Correct.. But, as regards SSL session caching, the cache initialization is done only in the 2nd phase, and all the children are supposed to inherit the segment-id from that point onwards.. (somebody, pl. correct me if i'm wrong)
-Madhu -----Original Message----- From: William A. Rowe, Jr. [mailto:[EMAIL PROTECTED] Sent: Monday, January 07, 2002 4:33 PM To: MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1); 'Aaron Bannert'; [email protected] Subject: Re: [PROPOSAL] apr_shm_t, a new shared memory API to replace old From: "MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1)" <[EMAIL PROTECTED]> Sent: Monday, January 07, 2002 6:14 PM > Wow !!. This is good.. Does this mean that we'll have a *full* > implementation of the SHMEM atleast now ??.. I'm eagarly waiting for it > :-).. A full implementation of only shm ... if you want 'memory management', you pass the base pointer and size to rmm :) I noted one thing that perhaps was forgotten. We have two forms of attach(), one that you pass the key into, and the second (better called reattach()) that you will pass the apr_shm_t into, obtained from the parent, that the child must call after fork to have a true shm segment once again. If it wasn't called, I believe the child gets a private copy of the shm (which is no fun.) Unless I'm missing something obvious...
