Till,
This is of course highly sensitive to the number of bytes in CAS's put
notify structure and also the db put notify structure. Looking closely
at this, it appears that the problem is present in R3.14 but not in
R3.13. No question that in the future we would like to avoid this
pitfall altogether.
To make the code more robust I placed the CA put notify structure on a
free list. This structure now includes a union of all DBF_XXX types
including DBF_STRING. If the data being written fits it is stored in the
union otherwise arrays are allocated once for each PV with calloc until
the client disconnects.
This will show up in the next release of R3.14.
Jeff
> -----Original Message-----
> From: Till Straumann [mailto:[email protected]]
> Sent: Friday, September 20, 2002 11:24 AM
> To: Tech Talk (EPICS)
> Subject: Another SPARC alignment bug in 3.14_beta1
>
> Hi there.
>
> You probably fixed this one already - RSRV on SPARC
> dies with SIGBUS on putNotify due to a misaligned buffer.
>
> Just to make sure, there's a patch attached (3.14_beta1)...
>
> -- Till
- References:
- Another SPARC alignment bug in 3.14_beta1 Till Straumann
- Navigate by Date:
- Prev:
Re: Knob boxes Glen Wright
- Next:
registerRecordDeviceDriver in iocsh Bill Nolan
- 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:
Another SPARC alignment bug in 3.14_beta1 Till Straumann
- Next:
Making CA calls from multi-threaded programs??? Leonard Reder
- 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
|