Am 30.10.2014 10:58, schrieb Peter Maydell: > On 30 October 2014 09:50, Christian Borntraeger <borntrae...@de.ibm.com> > wrote: >> Yes, I will try to get some of this fixed in valgrind as well. This will >> take a little longer though because the code changes are bigger than just >> 1 line of code. Given that valgrind has around 1 release/year, this patch >> set is certainly a nice band-aid that is useful for todays development. > > I guess these patches also mean we're not going to get valgrind warnings > if we forget to initialize a necessary field in the struct.
Yes, thats correct. The alternative would be to not use an designated initializer and instead memset reserved and pad field. > but I suppose we can live with that. > > -- PMM >