pcmcia subsystem completely broken

[Iain Barker]

Daniel Ritz wrote:

>there is something missing in the dmesg output of your 2.6.16 kernel:
>	Yenta TI: socket 0001:10:13.0, mfunc 0x00001002, devctl 0x60
>this means your kernel is compiled w/o TI initialization code. 

Hi Daniel,

If you look further down the mail, you will see that I load yenta_socket as a module, and that does run the TI initialisation code:

	# modprobe yenta_socket
	Linux Kernel Card Services
	  options:  [pci] [cardbus]
	PCI: enabling device 0000:00:00.0 (0146 -> 0147)
	Yenta: CardBus bridge found at 0000:00:00.0 [0000:0000]
	yenta 0000:00:00.0: Preassigned resource 0 busy, reconfiguring...
	yenta 0000:00:00.0: Preassigned resource 1 busy, reconfiguring...
	Yenta: Using CSCINT to route CSC interrupts to PCI
	Yenta: Routing CardBus interrupts to PCI
	Yenta TI: socket 0000:00:00.0, mfunc 0x00001000, devctl 0x66
	Yenta TI: socket 0000:00:00.0 probing PCI interrupt failed, trying to fix
	Yenta TI: socket 0000:00:00.0 falling back to parallel PCI interrupts
	Yenta TI: socket 0000:00:00.0 parallel PCI interrupts ok
	Yenta: ISA IRQ mask 0x0000, PCI irq 39
	Socket status: 30000047
	#

> what does a manual "pccardctl insert" do?

As this system is 2.6.11 kernel, it uses the pcmcia-cs package not pcmciautils.
The corresponding command would be "cardcrl insert" I believe.

	# cardctl insert
	DEV: registering device: ID = '0.0'
	cardmgr[246]: socket 0: ATA/IDE bus pcmcia: add device 0.0
	Fixed Disk
	bound device '0.0' to driver 'ide-cs'
	0.0: GetNextTuple: No more items
	cardmgr[246]: get dev info on socket 0 failed: No such device


So, the same problem!

GetNextTuple is attempting to scan the available IO resources specified in the config.opts file, but is not able to find a match.

My assumption is that this fails because the resources specified in that file are all 32-bit (i.e. 0x20000000 and above).


thanks,
	Iain