add NOSHARE flag #336
add NOSHARE flag #336
Conversation
lunatik.h
Outdated
| #define LUNATIK_VERSION "Lunatik 3.7" | ||
|
|
||
| #define LUNATIK_CLASS_SLEEPABLE 0x01 | ||
| #define LUNATIK_CLASS_NOSHARE 0x02 |
There was a problem hiding this comment.
how do you handle allocations? we could have "single" mode objects that might or might not sleep on allocation, besides the fact they aren't shared. Perhaps, a better idea is to have two bools instead of flags, sleep and shared. What do you think?
Physic69
force-pushed
the
lunatikclass
branch
from
3fff47b to
46a783a
Compare
| #define lunatik_object_issleepable(obj) ((obj)->sleep) | ||
| #define lunatik_class_issleepable(cls) ((cls)->sleep) |
There was a problem hiding this comment.
| #define lunatik_object_issleepable(obj) ((obj)->sleep) | |
| #define lunatik_class_issleepable(cls) ((cls)->sleep) | |
| #define lunatik_issleepable(o) ((o)->sleep) |
There was a problem hiding this comment.
btw, I wouldn't change this on the same PR.. it seems only cosmetic and not require to the main change your are adding.. let's keep the patches as self-contained as possible..
| const luaL_Reg *methods; | ||
| void (*release)(void *); | ||
| bool sleep; | ||
| bool shared; |
There was a problem hiding this comment.
I think we should invert this flag.. so we don't need to change all modules and just the places we really need (such as pointer)..
| bool shared; | |
| bool single; |
There was a problem hiding this comment.
actually.. I wouldn't add this to class now, we don't need classes that are exclusively "single", right? I would just have it on object
| if (!object->shared) | ||
| luaL_error(L, "%s objects cannot be shared across runtimes", object->class->name); |
There was a problem hiding this comment.
it's not enough, right? we probably need to add single flag to the API for creating new objects..
There was a problem hiding this comment.
also, avoiding creating mutex/spinlock and refcount unnecessarily..
| union { | ||
| struct mutex mutex; | ||
| spinlock_t spin; | ||
| }; |
There was a problem hiding this comment.
maybe we should have a substruct allocated dynamically for kref and lock.. but let's worry about this after having the first version settled..
| #define LUADATA_OPT_NONE 0x00 | ||
| #define LUADATA_OPT_READONLY 0x01 | ||
| #define LUADATA_OPT_FREE 0x02 | ||
| #define LUADATA_OPT_SINGLE 0x04 |
There was a problem hiding this comment.
you can just use the obj flag, right?
| static int luadata_lnew(lua_State *L) | ||
| { | ||
| size_t size = (size_t)luaL_checkinteger(L, 1); | ||
| bool single = lua_toboolean(L, 2); /* false -> sharable */ |
There was a problem hiding this comment.
| bool single = lua_toboolean(L, 2); /* false -> sharable */ | |
| bool single = lua_toboolean(L, 2); |
| #define lunatik_class_issleepable(cls) ((cls)->sleep) | ||
|
|
||
| #define lunatik_locker(o, mutex_op, spin_op) \ | ||
| do { \ |
There was a problem hiding this comment.
it should check if it's single, right? it should lock in that case.. also, refcount shouldn't be applied if single
2 participants
No description provided.