diff --git a/src/arch/x86_64/main.c b/src/arch/x86_64/main.c
index c052de09..3275beb6 100644
--- a/src/arch/x86_64/main.c
+++ b/src/arch/x86_64/main.c
@@ -42,7 +42,8 @@ void kmain(){
 	arch_mmu_init();
 	
 	vmm_init();	
-
+	
+	alloc_init();
 }
 
 
diff --git a/src/include/kernel/alloc.h b/src/include/kernel/alloc.h
new file mode 100644
index 00000000..b23e2e75
--- /dev/null
+++ b/src/include/kernel/alloc.h
@@ -0,0 +1,11 @@
+#ifndef _ALLOC_H_INCLUDE
+#define _ALLOC_H_INCLUDE
+
+#include <stddef.h>
+
+void* alloc(size_t);
+void  free(void*);
+
+void alloc_init();
+
+#endif
diff --git a/src/include/kernel/slab.h b/src/include/kernel/slab.h
new file mode 100644
index 00000000..ac30ecce
--- /dev/null
+++ b/src/include/kernel/slab.h
@@ -0,0 +1,11 @@
+#ifndef _SLAB_H_INCLUDE
+#define _SLAB_H_INCLUDE
+
+#include <stddef.h>
+
+void* slab_alloc(size_t);
+void  slab_free(void*);
+
+void slab_init();
+
+#endif
diff --git a/src/sys/mm/alloc.c b/src/sys/mm/alloc.c
new file mode 100644
index 00000000..2bae5a6a
--- /dev/null
+++ b/src/sys/mm/alloc.c
@@ -0,0 +1,14 @@
+#include <kernel/alloc.h>
+#include <kernel/slab.h>
+
+void* alloc(size_t size){
+	return slab_alloc(size);
+}
+
+void free(void* addr){
+	slab_free(addr);
+}
+
+void alloc_init(){
+	slab_init();
+}
diff --git a/src/sys/mm/slab.c b/src/sys/mm/slab.c
new file mode 100644
index 00000000..1ddda50e
--- /dev/null
+++ b/src/sys/mm/slab.c
@@ -0,0 +1,184 @@
+#include <kernel/slab.h>
+#include <arch/spinlock.h>
+#include <arch/mmu.h>
+#include <kernel/pmm.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <string.h>
+
+#define SLAB_SIZE PAGE_SIZE
+
+// TODO have the slabs be in virtual memory rather than in the hhdm
+
+// the layout of a slab in this implementation:
+// header (slab_t)
+// body is made of group several groups
+// a group is:
+// bitmap for the next 8 entries
+// the 8 entries
+
+
+typedef struct _slab {
+	int lock;
+	uint16_t freecount;
+	uint16_t firstfree;
+	uint16_t entrysize;
+	struct _slab* next;
+} __attribute__((packed))slab_t;
+
+typedef struct {
+	size_t entrysize;
+	size_t groupsize;
+	size_t groupcount;
+	slab_t* first;
+} slabdesc;
+
+static slabdesc slabs[7];
+
+static inline slabdesc* getslabdesc(size_t size){
+
+	slabdesc *desc = NULL;
+
+	for(size_t i = 0; i < 7 && desc == NULL; ++i){
+		if(slabs[i].entrysize >= size)
+			desc = &slabs[i];
+	}
+
+	return desc;
+
+}
+
+static void free(void* addr){
+	
+	// align down to page size
+	
+	slab_t *slab = (size_t)addr & ~(0xFFF);
+	spinlock_acquire(&slab->lock);
+	
+	// find group that contains the address
+
+	uint8_t* group = (void*)slab + sizeof(slab_t);
+
+	size_t groupsize = slab->entrysize*8 + 1;
+
+	for(;group < slab + SLAB_SIZE; group += groupsize){
+		if(group + groupsize <= addr)
+			continue;
+
+		size_t first = group+1;
+		
+		size_t bit = ((size_t)addr - first) / slab->entrysize;
+		
+		*group &= ~(1 << bit);
+
+		break;
+
+	}
+	
+	++slab->freecount;
+
+	spinlock_release(&slab->lock);
+}
+
+static void* allocin(slabdesc* desc, slab_t* slab){
+	
+	uint8_t* groups = (void*)slab + sizeof(slab_t);
+	void* allocval = NULL;
+
+	for(size_t i = 0; i < desc->groupcount; ++i, groups += desc->groupsize){
+		if(*groups == 0xFF) continue;
+
+
+		// group has a free member
+		
+		size_t bit = 0;
+
+		for(; bit < 8; ++bit){
+			if((*groups >> bit) & 1)
+				continue;
+			break;
+		}
+		
+		*groups |= (1 << bit);
+
+		allocval = groups + 1 + desc->entrysize * bit;
+
+		break;
+		
+	}
+
+	if(allocval){
+		--slab->freecount;
+		memset(allocval, 0, desc->entrysize);
+	}
+
+	
+	return allocval;
+	
+}
+
+slab_t* newslab(slabdesc *desc){
+	slab_t* slab = pmm_hhdmalloc(SLAB_SIZE / PAGE_SIZE);
+	if(!slab) return NULL;
+	memset(slab, 0, SLAB_SIZE);
+	slab->entrysize = desc->entrysize;
+	slab->freecount = desc->groupcount * 8;
+	return slab;
+}
+
+void* slab_alloc(size_t size){
+	slabdesc* desc = getslabdesc(size);
+	if(!desc) return NULL;
+	
+	slab_t* slab = desc->first;
+	slab_t* last;
+	
+	while(slab){
+		spinlock_acquire(&slab->lock);
+		if(slab->freecount == 0){
+			spinlock_release(&slab->lock);
+			last = slab;
+			slab = slab->next;
+			continue;
+		}
+		
+		break;
+
+	}
+	
+	if(!slab){
+		slab = newslab(desc);
+		if(!slab){
+			spinlock_release(&last->lock);
+			return NULL;
+		}
+		spinlock_acquire(&slab->lock);
+		last->next = slab;
+		spinlock_release(&last->lock);
+	}
+
+	void* alloc = allocin(desc, slab);
+
+	spinlock_release(&slab->lock);
+
+	return alloc;
+
+}
+
+void slab_free(void* addr){
+	
+	free(addr);
+
+}
+
+void slab_init(){
+	for(size_t i = 0; i < 7; ++i){
+		slabs[i].entrysize = 1 << (i + 3);
+		slabs[i].groupsize = slabs[i].entrysize * 8 + 1;
+		slabs[i].groupcount = (SLAB_SIZE - sizeof(slab_t)) / slabs[i].groupsize;
+		slab_t* slab = newslab(&slabs[i]);
+		if(!slab) _panic("Out of memory!", 0);
+		slabs[i].first = slab;
+		
+	}
+}