Next-in-Thread Next Message

None Sounds like an interrupt problem; need more messages 

Forum: PCMCIA Wireless Network Adapter Issues
Re: Question Problem with Wavelan Turbo 11Mbps (Alex Lam)
Re: More More info... (Alex Lam)
Date: 2000, Jul 31
From: David Hinds <dhinds@pcmcia.sourceforge.org>

The messages from the pcmcia_core and i82365 modules are important.

-- Dave

Next-in-Thread Next Message

Messages Inline: 0 1

More Very similar problem here

Re: None Sounds like an interrupt problem; need more messages (David Hinds)
Date: 2000, Aug 02
From: Elliot Lee sopwith

I'm having a very similar problem.

Kernel: 2.2.16-17 (the one that comes with Red Hat pinstripe beta), with associated pcmcia-cs 3.1.19.

Hardware: ActionTec PC-750 PCI <-> PCMCIA adapter thing in a generic Pentium II on one end, a Casio Fiva MPC-501 on the other end.

In both machines, Lucent WaveLAN/IEEE (turbo 11Mb Gold). I haven't any idea what the firmware version is.

Symptoms: 
Everything seems to load fine, ifconfig & iwconfig work fine, but no data gets through.

Matt Wilson (a coworker) had a known-to-work-in-other-situations card (silver version of the above) in his laptop. He got things to the point where pinging the wireless broadcast IP address would cause the Rx/Tx light to flash on the desktop machine, but no interrupts were coming through on the desktop.

[root@test152 wireless]# grep wvlan /proc/interrupts 
  4:          4          XT-PIC  wvlan_cs

When trying to send pings from the desktop machine, a bunch of
'wvlan0 Tx timed out! Resetting card' msgs would be generated. Matt tried a bunch of different IRQ's on the desktop box to see if changing that would help - no joy.

Communication from Matt's laptop to the Fiva did not work either, although the 'wvlan0 Tx timed out! Resetting card' message did not show up on the Fiva when pinging from it.

(The Fiva also crashes hard if the WaveLAN card is removed/inserted while pcmcia service is running, dunno if this is related, but would be happy just to get the WaveLAN working soon. :-)

Syslog info on wvlan_cs load:
Aug  2 20:12:41 test152 kernel: wvlan_cs: WaveLAN/IEEE PCMCIA driver v1.0.4
Aug  2 20:12:41 test152 kernel: wvlan_cs: (c) Andreas Neuhaus <andy@fasta.fh-dortmund.de>
Aug  2 20:12:41 test152 kernel: wvlan_cs: index 0x01: Vcc 5.0, irq 4, io 0x0100-0x013f
Aug  2 20:12:41 test152 kernel: wvlan_cs: Registered netdevice wvlan0

ActionTec PC-750 'lspci -v' output on the desktop box:
00:0b.0 CardBus bridge: Texas Instruments PCI1225 (rev 01)
	Subsystem: Action Tec Electronics Inc: Unknown device 0293
	Flags: bus master, medium devsel, latency 32, IRQ 10
	Memory at <ignored> (32-bit, non-prefetchable)
	Bus: primary=00, secondary=20, subordinate=22, sec-latency=176
	I/O window 0: 00000000-00000003
	I/O window 1: 00000000-00000003
	16-bit legacy interface ports at 0001

00:0b.1 CardBus bridge: Texas Instruments PCI1225 (rev 01)
	Subsystem: Action Tec Electronics Inc: Unknown device 0293
	Flags: bus master, medium devsel, latency 32, IRQ 9
	Memory at <ignored> (32-bit, non-prefetchable)
	Bus: primary=00, secondary=23, subordinate=25, sec-latency=176
	I/O window 0: 00000000-00000003
	I/O window 1: 00000000-00000003
	16-bit legacy interface ports at 0001

None Your case is much easier

Re: More Very similar problem here (Elliot Lee)
Date: 2000, Aug 03
From: David Hinds <dhinds@pcmcia.sourceforge.org>

Your lspci output shows a PCI interrupt for the CardBus bridge.  You
just need to specify:

  PCIC_OPTS="irq_mode=0"

to force the i82365 module to use only PCI interrupts.

-- Dave

Sad Nope.

Re: None Your case is much easier (David Hinds)
Date: 2000, Aug 03
From: Elliot Lee sopwith

No go (I had already tried this while looking at back msgs).

When starting up with irq_mode=0:

Aug  2 21:27:14 test152 kernel: PCI routing table version 1.0 at 0xf0a90
Aug  2 21:27:14 test152 kernel: Intel PCIC probe: 
Aug  2 21:27:14 test152 kernel:   TI 1225 rev 01 PCI-to-CardBus at slot 00:0b, mem 0x68000000
Aug  2 21:27:14 test152 kernel:     host opts [0]: [pci only] [pci irq 10] [lat 32/176] [bus 32/34]
Aug  2 21:27:14 test152 kernel:     host opts [1]: [pci only] [pci irq 9] [lat 32/176] [bus 35/37]
Aug  2 21:27:14 test152 kernel:     PCI irq 10 test failed
Aug  2 21:27:14 test152 kernel:     *NO* card interrupts, polling interval = 1000 ms
Aug  2 21:27:14 test152 cardmgr[747]: starting, version is 3.1.19
Aug  2 21:27:14 test152 cardmgr[747]: watching 2 sockets
Aug  2 21:27:14 test152 kernel: cs: IO port probe 0x0c00-0x0cff: excluding 0xcf8-0xcff
Aug  2 21:27:14 test152 kernel: cs: IO port probe 0x0800-0x08ff: clean.
Aug  2 21:27:14 test152 kernel: cs: IO port probe 0x0200-0x04ff: excluding 0x200-0x207 0x220-0x22f 0x290-0x297 0x330-0x337 0x388-0x38f 0x4d0-0x4d7
Aug  2 21:27:15 test152 cardmgr[747]: initializing socket 0
Aug  2 21:27:15 test152 kernel: cs: memory probe 0xa0000000-0xa0ffffff: clean.
Aug  2 21:27:15 test152 cardmgr[747]: socket 0: Lucent Technologies WaveLAN/IEEE Adapter
Aug  2 21:27:15 test152 cardmgr[747]:   product info: "Lucent Technologies", "WaveLAN/IEEE", "Version 01.01", ""
Aug  2 21:27:15 test152 cardmgr[747]:   manfid: 0x0156, 0x0002  function: 6 (network)
Aug  2 21:27:15 test152 cardmgr[747]: executing: 'modprobe wvlan_cs port_type=3 channel=1'
Aug  2 21:27:15 test152 kernel: wvlan_cs: WaveLAN/IEEE PCMCIA driver v1.0.4
Aug  2 21:27:15 test152 kernel: wvlan_cs: (c) Andreas Neuhaus <andy@fasta.fh-dortmund.de>
Aug  2 21:27:15 test152 kernel: wvlan_cs: RequestIRQ: Resource in use

Dunno if it is relevant, but once I have started & stopped pcmcia with irq_mode=0, this error continues to happen subsequent times when I start pcmcia without rebooting.

None Re: Sad: Nope.

Re: Sad Nope. (Elliot Lee)
Date: 2000, Aug 03
From: David Hinds <dhinds@pcmcia.sourceforge.org>

Hmmm.  Try each of these combinations:

  PCIC_OPTS="irq_mode=3 irq_list=0"

  PCIC_OPTS="irq_mode=0 p2cclk=1"

  PCIC_OPTS="irq_mode=3 irq_list=0 p2cclk=1"

-- Dave

News Yay! Works

Re: None Re: Sad: Nope. (David Hinds)
Date: 2000, Aug 03
From: Elliot Lee sopwith

The first one on that list (irq_mode=3 irq_list=0) worked for me, thanks much.

I want to say something about it being nice to do this stuff automatically, but I'm afraid I'd just be showing utter ignorance of the complexities of the hardware involved. :)
Sounds like an interrupt problem; need more messages


Add Message to: "Sounds like an interrupt problem; need more messages"

Members Subscribe Admin Mode Show Frames Help for HyperNews at pcmcia-cs.sourceforge.net 1.10