From f59533ee9807832ea5ca7dd5492592c8991a9f34 Mon Sep 17 00:00:00 2001 From: Rob Austein Date: Sun, 28 May 2017 12:11:31 -0400 Subject: Further keystore cleanup and consolidation. Still not yet expected to compile, much less run, but getting closer. --- hal_internal.h | 20 ++------------------ 1 file changed, 2 insertions(+), 18 deletions(-) (limited to 'hal_internal.h') diff --git a/hal_internal.h b/hal_internal.h index 89dfbbb..667c5a4 100644 --- a/hal_internal.h +++ b/hal_internal.h @@ -479,24 +479,8 @@ typedef struct { typedef struct hal_ks hal_ks_t; -#error Minor re-think needed on keystore init - -// I like current hal_ks_*_init() setup method, but how does anything -// get a handle on the keystore? Make the keystores global variables? -// Right now they're hidden in larger, driver-specific structures, but -// it would be easy enough to break them out. Have already forgotten -// how the old API handled this, except that it involved an init -// method via the driver. Init is going to be special in any case, -// since we can't dispatch through a driver pointer in the keystore -// object, so either we expose the keystore and the driver or we -// expose the keystore and the init function. The latter may be -// simpler. -// -// Another variation would be to keep the current nesting, add global -// pointer variables for the keystores, and have the init functions -// set the pointers. Only real advantage there is that it would give -// us an easy test for whether the keystore had been initialized...but -// we already have several of those, not clear what value another adds. +extern hal_ks_t * const hal_ks_token; +extern hal_ks_t * const hal_ks_volatile; /* * RPC lowest-level send and receive routines. These are blocking, and -- cgit v1.2.3