernet physical
configuration: broadcast=yes driver=bridge driverversion=2.3 firmware=N/A link=yes multicast=yes
fitz@fitz-laptop:~$ lspci -nn
00:00.0 Host bridge [0600]: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX Host bridge [8086:7190] (rev 03)
00:01.0 PCI bridge [0604]: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX AGP bridge [8086:7191] (rev 03)
00:02.0 CardBus bridge [0607]: Texas Instruments PCI1450 [104c:ac1b] (rev 03)
00:02.1 CardBus bridge [0607]: Texas Instruments PCI1450 [104c:ac1b] (rev 03)
00:03.0 Ethernet controller [0200]: 3Com Corporation 3c556B CardBus [Tornado] [10b7:6056] (rev 20)
00:03.1 Communication controller [0780]: 3Com Corporation Mini PCI 56k Winmodem [10b7:1007] (rev 20)
00:05.0 Multimedia audio controller [0401]: Cirrus Logic CS 4614/22/24/30 [CrystalClear SoundFusion Audio Accelerator] [1013:6003] (rev 01)
00:07.0 Bridge [0680]: Intel Corporation 82371AB/EB/MB PIIX4 ISA [8086:7110] (rev 02)
00:07.1 IDE interface [0101]: Intel Corporation 82371AB/EB/MB PIIX4 IDE [8086:7111] (rev 01)
00:07.2 USB Controller [0c03]: Intel Corporation 82371AB/EB/MB PIIX4 USB [8086:7112] (rev 01)
00:07.3 Bridge [0680]: Intel Corporation 82371AB/EB/MB PIIX4 ACPI [8086:7113] (rev 03)
01:00.0 VGA compatible controller [0300]: S3 Inc. 86C270-294 Savage/IX-MV [5333:8c12] (rev 13)
06:00.0 Network controller [0280]: Texas Instruments ACX 111 54Mbps Wireless Interface [104c:9066]
fitz@fitz-laptop:~$ dmseg | grep -e iwl -e wlan -e rt -e ath
bash: dmseg: command not found
fitz@fitz-laptop:~$ dmesg | grep -e iwl -e wlan -e rt -e ath
[ 0.000000] Movable zone start PFN for each node
[ 0.000000] ACPI: PM-Timer IO Port: 0x1008
[ 0.000000] Allocating PCI resources starting at 30000000 (gap: 20000000:dff80000)
[ 0.000000] Enabling unmasked SIMD FPU exception support... done.
[ 0.004000] virtual kernel memory layout:
[ 0.881353] Booting paravirtualized kernel on bare hardware
[ 1.365536] ACPI: (supports S0 S1 S3 S4 S5)
[ 1.626...

[Mika Pflüger (mikapflueger)]

Dunbrokin, yours seems to be another bug.

Because:
1. you - at least sometimes - see your interface through ifconfig.
2. your card uses another driver than mine
3. the driver reports problems due to 32bit I/O, are you running a 64 bit version? - if yes the problem could be the driver not beeing compatible with 64 bit kernel. Actually, the I/O timing problems the driver predicts do happen, thus this could be the issue

Therefore, maybe it is better if you report another bug.
Remember to _attach_ in your bug report the files generated by the following comands:
uname -a > uname-a.log
cat /proc/version_signature > proc_version_signature.log
sudo lspci -vvnn > lspci-vvnn.log (those are 2 v's not a w)
dmesg > dmesg.log
sudo lspcmcia -vv > lspcmcia-vv.log
That means, that you have to attach five different files.

Additionally, you say, that the card worked under kde 4.1 - unfortunately, hardware-issues are usually independent from the window manager / desktop suite and therefore you should maybe add the information, wich distro/version you were running (hardy?), the best would be to attach the output of dmesg -a > dmesg-a.log.workedBefore

It would be really kind if you could take the time to gather these information and add it do a new bug report. Hopefully some developers can then fix the bug, so that you and everybody else with a similar hardware configuration can enjoy intrepid final.

Thanks, Mika

[Mika Pflüger (mikapflueger)]

Hi,

tested again with the same hardware the Ubuntu Intrepid RC (from the i386-alternate CD) and got the same errors.
Additionally I now tried to install with network access, but the debian-installer has the same Kernel and therefore the debain-installer gets similar errors - I attach the output of the usual logs inside the debian-installer (switching to a console via alt+F2). As far as I can see there is nothing new about these logs.
Actually quite nasty is, that you now fail completely when you try to use the minimal install CD, as no network connection works during installation.

[Mika Pflüger (mikapflueger)]

I get the same bug for intrepid final. So I guess I will not update from 8.04.

[Mika Pflüger (mikapflueger)]

Same problem with lenny and a 2.6.26er kernel. Unfortunately I failed epic in reporting the bug to debians bts (reportbug to debian inside ubuntu does not work, chrooting into a debian does not work, I copied the conf-file of a debian reportbug over my ubuntu reportbug, reportbug works but fails to connect to my smtp-server (auth issue), I use blank as smtp-server as suggested, reportbug is happy - but no bug report shows up and I don't get the confirmation. I send the bug-report by hand via claws-mail and my smtp-server, same result. i don't like debian's bts) so that I can't provide a link to the bug report there. Interestingly, debian's debian-isntaller works, as it uses the 2.6.24er kernel - so a debian lenny can actually be installed, but not used afterwards.

[Adam Ehlers Nyholm Thomsen (adament)]

Just ran into what I believe is the exact same problem(It's the same pcmcia card and the same error) when trying to install Intrepid. Just found this fix http://groups.google.com/group/fa.linux.kernel/browse_thread/thread/cdeb23fce2818e04/e8f76bce2750eac2?lnk=gst&q=axnet_cs#e8f76bce2750eac2 from what I guess it might very well be related. I assume that this means it will be fixed for the 2.6.28 kernel. I will hopefully be compiling a new kernel tonight with the mentioned patch, I will report back with results.

[Mika Pflüger (mikapflueger)]

Hi,

thanks for the pointer to the thread. I think I won't compile my own
kernel anyway, but it would be nice, if this would go into the .28
kernel, because it is quite annoying - nowadays every single Linux
distro uses a .26 or .27 kernel and they all fail...

Let's hope this will be fixed.

[Steve Beattie (sbeattie)]

This bug was reported in the Intrepid development cycle; removing regression-potential and marking as regression-release.

[Mika Pflüger (mikapflueger)]

Like adament said, this patch went into the 2.6.28er kernel and is now in Jaunty Jackalope alpha 2. In Jaunty Jackalope Alpha 2 the card works like before, no problems at all.
Fortunately, the debian-installer uses a .28er kernel, too. This means, Jaunty Jackalope is fully usable with the Netgear FA 411.
However, the bug is still present (and I guess will not be solved) in Intrepid. Unless somebody backports a .28 kernel to Intrepid, Intrepid will remain unusable with machines with a Netgear FA 411.
As the upcoming Jaunty release will not be affected anymore and 8.10 is not an LTS release, I decrease priority.

[Mika Pflüger (mikapflueger)]

Sorry, I wanted to say: "...I would decrease priority."

[Leann Ogasawara (leannogasawara)]

Thanks for the udpate Mika. Per your comment, I'm marking this Fix Released for Jaunty. It'll also open an Intrepid nomination. I'm including the upstream git commit id for reference:

ogasawara@yoji:~/ubuntu-jaunty$ git log 208fbec5bec1de4fce48aab41efde11ba25ab04c
commit 208fbec5bec1de4fce48aab41efde11ba25ab04c
Author: Cord Walter <email address hidden>
Date: Thu Nov 20 13:46:57 2008 +0000

    axnet_cs / pcnet_cs: moving PCMCIA_DEVICE_PROD_ID for Netgear FA411

[Mika Pflüger (mikapflueger)]

Thanks for your efforts, Leann.

[Leann Ogasawara (leannogasawara)]

Seems this patch was committed to Intrepid as part of the stable kernel updates for 2.6.27-11.21 and should be available in the latest 2.6.27-14.33 kernel from intrepid-updates. See https://edge.launchpad.net/ubuntu/+source/linux . Marking this Fix Released.

ogasawara@yoji:~/ubuntu-intrepid$ git log -p 31b4568065605598ca82b47351fbe7a7c13bc21a
commit 31b4568065605598ca82b47351fbe7a7c13bc21a
Author: Cord Walter <email address hidden>
Date: Thu Nov 20 13:46:57 2008 +0000

    axnet_cs / pcnet_cs: moving PCMCIA_DEVICE_PROD_ID for Netgear FA411

    commit 208fbec5bec1de4fce48aab41efde11ba25ab04c upstream.