Nick Patavalis wrote:
> The crash happens durring the conversion from net to host
> format. Specifically when the "cac_dbr_cvrt[]" array is de-referenced
> and the outcome (function ptr) executed. When the crash happens the
> code reaches there with:
> piiu->curMsg.m_type == 37
> Since there are but 34 function pointers in the "cac_dbr_cvrt[]"
> array (as initialized in "convert.c"), crashing is unavoidable!
As Mark Rivers mentioned, there was a problem in R3.13 occurring
when all of the following were true (fixed in R3.13.beta5).
o IEEE native floating point format
o double precision operand
o little endian host
o source and destination address for conversion are identical
However, on the surface this does not appear to be the same
problem because:
o The data item that is incorrect is a 16 bit unsigned integer
o Many other programs are working {medm, probe, ...} on this
platform (and they use data type codes).
Possible causes of out of range data type code:
o some code is writing on the CA client's internal data structures
o CA protocol is being read missaligned. We have not seen this.
Causes we dont expect:
o ALH using the wrong data type code (this is checked by
the client lib and also by the server)
I dont have much to go on. Perhaps the floating point problem
is causing corruption to occur (seems unlikely).
We will need to reproduce the problem. Since I dont
have a lynux system set up here perhaps you can provide
a remote login (and access to -g compiled object code).
Do the symptoms (location of the crash) change if you
rearrange the configuration for ALH. Has anyone else used ALH
under lynux before?
> I just looked at the APS, WWW server, and the most recent release is
> R3.13.0beta6. I supose this won't do the trick....
>
beta6 has the fix for the double precision/ieee/liddle endian
problem.
> I also wonder if there is any documention on channel access other than
> the the... elementary "channel access reference", which only contains a
> short description of the function calls.
A CA tutorial that is less ... elementary is in the works :)
> Is there a Channel Access
> _Protocol_ reference? If so where I can look for it?
No such critter exists. Depending on funding, a simple protocol
document may be a product of the design cycle for CA V4.
Jeff
--
______________________________________________________________________
Jeffrey O. Hill Internet [email protected]
LANL MS H820 Voice 505 665 1831
Los Alamos, NM 87545 USA FAX 505 665 5107
- Replies:
- Re: Problems with Alh & Channel Access (R3.13.0b4). Nick Patavalis
- References:
- Problems with Alh & Channel Access (R3.13.0b4). Nick Patavalis
- Navigate by Date:
- Prev:
Solaris 2.5.1 include problem Steve Lewis
- Next:
Problems with Alh & Channel Access (R3.13.0b4). Nick Patavalis
- Index:
1994
1995
1996
<1997>
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
Problems with Alh & Channel Access (R3.13.0b4). Nick Patavalis
- Next:
Re: Problems with Alh & Channel Access (R3.13.0b4). Nick Patavalis
- Index:
1994
1995
1996
<1997>
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
|