[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Feral ISP driver: dma map request problem

The system is a XP-1000 

On Friday 09 January 2004 21:00, Kelledin wrote:
> On Friday 09 January 2004 11:06 am, Kelledin wrote:
> > Hmm.  It looks like the driver tried to do mapping as if the
> > machine had a 32-bit address space.  IIRC what it's supposed
> > to do is when it grabs some memory for PCI DMA mapping, it's
> > supposed to pass a flag signifying that the DMA region should
> > be accessible via 32-bit physical addressing.  Otherwise PCI
> > devices would likely have some trouble getting to it!
> >
> > Apparently that flag didn't get passed somewhere, and the
> > prospective DMA buffer got allocated way high in the 64-bit
> > address space.  So it couldn't be set up as a DMA buffer, and
> > the kernel's PCI DMA mapper spit it back.  By my
> > understanding, that's a fairly common mistake when porting
> > kernel drivers to 64-bit architectures.
> >
> > I'll take a look at the source code this afternoon and see
> > what I can spot.
> Apparently this type of DMA mapping doesn't work quite like I
> thought it did, or I'm missing something somewhere.
> Also, I'm compelled to wonder...what's your system type?  Some
> revisions of the EB164 series core-logic have some limitations
> as to available DMA methods.  You might be hitting one of those
> limitations here.

Thank you for your help. 
I'm not sure about the exact type of the system so I've attched parts of the 
boot message that might give a hint.
Its a XP-1000 Professional workstation, Model No E2F6W-V2 (which seems to 
indicate a 500 Mhz Prozessor )
We also have other XP1000 with Model No E2F6W-32 (also 500 Mhz ?) and 
E2G6W-CA.(667 Mhz ??) Do you think any of these systems would behave better ?

Could  you give me a some advice which alternatives would make sense if the 
controller issue can not be resolved ?
Could a separate SCSI Controller circumvent the problem and lead to a more 
stable system ?
Or could one try a large IDE drive ( we use the machine as SAMBA file server 
and need at least 70 GB drive space) since the XP1000 has ATAPI support ?
Thanks again


Here are some of the boot messages with respect to system and scsi system

Linux version 2.4.22acl (root@saphir) (gcc version 3.0.4) #1 Thu Nov 6 
17:31:08 CET 2003
Booting on Tsunami variation Monet using machine vector Monet from SRM
PCI: Bus 2, bridge: Digital Equipment Corporation DECchip 21152
  IO window: disabled.
  MEM window: disabled.
  PREFETCH window: disabled.
SMC37c669 Super I/O Controller found @ 0x3f0
---- with kernel driver qlogic
SCSI subsystem driver Revision: 1.00
qlogicisp : new isp1020 revision ID (6)
scsi0 : QLogic ISP1020 SCSI on PCI bus 01 device 30 irq 47 MEM base 
  Vendor: IBM       Model: IC35L036UWDY10-0  Rev: S23C
  Type:   Direct-Access                      ANSI SCSI revision: 03
  Vendor: IBM       Model: IC35L036UWDY10-0  Rev: S23C
  Type:   Direct-Access                      ANSI SCSI revision: 03
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
Attached scsi disk sdb at scsi0, channel 0, id 1, lun 0
---- with Feral ISP driver ---
SCSI subsystem driver Revision: 1.00
ISP SCSI and Fibre Channel Host Adapter Driver
      Linux Platform Version 2.1
      Common Core Code Version 2.7
      Built on Jan  2 2004, 14:08:49
isp0: Ultra Mode Capable
isp0: Board Type 1040C, Chip Revision 0x6, loaded F/W Revision 4.66.0
isp0: Last F/W revision was 5.57.1
scsi0 : Driver for a Qlogic ISP 1020/1040 Host Adapter

Dr. Ulrich Harttig
Protein Struktur Fabrik

Max-Planck-Institut fuer molekulare Genetik
Heubnerweg 6 D, 14059 Berlin, Germany
Tel    (+49 30) 32639-2814
WWW    www.proteinstrukturfabrik.de/~harttig

Reply to: