Re: [pcp] fetch/desc conflict defeats dynamic pmns, papi pmda case study

[Ken McDonell <kenj@xxxxxxxxxxxxxxxx>]

On 20/11/14 09:01, Ken McDonell wrote:
> > ...
>
> You could try the patch below ... the e_ext field of pmdaExt is already a 
> pointer to a private helper structure that holds the pmda_interface (version) 
> number that is needed for similar situations ... extending this to also have a 
> back pointer seems safe for everyone.
I've cleaned this patch up and replaced the pmda_interface field with a 
back pointer to the pmdaInterface struct from where the real (as opposed 
to a copy of the) pmda_interface value can be found.
This breaks nothing (all PMDAs been through QA) and allows Frank or any 
other PMDA builder to get access to the full pmdaInterface struct from 
within any of the callbacks.
Commit comming soon ...