diff --git a/src/venus/vkr_ring.c b/src/venus/vkr_ring.c
index 56e4f17..33424e6 100644
--- a/src/venus/vkr_ring.c
+++ b/src/venus/vkr_ring.c
@@ -68,7 +68,7 @@ vkr_ring_create(const struct vkr_ring_layout *layout,
       return NULL;
 
 #define ring_attach_shared(member)                                                       \
-   ring->shared.member = (void *)((uint8_t *)shared + layout->member.offset)
+   ring->shared.member = (void *)((uint8_t *)shared + layout->member.begin)
    ring_attach_shared(head);
    ring_attach_shared(tail);
    ring_attach_shared(status);
@@ -76,10 +76,10 @@ vkr_ring_create(const struct vkr_ring_layout *layout,
    ring_attach_shared(extra);
 #undef ring_attach_shared
 
-   assert(layout->buffer.size && util_is_power_of_two(layout->buffer.size));
-   ring->buffer_size = layout->buffer.size;
-   ring->buffer_mask = layout->buffer.size - 1;
-   ring->extra_size = layout->extra.size;
+   ring->buffer_size = vkr_region_size(&layout->buffer);
+   assert(ring->buffer_size && util_is_power_of_two(ring->buffer_size));
+   ring->buffer_mask = ring->buffer_size - 1;
+   ring->extra_size = vkr_region_size(&layout->extra);
 
    /* we will manage head and status, and we expect them to be 0 initially */
    if (*ring->shared.head || *ring->shared.status) {
diff --git a/src/venus/vkr_ring.h b/src/venus/vkr_ring.h
index 9e53d64..d14d641 100644
--- a/src/venus/vkr_ring.h
+++ b/src/venus/vkr_ring.h
@@ -8,20 +8,13 @@
 
 #include "vkr_common.h"
 
-struct virgl_context;
-
-struct memory_region {
-   size_t offset;
-   size_t size;
-};
-
 /* the layout of a ring in a virgl_resource */
 struct vkr_ring_layout {
-   struct memory_region head;
-   struct memory_region tail;
-   struct memory_region status;
-   struct memory_region buffer;
-   struct memory_region extra;
+   struct vkr_region head;
+   struct vkr_region tail;
+   struct vkr_region status;
+   struct vkr_region buffer;
+   struct vkr_region extra;
 };
 
 static_assert(ATOMIC_INT_LOCK_FREE == 2 && sizeof(atomic_uint) == 4,
diff --git a/src/venus/vkr_transport.c b/src/venus/vkr_transport.c
index dda963b..39de464 100644
--- a/src/venus/vkr_transport.c
+++ b/src/venus/vkr_transport.c
@@ -155,62 +155,62 @@ lookup_ring(struct vkr_context *ctx, uint64_t ring_id)
 static bool
 vkr_ring_layout_init(struct vkr_ring_layout *layout, const VkRingCreateInfoMESA *info)
 {
+   /* clang-format off */
    *layout = (struct vkr_ring_layout){
-      .head = { info->headOffset, sizeof(uint32_t) },
-      .tail = { info->tailOffset, sizeof(uint32_t) },
-      .status = { info->statusOffset, sizeof(uint32_t) },
-      .buffer = { info->bufferOffset, info->bufferSize },
-      .extra = { info->extraOffset, info->extraSize },
+      .head   = VKR_REGION_INIT(info->headOffset, sizeof(uint32_t)),
+      .tail   = VKR_REGION_INIT(info->tailOffset, sizeof(uint32_t)),
+      .status = VKR_REGION_INIT(info->statusOffset, sizeof(uint32_t)),
+      .buffer = VKR_REGION_INIT(info->bufferOffset, info->bufferSize),
+      .extra  = VKR_REGION_INIT(info->extraOffset, info->extraSize),
    };
 
-   const struct memory_region *regions[] = {
+   const struct vkr_region res_region = VKR_REGION_INIT(0, info->size);
+   const struct vkr_region *regions[] = {
       &layout->head,
       &layout->tail,
       &layout->status,
       &layout->buffer,
       &layout->extra,
    };
+   /* clang-format on */
 
    for (size_t i = 0; i < ARRAY_SIZE(regions); i++) {
-      if (regions[i]->offset > info->size ||
-          regions[i]->size > info->size - regions[i]->offset) {
-         vkr_log("ring buffer control variable (offset=%lu, size=%lu) placed"
+      const struct vkr_region *region = regions[i];
+
+      if (!vkr_region_is_valid(region) || !vkr_region_is_within(region, &res_region)) {
+         vkr_log("ring buffer control variable (begin=%lu, end=%lu) placed"
                  " out-of-bounds in shared memory layout",
-                 regions[i]->offset, regions[i]->size);
+                 region->begin, region->end);
          return false;
       }
 
-      if (regions[i]->offset & 0x3) {
-         vkr_log("ring buffer control variable (offset=%lu, size=%lu) must be"
+      if (!vkr_region_is_aligned(region, 4)) {
+         vkr_log("ring buffer control variable (begin=%lu, end=%lu) must be"
                  " 32-bit aligned in shared memory layout",
-                 regions[i]->offset, regions[i]->size);
+                 region->begin, region->end);
          return false;
       }
    }
 
    /* assumes region->size == 0 is valid */
    for (size_t i = 0; i < ARRAY_SIZE(regions); i++) {
-      if (!regions[i]->size)
-         continue;
+      const struct vkr_region *region = regions[i];
 
       for (size_t j = i + 1; j < ARRAY_SIZE(regions); j++) {
-         if (!regions[j]->size)
-            continue;
-
-         if (regions[i]->offset < regions[j]->offset + regions[j]->size &&
-             regions[j]->offset < regions[i]->offset + regions[i]->size) {
-            vkr_log("ring buffer control variable (offset=%lu, size=%lu)"
-                    " overlaps with control variable (offset=%lu, size=%lu)",
-                    regions[j]->offset, regions[j]->size, regions[i]->offset,
-                    regions[i]->size);
+         const struct vkr_region *other = regions[j];
+
+         if (!vkr_region_is_disjoint(region, other)) {
+            vkr_log("ring buffer control variable (begin=%lu, end=%lu)"
+                    " overlaps with control variable (begin=%lu, end=%lu)",
+                    other->begin, other->end, region->begin, region->end);
             return false;
          }
       }
    }
 
-   if (!layout->buffer.size || !util_is_power_of_two(layout->buffer.size)) {
-      vkr_log("ring buffer size (%lu) must be a power of two",
-              layout->buffer.size);
+   const size_t buf_size = vkr_region_size(&layout->buffer);
+   if (!buf_size || !util_is_power_of_two(buf_size)) {
+      vkr_log("ring buffer size (%lu) must be a power of two", buf_size);
       return false;
    }