Re: cardbus troubles

On Sunday, 27 Mar 2005 15:08 +0200, Milan Obuch wrote:

> > The card is now recognized:
> > Mar 27 14:04:25 makar kernel: pcib2: device cardbus1 requested decoded
> > memory range 0xc0200000-0xcfffffff Mar 27 14:04:25 makar kernel: pcib2:
> > device cardbus1 requested decoded I/O range 0x4000-0x8fff Mar 27 14:04:25
> > makar kernel: found->   vendor=0x1415, dev=0x950b, revid=0x00 Mar 27
> > 14:04:25 makar kernel: bus=6, slot=0, func=0
> > Mar 27 14:04:25 makar kernel: class=07-00-06, hdrtype=0x00, mfdev=0
> > Mar 27 14:04:25 makar kernel: cmdreg=0x0000, statreg=0x0290, cachelnsz=0
> > (dwords) Mar 27 14:04:25 makar kernel: lattimer=0x00 (0 ns), mingnt=0x00 (0
> > ns), maxlat=0x00 (0 ns) Mar 27 14:04:25 makar kernel: intpin=a, irq=11
> > Mar 27 14:04:25 makar kernel: powerspec 1  supports D0 D2 D3  current D0
> > Mar 27 14:04:25 makar kernel: puc0: <Oxford Semiconductor OXCB950 UART>
> > port 0x4000-0x400f,0x4010-0x4017 mem
> > 0xc0240000-0xc0240fff,0xc0241000-0xc0241fff,0xc0242000-0xc0242fff irq 11 at
> > device 0.0 on cardbus1 Mar 27 14:04:25 makar kernel: pcib2: device puc0
> > requested decoded I/O range 0x4010-0x4017 Mar 27 14:04:25 makar kernel:
> > sio4: <Oxford Semiconductor OXCB950 UART> on puc0 Mar 27 14:04:25 makar
> > kernel: sio4: type 16550A
> > Mar 27 14:04:25 makar kernel: sio4: unable to activate interrupt in fast
> > mode - using normal mode
> >
> 
> This is not bad, maybe shared interrupt, normally does not cause problems.
> I see it on some machines too, so this is only diagnostics.
> 
> > I hooked up a Sun machine to the serial and connected with tip to
> > it (/etc/remote settings are correct, tested on a different
> > machine). I didn't see anythin in tip, but I got this errors:
> > Mar 27 14:22:07 makar kernel: sio4: 2 more silo overflows (total 8)
> > Mar 27 14:22:08 makar kernel: sio4: 1 more silo overflow (total 9)
> > Mar 27 14:22:08 makar kernel: sio4: 24 more interrupt-level buffer
> > overflows (total 481) Mar 27 14:22:09 makar kernel: sio4: 1 more silo
> > overflow (total 10) Mar 27 14:22:09 makar kernel: sio4: 91 more
> > interrupt-level buffer overflows (total 572) Mar 27 14:22:16 makar kernel:
> > sio4: 29 more interrupt-level buffer overflows (total 601) Mar 27 14:22:24
> > makar kernel: sio4: 1 more silo overflow (total 11) Mar 27 14:22:24 makar
> > kernel: sio4: 88 more interrupt-level buffer overflows (total 689) Mar 27
> > 14:22:37 makar kernel: sio4: 1 more silo overflow (total 12) Mar 27
> > 14:22:37 makar kernel: sio4: 21 more interrupt-level buffer overflows
> > (total 710) Mar 27 14:25:51 makar kernel: acpi_ec0: info: new max delay is
> > 70 us
> >
> 
> Overflows are usually caused by inability to read data, which corresponds with 
> what are you seeing - no data read.
> 
> > If I remove the card:
> > Mar 27 14:26:33 makar kernel: cardbus1: release_all_resource: Resource
> > still owned by child, oops. (type=1, rid=0, addr=b) Mar 27 14:26:33 makar
> > kernel: cardbus1: release_all_resource: Resource still owned by child,
> > oops. (type=4, rid=16, addr=4010) Mar 27 14:26:33 makar kernel: sio4: still
> > open, forcing close
> > Mar 27 14:26:33 makar kernel: sio4: detached
> >
> > If I reconnect the card, the kernel panics.
> >
> 
> Maybe card definition needs some tweaking, but I do not know enough to tell 
> exactly. Maybe it is in cardbus - no idea, I have not that much experience 
> here.
> As said, maybe someone knows better. I will try to read some docs, but no idea 
> just now.

Is any developer interested to work on that? I'm also willing to
provide the necessary hardware.


Ihsan...

-- 
Swiss Unix User Group: http://www.suug.ch/
Software Packages for Solaris: http://www.blastwave.org/