xircom card bus with 2.4.20 link trouble

Hello,

Since I run the 2.4.20 kernel (+bootsplash patch), (I didn't try without this patch but I think it is not the problem), my Xircom Cardbus don't work any more! In fact, I put the card into the pcmcia slot, the pc detect it, load the correct module (xirc2ps_cb), start eth1 but the led link (both side) blinks.
I have try with a hub 10, a ethernet card 10 and an ethernet card 10/100. I have so try 2 cables... but same problem.
I have try to impose the ethernet type of link mii-tool -A10BaseT-HD and -F10BaseT-HD.

Thanks

Arnaud

extract of /var/messages (10mbps):
------

Jun 20 08:34:49 portable kernel: cs: memory probe 0xa0000000-0xa0ffffff:
clean.
Jun 20 08:34:49 portable kernel: xirc2ps_cs.c 1.31 1998/12/09 19:32:55
(dd9jn+kvh)
Jun 20 08:34:53 portable kernel: eth1: autonegotiation failed; using
10mbs
Jun 20 08:34:53 portable kernel: eth1: MII selected
Jun 20 08:34:53 portable kernel: eth1: media 10BaseT, silicon revision 0
Jun 20 08:34:53 portable kernel: eth1: Xircom: port 0x300, irq 3, hwaddr
00:80:C7:CC:1F:A
2
Jun 20 08:34:54 portable kernel: eth1: autonegotiation failed; using
10mbs
Jun 20 08:34:54 portable kernel: eth1: MII selected
Jun 20 08:34:54 portable kernel: eth1: media 10BaseT, silicon revision 0
Jun 20 08:34:59 portable kernel: eth1: autonegotiation failed; using
10mbs
Jun 20 08:34:59 portable kernel: eth1: MII selected

....
-----

short extract of /var/messages (100mbps):
-----

Jun 20 10:14:52 portable kernel: eth1: MII link partner: 05e1
Jun 20 10:14:52 portable kernel: eth1: MII selected
Jun 20 10:14:52 portable kernel: eth1: media 100BaseT, silicon revision

5
-----

----

Arnaud Ligot <[email protected]>

/me auto slaps

with the 2.4.20, it work but it don't work with 2.4.21. Sorry

Arnaud Ligot <[email protected]>

against me ;-)

with a lot of tries... there are is a """solution"""
- put the card
- the led link blink
- you MUST define an IP with ifconfig
- the led link stop blinking !!!

is that a bug ?

thanks

Arnaud

Arnaud Ligot <[email protected]>

Mine doesn't work either, I have a RBEM56G-100.

Bought this off ebay for quite cheap to replace my REM56G-100.

It detects it, etc etc, but never actually works.

I put the REM56G-100 back in, insmod'd the driver, and back in business.

I haven't tried a 2.5.x kernel, nor a 2.4.21(it's on my list of things to do)

I was using the xircom_cb driver though. The xircom_tulip_cb didn't work either.

Mike

I have a CEM56-100

the card works with 2.4.20 nice and works if you set an ip address with the 2.4.21

I use the xirc2ps_cs module. I haven't force cardmgr to use so I think it's the good module for the card

Arnaud Ligot <[email protected]>

Going to move to 2.4.21 in a few minutes(just booting up the laptop so i can copy the vmlinuz image over to it)

If it's a CEM56-100, that's not a cardbus card!

That's the regular pcmcia version. That would explain why your xirc2ps_cs is working. You have the same card as I do, with a differing model name. Never did figure out what's different between the REM and CEM.

Mike

a day my father give me this card and said me it's a cardbus card... I accept and since this day, I think it's a cardbus... but if you sais it's a regular pcmcia... why not ;-)

++
Arnaud

Arnaud Ligot <[email protected]>

7m33s+23m29s+6m58s+35s minutes and seconds later, I have a working 2.4.21

PCMCIA Card came up just fine, cardbus continues to refuse to work.

Now if only DMA mode would work on the Opti 621 chipset inside this Omnibook 5500/5700 hybrid laptop :) Then it wouldn't take 38 minutes to compile. I knew I should have done this on the p4/2.53 that sits right next to it!

But anyways, that's the way it goes

Mike

It should do - drivers/ide/pci/opti621.c

Indeed, I have it compiled in even.

Let me see if hte laptop is turned on currently.

Unfortunately, no. But anyways, it does the classic error of {your drive} {shouldn't do this}

and then disables DMA mode, and goes back to working.

It's an HP Omnibook 5700, and it's a Toshiba 1.3 gig drive in there. I don't remember if my Omnibook 5500 did it, and it's a pain to take the donor screen back off the 5700 and putting it on the 5500 to try.

I had to disable "automatically enable DMA" for the laptop not to jam up for a minute while it figures out the DMA mode isn't going to work.

It doesn't work in windows either for that matter, if I remember right the option is completely greyed out. Blacklist on windows?

I will check the exact lspci and hdparm -i when I get to the office tomorrow.

Mike

# hdparm -i /dev/hda

/dev/hda:

Model=TOSHIBA MK1301MAV, FwRev=B0.03 A, SerialNo=86O37347 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% } RawCHS=2633/16/63, TrkSize=40257, SectSize=639, ECCbytes=21 BuffType=DualPortCache, BuffSize=128kB, MaxMultSect=16, MultSect=16 CurCHS=2633/16/63, CurSects=2654064, LBA=yes, LBAsects=2654280 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: sdma0 sdma1 sdma2 mdma0 mdma1 *mdma2 AdvancedPM=no
Drive Supports : Reserved : ATA-1 ATA-2 ATA-3

# lspci -vvv -s 00:14.0

00:14.0 IDE interface: OPTi Inc. 82C621 (rev 12) (prog-if 80 [Master])
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR- FastB2B-
Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 0
Region 4: I/O ports at 3000 [size=16]

and when i hdparm -d 1 /dev/hda

blk: queue c0273700, I/O limit 4095Mb (mask 0xffffffff)
hda: dma_timer_expiry: dma status == 0x21
hda: timeout waiting for DMA
hda: timeout waiting for DMA
hda: (__ide_dma_test_irq) called while not waiting
hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }

It doesn't work with sdma0 or mdma0 either.

/dev/hda:
Timing buffer-cache reads: 128 MB in 4.38 seconds = 29.22 MB/sec
Timing buffered disk reads: 64 MB in 20.09 seconds = 3.19 MB/sec

No speed demon there!

Mike