diff --git a/configure b/configure
index 2d2a498..f08327e 100755
--- a/configure
+++ b/configure
@@ -328,9 +328,11 @@
 archipelago="no"
 gtk=""
 gtkabi=""
+gtk_gl="no"
 gnutls=""
 gnutls_hash=""
 vte=""
+virglrenderer=""
 tpm="yes"
 libssh2=""
 vhdx=""
@@ -1122,6 +1124,10 @@
   ;;
   --enable-vte) vte="yes"
   ;;
+  --disable-virglrenderer) virglrenderer="no"
+  ;;
+  --enable-virglrenderer) virglrenderer="yes"
+  ;;
   --disable-tpm) tpm="no"
   ;;
   --enable-tpm) tpm="yes"
@@ -3231,6 +3237,9 @@
     opengl_cflags="$($pkg_config --cflags $opengl_pkgs) $x11_cflags"
     opengl_libs="$($pkg_config --libs $opengl_pkgs) $x11_libs"
     opengl=yes
+    if test "$gtk" = "yes" && $pkg_config --exists "$gtkpackage >= 3.16"; then
+        gtk_gl="yes"
+    fi
   else
     if test "$opengl" = "yes" ; then
       feature_not_found "opengl" "Please install opengl (mesa) devel pkgs: $opengl_pkgs"
@@ -3967,6 +3976,27 @@
 fi
 
 ##########################################
+# virgl renderer probe
+
+if test "$virglrenderer" != "no" ; then
+  cat > $TMPC << EOF
+#include <virglrenderer.h>
+int main(void) { virgl_renderer_poll(); return 0; }
+EOF
+  virgl_cflags=$($pkg_config --cflags virglrenderer 2>/dev/null)
+  virgl_libs=$($pkg_config --libs virglrenderer 2>/dev/null)
+  if $pkg_config virglrenderer >/dev/null 2>&1 && \
+     compile_prog "$virgl_cflags" "$virgl_libs" ; then
+    virglrenderer="yes"
+  else
+    if test "$virglrenderer" = "yes" ; then
+      feature_not_found "virglrenderer"
+    fi
+    virglrenderer="no"
+  fi
+fi
+
+##########################################
 # check if we have fdatasync
 
 fdatasync=no
@@ -4576,6 +4606,7 @@
 echo "pixman            $pixman"
 echo "SDL support       $sdl"
 echo "GTK support       $gtk"
+echo "GTK GL support    $gtk_gl"
 echo "GNUTLS support    $gnutls"
 echo "GNUTLS hash       $gnutls_hash"
 echo "GNUTLS gcrypt     $gnutls_gcrypt"
@@ -4583,6 +4614,7 @@
 echo "libtasn1          $tasn1"
 echo "VTE support       $vte"
 echo "curses support    $curses"
+echo "virgl support     $virglrenderer"
 echo "curl support      $curl"
 echo "mingw32 support   $mingw32"
 echo "Audio drivers     $audio_drv_list"
@@ -4934,6 +4966,9 @@
   echo "CONFIG_GTK=y" >> $config_host_mak
   echo "CONFIG_GTKABI=$gtkabi" >> $config_host_mak
   echo "GTK_CFLAGS=$gtk_cflags" >> $config_host_mak
+  if test "$gtk_gl" = "yes" ; then
+    echo "CONFIG_GTK_GL=y" >> $config_host_mak
+  fi
 fi
 if test "$gnutls" = "yes" ; then
   echo "CONFIG_GNUTLS=y" >> $config_host_mak
@@ -4955,6 +4990,11 @@
   echo "CONFIG_VTE=y" >> $config_host_mak
   echo "VTE_CFLAGS=$vte_cflags" >> $config_host_mak
 fi
+if test "$virglrenderer" = "yes" ; then
+  echo "CONFIG_VIRGL=y" >> $config_host_mak
+  echo "VIRGL_CFLAGS=$virgl_cflags" >> $config_host_mak
+  echo "VIRGL_LIBS=$virgl_libs" >> $config_host_mak
+fi
 if test "$xen" = "yes" ; then
   echo "CONFIG_XEN_BACKEND=y" >> $config_host_mak
   echo "CONFIG_XEN_CTRL_INTERFACE_VERSION=$xen_ctrl_version" >> $config_host_mak
diff --git a/hw/display/Makefile.objs b/hw/display/Makefile.objs
index dd8ea76..f0cf431 100644
--- a/hw/display/Makefile.objs
+++ b/hw/display/Makefile.objs
@@ -35,6 +35,10 @@
 
 common-obj-$(CONFIG_QXL) += qxl.o qxl-logger.o qxl-render.o
 
-obj-$(CONFIG_VIRTIO) += virtio-gpu.o
+obj-$(CONFIG_VIRTIO) += virtio-gpu.o virtio-gpu-3d.o
 obj-$(CONFIG_VIRTIO_PCI) += virtio-gpu-pci.o
 obj-$(CONFIG_VIRTIO_VGA) += virtio-vga.o
+virtio-gpu.o-cflags := $(VIRGL_CFLAGS)
+virtio-gpu.o-libs += $(VIRGL_LIBS)
+virtio-gpu-3d.o-cflags := $(VIRGL_CFLAGS)
+virtio-gpu-3d.o-libs += $(VIRGL_LIBS)
diff --git a/hw/display/virtio-gpu-3d.c b/hw/display/virtio-gpu-3d.c
new file mode 100644
index 0000000..28dccfd
--- /dev/null
+++ b/hw/display/virtio-gpu-3d.c
@@ -0,0 +1,598 @@
+/*
+ * Virtio GPU Device
+ *
+ * Copyright Red Hat, Inc. 2013-2014
+ *
+ * Authors:
+ *     Dave Airlie <airlied@redhat.com>
+ *     Gerd Hoffmann <kraxel@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+
+#include "qemu-common.h"
+#include "qemu/iov.h"
+#include "trace.h"
+#include "hw/virtio/virtio.h"
+#include "hw/virtio/virtio-gpu.h"
+
+#ifdef CONFIG_VIRGL
+
+#include "virglrenderer.h"
+
+static struct virgl_renderer_callbacks virtio_gpu_3d_cbs;
+
+static void virgl_cmd_create_resource_2d(VirtIOGPU *g,
+                                         struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_create_2d c2d;
+    struct virgl_renderer_resource_create_args args;
+
+    VIRTIO_GPU_FILL_CMD(c2d);
+    trace_virtio_gpu_cmd_res_create_2d(c2d.resource_id, c2d.format,
+                                       c2d.width, c2d.height);
+
+    args.handle = c2d.resource_id;
+    args.target = 2;
+    args.format = c2d.format;
+    args.bind = (1 << 1);
+    args.width = c2d.width;
+    args.height = c2d.height;
+    args.depth = 1;
+    args.array_size = 1;
+    args.last_level = 0;
+    args.nr_samples = 0;
+    args.flags = VIRTIO_GPU_RESOURCE_FLAG_Y_0_TOP;
+    virgl_renderer_resource_create(&args, NULL, 0);
+}
+
+static void virgl_cmd_create_resource_3d(VirtIOGPU *g,
+                                         struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_create_3d c3d;
+    struct virgl_renderer_resource_create_args args;
+
+    VIRTIO_GPU_FILL_CMD(c3d);
+    trace_virtio_gpu_cmd_res_create_3d(c3d.resource_id, c3d.format,
+                                       c3d.width, c3d.height, c3d.depth);
+
+    args.handle = c3d.resource_id;
+    args.target = c3d.target;
+    args.format = c3d.format;
+    args.bind = c3d.bind;
+    args.width = c3d.width;
+    args.height = c3d.height;
+    args.depth = c3d.depth;
+    args.array_size = c3d.array_size;
+    args.last_level = c3d.last_level;
+    args.nr_samples = c3d.nr_samples;
+    args.flags = c3d.flags;
+    virgl_renderer_resource_create(&args, NULL, 0);
+}
+
+static void virgl_cmd_resource_unref(VirtIOGPU *g,
+                                     struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_unref unref;
+
+    VIRTIO_GPU_FILL_CMD(unref);
+    trace_virtio_gpu_cmd_res_unref(unref.resource_id);
+
+    virgl_renderer_resource_unref(unref.resource_id);
+}
+
+static void virgl_cmd_context_create(VirtIOGPU *g,
+                                     struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_ctx_create cc;
+
+    VIRTIO_GPU_FILL_CMD(cc);
+    trace_virtio_gpu_cmd_ctx_create(cc.hdr.ctx_id,
+                                    cc.debug_name);
+
+    virgl_renderer_context_create(cc.hdr.ctx_id, cc.nlen,
+                                  cc.debug_name);
+}
+
+static void virgl_cmd_context_destroy(VirtIOGPU *g,
+                                      struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_ctx_destroy cd;
+
+    VIRTIO_GPU_FILL_CMD(cd);
+    trace_virtio_gpu_cmd_ctx_destroy(cd.hdr.ctx_id);
+
+    virgl_renderer_context_destroy(cd.hdr.ctx_id);
+}
+
+static void virtio_gpu_rect_update(VirtIOGPU *g, int idx, int x, int y,
+                                int width, int height)
+{
+    if (!g->scanout[idx].con) {
+        return;
+    }
+
+    dpy_gl_update(g->scanout[idx].con, x, y, width, height);
+}
+
+static void virgl_cmd_resource_flush(VirtIOGPU *g,
+                                     struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_flush rf;
+    int i;
+
+    VIRTIO_GPU_FILL_CMD(rf);
+    trace_virtio_gpu_cmd_res_flush(rf.resource_id,
+                                   rf.r.width, rf.r.height, rf.r.x, rf.r.y);
+
+    for (i = 0; i < VIRTIO_GPU_MAX_SCANOUT; i++) {
+        if (g->scanout[i].resource_id != rf.resource_id) {
+            continue;
+        }
+        virtio_gpu_rect_update(g, i, rf.r.x, rf.r.y, rf.r.width, rf.r.height);
+    }
+}
+
+static void virgl_cmd_set_scanout(VirtIOGPU *g,
+                                  struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_set_scanout ss;
+    struct virgl_renderer_resource_info info;
+    int ret;
+
+    VIRTIO_GPU_FILL_CMD(ss);
+    trace_virtio_gpu_cmd_set_scanout(ss.scanout_id, ss.resource_id,
+                                     ss.r.width, ss.r.height, ss.r.x, ss.r.y);
+
+    if (ss.scanout_id >= VIRTIO_GPU_MAX_SCANOUT) {
+        qemu_log_mask(LOG_GUEST_ERROR, "%s: illegal scanout id specified %d",
+                      __func__, ss.scanout_id);
+        cmd->error = VIRTIO_GPU_RESP_ERR_INVALID_SCANOUT_ID;
+        return;
+    }
+    g->enable = 1;
+
+    memset(&info, 0, sizeof(info));
+
+    if (ss.resource_id && ss.r.width && ss.r.height) {
+        ret = virgl_renderer_resource_get_info(ss.resource_id, &info);
+        if (ret == -1) {
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "%s: illegal resource specified %d\n",
+                          __func__, ss.resource_id);
+            cmd->error = VIRTIO_GPU_RESP_ERR_INVALID_RESOURCE_ID;
+            return;
+        }
+        qemu_console_resize(g->scanout[ss.scanout_id].con,
+                            ss.r.width, ss.r.height);
+        virgl_renderer_force_ctx_0();
+        dpy_gl_scanout(g->scanout[ss.scanout_id].con, info.tex_id,
+                       info.flags & 1 /* FIXME: Y_0_TOP */,
+                       ss.r.x, ss.r.y, ss.r.width, ss.r.height);
+    } else {
+        if (ss.scanout_id != 0) {
+            dpy_gfx_replace_surface(g->scanout[ss.scanout_id].con, NULL);
+        }
+        dpy_gl_scanout(g->scanout[ss.scanout_id].con, 0, false,
+                       0, 0, 0, 0);
+    }
+    g->scanout[ss.scanout_id].resource_id = ss.resource_id;
+}
+
+static void virgl_cmd_submit_3d(VirtIOGPU *g,
+                                struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_cmd_submit cs;
+    void *buf;
+    size_t s;
+
+    VIRTIO_GPU_FILL_CMD(cs);
+    trace_virtio_gpu_cmd_ctx_submit(cs.hdr.ctx_id, cs.size);
+
+    buf = g_malloc(cs.size);
+    s = iov_to_buf(cmd->elem.out_sg, cmd->elem.out_num,
+                   sizeof(cs), buf, cs.size);
+    if (s != cs.size) {
+        qemu_log_mask(LOG_GUEST_ERROR, "%s: size mismatch (%zd/%d)",
+                      __func__, s, cs.size);
+        cmd->error = VIRTIO_GPU_RESP_ERR_INVALID_PARAMETER;
+        return;
+    }
+
+    if (virtio_gpu_stats_enabled(g->conf)) {
+        g->stats.req_3d++;
+        g->stats.bytes_3d += cs.size;
+    }
+
+    virgl_renderer_submit_cmd(buf, cs.hdr.ctx_id, cs.size / 4);
+
+    g_free(buf);
+}
+
+static void virgl_cmd_transfer_to_host_2d(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_transfer_to_host_2d t2d;
+    struct virtio_gpu_box box;
+
+    VIRTIO_GPU_FILL_CMD(t2d);
+    trace_virtio_gpu_cmd_res_xfer_toh_2d(t2d.resource_id);
+
+    box.x = t2d.r.x;
+    box.y = t2d.r.y;
+    box.z = 0;
+    box.w = t2d.r.width;
+    box.h = t2d.r.height;
+    box.d = 1;
+
+    virgl_renderer_transfer_write_iov(t2d.resource_id,
+                                      0,
+                                      0,
+                                      0,
+                                      0,
+                                      (struct virgl_box *)&box,
+                                      t2d.offset, NULL, 0);
+}
+
+static void virgl_cmd_transfer_to_host_3d(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_transfer_host_3d t3d;
+
+    VIRTIO_GPU_FILL_CMD(t3d);
+    trace_virtio_gpu_cmd_res_xfer_toh_3d(t3d.resource_id);
+
+    virgl_renderer_transfer_write_iov(t3d.resource_id,
+                                      t3d.hdr.ctx_id,
+                                      t3d.level,
+                                      t3d.stride,
+                                      t3d.layer_stride,
+                                      (struct virgl_box *)&t3d.box,
+                                      t3d.offset, NULL, 0);
+}
+
+static void
+virgl_cmd_transfer_from_host_3d(VirtIOGPU *g,
+                                struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_transfer_host_3d tf3d;
+
+    VIRTIO_GPU_FILL_CMD(tf3d);
+    trace_virtio_gpu_cmd_res_xfer_fromh_3d(tf3d.resource_id);
+
+    virgl_renderer_transfer_read_iov(tf3d.resource_id,
+                                     tf3d.hdr.ctx_id,
+                                     tf3d.level,
+                                     tf3d.stride,
+                                     tf3d.layer_stride,
+                                     (struct virgl_box *)&tf3d.box,
+                                     tf3d.offset, NULL, 0);
+}
+
+
+static void virgl_resource_attach_backing(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_attach_backing att_rb;
+    struct iovec *res_iovs;
+    int ret;
+
+    VIRTIO_GPU_FILL_CMD(att_rb);
+    trace_virtio_gpu_cmd_res_back_attach(att_rb.resource_id);
+
+    ret = virtio_gpu_create_mapping_iov(&att_rb, cmd, &res_iovs);
+    if (ret != 0) {
+        cmd->error = VIRTIO_GPU_RESP_ERR_UNSPEC;
+        return;
+    }
+
+    virgl_renderer_resource_attach_iov(att_rb.resource_id,
+                                       res_iovs, att_rb.nr_entries);
+}
+
+static void virgl_resource_detach_backing(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_resource_detach_backing detach_rb;
+    struct iovec *res_iovs = NULL;
+    int num_iovs = 0;
+
+    VIRTIO_GPU_FILL_CMD(detach_rb);
+    trace_virtio_gpu_cmd_res_back_detach(detach_rb.resource_id);
+
+    virgl_renderer_resource_detach_iov(detach_rb.resource_id,
+                                       &res_iovs,
+                                       &num_iovs);
+    if (res_iovs == NULL || num_iovs == 0) {
+        return;
+    }
+    virtio_gpu_cleanup_mapping_iov(res_iovs, num_iovs);
+}
+
+
+static void virgl_cmd_ctx_attach_resource(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_ctx_resource att_res;
+
+    VIRTIO_GPU_FILL_CMD(att_res);
+    trace_virtio_gpu_cmd_ctx_res_attach(att_res.hdr.ctx_id,
+                                        att_res.resource_id);
+
+    virgl_renderer_ctx_attach_resource(att_res.hdr.ctx_id, att_res.resource_id);
+}
+
+static void virgl_cmd_ctx_detach_resource(VirtIOGPU *g,
+                                          struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_ctx_resource det_res;
+
+    VIRTIO_GPU_FILL_CMD(det_res);
+    trace_virtio_gpu_cmd_ctx_res_detach(det_res.hdr.ctx_id,
+                                        det_res.resource_id);
+
+    virgl_renderer_ctx_detach_resource(det_res.hdr.ctx_id, det_res.resource_id);
+}
+
+static void virgl_cmd_get_capset_info(VirtIOGPU *g,
+                                      struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_get_capset_info info;
+    struct virtio_gpu_resp_capset_info resp;
+
+    VIRTIO_GPU_FILL_CMD(info);
+
+    if (info.capset_index == 0) {
+        resp.capset_id = VIRTIO_GPU_CAPSET_VIRGL;
+        virgl_renderer_get_cap_set(resp.capset_id,
+                                   &resp.capset_max_version,
+                                   &resp.capset_max_size);
+    } else {
+        resp.capset_max_version = 0;
+        resp.capset_max_size = 0;
+    }
+    resp.hdr.type = VIRTIO_GPU_RESP_OK_CAPSET_INFO;
+    virtio_gpu_ctrl_response(g, cmd, &resp.hdr, sizeof(resp));
+}
+
+static void virgl_cmd_get_capset(VirtIOGPU *g,
+                                 struct virtio_gpu_ctrl_command *cmd)
+{
+    struct virtio_gpu_get_capset gc;
+    struct virtio_gpu_resp_capset *resp;
+    uint32_t max_ver, max_size;
+    VIRTIO_GPU_FILL_CMD(gc);
+
+    virgl_renderer_get_cap_set(gc.capset_id, &max_ver,
+                               &max_size);
+    resp = g_malloc(sizeof(*resp) + max_size);
+
+    resp->hdr.type = VIRTIO_GPU_RESP_OK_CAPSET;
+    virgl_renderer_fill_caps(gc.capset_id,
+                             gc.capset_version,
+                             (void *)resp->capset_data);
+    virtio_gpu_ctrl_response(g, cmd, &resp->hdr, sizeof(*resp) + max_size);
+    g_free(resp);
+}
+
+void virtio_gpu_virgl_process_cmd(VirtIOGPU *g,
+                                      struct virtio_gpu_ctrl_command *cmd)
+{
+    VIRTIO_GPU_FILL_CMD(cmd->cmd_hdr);
+
+    virgl_renderer_force_ctx_0();
+    switch (cmd->cmd_hdr.type) {
+    case VIRTIO_GPU_CMD_CTX_CREATE:
+        virgl_cmd_context_create(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_CTX_DESTROY:
+        virgl_cmd_context_destroy(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_RESOURCE_CREATE_2D:
+        virgl_cmd_create_resource_2d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_RESOURCE_CREATE_3D:
+        virgl_cmd_create_resource_3d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_SUBMIT_3D:
+        virgl_cmd_submit_3d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_TRANSFER_TO_HOST_2D:
+        virgl_cmd_transfer_to_host_2d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_TRANSFER_TO_HOST_3D:
+        virgl_cmd_transfer_to_host_3d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D:
+        virgl_cmd_transfer_from_host_3d(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_RESOURCE_ATTACH_BACKING:
+        virgl_resource_attach_backing(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_RESOURCE_DETACH_BACKING:
+        virgl_resource_detach_backing(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_SET_SCANOUT:
+        virgl_cmd_set_scanout(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_RESOURCE_FLUSH:
+        virgl_cmd_resource_flush(g, cmd);
+       break;
+    case VIRTIO_GPU_CMD_RESOURCE_UNREF:
+        virgl_cmd_resource_unref(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_CTX_ATTACH_RESOURCE:
+        /* TODO add security */
+        virgl_cmd_ctx_attach_resource(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_CTX_DETACH_RESOURCE:
+        /* TODO add security */
+        virgl_cmd_ctx_detach_resource(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_GET_CAPSET_INFO:
+        virgl_cmd_get_capset_info(g, cmd);
+        break;
+    case VIRTIO_GPU_CMD_GET_CAPSET:
+        virgl_cmd_get_capset(g, cmd);
+        break;
+
+    case VIRTIO_GPU_CMD_GET_DISPLAY_INFO:
+        virtio_gpu_get_display_info(g, cmd);
+        break;
+    default:
+        cmd->error = VIRTIO_GPU_RESP_ERR_UNSPEC;
+        break;
+    }
+
+    if (cmd->finished) {
+        return;
+    }
+    if (cmd->error) {
+        fprintf(stderr, "%s: ctrl 0x%x, error 0x%x\n", __func__,
+                cmd->cmd_hdr.type, cmd->error);
+        virtio_gpu_ctrl_response_nodata(g, cmd, cmd->error);
+        return;
+    }
+    if (!(cmd->cmd_hdr.flags & VIRTIO_GPU_FLAG_FENCE)) {
+        virtio_gpu_ctrl_response_nodata(g, cmd, VIRTIO_GPU_RESP_OK_NODATA);
+        return;
+    }
+
+    trace_virtio_gpu_fence_ctrl(cmd->cmd_hdr.fence_id, cmd->cmd_hdr.type);
+    virgl_renderer_create_fence(cmd->cmd_hdr.fence_id, cmd->cmd_hdr.type);
+}
+
+static void virgl_write_fence(void *opaque, uint32_t fence)
+{
+    VirtIOGPU *g = opaque;
+    struct virtio_gpu_ctrl_command *cmd, *tmp;
+
+    QTAILQ_FOREACH_SAFE(cmd, &g->fenceq, next, tmp) {
+        /*
+	 * the guest can end up emitting fences out of order
+	 * so we should check all fenced cmds not just the first one.
+	 */
+        if (cmd->cmd_hdr.fence_id > fence) {
+            continue;
+        }
+        trace_virtio_gpu_fence_resp(cmd->cmd_hdr.fence_id);
+        virtio_gpu_ctrl_response_nodata(g, cmd, VIRTIO_GPU_RESP_OK_NODATA);
+        QTAILQ_REMOVE(&g->fenceq, cmd, next);
+        g_free(cmd);
+        g->inflight--;
+        if (virtio_gpu_stats_enabled(g->conf)) {
+            fprintf(stderr, "inflight: %3d (-)\r", g->inflight);
+        }
+    }
+}
+
+static virgl_renderer_gl_context
+virgl_create_context(void *opaque, int scanout_idx,
+                     struct virgl_renderer_gl_ctx_param *params)
+{
+    VirtIOGPU *g = opaque;
+    QEMUGLContext ctx;
+    QEMUGLParams qparams;
+
+    qparams.major_ver = params->major_ver;
+    qparams.minor_ver = params->minor_ver;
+
+    ctx = dpy_gl_ctx_create(g->scanout[scanout_idx].con, &qparams);
+    return (virgl_renderer_gl_context)ctx;
+}
+
+static void virgl_destroy_context(void *opaque, virgl_renderer_gl_context ctx)
+{
+    VirtIOGPU *g = opaque;
+    QEMUGLContext qctx = (QEMUGLContext)ctx;
+
+    dpy_gl_ctx_destroy(g->scanout[0].con, qctx);
+}
+
+static int virgl_make_context_current(void *opaque, int scanout_idx,
+                                      virgl_renderer_gl_context ctx)
+{
+    VirtIOGPU *g = opaque;
+    QEMUGLContext qctx = (QEMUGLContext)ctx;
+
+    return dpy_gl_ctx_make_current(g->scanout[scanout_idx].con, qctx);
+}
+
+static struct virgl_renderer_callbacks virtio_gpu_3d_cbs = {
+    .version             = 1,
+    .write_fence         = virgl_write_fence,
+    .create_gl_context   = virgl_create_context,
+    .destroy_gl_context  = virgl_destroy_context,
+    .make_current        = virgl_make_context_current,
+};
+
+static void virtio_gpu_print_stats(void *opaque)
+{
+    VirtIOGPU *g = opaque;
+
+    if (g->stats.requests) {
+        fprintf(stderr, "stats: vq req %4d, %3d -- 3D %4d (%5d)\n",
+                g->stats.requests,
+                g->stats.max_inflight,
+                g->stats.req_3d,
+                g->stats.bytes_3d);
+        g->stats.requests     = 0;
+        g->stats.max_inflight = 0;
+        g->stats.req_3d       = 0;
+        g->stats.bytes_3d     = 0;
+    } else {
+        fprintf(stderr, "stats: idle\r");
+    }
+    timer_mod(g->print_stats, qemu_clock_get_ms(QEMU_CLOCK_VIRTUAL) + 1000);
+}
+
+static void virtio_gpu_fence_poll(void *opaque)
+{
+    VirtIOGPU *g = opaque;
+
+    virgl_renderer_poll();
+    if (g->inflight) {
+        timer_mod(g->fence_poll, qemu_clock_get_ms(QEMU_CLOCK_VIRTUAL) + 10);
+    }
+}
+
+void virtio_gpu_virgl_fence_poll(VirtIOGPU *g)
+{
+    virtio_gpu_fence_poll(g);
+}
+
+void virtio_gpu_virgl_reset(VirtIOGPU *g)
+{
+    int i;
+
+    /* virgl_renderer_reset() ??? */
+    for (i = 0; i < g->conf.max_outputs; i++) {
+        if (i != 0) {
+            dpy_gfx_replace_surface(g->scanout[i].con, NULL);
+        }
+        dpy_gl_scanout(g->scanout[i].con, 0, false, 0, 0, 0, 0);
+    }
+}
+
+int virtio_gpu_virgl_init(VirtIOGPU *g)
+{
+    int ret;
+
+    ret = virgl_renderer_init(g, 0, &virtio_gpu_3d_cbs);
+    if (ret != 0) {
+        return ret;
+    }
+
+    g->fence_poll = timer_new_ms(QEMU_CLOCK_VIRTUAL,
+                                 virtio_gpu_fence_poll, g);
+
+    if (virtio_gpu_stats_enabled(g->conf)) {
+        g->print_stats = timer_new_ms(QEMU_CLOCK_VIRTUAL,
+                                      virtio_gpu_print_stats, g);
+        timer_mod(g->print_stats, qemu_clock_get_ms(QEMU_CLOCK_VIRTUAL) + 1000);
+    }
+    return 0;
+}
+
+#endif /* CONFIG_VIRGL */
diff --git a/hw/display/virtio-gpu-pci.c b/hw/display/virtio-gpu-pci.c
index 5bc62cf..eef137f 100644
--- a/hw/display/virtio-gpu-pci.c
+++ b/hw/display/virtio-gpu-pci.c
@@ -6,8 +6,8 @@
  * Authors:
  *  Dave Airlie
  *
- * This work is licensed under the terms of the GNU GPL, version 2.  See
- * the COPYING file in the top-level directory.
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
  *
  */
 #include "hw/pci/pci.h"
diff --git a/hw/display/virtio-gpu.c b/hw/display/virtio-gpu.c
index a67d927..a836ce3 100644
--- a/hw/display/virtio-gpu.c
+++ b/hw/display/virtio-gpu.c
@@ -7,7 +7,7 @@
  *     Dave Airlie <airlied@redhat.com>
  *     Gerd Hoffmann <kraxel@redhat.com>
  *
- * This work is licensed under the terms of the GNU GPL, version 2.
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
  * See the COPYING file in the top-level directory.
  */
 
@@ -22,6 +22,23 @@
 static struct virtio_gpu_simple_resource*
 virtio_gpu_find_resource(VirtIOGPU *g, uint32_t resource_id);
 
+#ifdef CONFIG_VIRGL
+#include "virglrenderer.h"
+#define VIRGL(_g, _virgl, _simple, ...)                     \
+    do {                                                    \
+        if (_g->use_virgl_renderer) {                       \
+            _virgl(__VA_ARGS__);                            \
+        } else {                                            \
+            _simple(__VA_ARGS__);                           \
+        }                                                   \
+    } while (0)
+#else
+#define VIRGL(_g, _virgl, _simple, ...)                 \
+    do {                                                \
+        _simple(__VA_ARGS__);                           \
+    } while (0)
+#endif
+
 static void update_cursor_data_simple(VirtIOGPU *g,
                                       struct virtio_gpu_scanout *s,
                                       uint32_t resource_id)
@@ -45,16 +62,49 @@
            pixels * sizeof(uint32_t));
 }
 
+#ifdef CONFIG_VIRGL
+
+static void update_cursor_data_virgl(VirtIOGPU *g,
+                                     struct virtio_gpu_scanout *s,
+                                     uint32_t resource_id)
+{
+    uint32_t width, height;
+    uint32_t pixels, *data;
+
+    data = virgl_renderer_get_cursor_data(resource_id, &width, &height);
+    if (!data) {
+        return;
+    }
+
+    if (width != s->current_cursor->width ||
+        height != s->current_cursor->height) {
+        return;
+    }
+
+    pixels = s->current_cursor->width * s->current_cursor->height;
+    memcpy(s->current_cursor->data, data, pixels * sizeof(uint32_t));
+    free(data);
+}
+
+#endif
+
 static void update_cursor(VirtIOGPU *g, struct virtio_gpu_update_cursor *cursor)
 {
     struct virtio_gpu_scanout *s;
+    bool move = cursor->hdr.type != VIRTIO_GPU_CMD_MOVE_CURSOR;
 
     if (cursor->pos.scanout_id >= g->conf.max_outputs) {
         return;
     }
     s = &g->scanout[cursor->pos.scanout_id];
 
-    if (cursor->hdr.type != VIRTIO_GPU_CMD_MOVE_CURSOR) {
+    trace_virtio_gpu_update_cursor(cursor->pos.scanout_id,
+                                   cursor->pos.x,
+                                   cursor->pos.y,
+                                   move ? "move" : "update",
+                                   cursor->resource_id);
+
+    if (move) {
         if (!s->current_cursor) {
             s->current_cursor = cursor_alloc(64, 64);
         }
@@ -63,7 +113,8 @@
         s->current_cursor->hot_y = cursor->hot_y;
 
         if (cursor->resource_id > 0) {
-            update_cursor_data_simple(g, s, cursor->resource_id);
+            VIRGL(g, update_cursor_data_virgl, update_cursor_data_simple,
+                  g, s, cursor->resource_id);
         }
         dpy_cursor_define(s->con, s->current_cursor);
     }
@@ -92,9 +143,23 @@
 static uint64_t virtio_gpu_get_features(VirtIODevice *vdev, uint64_t features,
                                         Error **errp)
 {
+    VirtIOGPU *g = VIRTIO_GPU(vdev);
+
+    if (virtio_gpu_virgl_enabled(g->conf)) {
+        features |= (1 << VIRTIO_GPU_FEATURE_VIRGL);
+    }
     return features;
 }
 
+static void virtio_gpu_set_features(VirtIODevice *vdev, uint64_t features)
+{
+    static const uint32_t virgl = (1 << VIRTIO_GPU_FEATURE_VIRGL);
+    VirtIOGPU *g = VIRTIO_GPU(vdev);
+
+    g->use_virgl_renderer = ((features & virgl) == virgl);
+    trace_virtio_gpu_features(g->use_virgl_renderer);
+}
+
 static void virtio_gpu_notify_event(VirtIOGPU *g, uint32_t event_type)
 {
     g->virtio_config.events_read |= event_type;
@@ -563,7 +628,6 @@
                           __func__, ab->resource_id, i);
             virtio_gpu_cleanup_mapping_iov(*iov, i);
             g_free(ents);
-            g_free(*iov);
             *iov = NULL;
             return -1;
         }
@@ -580,12 +644,12 @@
         cpu_physical_memory_unmap(iov[i].iov_base, iov[i].iov_len, 1,
                                   iov[i].iov_len);
     }
+    g_free(iov);
 }
 
 static void virtio_gpu_cleanup_mapping(struct virtio_gpu_simple_resource *res)
 {
     virtio_gpu_cleanup_mapping_iov(res->iov, res->iov_cnt);
-    g_free(res->iov);
     res->iov = NULL;
     res->iov_cnt = 0;
 }
@@ -699,25 +763,43 @@
         return;
     }
 
+#ifdef CONFIG_VIRGL
+    if (!g->renderer_inited && g->use_virgl_renderer) {
+        virtio_gpu_virgl_init(g);
+        g->renderer_inited = true;
+    }
+#endif
+
     cmd = g_new(struct virtio_gpu_ctrl_command, 1);
     while (virtqueue_pop(vq, &cmd->elem)) {
         cmd->vq = vq;
         cmd->error = 0;
         cmd->finished = false;
-        g->stats.requests++;
+        if (virtio_gpu_stats_enabled(g->conf)) {
+            g->stats.requests++;
+        }
 
-        virtio_gpu_simple_process_cmd(g, cmd);
+        VIRGL(g, virtio_gpu_virgl_process_cmd, virtio_gpu_simple_process_cmd,
+              g, cmd);
         if (!cmd->finished) {
             QTAILQ_INSERT_TAIL(&g->fenceq, cmd, next);
-            g->stats.inflight++;
-            if (g->stats.max_inflight < g->stats.inflight) {
-                g->stats.max_inflight = g->stats.inflight;
+            g->inflight++;
+            if (virtio_gpu_stats_enabled(g->conf)) {
+                if (g->stats.max_inflight < g->inflight) {
+                    g->stats.max_inflight = g->inflight;
+                }
+                fprintf(stderr, "inflight: %3d (+)\r", g->inflight);
             }
-            fprintf(stderr, "inflight: %3d (+)\r", g->stats.inflight);
             cmd = g_new(struct virtio_gpu_ctrl_command, 1);
         }
     }
     g_free(cmd);
+
+#ifdef CONFIG_VIRGL
+    if (g->use_virgl_renderer) {
+        virtio_gpu_virgl_fence_poll(g);
+    }
+#endif
 }
 
 static void virtio_gpu_ctrl_bh(void *opaque)
@@ -804,6 +886,7 @@
 {
     VirtIODevice *vdev = VIRTIO_DEVICE(qdev);
     VirtIOGPU *g = VIRTIO_GPU(qdev);
+    bool have_virgl;
     int i;
 
     g->config_size = sizeof(struct virtio_gpu_config);
@@ -814,8 +897,25 @@
     g->req_state[0].width = 1024;
     g->req_state[0].height = 768;
 
-    g->ctrl_vq   = virtio_add_queue(vdev, 64, virtio_gpu_handle_ctrl_cb);
-    g->cursor_vq = virtio_add_queue(vdev, 16, virtio_gpu_handle_cursor_cb);
+    g->use_virgl_renderer = false;
+#if !defined(CONFIG_VIRGL) || defined(HOST_WORDS_BIGENDIAN)
+    have_virgl = false;
+#else
+    have_virgl = display_opengl;
+#endif
+    if (!have_virgl) {
+        g->conf.flags &= ~(1 << VIRTIO_GPU_FLAG_VIRGL_ENABLED);
+    }
+
+    if (virtio_gpu_virgl_enabled(g->conf)) {
+        /* use larger control queue in 3d mode */
+        g->ctrl_vq   = virtio_add_queue(vdev, 256, virtio_gpu_handle_ctrl_cb);
+        g->cursor_vq = virtio_add_queue(vdev, 16, virtio_gpu_handle_cursor_cb);
+        g->virtio_config.num_capsets = 1;
+    } else {
+        g->ctrl_vq   = virtio_add_queue(vdev, 64, virtio_gpu_handle_ctrl_cb);
+        g->cursor_vq = virtio_add_queue(vdev, 16, virtio_gpu_handle_cursor_cb);
+    }
 
     g->ctrl_bh = qemu_bh_new(virtio_gpu_ctrl_bh, g);
     g->cursor_bh = qemu_bh_new(virtio_gpu_cursor_bh, g);
@@ -869,10 +969,23 @@
         g->scanout[i].ds = NULL;
     }
     g->enabled_output_bitmask = 1;
+
+#ifdef CONFIG_VIRGL
+    if (g->use_virgl_renderer) {
+        virtio_gpu_virgl_reset(g);
+        g->use_virgl_renderer = 0;
+    }
+#endif
 }
 
 static Property virtio_gpu_properties[] = {
     DEFINE_PROP_UINT32("max_outputs", VirtIOGPU, conf.max_outputs, 1),
+#ifdef CONFIG_VIRGL
+    DEFINE_PROP_BIT("virgl", VirtIOGPU, conf.flags,
+                    VIRTIO_GPU_FLAG_VIRGL_ENABLED, true),
+    DEFINE_PROP_BIT("stats", VirtIOGPU, conf.flags,
+                    VIRTIO_GPU_FLAG_STATS_ENABLED, false),
+#endif
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -885,6 +998,7 @@
     vdc->get_config = virtio_gpu_get_config;
     vdc->set_config = virtio_gpu_set_config;
     vdc->get_features = virtio_gpu_get_features;
+    vdc->set_features = virtio_gpu_set_features;
 
     vdc->reset = virtio_gpu_reset;
 
@@ -917,3 +1031,14 @@
 QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resource_attach_backing) != 32);
 QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resource_detach_backing) != 32);
 QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resp_display_info)       != 408);
+
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_transfer_host_3d)        != 72);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resource_create_3d)      != 72);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_ctx_create)              != 96);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_ctx_destroy)             != 24);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_ctx_resource)            != 32);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_cmd_submit)              != 32);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_get_capset_info)         != 32);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resp_capset_info)        != 40);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_get_capset)              != 32);
+QEMU_BUILD_BUG_ON(sizeof(struct virtio_gpu_resp_capset)             != 24);
diff --git a/include/hw/virtio/virtio-gpu.h b/include/hw/virtio/virtio-gpu.h
index 8896761..9b279d7 100644
--- a/include/hw/virtio/virtio-gpu.h
+++ b/include/hw/virtio/virtio-gpu.h
@@ -56,8 +56,19 @@
     int x, y;
 };
 
+enum virtio_gpu_conf_flags {
+    VIRTIO_GPU_FLAG_VIRGL_ENABLED = 1,
+    VIRTIO_GPU_FLAG_STATS_ENABLED,
+};
+
+#define virtio_gpu_virgl_enabled(_cfg) \
+    (_cfg.flags & (1 << VIRTIO_GPU_FLAG_VIRGL_ENABLED))
+#define virtio_gpu_stats_enabled(_cfg) \
+    (_cfg.flags & (1 << VIRTIO_GPU_FLAG_STATS_ENABLED))
+
 struct virtio_gpu_conf {
     uint32_t max_outputs;
+    uint32_t flags;
 };
 
 struct virtio_gpu_ctrl_command {
@@ -92,11 +103,13 @@
     int enabled_output_bitmask;
     struct virtio_gpu_config virtio_config;
 
+    bool use_virgl_renderer;
+    bool renderer_inited;
     QEMUTimer *fence_poll;
     QEMUTimer *print_stats;
 
+    uint32_t inflight;
     struct {
-        uint32_t inflight;
         uint32_t max_inflight;
         uint32_t requests;
         uint32_t req_3d;
@@ -139,4 +152,11 @@
                                   struct iovec **iov);
 void virtio_gpu_cleanup_mapping_iov(struct iovec *iov, uint32_t count);
 
+/* virtio-gpu-3d.c */
+void virtio_gpu_virgl_process_cmd(VirtIOGPU *g,
+                                  struct virtio_gpu_ctrl_command *cmd);
+void virtio_gpu_virgl_fence_poll(VirtIOGPU *g);
+void virtio_gpu_virgl_reset(VirtIOGPU *g);
+int virtio_gpu_virgl_init(VirtIOGPU *g);
+
 #endif
diff --git a/include/standard-headers/linux/virtio_gpu.h b/include/standard-headers/linux/virtio_gpu.h
index 72ef815..76e5e52 100644
--- a/include/standard-headers/linux/virtio_gpu.h
+++ b/include/standard-headers/linux/virtio_gpu.h
@@ -40,6 +40,8 @@
 
 #include "standard-headers/linux/types.h"
 
+#define VIRTIO_GPU_FEATURE_VIRGL 0
+
 enum virtio_gpu_ctrl_type {
 	VIRTIO_GPU_UNDEFINED = 0,
 
@@ -52,6 +54,18 @@
 	VIRTIO_GPU_CMD_TRANSFER_TO_HOST_2D,
 	VIRTIO_GPU_CMD_RESOURCE_ATTACH_BACKING,
 	VIRTIO_GPU_CMD_RESOURCE_DETACH_BACKING,
+	VIRTIO_GPU_CMD_GET_CAPSET_INFO,
+	VIRTIO_GPU_CMD_GET_CAPSET,
+
+	/* 3d commands */
+	VIRTIO_GPU_CMD_CTX_CREATE = 0x0200,
+	VIRTIO_GPU_CMD_CTX_DESTROY,
+	VIRTIO_GPU_CMD_CTX_ATTACH_RESOURCE,
+	VIRTIO_GPU_CMD_CTX_DETACH_RESOURCE,
+	VIRTIO_GPU_CMD_RESOURCE_CREATE_3D,
+	VIRTIO_GPU_CMD_TRANSFER_TO_HOST_3D,
+	VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D,
+	VIRTIO_GPU_CMD_SUBMIT_3D,
 
 	/* cursor commands */
 	VIRTIO_GPU_CMD_UPDATE_CURSOR = 0x0300,
@@ -60,6 +74,8 @@
 	/* success responses */
 	VIRTIO_GPU_RESP_OK_NODATA = 0x1100,
 	VIRTIO_GPU_RESP_OK_DISPLAY_INFO,
+	VIRTIO_GPU_RESP_OK_CAPSET_INFO,
+	VIRTIO_GPU_RESP_OK_CAPSET,
 
 	/* error responses */
 	VIRTIO_GPU_RESP_ERR_UNSPEC = 0x1200,
@@ -180,13 +196,107 @@
 	} pmodes[VIRTIO_GPU_MAX_SCANOUTS];
 };
 
+/* data passed in the control vq, 3d related */
+
+struct virtio_gpu_box {
+	uint32_t x, y, z;
+	uint32_t w, h, d;
+};
+
+/* VIRTIO_GPU_CMD_TRANSFER_TO_HOST_3D, VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D */
+struct virtio_gpu_transfer_host_3d {
+	struct virtio_gpu_ctrl_hdr hdr;
+	struct virtio_gpu_box box;
+	uint64_t offset;
+	uint32_t resource_id;
+	uint32_t level;
+	uint32_t stride;
+	uint32_t layer_stride;
+};
+
+/* VIRTIO_GPU_CMD_RESOURCE_CREATE_3D */
+#define VIRTIO_GPU_RESOURCE_FLAG_Y_0_TOP (1 << 0)
+struct virtio_gpu_resource_create_3d {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t resource_id;
+	uint32_t target;
+	uint32_t format;
+	uint32_t bind;
+	uint32_t width;
+	uint32_t height;
+	uint32_t depth;
+	uint32_t array_size;
+	uint32_t last_level;
+	uint32_t nr_samples;
+	uint32_t flags;
+	uint32_t padding;
+};
+
+/* VIRTIO_GPU_CMD_CTX_CREATE */
+struct virtio_gpu_ctx_create {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t nlen;
+	uint32_t padding;
+	char debug_name[64];
+};
+
+/* VIRTIO_GPU_CMD_CTX_DESTROY */
+struct virtio_gpu_ctx_destroy {
+	struct virtio_gpu_ctrl_hdr hdr;
+};
+
+/* VIRTIO_GPU_CMD_CTX_ATTACH_RESOURCE, VIRTIO_GPU_CMD_CTX_DETACH_RESOURCE */
+struct virtio_gpu_ctx_resource {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t resource_id;
+	uint32_t padding;
+};
+
+/* VIRTIO_GPU_CMD_SUBMIT_3D */
+struct virtio_gpu_cmd_submit {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t size;
+	uint32_t padding;
+};
+
+#define VIRTIO_GPU_CAPSET_VIRGL 1
+
+/* VIRTIO_GPU_CMD_GET_CAPSET_INFO */
+struct virtio_gpu_get_capset_info {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t capset_index;
+	uint32_t padding;
+};
+
+/* VIRTIO_GPU_RESP_OK_CAPSET_INFO */
+struct virtio_gpu_resp_capset_info {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t capset_id;
+	uint32_t capset_max_version;
+	uint32_t capset_max_size;
+	uint32_t padding;
+};
+
+/* VIRTIO_GPU_CMD_GET_CAPSET */
+struct virtio_gpu_get_capset {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint32_t capset_id;
+	uint32_t capset_version;
+};
+
+/* VIRTIO_GPU_RESP_OK_CAPSET */
+struct virtio_gpu_resp_capset {
+	struct virtio_gpu_ctrl_hdr hdr;
+	uint8_t capset_data[];
+};
+
 #define VIRTIO_GPU_EVENT_DISPLAY (1 << 0)
 
 struct virtio_gpu_config {
 	uint32_t events_read;
 	uint32_t events_clear;
 	uint32_t num_scanouts;
-	uint32_t reserved;
+	uint32_t num_capsets;
 };
 
 /* simple formats for fbcon/X use */
diff --git a/include/ui/console.h b/include/ui/console.h
index 047a2b4..d887f91 100644
--- a/include/ui/console.h
+++ b/include/ui/console.h
@@ -157,6 +157,14 @@
 void cursor_get_mono_image(QEMUCursor *c, int foreground, uint8_t *mask);
 void cursor_get_mono_mask(QEMUCursor *c, int transparent, uint8_t *mask);
 
+typedef void *QEMUGLContext;
+typedef struct QEMUGLParams QEMUGLParams;
+
+struct QEMUGLParams {
+    int major_ver;
+    int minor_ver;
+};
+
 typedef struct DisplayChangeListenerOps {
     const char *dpy_name;
 
@@ -183,6 +191,21 @@
                           int x, int y, int on);
     void (*dpy_cursor_define)(DisplayChangeListener *dcl,
                               QEMUCursor *cursor);
+
+    QEMUGLContext (*dpy_gl_ctx_create)(DisplayChangeListener *dcl,
+                                       QEMUGLParams *params);
+    void (*dpy_gl_ctx_destroy)(DisplayChangeListener *dcl,
+                               QEMUGLContext ctx);
+    int (*dpy_gl_ctx_make_current)(DisplayChangeListener *dcl,
+                                   QEMUGLContext ctx);
+    QEMUGLContext (*dpy_gl_ctx_get_current)(DisplayChangeListener *dcl);
+
+    void (*dpy_gl_scanout)(DisplayChangeListener *dcl,
+                           uint32_t backing_id, bool backing_y_0_top,
+                           uint32_t x, uint32_t y, uint32_t w, uint32_t h);
+    void (*dpy_gl_update)(DisplayChangeListener *dcl,
+                          uint32_t x, uint32_t y, uint32_t w, uint32_t h);
+
 } DisplayChangeListenerOps;
 
 struct DisplayChangeListener {
@@ -244,6 +267,20 @@
 bool dpy_gfx_check_format(QemuConsole *con,
                           pixman_format_code_t format);
 
+void dpy_gl_scanout(QemuConsole *con,
+                    uint32_t backing_id, bool backing_y_0_top,
+                    uint32_t x, uint32_t y, uint32_t w, uint32_t h);
+void dpy_gl_update(QemuConsole *con,
+                   uint32_t x, uint32_t y, uint32_t w, uint32_t h);
+
+QEMUGLContext dpy_gl_ctx_create(QemuConsole *con,
+                                QEMUGLParams *params);
+void dpy_gl_ctx_destroy(QemuConsole *con, QEMUGLContext ctx);
+int dpy_gl_ctx_make_current(QemuConsole *con, QEMUGLContext ctx);
+QEMUGLContext dpy_gl_ctx_get_current(QemuConsole *con);
+
+bool console_has_gl(QemuConsole *con);
+
 static inline int surface_stride(DisplaySurface *s)
 {
     return pixman_image_get_stride(s->image);
diff --git a/include/ui/egl-context.h b/include/ui/egl-context.h
new file mode 100644
index 0000000..f004ce1
--- /dev/null
+++ b/include/ui/egl-context.h
@@ -0,0 +1,14 @@
+#ifndef EGL_CONTEXT_H
+#define EGL_CONTEXT_H
+
+#include "ui/console.h"
+#include "ui/egl-helpers.h"
+
+QEMUGLContext qemu_egl_create_context(DisplayChangeListener *dcl,
+                                      QEMUGLParams *params);
+void qemu_egl_destroy_context(DisplayChangeListener *dcl, QEMUGLContext ctx);
+int qemu_egl_make_context_current(DisplayChangeListener *dcl,
+                                  QEMUGLContext ctx);
+QEMUGLContext qemu_egl_get_current_context(DisplayChangeListener *dcl);
+
+#endif /* EGL_CONTEXT_H */
diff --git a/include/ui/gtk.h b/include/ui/gtk.h
index 0359333..bf289cf 100644
--- a/include/ui/gtk.h
+++ b/include/ui/gtk.h
@@ -20,6 +20,7 @@
 
 #if defined(CONFIG_OPENGL)
 #include "ui/egl-helpers.h"
+#include "ui/egl-context.h"
 #endif
 
 /* Compatibility define to let us build on both Gtk2 and Gtk3 */
@@ -46,6 +47,11 @@
     EGLContext ectx;
     EGLSurface esurface;
     int glupdates;
+    int x, y, w, h;
+    GLuint tex_id;
+    GLuint fbo_id;
+    bool y0_top;
+    bool scanout_mode;
 #endif
 } VirtualGfxConsole;
 
@@ -90,6 +96,39 @@
 void gd_egl_refresh(DisplayChangeListener *dcl);
 void gd_egl_switch(DisplayChangeListener *dcl,
                    DisplaySurface *surface);
+QEMUGLContext gd_egl_create_context(DisplayChangeListener *dcl,
+                                    QEMUGLParams *params);
+void gd_egl_scanout(DisplayChangeListener *dcl,
+                    uint32_t backing_id, bool backing_y_0_top,
+                    uint32_t x, uint32_t y,
+                    uint32_t w, uint32_t h);
+void gd_egl_scanout_flush(DisplayChangeListener *dcl,
+                          uint32_t x, uint32_t y, uint32_t w, uint32_t h);
 void gtk_egl_init(void);
+int gd_egl_make_current(DisplayChangeListener *dcl,
+                        QEMUGLContext ctx);
+
+/* ui/gtk-gl-area.c */
+void gd_gl_area_init(VirtualConsole *vc);
+void gd_gl_area_draw(VirtualConsole *vc);
+void gd_gl_area_update(DisplayChangeListener *dcl,
+                       int x, int y, int w, int h);
+void gd_gl_area_refresh(DisplayChangeListener *dcl);
+void gd_gl_area_switch(DisplayChangeListener *dcl,
+                       DisplaySurface *surface);
+QEMUGLContext gd_gl_area_create_context(DisplayChangeListener *dcl,
+                                        QEMUGLParams *params);
+void gd_gl_area_destroy_context(DisplayChangeListener *dcl,
+                                QEMUGLContext ctx);
+void gd_gl_area_scanout(DisplayChangeListener *dcl,
+                        uint32_t backing_id, bool backing_y_0_top,
+                        uint32_t x, uint32_t y,
+                        uint32_t w, uint32_t h);
+void gd_gl_area_scanout_flush(DisplayChangeListener *dcl,
+                              uint32_t x, uint32_t y, uint32_t w, uint32_t h);
+void gtk_gl_area_init(void);
+QEMUGLContext gd_gl_area_get_current_context(DisplayChangeListener *dcl);
+int gd_gl_area_make_current(DisplayChangeListener *dcl,
+                            QEMUGLContext ctx);
 
 #endif /* UI_GTK_H */
diff --git a/include/ui/shader.h b/include/ui/shader.h
index 8509596..f7d8618 100644
--- a/include/ui/shader.h
+++ b/include/ui/shader.h
@@ -3,7 +3,9 @@
 
 #include <epoxy/gl.h>
 
-void qemu_gl_run_texture_blit(GLint texture_blit_prog);
+GLuint qemu_gl_init_texture_blit(GLint texture_blit_prog);
+void qemu_gl_run_texture_blit(GLint texture_blit_prog,
+                              GLint texture_blit_vao);
 
 GLuint qemu_gl_create_compile_shader(GLenum type, const GLchar *src);
 GLuint qemu_gl_create_link_program(GLuint vert, GLuint frag);
diff --git a/trace-events b/trace-events
index b813ae4..a0ddc6b 100644
--- a/trace-events
+++ b/trace-events
@@ -1178,6 +1178,7 @@
 vmware_setmode(uint32_t w, uint32_t h, uint32_t bpp) "%dx%d @ %d bpp"
 
 # hw/display/virtio-gpu.c
+virtio_gpu_features(bool virgl) "virgl %d"
 virtio_gpu_cmd_get_display_info(void) ""
 virtio_gpu_cmd_get_caps(void) ""
 virtio_gpu_cmd_set_scanout(uint32_t id, uint32_t res, uint32_t w, uint32_t h, uint32_t x, uint32_t y) "id %d, res 0x%x, w %d, h %d, x %d, y %d"
@@ -1187,7 +1188,15 @@
 virtio_gpu_cmd_res_back_attach(uint32_t res) "res 0x%x"
 virtio_gpu_cmd_res_back_detach(uint32_t res) "res 0x%x"
 virtio_gpu_cmd_res_xfer_toh_2d(uint32_t res) "res 0x%x"
+virtio_gpu_cmd_res_xfer_toh_3d(uint32_t res) "res 0x%x"
+virtio_gpu_cmd_res_xfer_fromh_3d(uint32_t res) "res 0x%x"
 virtio_gpu_cmd_res_flush(uint32_t res, uint32_t w, uint32_t h, uint32_t x, uint32_t y) "res 0x%x, w %d, h %d, x %d, y %d"
+virtio_gpu_cmd_ctx_create(uint32_t ctx, const char *name) "ctx 0x%x, name %s"
+virtio_gpu_cmd_ctx_destroy(uint32_t ctx) "ctx 0x%x"
+virtio_gpu_cmd_ctx_res_attach(uint32_t ctx, uint32_t res) "ctx 0x%x, res 0x%x"
+virtio_gpu_cmd_ctx_res_detach(uint32_t ctx, uint32_t res) "ctx 0x%x, res 0x%x"
+virtio_gpu_cmd_ctx_submit(uint32_t ctx, uint32_t size) "ctx 0x%x, size %d"
+virtio_gpu_update_cursor(uint32_t scanout, uint32_t x, uint32_t y, const char *type, uint32_t res) "scanout %d, x %d, y %d, %s, res 0x%x"
 virtio_gpu_fence_ctrl(uint64_t fence, uint32_t type) "fence 0x%" PRIx64 ", type 0x%x"
 virtio_gpu_fence_resp(uint64_t fence) "fence 0x%" PRIx64
 
diff --git a/ui/Makefile.objs b/ui/Makefile.objs
index 0034fbb..728393c 100644
--- a/ui/Makefile.objs
+++ b/ui/Makefile.objs
@@ -31,11 +31,17 @@
 common-obj-y += shader.o
 common-obj-y += console-gl.o
 common-obj-y += egl-helpers.o
+common-obj-y += egl-context.o
+ifeq ($(CONFIG_GTK_GL),y)
+common-obj-$(CONFIG_GTK) += gtk-gl-area.o
+else
 common-obj-$(CONFIG_GTK) += gtk-egl.o
 endif
+endif
 
 gtk.o-cflags := $(GTK_CFLAGS) $(VTE_CFLAGS)
 gtk-egl.o-cflags := $(GTK_CFLAGS) $(VTE_CFLAGS) $(OPENGL_CFLAGS)
+gtk-gl-area.o-cflags := $(GTK_CFLAGS) $(VTE_CFLAGS) $(OPENGL_CFLAGS)
 shader.o-cflags += $(OPENGL_CFLAGS)
 console-gl.o-cflags += $(OPENGL_CFLAGS)
 egl-helpers.o-cflags += $(OPENGL_CFLAGS)
diff --git a/ui/console-gl.c b/ui/console-gl.c
index cb45cf8..baf397b 100644
--- a/ui/console-gl.c
+++ b/ui/console-gl.c
@@ -33,6 +33,7 @@
 
 struct ConsoleGLState {
     GLint texture_blit_prog;
+    GLint texture_blit_vao;
 };
 
 /* ---------------------------------------------------------------------- */
@@ -47,6 +48,9 @@
         exit(1);
     }
 
+    gls->texture_blit_vao =
+        qemu_gl_init_texture_blit(gls->texture_blit_prog);
+
     return gls;
 }
 
@@ -131,7 +135,8 @@
     glClearColor(0.1f, 0.1f, 0.1f, 0.0f);
     glClear(GL_COLOR_BUFFER_BIT);
 
-    qemu_gl_run_texture_blit(gls->texture_blit_prog);
+    qemu_gl_run_texture_blit(gls->texture_blit_prog,
+                             gls->texture_blit_vao);
 }
 
 void surface_gl_destroy_texture(ConsoleGLState *gls,
diff --git a/ui/console.c b/ui/console.c
index 75fc492..31f0d35 100644
--- a/ui/console.c
+++ b/ui/console.c
@@ -121,6 +121,7 @@
     DisplayState *ds;
     DisplaySurface *surface;
     int dcls;
+    DisplayChangeListener *gl;
 
     /* Graphic console state.  */
     Object *device;
@@ -1332,6 +1333,11 @@
     g_free(surface);
 }
 
+bool console_has_gl(QemuConsole *con)
+{
+    return con->gl != NULL;
+}
+
 void register_displaychangelistener(DisplayChangeListener *dcl)
 {
     static const char nodev[] =
@@ -1339,6 +1345,17 @@
     static DisplaySurface *dummy;
     QemuConsole *con;
 
+    if (dcl->ops->dpy_gl_ctx_create) {
+        /* display has opengl support */
+        assert(dcl->con);
+        if (dcl->con->gl) {
+            fprintf(stderr, "can't register two opengl displays (%s, %s)\n",
+                    dcl->ops->dpy_name, dcl->con->gl->ops->dpy_name);
+            exit(1);
+        }
+        dcl->con->gl = dcl;
+    }
+
     trace_displaychangelistener_register(dcl, dcl->ops->dpy_name);
     dcl->ds = get_alloc_displaystate();
     QLIST_INSERT_HEAD(&dcl->ds->listeners, dcl, next);
@@ -1417,9 +1434,13 @@
 {
     DisplayState *s = con->ds;
     DisplayChangeListener *dcl;
-    int width = surface_width(con->surface);
-    int height = surface_height(con->surface);
+    int width = w;
+    int height = h;
 
+    if (con->surface) {
+        width = surface_width(con->surface);
+        height = surface_height(con->surface);
+    }
     x = MAX(x, 0);
     y = MAX(y, 0);
     x = MIN(x, width);
@@ -1619,6 +1640,48 @@
     return false;
 }
 
+QEMUGLContext dpy_gl_ctx_create(QemuConsole *con,
+                                struct QEMUGLParams *qparams)
+{
+    assert(con->gl);
+    return con->gl->ops->dpy_gl_ctx_create(con->gl, qparams);
+}
+
+void dpy_gl_ctx_destroy(QemuConsole *con, QEMUGLContext ctx)
+{
+    assert(con->gl);
+    con->gl->ops->dpy_gl_ctx_destroy(con->gl, ctx);
+}
+
+int dpy_gl_ctx_make_current(QemuConsole *con, QEMUGLContext ctx)
+{
+    assert(con->gl);
+    return con->gl->ops->dpy_gl_ctx_make_current(con->gl, ctx);
+}
+
+QEMUGLContext dpy_gl_ctx_get_current(QemuConsole *con)
+{
+    assert(con->gl);
+    return con->gl->ops->dpy_gl_ctx_get_current(con->gl);
+}
+
+void dpy_gl_scanout(QemuConsole *con,
+                    uint32_t backing_id, bool backing_y_0_top,
+                    uint32_t x, uint32_t y, uint32_t width, uint32_t height)
+{
+    assert(con->gl);
+    con->gl->ops->dpy_gl_scanout(con->gl, backing_id,
+                                 backing_y_0_top,
+                                 x, y, width, height);
+}
+
+void dpy_gl_update(QemuConsole *con,
+                   uint32_t x, uint32_t y, uint32_t w, uint32_t h)
+{
+    assert(con->gl);
+    con->gl->ops->dpy_gl_update(con->gl, x, y, w, h);
+}
+
 /***********************************************************/
 /* register display */
 
diff --git a/ui/egl-context.c b/ui/egl-context.c
new file mode 100644
index 0000000..40102e3
--- /dev/null
+++ b/ui/egl-context.c
@@ -0,0 +1,34 @@
+#include "qemu-common.h"
+#include "ui/egl-context.h"
+
+QEMUGLContext qemu_egl_create_context(DisplayChangeListener *dcl,
+                                      QEMUGLParams *params)
+{
+   EGLContext ctx;
+   EGLint ctx_att[] = {
+      EGL_CONTEXT_CLIENT_VERSION, params->major_ver,
+      EGL_CONTEXT_MINOR_VERSION_KHR, params->minor_ver,
+      EGL_NONE
+   };
+
+   ctx = eglCreateContext(qemu_egl_display, qemu_egl_config,
+                          eglGetCurrentContext(), ctx_att);
+   return ctx;
+}
+
+void qemu_egl_destroy_context(DisplayChangeListener *dcl, QEMUGLContext ctx)
+{
+    eglDestroyContext(qemu_egl_display, ctx);
+}
+
+int qemu_egl_make_context_current(DisplayChangeListener *dcl,
+                                  QEMUGLContext ctx)
+{
+   return eglMakeCurrent(qemu_egl_display,
+                         EGL_NO_SURFACE, EGL_NO_SURFACE, ctx);
+}
+
+QEMUGLContext qemu_egl_get_current_context(DisplayChangeListener *dcl)
+{
+    return eglGetCurrentContext();
+}
diff --git a/ui/gtk-egl.c b/ui/gtk-egl.c
index 15b41f2..500c42c 100644
--- a/ui/gtk-egl.c
+++ b/ui/gtk-egl.c
@@ -21,6 +21,29 @@
 
 #include "sysemu/sysemu.h"
 
+static void gtk_egl_set_scanout_mode(VirtualConsole *vc, bool scanout)
+{
+    if (vc->gfx.scanout_mode == scanout) {
+        return;
+    }
+
+    vc->gfx.scanout_mode = scanout;
+    if (!vc->gfx.scanout_mode) {
+        if (vc->gfx.fbo_id) {
+            glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT,
+                                      GL_COLOR_ATTACHMENT0_EXT,
+                                      GL_TEXTURE_2D, 0, 0);
+            glBindFramebuffer(GL_FRAMEBUFFER_EXT, 0);
+            glDeleteFramebuffers(1, &vc->gfx.fbo_id);
+            vc->gfx.fbo_id = 0;
+        }
+        if (vc->gfx.surface) {
+            surface_gl_destroy_texture(vc->gfx.gls, vc->gfx.ds);
+            surface_gl_create_texture(vc->gfx.gls, vc->gfx.ds);
+        }
+    }
+}
+
 /** DisplayState Callbacks (opengl version) **/
 
 void gd_egl_init(VirtualConsole *vc)
@@ -50,19 +73,26 @@
     GdkWindow *window;
     int ww, wh;
 
-    if (!vc->gfx.gls || !vc->gfx.ds) {
+    if (!vc->gfx.gls) {
         return;
     }
 
-    eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
-                   vc->gfx.esurface, vc->gfx.ectx);
+    if (vc->gfx.scanout_mode) {
+        gd_egl_scanout_flush(&vc->gfx.dcl, 0, 0, vc->gfx.w, vc->gfx.h);
+    } else {
+        if (!vc->gfx.ds) {
+            return;
+        }
+        eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
+                       vc->gfx.esurface, vc->gfx.ectx);
 
-    window = gtk_widget_get_window(vc->gfx.drawing_area);
-    gdk_drawable_get_size(window, &ww, &wh);
-    surface_gl_setup_viewport(vc->gfx.gls, vc->gfx.ds, ww, wh);
-    surface_gl_render_texture(vc->gfx.gls, vc->gfx.ds);
+        window = gtk_widget_get_window(vc->gfx.drawing_area);
+        gdk_drawable_get_size(window, &ww, &wh);
+        surface_gl_setup_viewport(vc->gfx.gls, vc->gfx.ds, ww, wh);
+        surface_gl_render_texture(vc->gfx.gls, vc->gfx.ds);
 
-    eglSwapBuffers(qemu_egl_display, vc->gfx.esurface);
+        eglSwapBuffers(qemu_egl_display, vc->gfx.esurface);
+    }
 }
 
 void gd_egl_update(DisplayChangeListener *dcl,
@@ -99,6 +129,7 @@
 
     if (vc->gfx.glupdates) {
         vc->gfx.glupdates = 0;
+        gtk_egl_set_scanout_mode(vc, false);
         gd_egl_draw(vc);
     }
 }
@@ -128,6 +159,81 @@
     }
 }
 
+QEMUGLContext gd_egl_create_context(DisplayChangeListener *dcl,
+                                    QEMUGLParams *params)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
+                   vc->gfx.esurface, vc->gfx.ectx);
+    return qemu_egl_create_context(dcl, params);
+}
+
+void gd_egl_scanout(DisplayChangeListener *dcl,
+                    uint32_t backing_id, bool backing_y_0_top,
+                    uint32_t x, uint32_t y,
+                    uint32_t w, uint32_t h)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    vc->gfx.x = x;
+    vc->gfx.y = y;
+    vc->gfx.w = w;
+    vc->gfx.h = h;
+    vc->gfx.tex_id = backing_id;
+    vc->gfx.y0_top = backing_y_0_top;
+
+    eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
+                   vc->gfx.esurface, vc->gfx.ectx);
+
+    if (vc->gfx.tex_id == 0 || vc->gfx.w == 0 || vc->gfx.h == 0) {
+        gtk_egl_set_scanout_mode(vc, false);
+        return;
+    }
+
+    gtk_egl_set_scanout_mode(vc, true);
+    if (!vc->gfx.fbo_id) {
+        glGenFramebuffers(1, &vc->gfx.fbo_id);
+    }
+
+    glBindFramebuffer(GL_FRAMEBUFFER_EXT, vc->gfx.fbo_id);
+    glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT0_EXT,
+                              GL_TEXTURE_2D, vc->gfx.tex_id, 0);
+}
+
+void gd_egl_scanout_flush(DisplayChangeListener *dcl,
+                          uint32_t x, uint32_t y, uint32_t w, uint32_t h)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+    GdkWindow *window;
+    int ww, wh, y1, y2;
+
+    if (!vc->gfx.scanout_mode) {
+        return;
+    }
+    if (!vc->gfx.fbo_id) {
+        return;
+    }
+
+    eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
+                   vc->gfx.esurface, vc->gfx.ectx);
+
+    glBindFramebuffer(GL_READ_FRAMEBUFFER, vc->gfx.fbo_id);
+    glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);
+
+    window = gtk_widget_get_window(vc->gfx.drawing_area);
+    gdk_drawable_get_size(window, &ww, &wh);
+    glViewport(0, 0, ww, wh);
+    y1 = vc->gfx.y0_top ? 0 : vc->gfx.h;
+    y2 = vc->gfx.y0_top ? vc->gfx.h : 0;
+    glBlitFramebuffer(0, y1, vc->gfx.w, y2,
+                      0, 0, ww, wh,
+                      GL_COLOR_BUFFER_BIT, GL_NEAREST);
+    glBindFramebuffer(GL_FRAMEBUFFER_EXT, vc->gfx.fbo_id);
+
+    eglSwapBuffers(qemu_egl_display, vc->gfx.esurface);
+}
+
 void gtk_egl_init(void)
 {
     GdkDisplay *gdk_display = gdk_display_get_default();
@@ -139,3 +245,12 @@
 
     display_opengl = 1;
 }
+
+int gd_egl_make_current(DisplayChangeListener *dcl,
+                        QEMUGLContext ctx)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    return eglMakeCurrent(qemu_egl_display, vc->gfx.esurface,
+                          vc->gfx.esurface, ctx);
+}
diff --git a/ui/gtk-gl-area.c b/ui/gtk-gl-area.c
new file mode 100644
index 0000000..dec3edb
--- /dev/null
+++ b/ui/gtk-gl-area.c
@@ -0,0 +1,223 @@
+/*
+ * GTK UI -- glarea opengl code.
+ *
+ * Requires 3.16+ (GtkGLArea widget).
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+
+#include "qemu-common.h"
+
+#include "trace.h"
+
+#include "ui/console.h"
+#include "ui/gtk.h"
+#include "ui/egl-helpers.h"
+
+#include "sysemu/sysemu.h"
+
+static void gtk_gl_area_set_scanout_mode(VirtualConsole *vc, bool scanout)
+{
+    if (vc->gfx.scanout_mode == scanout) {
+        return;
+    }
+
+    vc->gfx.scanout_mode = scanout;
+    if (!vc->gfx.scanout_mode) {
+        if (vc->gfx.fbo_id) {
+            glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT,
+                                      GL_COLOR_ATTACHMENT0_EXT,
+                                      GL_TEXTURE_2D, 0, 0);
+            glBindFramebuffer(GL_FRAMEBUFFER_EXT, 0);
+            glDeleteFramebuffers(1, &vc->gfx.fbo_id);
+            vc->gfx.fbo_id = 0;
+        }
+        if (vc->gfx.surface) {
+            surface_gl_destroy_texture(vc->gfx.gls, vc->gfx.ds);
+            surface_gl_create_texture(vc->gfx.gls, vc->gfx.ds);
+        }
+    }
+}
+
+/** DisplayState Callbacks (opengl version) **/
+
+void gd_gl_area_draw(VirtualConsole *vc)
+{
+    int ww, wh, y1, y2;
+
+    if (!vc->gfx.gls) {
+        return;
+    }
+
+    gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+    ww = gtk_widget_get_allocated_width(vc->gfx.drawing_area);
+    wh = gtk_widget_get_allocated_height(vc->gfx.drawing_area);
+
+    if (vc->gfx.scanout_mode) {
+        if (!vc->gfx.fbo_id) {
+            return;
+        }
+
+        glBindFramebuffer(GL_READ_FRAMEBUFFER, vc->gfx.fbo_id);
+        /* GtkGLArea sets GL_DRAW_FRAMEBUFFER for us */
+
+        glViewport(0, 0, ww, wh);
+        y1 = vc->gfx.y0_top ? 0 : vc->gfx.h;
+        y2 = vc->gfx.y0_top ? vc->gfx.h : 0;
+        glBlitFramebuffer(0, y1, vc->gfx.w, y2,
+                          0, 0, ww, wh,
+                          GL_COLOR_BUFFER_BIT, GL_NEAREST);
+    } else {
+        if (!vc->gfx.ds) {
+            return;
+        }
+        gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+
+        surface_gl_setup_viewport(vc->gfx.gls, vc->gfx.ds, ww, wh);
+        surface_gl_render_texture(vc->gfx.gls, vc->gfx.ds);
+    }
+}
+
+void gd_gl_area_update(DisplayChangeListener *dcl,
+                   int x, int y, int w, int h)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    if (!vc->gfx.gls || !vc->gfx.ds) {
+        return;
+    }
+
+    gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+    surface_gl_update_texture(vc->gfx.gls, vc->gfx.ds, x, y, w, h);
+    vc->gfx.glupdates++;
+}
+
+void gd_gl_area_refresh(DisplayChangeListener *dcl)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    if (!vc->gfx.gls) {
+        if (!gtk_widget_get_realized(vc->gfx.drawing_area)) {
+            return;
+        }
+        gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+        vc->gfx.gls = console_gl_init_context();
+        if (vc->gfx.ds) {
+            surface_gl_create_texture(vc->gfx.gls, vc->gfx.ds);
+        }
+    }
+
+    graphic_hw_update(dcl->con);
+
+    if (vc->gfx.glupdates) {
+        vc->gfx.glupdates = 0;
+        gtk_gl_area_set_scanout_mode(vc, false);
+        gtk_gl_area_queue_render(GTK_GL_AREA(vc->gfx.drawing_area));
+    }
+}
+
+void gd_gl_area_switch(DisplayChangeListener *dcl,
+                       DisplaySurface *surface)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+    bool resized = true;
+
+    trace_gd_switch(vc->label, surface_width(surface), surface_height(surface));
+
+    if (vc->gfx.ds &&
+        surface_width(vc->gfx.ds) == surface_width(surface) &&
+        surface_height(vc->gfx.ds) == surface_height(surface)) {
+        resized = false;
+    }
+
+    if (vc->gfx.gls) {
+        gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+        surface_gl_destroy_texture(vc->gfx.gls, vc->gfx.ds);
+        surface_gl_create_texture(vc->gfx.gls, surface);
+    }
+    vc->gfx.ds = surface;
+
+    if (resized) {
+        gd_update_windowsize(vc);
+    }
+}
+
+QEMUGLContext gd_gl_area_create_context(DisplayChangeListener *dcl,
+                                        QEMUGLParams *params)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+    GdkWindow *window;
+    GdkGLContext *ctx;
+    GError *err = NULL;
+
+    gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+    window = gtk_widget_get_window(vc->gfx.drawing_area);
+    ctx = gdk_window_create_gl_context(window, &err);
+    gdk_gl_context_set_required_version(ctx,
+                                        params->major_ver,
+                                        params->minor_ver);
+    gdk_gl_context_realize(ctx, &err);
+    return ctx;
+}
+
+void gd_gl_area_destroy_context(DisplayChangeListener *dcl, QEMUGLContext ctx)
+{
+    /* FIXME */
+}
+
+void gd_gl_area_scanout(DisplayChangeListener *dcl,
+                        uint32_t backing_id, bool backing_y_0_top,
+                        uint32_t x, uint32_t y,
+                        uint32_t w, uint32_t h)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    vc->gfx.x = x;
+    vc->gfx.y = y;
+    vc->gfx.w = w;
+    vc->gfx.h = h;
+    vc->gfx.tex_id = backing_id;
+    vc->gfx.y0_top = backing_y_0_top;
+
+    gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
+
+    if (vc->gfx.tex_id == 0 || vc->gfx.w == 0 || vc->gfx.h == 0) {
+        gtk_gl_area_set_scanout_mode(vc, false);
+        return;
+    }
+
+    gtk_gl_area_set_scanout_mode(vc, true);
+    if (!vc->gfx.fbo_id) {
+        glGenFramebuffers(1, &vc->gfx.fbo_id);
+    }
+
+    glBindFramebuffer(GL_FRAMEBUFFER_EXT, vc->gfx.fbo_id);
+    glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT0_EXT,
+                              GL_TEXTURE_2D, vc->gfx.tex_id, 0);
+}
+
+void gd_gl_area_scanout_flush(DisplayChangeListener *dcl,
+                          uint32_t x, uint32_t y, uint32_t w, uint32_t h)
+{
+    VirtualConsole *vc = container_of(dcl, VirtualConsole, gfx.dcl);
+
+    gtk_gl_area_queue_render(GTK_GL_AREA(vc->gfx.drawing_area));
+}
+
+void gtk_gl_area_init(void)
+{
+    display_opengl = 1;
+}
+
+QEMUGLContext gd_gl_area_get_current_context(DisplayChangeListener *dcl)
+{
+    return gdk_gl_context_get_current();
+}
+
+int gd_gl_area_make_current(DisplayChangeListener *dcl,
+                            QEMUGLContext ctx)
+{
+    gdk_gl_context_make_current(ctx);
+    return 0;
+}
diff --git a/ui/gtk.c b/ui/gtk.c
index 3057cdc..2947838 100644
--- a/ui/gtk.c
+++ b/ui/gtk.c
@@ -367,6 +367,12 @@
     GtkWidget *area = vc->gfx.drawing_area;
     int ww, wh;
     gdk_drawable_get_size(gtk_widget_get_window(area), &ww, &wh);
+#if defined(CONFIG_GTK_GL)
+    if (vc->gfx.gls) {
+        gtk_gl_area_queue_render(GTK_GL_AREA(vc->gfx.drawing_area));
+        return;
+    }
+#endif
     gtk_widget_queue_draw_area(area, 0, 0, ww, wh);
 }
 
@@ -607,6 +613,27 @@
 
 /** DisplayState Callbacks (opengl version) **/
 
+#if defined(CONFIG_GTK_GL)
+
+static const DisplayChangeListenerOps dcl_gl_area_ops = {
+    .dpy_name             = "gtk-egl",
+    .dpy_gfx_update       = gd_gl_area_update,
+    .dpy_gfx_switch       = gd_gl_area_switch,
+    .dpy_gfx_check_format = console_gl_check_format,
+    .dpy_refresh          = gd_gl_area_refresh,
+    .dpy_mouse_set        = gd_mouse_set,
+    .dpy_cursor_define    = gd_cursor_define,
+
+    .dpy_gl_ctx_create       = gd_gl_area_create_context,
+    .dpy_gl_ctx_destroy      = gd_gl_area_destroy_context,
+    .dpy_gl_ctx_make_current = gd_gl_area_make_current,
+    .dpy_gl_ctx_get_current  = gd_gl_area_get_current_context,
+    .dpy_gl_scanout          = gd_gl_area_scanout,
+    .dpy_gl_update           = gd_gl_area_scanout_flush,
+};
+
+#else
+
 static const DisplayChangeListenerOps dcl_egl_ops = {
     .dpy_name             = "gtk-egl",
     .dpy_gfx_update       = gd_egl_update,
@@ -615,9 +642,17 @@
     .dpy_refresh          = gd_egl_refresh,
     .dpy_mouse_set        = gd_mouse_set,
     .dpy_cursor_define    = gd_cursor_define,
+
+    .dpy_gl_ctx_create       = gd_egl_create_context,
+    .dpy_gl_ctx_destroy      = qemu_egl_destroy_context,
+    .dpy_gl_ctx_make_current = gd_egl_make_current,
+    .dpy_gl_ctx_get_current  = qemu_egl_get_current_context,
+    .dpy_gl_scanout          = gd_egl_scanout,
+    .dpy_gl_update           = gd_egl_scanout_flush,
 };
 
-#endif
+#endif /* CONFIG_GTK_GL */
+#endif /* CONFIG_OPENGL */
 
 /** QEMU Events **/
 
@@ -667,6 +702,39 @@
     return TRUE;
 }
 
+static void gd_set_ui_info(VirtualConsole *vc, gint width, gint height)
+{
+    QemuUIInfo info;
+
+    memset(&info, 0, sizeof(info));
+    info.width = width;
+    info.height = height;
+    dpy_set_ui_info(vc->gfx.dcl.con, &info);
+}
+
+#if defined(CONFIG_GTK_GL)
+
+static gboolean gd_render_event(GtkGLArea *area, GdkGLContext *context,
+                                void *opaque)
+{
+    VirtualConsole *vc = opaque;
+
+    if (vc->gfx.gls) {
+        gd_gl_area_draw(vc);
+    }
+    return TRUE;
+}
+
+static void gd_resize_event(GtkGLArea *area,
+                            gint width, gint height, gpointer *opaque)
+{
+    VirtualConsole *vc = (void *)opaque;
+
+    gd_set_ui_info(vc, width, height);
+}
+
+#endif
+
 static gboolean gd_draw_event(GtkWidget *widget, cairo_t *cr, void *opaque)
 {
     VirtualConsole *vc = opaque;
@@ -677,8 +745,13 @@
 
 #if defined(CONFIG_OPENGL)
     if (vc->gfx.gls) {
+#if defined(CONFIG_GTK_GL)
+        /* invoke render callback please */
+        return FALSE;
+#else
         gd_egl_draw(vc);
         return TRUE;
+#endif
     }
 #endif
 
@@ -1466,12 +1539,8 @@
                              GdkEventConfigure *cfg, gpointer opaque)
 {
     VirtualConsole *vc = opaque;
-    QemuUIInfo info;
 
-    memset(&info, 0, sizeof(info));
-    info.width = cfg->width;
-    info.height = cfg->height;
-    dpy_set_ui_info(vc->gfx.dcl.con, &info);
+    gd_set_ui_info(vc, cfg->width, cfg->height);
     return FALSE;
 }
 
@@ -1628,6 +1697,15 @@
 #if GTK_CHECK_VERSION(3, 0, 0)
     g_signal_connect(vc->gfx.drawing_area, "draw",
                      G_CALLBACK(gd_draw_event), vc);
+#if defined(CONFIG_GTK_GL)
+    if (display_opengl) {
+        /* wire up GtkGlArea events */
+        g_signal_connect(vc->gfx.drawing_area, "render",
+                         G_CALLBACK(gd_render_event), vc);
+        g_signal_connect(vc->gfx.drawing_area, "resize",
+                         G_CALLBACK(gd_resize_event), vc);
+    }
+#endif
 #else
     g_signal_connect(vc->gfx.drawing_area, "expose-event",
                      G_CALLBACK(gd_expose_event), vc);
@@ -1736,7 +1814,37 @@
     vc->gfx.scale_x = 1.0;
     vc->gfx.scale_y = 1.0;
 
-    vc->gfx.drawing_area = gtk_drawing_area_new();
+#if defined(CONFIG_OPENGL)
+    if (display_opengl) {
+#if defined(CONFIG_GTK_GL)
+        vc->gfx.drawing_area = gtk_gl_area_new();
+        vc->gfx.dcl.ops = &dcl_gl_area_ops;
+#else
+        vc->gfx.drawing_area = gtk_drawing_area_new();
+        /*
+         * gtk_widget_set_double_buffered() was deprecated in 3.14.
+         * It is required for opengl rendering on X11 though.  A
+         * proper replacement (native opengl support) is only
+         * available in 3.16+.  Silence the warning if possible.
+         */
+#ifdef CONFIG_PRAGMA_DIAGNOSTIC_AVAILABLE
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
+#endif
+        gtk_widget_set_double_buffered(vc->gfx.drawing_area, FALSE);
+#ifdef CONFIG_PRAGMA_DIAGNOSTIC_AVAILABLE
+#pragma GCC diagnostic pop
+#endif
+        vc->gfx.dcl.ops = &dcl_egl_ops;
+#endif /* CONFIG_GTK_GL */
+    } else
+#endif
+    {
+        vc->gfx.drawing_area = gtk_drawing_area_new();
+        vc->gfx.dcl.ops = &dcl_ops;
+    }
+
+
     gtk_widget_add_events(vc->gfx.drawing_area,
                           GDK_POINTER_MOTION_MASK |
                           GDK_BUTTON_PRESS_MASK |
@@ -1754,29 +1862,6 @@
     gtk_notebook_append_page(GTK_NOTEBOOK(s->notebook),
                              vc->tab_item, gtk_label_new(vc->label));
 
-#if defined(CONFIG_OPENGL)
-    if (display_opengl) {
-        /*
-         * gtk_widget_set_double_buffered() was deprecated in 3.14.
-         * It is required for opengl rendering on X11 though.  A
-         * proper replacement (native opengl support) is only
-         * available in 3.16+.  Silence the warning if possible.
-         */
-#ifdef CONFIG_PRAGMA_DIAGNOSTIC_AVAILABLE
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
-#endif
-        gtk_widget_set_double_buffered(vc->gfx.drawing_area, FALSE);
-#ifdef CONFIG_PRAGMA_DIAGNOSTIC_AVAILABLE
-#pragma GCC diagnostic pop
-#endif
-        vc->gfx.dcl.ops = &dcl_egl_ops;
-    } else
-#endif
-    {
-        vc->gfx.dcl.ops = &dcl_ops;
-    }
-
     vc->gfx.dcl.con = con;
     register_displaychangelistener(&vc->gfx.dcl);
 
@@ -2059,8 +2144,12 @@
         break;
     case 1: /* on */
 #if defined(CONFIG_OPENGL)
+#if defined(CONFIG_GTK_GL)
+        gtk_gl_area_init();
+#else
         gtk_egl_init();
 #endif
+#endif
         break;
     default:
         g_assert_not_reached();
diff --git a/ui/sdl2-2d.c b/ui/sdl2-2d.c
index d0b340f..191ee3b 100644
--- a/ui/sdl2-2d.c
+++ b/ui/sdl2-2d.c
@@ -45,10 +45,23 @@
         return;
     }
 
+    /*
+     * SDL2 seems to do some double-buffering, and trying to only
+     * update the changed areas results in only one of the two buffers
+     * being updated.  Which flickers alot.  So lets not try to be
+     * clever do a full update every time ...
+     */
+#if 0
     rect.x = x;
     rect.y = y;
     rect.w = w;
     rect.h = h;
+#else
+    rect.x = 0;
+    rect.y = 0;
+    rect.w = surface_width(surf);
+    rect.h = surface_height(surf);
+#endif
 
     SDL_UpdateTexture(scon->texture, NULL, surface_data(surf),
                       surface_stride(surf));
diff --git a/ui/shader.c b/ui/shader.c
index 52a4632..0588655 100644
--- a/ui/shader.c
+++ b/ui/shader.c
@@ -29,21 +29,42 @@
 
 /* ---------------------------------------------------------------------- */
 
-void qemu_gl_run_texture_blit(GLint texture_blit_prog)
+GLuint qemu_gl_init_texture_blit(GLint texture_blit_prog)
 {
-    GLfloat in_position[] = {
+    static const GLfloat in_position[] = {
         -1, -1,
         1,  -1,
         -1,  1,
         1,   1,
     };
     GLint l_position;
+    GLuint vao, buffer;
 
-    glUseProgram(texture_blit_prog);
+    glGenVertexArrays(1, &vao);
+    glBindVertexArray(vao);
+
+    /* this is the VBO that holds the vertex data */
+    glGenBuffers(1, &buffer);
+    glBindBuffer(GL_ARRAY_BUFFER, buffer);
+    glBufferData(GL_ARRAY_BUFFER, sizeof(in_position), in_position,
+                 GL_STATIC_DRAW);
+
     l_position = glGetAttribLocation(texture_blit_prog, "in_position");
-    glVertexAttribPointer(l_position, 2, GL_FLOAT, GL_FALSE, 0, in_position);
+    glVertexAttribPointer(l_position, 2, GL_FLOAT, GL_FALSE, 0, 0);
     glEnableVertexAttribArray(l_position);
-    glDrawArrays(GL_TRIANGLE_STRIP, l_position, 4);
+
+    glBindBuffer(GL_ARRAY_BUFFER, 0);
+    glBindVertexArray(0);
+
+    return vao;
+}
+
+void qemu_gl_run_texture_blit(GLint texture_blit_prog,
+                              GLint texture_blit_vao)
+{
+    glUseProgram(texture_blit_prog);
+    glBindVertexArray(texture_blit_vao);
+    glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
 }
 
 /* ---------------------------------------------------------------------- */