[LinuxPPS] New test release

[Hal V. Engel]

On Monday 07 December 2009 02:37:27 pm clemens at dwf.com wrote:
> Well, I now have 2.6.32 + ntp-pps-v2.6.32-rc8.diff running on both a 32bit
>  and a 64bit platform.  There don't seem to be any problems, so I have no
>  idea what is going on with Hal's installation.
> 
> ---
> 
> With Hal's system:  is something connected via a Modem?  I dont understand
>  the Modem Status line.  and I have no idea what could cause a buffer
>  overflow. Seems it MIGHT be coming from refclock_ioctl, since that's the
>  last routine mentioned, but I dont see the string 'buffer overflow
>  detected' anywhere in the source.  If you run it again, does it die the
>  same way?
> 

There is no modem but I do have other devices attached to various serial 
ports.  These include a reflective spectrophotometer and a color meter (I work 
on systems that involve color science).   I currently have the kernel 
configured for 5 ports since I have one port on the motherboard and an add in 
card with 4 ports.

# setserial -g /dev/ttyS*
/dev/ttyS0, UART: 16550A, Port: 0xe880, IRQ: 19
/dev/ttyS1, UART: 16550A, Port: 0x02f8, IRQ: 3, Flags: low_latency
/dev/ttyS2, UART: 16550A, Port: 0xe800, IRQ: 19
/dev/ttyS3, UART: 16550A, Port: 0xe480, IRQ: 19
/dev/ttyS4, UART: 16550A, Port: 0xec00, IRQ: 19

The Oncore is on /dev/ttyS1 which is the motherboard port.  I just ran 
some verification code against the spectrophotometer and the color meter and 
they appear to be working fine (the software found them and verified that they 
are the devices I expected).  I also ran WinOncore12 to see if it would find 
and communicate with the GPS and it had no problems.  So the ports appear to 
be OK and the GPS also appears to be OK.

If I rerun it I get the same result.

I also looked over the code where it appears to be blowing up and I don't see 
how there could be a buffer overflow issue in refclock_ioctl().

I am running 2.6.32 + ntp-pps-v2.6.32-rc8.dif on an x86_64 system.  The patch 
set applied cleanly with no messages.  The patched kernel also builds without 
any issues.

Hal