:This is a long post.  Kept it on Users (where it started), and because 
:I'm hoping others here might be able to tell us whether or not their 
:systems use the same Texas Instruments OHCI Compliant IEEE 1394 Host 
:Controller I have in my system (vendor=104c, dev=803a), and if they are 
:also having issues with it.
:...
:... this thread is a hit on the VAIO NR notebook side of the error
:... and the firewire thread from November 2004 on the kernel list 
:(between Matt & Gabor MICSKO) hits the details.
:       [[ 
:http://leaf.dragonflybsd.org/mailarchive/kernel/2004-11/msg00037.html 
:]]
:
:       fwohci0: EUI64 08:00:46:03:02:83:de:3e
:       NMI ISA a1, EISA ff
:       RAM parity error, likely hardware failure
:       
:       Fatal trap 19: non-maskable interrupt trap while in kernel mode
:       instruction pointer     = 0x8:0xc020e7be
:       stack pointer           = 0x10:0xc07789f8
:       frame pointer           = 0x10:0xc0778a14
:       code segment            = base 0x0, limit 0xfffff, type 0x1b
:                                               = DPL 0, pres 1, def32 1, gran 1
:       processor eflags                = interupt enabled, IOPL = 0
:       current process         = 0 (swapper)
:       current thread          - pri 12
:       
:       kernel: type 19 trap, code=0
:       Stopped at        fwphy_rddata+0x96:       testw   %bx, %bx
:...
:Any thoughts, or similar experiences?
:
:--David

    This looks like a case of the driver accessing on-device memory
    which has not yet been initialized.  That is, the parity error is
    not due to your RAM, but a PCI bus parity error due to bug in the
    device initialization sequence.

    Hmm.  One thing I see right off the bat is that the FreeBSD driver
    does have an extra DELAY(500) in the probe code.  Try this patch.

                                                -Matt

Index: fwohci.c
===================================================================
RCS file: /cvs/src/sys/bus/firewire/fwohci.c,v
retrieving revision 1.17
diff -u -p -r1.17 fwohci.c
--- fwohci.c    20 Apr 2008 13:44:25 -0000      1.17
+++ fwohci.c    17 May 2008 17:18:08 -0000
@@ -426,6 +426,8 @@  *    number of port supported by core-l
  *    It is not actually available port on your PC .
  */
        OWRITE(sc, OHCI_HCCCTL, OHCI_HCC_LPS);
+       DELAY(500);
+
        reg = fwphy_rddata(sc, FW_PHY_SPD_REG);
 
        if((reg >> 5) != 7 ){

Reply via email to