[patch] Minor PAGG attach/detach semantic change for 2.6.11

[Kingsley Cheung <kingsley@xxxxxxxxxx>]

Hi Erik,

While testing the propagation of pagg_attach errors to fork() I
noticed that the detach callback is called again for the client
responsible for the error.  Perhaps you may have a different opinion
on this, but IMHO this is unnecessary as the client passing the
failure error up should have already cleaned up any data in its pagg
structure during its attach callback.

I've atttached a patch that applies (with some fuzz) correctly to
2.6.11 that avoids this unncessary call to the client's detach
callback.  Please consider applying.

Thanks, 
-- 
                Kingsley

pagg.patch
Description: Text document