Use base allocator in critnib

[ldorau]

Requires:

• Add linear base allocator #177

[bratpiorka]

others - please review as I'm on vacation

[KFilipek]

Generally okay, but if the function umf_ba_linear_destroy needs an explanation every time that it also destroys the parent (the provider), it should have another name. If it's not the topic for this PR, please do it separately, I would be grateful.

[ldorau]

Generally okay, but if the function umf_ba_linear_destroy needs an explanation every time that it also destroys the parent (the provider), it should have another name. If it's not the topic for this PR, please do it separately, I would be grateful.

@KFilipek

1. It does NOT need an explanation every time - it is only my good will that I added those comments ;-) - I can remove them all.
2. It does NOT destroy any parent. umf_ba_linear_destroy() only destroys a mutex and destroys a pool itself (unmaps a range of memory) - so what would be a better name? - see its implementation (ba_os_free() is a wrapper for munmap()):

void umf_ba_linear_destroy(umf_ba_linear_pool_t *pool) {
    util_mutex_destroy_not_free(&pool->metadata.lock);
    ba_os_free(pool, pool->metadata.pool_size);
}

All issues fixed

[KFilipek]

Generally okay, but if the function umf_ba_linear_destroy needs an explanation every time that it also destroys the parent (the provider), it should have another name. If it's not the topic for this PR, please do it separately, I would be grateful.

@KFilipek

1. It does NOT need an explanation every time - it is only my good will that I added those comments ;-) - I can remove them all.
2. It does NOT destroy any parent. umf_ba_linear_destroy() only destroys a mutex and destroys a pool itself (unmaps a range of memory) - so what would be a better name? - see its implementation (ba_os_free() is a wrapper for munmap()):

void umf_ba_linear_destroy(umf_ba_linear_pool_t *pool) {
    util_mutex_destroy_not_free(&pool->metadata.lock);
    ba_os_free(pool, pool->metadata.pool_size);
}

Currently, I'm not sure what I've seen previously, but you're right about your changes. Sorry for that.

[ldorau]

@igchor Can we merge this PR?