Server reboots when silo overflows - FreeBSD 4.11

[Mike Tancsa]

At 11:42 AM 10/02/2005, Lefteris Tsintjelis wrote:
>Hi,
>
>There seems to be a serious problem with the puc device driver. When the 
>driver
>expiriences overflows the server simply reboots!
>
>device          puc
>options         PUC_FASTINTR
>
>I have tried to rebuild a new kernel without the PUC_FASTINTR option but 
>without
>much luck. It still keeps on rebooting right after a few silo overflow errors.
>puc0: <VScom PCI-200L> port 
>0xb000-0xb007,0xb400-0xb407,0xb800-0xb807,0xd000-0xd007 mem 
>0xea800000-0xea80003f irq 14 at device 12.0 on pci2
>sio4: configured irq 14 not in bitmap of probed irqs 0x8000
>sio4: type 16550A
>sio5: configured irq 14 not in bitmap of probed irqs 0x8000
>sio5: type 16550A



How do you have device sio configured in your kernel ?  Are you trying to 
specify the ports for the puc attached serial ports ?  All you need is

device          sio
device          puc

Also, enable crash dumps and build a debug kernel.  Add

makeoptions    DEBUG=-g
to your kernel config

and at bootup time add

dumpdev="/dev/da0s1b"           # Device name to crashdump to (or NO).
dumpdir="/var/crash"            # Directory where crash dumps are to be stored

Assuming your swap is on da0s1b

         ---Mike