mirror of
https://github.com/fergalmoran/ladybird.git
synced 2025-12-30 13:19:47 +00:00
Kernel: Implement kmalloc_good_size for the new kmalloc
This lets kmalloc-aware data structures like Vector and HashTable use up the extra wasted space we allocate in the slab heaps & heap chunks.
This commit is contained in:
@@ -22,6 +22,7 @@ static constexpr size_t CHUNK_SIZE = 32;
|
||||
#else
|
||||
static constexpr size_t CHUNK_SIZE = 64;
|
||||
#endif
|
||||
static_assert(is_power_of_two(CHUNK_SIZE));
|
||||
|
||||
static constexpr size_t INITIAL_KMALLOC_MEMORY_SIZE = 2 * MiB;
|
||||
|
||||
@@ -435,7 +436,13 @@ void kfree_sized(void* ptr, size_t size)
|
||||
|
||||
size_t kmalloc_good_size(size_t size)
|
||||
{
|
||||
return size;
|
||||
VERIFY(size > 0);
|
||||
// NOTE: There's no need to take the kmalloc lock, as the kmalloc slab-heaps (and their sizes) are constant
|
||||
for (auto const& slabheap : g_kmalloc_global->slabheaps) {
|
||||
if (size <= slabheap.slab_size())
|
||||
return slabheap.slab_size();
|
||||
}
|
||||
return round_up_to_power_of_two(size + Heap<CHUNK_SIZE>::AllocationHeaderSize, CHUNK_SIZE) - Heap<CHUNK_SIZE>::AllocationHeaderSize;
|
||||
}
|
||||
|
||||
void* kmalloc_aligned(size_t size, size_t alignment)
|
||||
|
||||
Reference in New Issue
Block a user