Hi Jeff,
You was right. I added _X86_ macrodefinition as compiler option
and rebuild the core. After this I got an access to my records from
UNIX via CA.
Below I print the content of CONFIG.Vx.sbs_pc6
which I used to build target iocCore:
------------------------------------------------------------
# Created by Korobov for SBS PC6
# CONFIG.Vx.sbs_pc6
#
# This file is maintained by the EPICS community.
# Vx GNU cross compiler suffix
CMPLR_SUFFIX = 386
# For Vx directories of form:
# $(VX_DIR)/$(HOST_ARCH).$(ARCH_CLASS)/bin
ARCH_CLASS = sbs_pc6
ARCH_DEP_CPPFLAGS = -DCPU=PENTIUM -DCPU_VARIANT=PENTIUM -D_X86_
ARCH_DEP_CFLAGS = -mpentium
ARCH_DEP_CXXFLAGS += -x 'c++'
ARCH_DEP_CFLAGS += -fno-builtin -fno-defer-pop
-----------------------------------------------------------
Thanks a lot,
V.Korobov
On Mon, 9 Sep 2002, Jeff Hill wrote:
>
> Vladis,
>
> Sounds like an big / little endian compile time switch problem.
>
> My best guess is that the following R3.13 preprocessor code
> in base/src/ca/net_convert.h is not selecting CA_FLOAT_IEEE
> and CA_FLOAT_IEEE - possibly because your C compiler isn't
> emitting one of _M_IX86 or _X86_ for your Pentium compatible
> processor. Please send your changes if this turns out to
> be the cause of your problem, and we will incorporate them
> into the next R3.13 release.
>
> This particular set of ifdefs
> was replaced by a portable, and hopefully compile time,
> endian test in R3.14. The default assumption in R3.14 is
> IEEE floating point, but this can be replaced by architecture
> specific host to network translation code in R3.14 if you are
> porting to an os that does not support it.
>
> The GNU gcc manuals include documentation on the various
> architecture specific built in preprocessor macros that
> may be defined by default by your gnu C compiler.
>
> /*
> * Here are the definitions for architecture dependent byte ordering
> * and floating point format
> */
> #if defined(VAX)
> # define CA_FLOAT_MIT
> # define CA_LITTLE_ENDIAN
> #elif defined(_M_IX86) || defined(_X86_)
> # define CA_FLOAT_IEEE
> # define CA_LITTLE_ENDIAN
> #elif ( defined (__ALPHA) || defined (__alpha) ) && ( defined (VMS) ||
> defined (__VMS) )
> # define CA_FLOAT_MIT
> # define CA_LITTLE_ENDIAN
> #elif (defined(__ALPHA) && defined(UNIX) || defined(__alpha)) &&
> defined(UNIX)
> # define CA_FLOAT_IEEE
> # define CA_LITTLE_ENDIAN
> #else
> # define CA_FLOAT_IEEE
> # define CA_BIG_ENDIAN
> #endif
>
> Jeff
>
> > -----Original Message-----
> > From: Vladis Korobov [mailto:[email protected]]
> > Sent: Monday, September 09, 2002 4:22 AM
> > To: [email protected]
> > Subject: "bad UDP msg ..." error message
> >
> > Hi all,
> >
> > I began to test SBS PC6 embedded computer. It has PENTIUM compatible
> CPU.
> > There is BSP for this target for Tornado II. I use R3.13.6 Epics
> release.
> > To build Epics I added file CONFIG.Vx.sbs_pc6 derived from
> > CONFIG.Vx.pc486
> > just by replacing appropriate compilation options (taken from BSP).
> >
> > This file looks like:
> >
> > # Created by Korobov for SBS PC6
> > # CONFIG.Vx.sbs_pc6
> > #
> > # This file is maintained by the EPICS community.
> >
> > # Vx GNU cross compiler suffix
> > CMPLR_SUFFIX = 386
> >
> > # For Vx directories of form:
> > # $(VX_DIR)/$(HOST_ARCH).$(ARCH_CLASS)/bin
> > ARCH_CLASS = sbs_pc6
> >
> > ARCH_DEP_CFLAGS = -DCPU=PENTIUM -mpentium -DCPU_VARIANT=PENTIUM
> > ARCH_DEP_CXXFLAGS += -x 'c++'
> > ARCH_DEP_CFLAGS += -fno-builtin -fno-defer-pop
> >
> > ----------------------------------------------------
> >
> > Epics core was built successfully. I created some test database and
> > downloaded it into the IOC. From IOC side it seems to work OK.
> > But when I try to access any records from outside after starting any
> > CA clients I get messages (on client side):
> >
> > ..../iocinf.c: bad UDP msg from epicsPC20.desy.de:5064
> >
> > What could be the cause of such behaviour.
> >
> > Thanks in advance for any help or ideas,
> >
> > V.Korobov
>
>
>
>
- References:
- RE: "bad UDP msg ..." error message Jeff Hill
- Navigate by Date:
- Prev:
Re: CA problem (Pedro Gigoux) Bernd Schoeneburg
- Next:
e2db Jim Thomas
- 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:
RE: "bad UDP msg ..." error message Jeff Hill
- Next:
CA problem Pedro Gigoux
- 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
|