Add tlsf_extend_pool function #18
Conversation
could work with sbrk to extend the pool size dynamically Change-Id: I4f2cda419f88a31bc478e74813ebcd0d1275617c
| } | ||
|
|
||
| pool_t tlsf_add_pool(tlsf_t tlsf, void* mem, size_t bytes) | ||
| pool_t tlsf_extend_pool(tlsf_t tlsf, void* mem, size_t bytes, size_t incr) |
There was a problem hiding this comment.
API is ambiguous. I'd prefer:
pool_t tlsf_extend_pool(tlsf_t tlsf, pool_t pool, size_t incr);
There was a problem hiding this comment.
Yes, I also think your prototype is better, but since the alignment may happen, it's impossible to recover the original size by finding the sentinal block. That's why the caller need pass the old size again.
There was a problem hiding this comment.
What about:
pool_t tlsf_resize_pool(tlsf_t tlsf, pool_t pool, size_t oldBytes, size_t newBytes)
There was a problem hiding this comment.
Yes, it's also my initial thinking, but it's very hard(if not possible) to reduce the running pool size since some memory may be using by user. That's why I:
- Name the new function tlsf_extend_pool
- Pass the increased bytes instead new size(newBytes may smaller than oldBytes)
There was a problem hiding this comment.
it's very hard(if not possible) to reduce the running pool size since some memory may be using by user.
Documentation will warn about dangerousness of reducing running pool size.
User anyway should know what he is doing.
There was a problem hiding this comment.
The function just try to extend the pool:
- The algo to reduce pool is totally different
- And the condition to call reduce is also different
so the function is better to indicate what it can do and what not to do explicitly.
| } | ||
|
|
||
| pool_t tlsf_add_pool(tlsf_t tlsf, void* mem, size_t bytes) | ||
| pool_t tlsf_extend_pool(tlsf_t tlsf, void* mem, size_t bytes, size_t incr) |
3 participants
could work with sbrk to extend the pool size dynamically