diff --git a/nx/source/applets/swkbd.c b/nx/source/applets/swkbd.c
index 595c5f76..b366db85 100644
--- a/nx/source/applets/swkbd.c
+++ b/nx/source/applets/swkbd.c
@@ -1,11 +1,11 @@
 #include <string.h>
-#include <malloc.h>
 #include <math.h>
 #include "libapplet_internal.h"
 #include "applets/swkbd.h"
 #include "services/vi.h"
 #include "runtime/hosversion.h"
 #include "runtime/util/utf.h"
+#include "../runtime/alloc.h"
 
 static Result _swkbdGetReplies(SwkbdInline* s);
 
@@ -85,7 +85,7 @@ Result swkbdCreate(SwkbdConfig* c, s32 max_dictwords) {
         c->workbuf_size = (c->workbuf_size + 0xfff) & ~0xfff;
     }
 
-    c->workbuf = (u8*)memalign(0x1000, c->workbuf_size);
+    c->workbuf = (u8*)__libnx_aligned_alloc(0x1000, c->workbuf_size);
     if (c->workbuf==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     if (R_SUCCEEDED(rc)) memset(c->workbuf, 0, c->workbuf_size);
 
@@ -93,7 +93,7 @@ Result swkbdCreate(SwkbdConfig* c, s32 max_dictwords) {
 }
 
 void swkbdClose(SwkbdConfig* c) {
-    free(c->workbuf);
+    __libnx_free(c->workbuf);
     memset(c, 0, sizeof(SwkbdConfig));
 }
 
@@ -279,14 +279,14 @@ Result swkbdShow(SwkbdConfig* c, char* out_string, size_t out_string_size) {
     memset(&storage, 0, sizeof(AppletStorage));
     memset(&customizedDictionarySet_storage, 0, sizeof(customizedDictionarySet_storage));
 
-    strbuf = (u16*)malloc(strbuf_size+2);
+    strbuf = (u16*)__libnx_alloc(strbuf_size+2);
     if (strbuf==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     if (strbuf) memset(strbuf, 0, strbuf_size+2);
     if (R_FAILED(rc)) return rc;
 
     rc = appletCreateLibraryApplet(&holder, AppletId_LibraryAppletSwkbd, LibAppletMode_AllForeground);
     if (R_FAILED(rc)) {
-        free(strbuf);
+        __libnx_free(strbuf);
         return rc;
     }
 
@@ -354,7 +354,7 @@ Result swkbdShow(SwkbdConfig* c, char* out_string, size_t out_string_size) {
     appletStorageCloseTmem(&storage);
     appletStorageCloseTmem(&customizedDictionarySet_storage);
 
-    free(strbuf);
+    __libnx_free(strbuf);
 
     return rc;
 }
@@ -412,18 +412,18 @@ Result swkbdInlineCreate(SwkbdInline* s) {
     swkbdInlineSetUtf8Mode(s, true);
 
     s->interactive_tmpbuf_size = 0x1000;
-    s->interactive_tmpbuf = (u8*)malloc(s->interactive_tmpbuf_size);
+    s->interactive_tmpbuf = (u8*)__libnx_alloc(s->interactive_tmpbuf_size);
     if (s->interactive_tmpbuf==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     if (R_SUCCEEDED(rc)) memset(s->interactive_tmpbuf, 0, s->interactive_tmpbuf_size);
 
     if (R_SUCCEEDED(rc)) {
         s->interactive_strbuf_size = 0x1001;
-        s->interactive_strbuf = (char*)malloc(s->interactive_strbuf_size);
+        s->interactive_strbuf = (char*)__libnx_alloc(s->interactive_strbuf_size);
         if (s->interactive_strbuf==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
         if (R_SUCCEEDED(rc)) memset(s->interactive_strbuf, 0, s->interactive_strbuf_size);
 
         if (R_FAILED(rc)) {
-            free(s->interactive_tmpbuf);
+            __libnx_free(s->interactive_tmpbuf);
             s->interactive_tmpbuf = NULL;
         }
     }
@@ -470,10 +470,10 @@ Result swkbdInlineClose(SwkbdInline* s) {
         appletHolderClose(&s->holder);
     }
 
-    free(s->interactive_tmpbuf);
+    __libnx_free(s->interactive_tmpbuf);
     s->interactive_tmpbuf = NULL;
     s->interactive_tmpbuf_size = 0;
-    free(s->interactive_strbuf);
+    __libnx_free(s->interactive_strbuf);
     s->interactive_strbuf = NULL;
     s->interactive_strbuf_size = 0;
 
diff --git a/nx/source/applets/web.c b/nx/source/applets/web.c
index c4f429df..cd8a1f00 100644
--- a/nx/source/applets/web.c
+++ b/nx/source/applets/web.c
@@ -1,5 +1,4 @@
 #include <string.h>
-#include <malloc.h>
 #include "libapplet_internal.h"
 #include "applets/web.h"
 #include "runtime/hosversion.h"
diff --git a/nx/source/audio/driver.c b/nx/source/audio/driver.c
index b8cb2f18..96706dc1 100644
--- a/nx/source/audio/driver.c
+++ b/nx/source/audio/driver.c
@@ -64,7 +64,7 @@ Result audrvCreate(AudioDriver* d, const AudioRendererConfig* config, int num_fi
     size_t etc_size = _audrvGetEtcSize(config);
     memset(d, 0, sizeof(AudioDriver));
 
-    d->etc = (AudioDriverEtc*)malloc(etc_size);
+    d->etc = (AudioDriverEtc*)__libnx_alloc(etc_size);
     if (!d->etc)
         goto _error0;
 
@@ -80,12 +80,12 @@ Result audrvCreate(AudioDriver* d, const AudioRendererConfig* config, int num_fi
     d->etc->sinks = (AudioDriverEtcSink*)(d->etc->mixes+config->num_mix_objs);
 
     d->etc->out_buf_size = audrenGetOutputParamSize(config);
-    d->etc->out_buf = memalign(AUDREN_OUTPUT_PARAM_ALIGNMENT, d->etc->out_buf_size);
+    d->etc->out_buf = __libnx_aligned_alloc(AUDREN_OUTPUT_PARAM_ALIGNMENT, d->etc->out_buf_size);
     if (!d->etc->out_buf)
         goto _error1;
 
     d->etc->in_buf_size = audrenGetInputParamSize(config);
-    d->etc->in_buf = memalign(AUDREN_INPUT_PARAM_ALIGNMENT, d->etc->in_buf_size);
+    d->etc->in_buf = __libnx_aligned_alloc(AUDREN_INPUT_PARAM_ALIGNMENT, d->etc->in_buf_size);
     if (!d->etc->in_buf)
         goto _error2;
 
@@ -93,9 +93,9 @@ Result audrvCreate(AudioDriver* d, const AudioRendererConfig* config, int num_fi
     return 0;
 
 _error2:
-    free(d->etc->out_buf);
+    __libnx_free(d->etc->out_buf);
 _error1:
-    free(d->etc);
+    __libnx_free(d->etc);
     d->etc = NULL;
 _error0:
     return MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
@@ -136,8 +136,8 @@ Result audrvUpdate(AudioDriver* d)
 
 void audrvClose(AudioDriver* d)
 {
-    free(d->etc->in_buf);
-    free(d->etc->out_buf);
-    free(d->etc);
+    __libnx_free(d->etc->in_buf);
+    __libnx_free(d->etc->out_buf);
+    __libnx_free(d->etc);
     memset(d, 0, sizeof(AudioDriver));
 }
diff --git a/nx/source/audio/driver_internal.h b/nx/source/audio/driver_internal.h
index 116fe1a4..f234bbfc 100644
--- a/nx/source/audio/driver_internal.h
+++ b/nx/source/audio/driver_internal.h
@@ -1,10 +1,10 @@
 #pragma once
-#include <malloc.h>
 #include <string.h>
 #include "types.h"
 #include "result.h"
 #include "services/audren.h"
 #include "audio/driver.h"
+#include "../runtime/alloc.h"
 
 typedef struct {
     int next_free;
diff --git a/nx/source/display/framebuffer.c b/nx/source/display/framebuffer.c
index 25043812..00a35fc8 100644
--- a/nx/source/display/framebuffer.c
+++ b/nx/source/display/framebuffer.c
@@ -11,6 +11,7 @@
 #include "display/native_window.h"
 #include "display/framebuffer.h"
 #include "nvidia/graphic_buffer.h"
+#include "../runtime/alloc.h"
 
 static const NvColorFormat g_nvColorFmtTable[] = {
     NvColorFormat_A8B8G8R8, // PIXEL_FORMAT_RGBA_8888
@@ -76,7 +77,7 @@ Result framebufferCreate(Framebuffer* fb, NWindow *win, u32 width, u32 height, u
     const u32 fb_size = width_aligned_bytes*height_aligned;
     const u32 buf_size = (num_fbs*fb_size + 0xFFF) &~ 0xFFF; // needs to be page aligned
 
-    fb->buf = aligned_alloc(0x1000, buf_size);
+    fb->buf = __libnx_aligned_alloc(0x1000, buf_size);
     if (!fb->buf)
         rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
 
@@ -119,10 +120,11 @@ Result framebufferMakeLinear(Framebuffer* fb)
         return MAKERESULT(Module_Libnx, LibnxError_AlreadyInitialized);
 
     u32 height = (fb->win->height + 7) &~ 7; // GOBs are 8 rows tall
-    fb->buf_linear = calloc(1, fb->stride*height);
+    fb->buf_linear = __libnx_alloc(fb->stride*height);
     if (!fb->buf_linear)
         return MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
 
+    memset(fb->buf_linear, 0, fb->stride*height);
     return 0;
 }
 
@@ -132,12 +134,12 @@ void framebufferClose(Framebuffer* fb)
         return;
 
     if (fb->buf_linear)
-        free(fb->buf_linear);
+        __libnx_free(fb->buf_linear);
 
     if (fb->buf) {
         nwindowReleaseBuffers(fb->win);
         nvMapClose(&fb->map);
-        free(fb->buf);
+        __libnx_free(fb->buf);
     }
 
     memset(fb, 0, sizeof(*fb));
diff --git a/nx/source/kernel/jit.c b/nx/source/kernel/jit.c
index 1028e3bb..452a3ad7 100644
--- a/nx/source/kernel/jit.c
+++ b/nx/source/kernel/jit.c
@@ -1,5 +1,4 @@
 // Copyright 2018 plutoo
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "runtime/env.h"
@@ -7,6 +6,7 @@
 #include "kernel/svc.h"
 #include "kernel/virtmem.h"
 #include "kernel/jit.h"
+#include "../runtime/alloc.h"
 
 Result jitCreate(Jit* j, size_t size)
 {
@@ -29,8 +29,7 @@ Result jitCreate(Jit* j, size_t size)
     }
 
     size = (size + 0xFFF) &~ 0xFFF;
-
-    void* src_addr = memalign(0x1000, size);
+    void* src_addr = __libnx_aligned_alloc(0x1000, size);
 
     if (src_addr == NULL)
         return MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
@@ -92,7 +91,7 @@ Result jitCreate(Jit* j, size_t size)
     }
 
     if (R_FAILED(rc)) {
-        free(j->src_addr);
+        __libnx_free(j->src_addr);
         j->src_addr = NULL;
     }
 
@@ -179,7 +178,7 @@ Result jitClose(Jit* j)
 
     if (R_SUCCEEDED(rc)) {
         if (j->src_addr != NULL) {
-            free(j->src_addr);
+            __libnx_free(j->src_addr);
             j->src_addr = NULL;
         }
     }
diff --git a/nx/source/kernel/shmem.c b/nx/source/kernel/shmem.c
index 99885b6f..5ed28da4 100644
--- a/nx/source/kernel/shmem.c
+++ b/nx/source/kernel/shmem.c
@@ -1,5 +1,4 @@
 // Copyright 2017 plutoo
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/svc.h"
diff --git a/nx/source/kernel/thread.c b/nx/source/kernel/thread.c
index af227449..2703a0f6 100644
--- a/nx/source/kernel/thread.c
+++ b/nx/source/kernel/thread.c
@@ -1,5 +1,4 @@
 // Copyright 2017 plutoo
-#include <malloc.h>
 #include <string.h>
 #include "types.h"
 #include "result.h"
@@ -11,6 +10,7 @@
 #include "runtime/env.h"
 #include "runtime/diag.h"
 #include "../internal.h"
+#include "../runtime/alloc.h"
 
 #define USER_TLS_BEGIN 0x108
 #define USER_TLS_END   (0x200 - sizeof(ThreadVars))
@@ -90,14 +90,6 @@ void __libnx_init_thread(void) {
     getThreadVars()->thread_ptr = &g_mainThread;
 }
 
-void* __attribute__((weak)) __libnx_thread_alloc(size_t size) {
-    return memalign(0x1000, size);
-}
-
-void __attribute__((weak)) __libnx_thread_free(void* mem) {
-    free(mem);
-}
-
 Result threadCreate(
     Thread* t, ThreadFunc entry, void* arg, void* stack_mem, size_t stack_sz,
     int prio, int cpuid)
@@ -109,7 +101,7 @@ Result threadCreate(
     bool owns_stack_mem;
     if (stack_mem == NULL) {
         // Allocate new memory, stack then reent then tls.
-        stack_mem = __libnx_thread_alloc(((stack_sz + reent_sz + tls_sz) + 0xFFF) & ~0xFFF);
+        stack_mem = __libnx_aligned_alloc(0x1000, stack_sz + reent_sz + tls_sz);
 
         owns_stack_mem = true;
     } else {
@@ -190,7 +182,7 @@ Result threadCreate(
 
     if (R_FAILED(rc)) {
         if (owns_stack_mem) {
-            __libnx_thread_free(stack_mem);
+            __libnx_free(stack_mem);
         }
     }
 
@@ -248,7 +240,7 @@ Result threadClose(Thread* t) {
 
     if (R_SUCCEEDED(rc)) {
         if (t->owns_stack_mem) {
-            __libnx_thread_free(t->stack_mem);
+            __libnx_free(t->stack_mem);
         }
         svcCloseHandle(t->handle);
     }
diff --git a/nx/source/kernel/tmem.c b/nx/source/kernel/tmem.c
index e5a9f7fe..cbf1aeb1 100644
--- a/nx/source/kernel/tmem.c
+++ b/nx/source/kernel/tmem.c
@@ -1,12 +1,12 @@
 // Copyright 2017 plutoo
 #include <string.h>
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/svc.h"
 #include "kernel/tmem.h"
 #include "kernel/virtmem.h"
 #include "runtime/diag.h"
+#include "../runtime/alloc.h"
 
 Result tmemCreate(TransferMemory* t, size_t size, Permission perm)
 {
@@ -16,7 +16,7 @@ Result tmemCreate(TransferMemory* t, size_t size, Permission perm)
     t->size = size;
     t->perm = perm;
     t->map_addr = NULL;
-    t->src_addr = memalign(0x1000, size);
+    t->src_addr = __libnx_aligned_alloc(0x1000, size);
 
     if (t->src_addr == NULL) {
         rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
@@ -30,7 +30,7 @@ Result tmemCreate(TransferMemory* t, size_t size, Permission perm)
     }
 
     if (R_FAILED(rc)) {
-        free(t->src_addr);
+        __libnx_free(t->src_addr);
         t->src_addr = NULL;
     }
 
@@ -114,7 +114,7 @@ Result tmemClose(TransferMemory* t)
         }
 
         if (t->src_addr != NULL) {
-            free(t->src_addr);
+            __libnx_free(t->src_addr);
         }
 
         t->src_addr = NULL;
diff --git a/nx/source/kernel/virtmem.c b/nx/source/kernel/virtmem.c
index 650a7332..931ac292 100644
--- a/nx/source/kernel/virtmem.c
+++ b/nx/source/kernel/virtmem.c
@@ -5,7 +5,7 @@
 #include "kernel/virtmem.h"
 #include "kernel/random.h"
 #include "runtime/diag.h"
-#include <stdlib.h>
+#include "../runtime/alloc.h"
 
 #define SEQUENTIAL_GUARD_REGION_SIZE 0x1000
 #define RANDOM_MAX_ATTEMPTS 0x200
@@ -220,7 +220,7 @@ void* virtmemFindCodeMemory(size_t size, size_t guard_size) {
 
 VirtmemReservation* virtmemAddReservation(void* mem, size_t size) {
     if (!mutexIsLockedByCurrentThread(&g_VirtmemMutex)) return NULL;
-    VirtmemReservation* rv = (VirtmemReservation*)malloc(sizeof(VirtmemReservation));
+    VirtmemReservation* rv = (VirtmemReservation*)__libnx_alloc(sizeof(VirtmemReservation));
     if (rv) {
         rv->region.start = (uintptr_t)mem;
         rv->region.end   = rv->region.start + size;
@@ -241,4 +241,5 @@ void virtmemRemoveReservation(VirtmemReservation* rv) {
         rv->prev->next = rv->next;
     else
         g_Reservations = rv->next;
+    __libnx_free(rv);
 }
diff --git a/nx/source/nvidia/address_space.c b/nx/source/nvidia/address_space.c
index 051c9aaa..e7079734 100644
--- a/nx/source/nvidia/address_space.c
+++ b/nx/source/nvidia/address_space.c
@@ -1,4 +1,3 @@
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/svc.h"
diff --git a/nx/source/nvidia/channel.c b/nx/source/nvidia/channel.c
index 1dfa9c46..12efedd7 100644
--- a/nx/source/nvidia/channel.c
+++ b/nx/source/nvidia/channel.c
@@ -1,4 +1,3 @@
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/svc.h"
diff --git a/nx/source/nvidia/gpu_channel.c b/nx/source/nvidia/gpu_channel.c
index fb87649b..78931698 100644
--- a/nx/source/nvidia/gpu_channel.c
+++ b/nx/source/nvidia/gpu_channel.c
@@ -1,4 +1,3 @@
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/svc.h"
diff --git a/nx/source/runtime/alloc.c b/nx/source/runtime/alloc.c
new file mode 100644
index 00000000..ec882180
--- /dev/null
+++ b/nx/source/runtime/alloc.c
@@ -0,0 +1,15 @@
+#include "alloc.h"
+#include <stdlib.h>
+
+void* __attribute__((weak)) __libnx_alloc(size_t size) {
+    return malloc(size);
+}
+
+void* __attribute__((weak)) __libnx_aligned_alloc(size_t alignment, size_t size) {
+    size = (size + alignment - 1) &~ (alignment - 1);
+    return aligned_alloc(alignment, size);
+}
+
+void __attribute__((weak)) __libnx_free(void* p) {
+    free(p);
+}
diff --git a/nx/source/runtime/alloc.h b/nx/source/runtime/alloc.h
new file mode 100644
index 00000000..7e820fa2
--- /dev/null
+++ b/nx/source/runtime/alloc.h
@@ -0,0 +1,6 @@
+#pragma once
+#include "types.h"
+
+void* __libnx_alloc(size_t size);
+void* __libnx_aligned_alloc(size_t alignment, size_t size);
+void __libnx_free(void* p);
diff --git a/nx/source/runtime/devices/fs_dev.c b/nx/source/runtime/devices/fs_dev.c
index e829a22a..f5590e92 100644
--- a/nx/source/runtime/devices/fs_dev.c
+++ b/nx/source/runtime/devices/fs_dev.c
@@ -14,6 +14,7 @@
 #include "runtime/env.h"
 #include "services/time.h"
 
+#include "../alloc.h"
 #include "path_buf.h"
 
 /*! @internal
@@ -324,7 +325,7 @@ static int _fsdevMountDevice(const char *name, FsFileSystem fs, fsdev_fsdevice *
     goto _fail;
 
   device->setup = 1;
-  device->cwd = __nx_fsdev_support_cwd ? malloc(FS_MAX_PATH) : NULL;
+  device->cwd = __nx_fsdev_support_cwd ? __libnx_alloc(FS_MAX_PATH) : NULL;
   if(device->cwd!=NULL)
   {
     device->cwd[0] = '/';
@@ -365,7 +366,7 @@ static int _fsdevUnmountDeviceStruct(fsdev_fsdevice *device)
   strncat(name, ":", sizeof(name)-strlen(name)-1);
 
   RemoveDevice(name);
-  free(device->cwd);
+  __libnx_free(device->cwd);
   fsFsClose(&device->fs);
 
   if(device->id == fsdev_fsdevice_cwd)
diff --git a/nx/source/runtime/devices/romfs_dev.c b/nx/source/runtime/devices/romfs_dev.c
index 8c19e3b2..baad1d77 100644
--- a/nx/source/runtime/devices/romfs_dev.c
+++ b/nx/source/runtime/devices/romfs_dev.c
@@ -14,6 +14,7 @@
 #include "runtime/env.h"
 #include "nro.h"
 
+#include "../alloc.h"
 #include "path_buf.h"
 
 typedef enum {
@@ -240,10 +241,10 @@ static romfs_mount* romfs_alloc(void)
 
 static void romfs_free(romfs_mount *mount)
 {
-    free(mount->fileTable);
-    free(mount->fileHashTable);
-    free(mount->dirTable);
-    free(mount->dirHashTable);
+    __libnx_free(mount->fileTable);
+    __libnx_free(mount->fileHashTable);
+    __libnx_free(mount->dirTable);
+    __libnx_free(mount->dirHashTable);
     _romfsResetMount(mount, mount->id);
 }
 
@@ -395,25 +396,25 @@ Result romfsMountCommon(const char *name, romfs_mount *mount)
     if (_romfs_read(mount, 0, &mount->header, sizeof(mount->header)) != sizeof(mount->header))
         goto fail_io;
 
-    mount->dirHashTable = (u32*)malloc(mount->header.dirHashTableSize);
+    mount->dirHashTable = (u32*)__libnx_alloc(mount->header.dirHashTableSize);
     if (!mount->dirHashTable)
         goto fail_oom;
     if (!_romfs_read_chk(mount, mount->header.dirHashTableOff, mount->dirHashTable, mount->header.dirHashTableSize))
         goto fail_io;
 
-    mount->dirTable = malloc(mount->header.dirTableSize);
+    mount->dirTable = __libnx_alloc(mount->header.dirTableSize);
     if (!mount->dirTable)
         goto fail_oom;
     if (!_romfs_read_chk(mount, mount->header.dirTableOff, mount->dirTable, mount->header.dirTableSize))
         goto fail_io;
 
-    mount->fileHashTable = (u32*)malloc(mount->header.fileHashTableSize);
+    mount->fileHashTable = (u32*)__libnx_alloc(mount->header.fileHashTableSize);
     if (!mount->fileHashTable)
         goto fail_oom;
     if (!_romfs_read_chk(mount, mount->header.fileHashTableOff, mount->fileHashTable, mount->header.fileHashTableSize))
         goto fail_io;
 
-    mount->fileTable = malloc(mount->header.fileTableSize);
+    mount->fileTable = __libnx_alloc(mount->header.fileTableSize);
     if (!mount->fileTable)
         goto fail_oom;
     if (!_romfs_read_chk(mount, mount->header.fileTableOff, mount->fileTable, mount->header.fileTableSize))
diff --git a/nx/source/runtime/devices/socket.c b/nx/source/runtime/devices/socket.c
index 02eddf3f..a2f0b4fe 100644
--- a/nx/source/runtime/devices/socket.c
+++ b/nx/source/runtime/devices/socket.c
@@ -4,7 +4,6 @@
 #include <errno.h>
 #include <alloca.h>
 #include <sys/iosupport.h>
-#include <malloc.h>
 
 #include <fcntl.h>
 #include <poll.h>
@@ -19,6 +18,7 @@
 #include "services/ssl.h"
 #include "runtime/devices/socket.h"
 #include "runtime/hosversion.h"
+#include "../alloc.h"
 
 __attribute__((weak)) size_t __nx_pollfd_sb_max_fds = 64;
 
@@ -308,7 +308,7 @@ int select(int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds, struc
     if(numfds <= __nx_pollfd_sb_max_fds)
         pollinfo = (struct pollfd *)alloca(numfds * sizeof(struct pollfd));
     else
-        pollinfo = (struct pollfd *)malloc(numfds * sizeof(struct pollfd));
+        pollinfo = (struct pollfd *)__libnx_alloc(numfds * sizeof(struct pollfd));
     if(pollinfo == NULL) {
         errno = ENOMEM;
         return -1;
@@ -379,7 +379,7 @@ int select(int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds, struc
 
 cleanup:
     if(numfds > __nx_pollfd_sb_max_fds)
-        free(pollinfo);
+        __libnx_free(pollinfo);
     return rc;
 }
 
@@ -396,7 +396,7 @@ int poll(struct pollfd *fds, nfds_t nfds, int timeout) {
     if(nfds <= __nx_pollfd_sb_max_fds)
         fds2 = (struct pollfd *)alloca(nfds * sizeof(struct pollfd));
     else
-        fds2 = (struct pollfd *)malloc(nfds * sizeof(struct pollfd));
+        fds2 = (struct pollfd *)__libnx_alloc(nfds * sizeof(struct pollfd));
     if(fds2 == NULL) {
         errno = ENOMEM;
         return -1;
@@ -426,7 +426,7 @@ int poll(struct pollfd *fds, nfds_t nfds, int timeout) {
     }
 
     if(nfds > __nx_pollfd_sb_max_fds)
-        free(fds2);
+        __libnx_free(fds2);
     return ret;
 }
 
@@ -629,7 +629,7 @@ int ioctl(int fd, int request, ...) {
                 return -1;
             }
 
-            struct bpf_program_serialized *prog_ser = (struct bpf_program_serialized *)malloc(sizeof(struct bpf_program_serialized));
+            struct bpf_program_serialized *prog_ser = (struct bpf_program_serialized *)__libnx_alloc(sizeof(struct bpf_program_serialized));
             if(prog_ser == NULL) {
                 errno = ENOMEM;
                 return -1;
@@ -640,7 +640,7 @@ int ioctl(int fd, int request, ...) {
 
             request = _IOC(request & IOC_DIRMASK, IOCGROUP(request), IOCBASECMD(request), sizeof(struct bpf_program_serialized));
             ret = bsdIoctl(fd, request, prog_ser);
-            free(prog_ser);
+            __libnx_free(prog_ser);
             return _socketParseBsdResult(NULL, ret);
         }
         default:
@@ -967,7 +967,7 @@ static int _mmsgInitCommon(u8 **buf, size_t *alignsize, struct mmsghdr *msgvec,
     }
 
     *alignsize = (bufsize+0xfff) & ~0xfff;
-    *buf = (u8*)memalign(0x1000, *alignsize);
+    *buf = (u8*)__libnx_aligned_alloc(0x1000, *alignsize);
     if (*buf == NULL) {
         errno = ENOMEM;
         return -1;
@@ -1018,7 +1018,7 @@ int sendmmsg(int sockfd, struct mmsghdr *msgvec, unsigned int vlen, int flags) {
         if (ret2==-1) ret = ret2;
     }
 
-    free(buf);
+    __libnx_free(buf);
 
     return ret;
 }
@@ -1062,7 +1062,7 @@ int recvmmsg(int sockfd, struct mmsghdr *msgvec, unsigned int vlen, int flags, s
         if (ret2==-1) ret = ret2;
     }
 
-    free(buf);
+    __libnx_free(buf);
 
     return ret;
 }
diff --git a/nx/source/runtime/devices/usb_comms.c b/nx/source/runtime/devices/usb_comms.c
index d0fcb933..18edd882 100644
--- a/nx/source/runtime/devices/usb_comms.c
+++ b/nx/source/runtime/devices/usb_comms.c
@@ -1,5 +1,4 @@
 #include <string.h>
-#include <malloc.h>
 #include "types.h"
 #include "result.h"
 #include "kernel/rwlock.h"
@@ -7,6 +6,7 @@
 #include "runtime/hosversion.h"
 #include "runtime/diag.h"
 #include "runtime/devices/usb_comms.h"
+#include "../alloc.h"
 
 #define TOTAL_INTERFACES 4
 
@@ -177,8 +177,8 @@ static void _usbCommsInterfaceFree(usbCommsInterface *interface)
     interface->endpoint_out = NULL;
     interface->interface = NULL;
 
-    free(interface->endpoint_in_buffer);
-    free(interface->endpoint_out_buffer);
+    __libnx_free(interface->endpoint_in_buffer);
+    __libnx_free(interface->endpoint_out_buffer);
     interface->endpoint_in_buffer = NULL;
     interface->endpoint_out_buffer = NULL;
 
@@ -258,11 +258,11 @@ static Result _usbCommsInterfaceInit5x(u32 intf_ind, const UsbCommsInterfaceInfo
     interface->initialized = 1;
 
     //The buffer for PostBufferAsync commands must be 0x1000-byte aligned.
-    interface->endpoint_in_buffer = memalign(0x1000, 0x1000);
+    interface->endpoint_in_buffer = __libnx_aligned_alloc(0x1000, 0x1000);
     if (interface->endpoint_in_buffer==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
 
     if (R_SUCCEEDED(rc)) {
-        interface->endpoint_out_buffer = memalign(0x1000, 0x1000);
+        interface->endpoint_out_buffer = __libnx_aligned_alloc(0x1000, 0x1000);
         if (interface->endpoint_out_buffer==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     }
 
@@ -360,11 +360,11 @@ static Result _usbCommsInterfaceInit1x(u32 intf_ind, const UsbCommsInterfaceInfo
     interface->initialized = 1;
 
     //The buffer for PostBufferAsync commands must be 0x1000-byte aligned.
-    interface->endpoint_in_buffer = memalign(0x1000, 0x1000);
+    interface->endpoint_in_buffer = __libnx_aligned_alloc(0x1000, 0x1000);
     if (interface->endpoint_in_buffer==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
 
     if (R_SUCCEEDED(rc)) {
-        interface->endpoint_out_buffer = memalign(0x1000, 0x1000);
+        interface->endpoint_out_buffer = __libnx_aligned_alloc(0x1000, 0x1000);
         if (interface->endpoint_out_buffer==NULL) rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     }
 
diff --git a/nx/source/runtime/newlib.c b/nx/source/runtime/newlib.c
index faee450e..57112aa6 100644
--- a/nx/source/runtime/newlib.c
+++ b/nx/source/runtime/newlib.c
@@ -18,6 +18,7 @@
 #include "services/time.h"
 #include "runtime/diag.h"
 #include "result.h"
+#include "alloc.h"
 
 #define THRD_MAIN_HANDLE ((struct __pthread_t*)~(uintptr_t)0)
 
@@ -183,7 +184,7 @@ int __syscall_thread_create(struct __pthread_t **thread, void* (*func)(void*), v
     if (R_FAILED(rc))
         return EPERM;
 
-    struct __pthread_t* t = (struct __pthread_t*)malloc(sizeof(struct __pthread_t));
+    struct __pthread_t* t = (struct __pthread_t*)__libnx_alloc(sizeof(struct __pthread_t));
     if (!t)
         return ENOMEM;
 
@@ -217,7 +218,7 @@ int __syscall_thread_create(struct __pthread_t **thread, void* (*func)(void*), v
 _error2:
     threadClose(&t->thr);
 _error1:
-    free(t);
+    __libnx_free(t);
     return ENOMEM;
 }
 
@@ -232,7 +233,7 @@ void* __syscall_thread_join(struct __pthread_t *thread)
 
     void* ret = thread->rc;
     threadClose(&thread->thr);
-    free(thread);
+    __libnx_free(thread);
 
     return ret;
 }
diff --git a/nx/source/runtime/resolver.c b/nx/source/runtime/resolver.c
index 36728b9f..7a35e3ad 100644
--- a/nx/source/runtime/resolver.c
+++ b/nx/source/runtime/resolver.c
@@ -14,6 +14,7 @@
 #include "services/nifm.h"
 #include "runtime/hosversion.h"
 #include "runtime/resolver.h"
+#include "alloc.h"
 
 __thread int h_errno;
 
@@ -126,7 +127,7 @@ static struct hostent *_resolverDeserializeHostent(const void *out_he_serialized
     pos_addresses = pos;
     pos += addrlen * nb_addresses;
 
-    he = malloc(
+    he = __libnx_alloc(
         sizeof(struct hostent)
         + name_size
         + 8 * (nb_aliases + 1 + nb_addresses + 1)
@@ -244,7 +245,7 @@ static struct addrinfo_serialized_hdr *_resolverSerializeAddrInfoList(size_t *ou
     if (reqsize > g_resolverAddrInfoHintsBufferSize)
         return NULL;
 
-    struct addrinfo_serialized_hdr *out = malloc(reqsize);
+    struct addrinfo_serialized_hdr *out = __libnx_alloc(reqsize);
     if (!out)
         return NULL;
 
@@ -268,7 +269,7 @@ static struct addrinfo *_resolverDeserializeAddrInfo(size_t *out_len, const stru
 
     size_t subsize1 = hdr->ai_addrlen ? ntohl(hdr->ai_addrlen) : 4;
     size_t subsize2 = strlen((const char *)hdr + sizeof(struct addrinfo_serialized_hdr) + subsize1) + 1;
-    struct addrinfo_node *node = malloc(sizeof(struct addrinfo_node) + subsize2);
+    struct addrinfo_node *node = __libnx_alloc(sizeof(struct addrinfo_node) + subsize2);
 
     *out_len = sizeof(struct addrinfo_serialized_hdr) + subsize1 + subsize2;
     if (!node)
@@ -348,13 +349,13 @@ static struct addrinfo *_resolverDeserializeAddrInfoList(struct addrinfo_seriali
 }
 
 void freehostent(struct hostent *he) {
-    free(he);
+    __libnx_free(he);
 }
 
 void freeaddrinfo(struct addrinfo *ai) {
     for (struct addrinfo *node = ai, *next; node; node = next) {
         next = node->ai_next;
-        free(node);
+        __libnx_free(node);
     }
 }
 
@@ -371,7 +372,7 @@ struct hostent *gethostbyname(const char *name) {
         return NULL;
     }
 
-    void *out_serialized = malloc(g_resolverHostByNameBufferSize);
+    void *out_serialized = __libnx_alloc(g_resolverHostByNameBufferSize);
     if (!out_serialized) {
         h_errno = NETDB_INTERNAL;
         errno = ENOMEM;
@@ -403,7 +404,7 @@ struct hostent *gethostbyname(const char *name) {
     if (h_errno == NETDB_SUCCESS)
         ret = _resolverDeserializeHostent(out_serialized);
 
-    free(out_serialized);
+    __libnx_free(out_serialized);
     return ret;
 }
 
@@ -426,7 +427,7 @@ struct hostent *gethostbyaddr(const void *addr, socklen_t len, int type) {
         return NULL;
     }
 
-    void *out_serialized = malloc(g_resolverHostByAddrBufferSize);
+    void *out_serialized = __libnx_alloc(g_resolverHostByAddrBufferSize);
     if (!out_serialized) {
         h_errno = NETDB_INTERNAL;
         errno = ENOMEM;
@@ -458,7 +459,7 @@ struct hostent *gethostbyaddr(const void *addr, socklen_t len, int type) {
     if (h_errno == NETDB_SUCCESS)
         ret = _resolverDeserializeHostent(out_serialized);
 
-    free(out_serialized);
+    __libnx_free(out_serialized);
     return ret;
 }
 
@@ -508,9 +509,9 @@ int getaddrinfo(const char *node, const char *service, const struct addrinfo *hi
         }
     }
 
-    struct addrinfo_serialized_hdr *out_serialized = malloc(g_resolverAddrInfoBufferSize);
+    struct addrinfo_serialized_hdr *out_serialized = __libnx_alloc(g_resolverAddrInfoBufferSize);
     if (!out_serialized) {
-        free(hints_serialized);
+        __libnx_free(hints_serialized);
         errno = ENOMEM;
         return EAI_FAIL;
     }
@@ -528,7 +529,7 @@ int getaddrinfo(const char *node, const char *service, const struct addrinfo *hi
         NULL);
     g_resolverResult = rc;
     g_resolverCancelHandle = 0;
-    free(hints_serialized);
+    __libnx_free(hints_serialized);
 
     if (R_FAILED(rc)) {
         if (R_MODULE(rc) == 21) // SM
@@ -548,7 +549,7 @@ int getaddrinfo(const char *node, const char *service, const struct addrinfo *hi
         }
     }
 
-    free(out_serialized);
+    __libnx_free(out_serialized);
     return ret;
 }
 
diff --git a/nx/source/runtime/ringcon.c b/nx/source/runtime/ringcon.c
index 36bba588..95a935d8 100644
--- a/nx/source/runtime/ringcon.c
+++ b/nx/source/runtime/ringcon.c
@@ -1,7 +1,7 @@
 #include <string.h>
-#include <malloc.h>
 #include "runtime/ringcon.h"
 #include "arm/counter.h"
+#include "alloc.h"
 
 static Result _ringconSetup(RingCon *c);
 
@@ -62,7 +62,7 @@ Result ringconCreate(RingCon *c, HidNpadIdType id) {
     memset(c, 0, sizeof(*c));
 
     c->workbuf_size = 0x1000;
-    c->workbuf = memalign(0x1000, c->workbuf_size);
+    c->workbuf = __libnx_aligned_alloc(0x1000, c->workbuf_size);
     if (c->workbuf == NULL)
         rc = MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     else
@@ -101,7 +101,7 @@ void ringconClose(RingCon *c) {
         hidbusFinalize(c->handle);
     }
 
-    free(c->workbuf);
+    __libnx_free(c->workbuf);
     c->workbuf = 0;
 }
 
diff --git a/nx/source/services/hid.c b/nx/source/services/hid.c
index dabf742a..f0ebc327 100644
--- a/nx/source/services/hid.c
+++ b/nx/source/services/hid.c
@@ -2,7 +2,6 @@
 #include "service_guard.h"
 #include <string.h>
 #include <stdlib.h>
-#include <malloc.h>
 #include <stdatomic.h>
 #include "kernel/shmem.h"
 #include "kernel/mutex.h"
@@ -11,6 +10,7 @@
 #include "services/hid.h"
 #include "runtime/hosversion.h"
 #include "runtime/diag.h"
+#include "../runtime/alloc.h"
 
 static Service g_hidSrv;
 static Service g_hidIAppletResource;
@@ -1410,7 +1410,7 @@ Result hidInitializeSevenSixAxisSensor(void) {
         if (R_FAILED(rc)) return rc;
     }
 
-    g_sevenSixAxisSensorBuffer = (u8*)memalign(0x1000, bufsize);
+    g_sevenSixAxisSensorBuffer = (u8*)__libnx_aligned_alloc(0x1000, bufsize);
     if (g_sevenSixAxisSensorBuffer == NULL)
         return MAKERESULT(Module_Libnx, LibnxError_OutOfMemory);
     memset(g_sevenSixAxisSensorBuffer, 0, bufsize);
@@ -1424,7 +1424,7 @@ Result hidInitializeSevenSixAxisSensor(void) {
         tmemClose(&g_sevenSixAxisSensorTmem0);
         tmemClose(&g_sevenSixAxisSensorTmem1);
 
-        free(g_sevenSixAxisSensorBuffer);
+        __libnx_free(g_sevenSixAxisSensorBuffer);
         g_sevenSixAxisSensorBuffer = NULL;
     }
 
@@ -1444,7 +1444,7 @@ Result hidFinalizeSevenSixAxisSensor(void) {
     tmemClose(&g_sevenSixAxisSensorTmem0);
     tmemClose(&g_sevenSixAxisSensorTmem1);
 
-    free(g_sevenSixAxisSensorBuffer);
+    __libnx_free(g_sevenSixAxisSensorBuffer);
     g_sevenSixAxisSensorBuffer = NULL;
 
     return rc;