diff --git a/gfx/wr/webrender/src/device/gl.rs b/gfx/wr/webrender/src/device/gl.rs
index bc4a295c15bc6..625ddb39c49a6 100644
--- a/gfx/wr/webrender/src/device/gl.rs
+++ b/gfx/wr/webrender/src/device/gl.rs
@@ -985,6 +985,8 @@ pub struct Capabilities {
     pub supports_render_target_invalidate: bool,
     /// Whether the driver can reliably upload data to R8 format textures.
     pub supports_r8_texture_upload: bool,
+    /// Whether the extension QCOM_tiled_rendering is supported.
+    pub supports_qcom_tiled_rendering: bool,
     /// Whether clip-masking is supported natively by the GL implementation
     /// rather than emulated in shaders.
     pub uses_native_clip_mask: bool,
@@ -1839,6 +1841,8 @@ impl Device {
             true
         };
 
+        let supports_qcom_tiled_rendering = supports_extension(&extensions, "GL_QCOM_tiled_rendering");
+
         // On some Adreno 3xx devices the vertex array object must be unbound and rebound after
         // an attached buffer has been orphaned.
         let requires_vao_rebind_after_orphaning = is_adreno_3xx;
@@ -1876,6 +1880,7 @@ impl Device {
                 prefers_clear_scissor,
                 supports_render_target_invalidate,
                 supports_r8_texture_upload,
+                supports_qcom_tiled_rendering,
                 uses_native_clip_mask,
                 uses_native_antialiasing,
                 supports_image_external_essl3,
diff --git a/gfx/wr/webrender/src/renderer/mod.rs b/gfx/wr/webrender/src/renderer/mod.rs
index af639e13d12f4..b44c107a6636d 100644
--- a/gfx/wr/webrender/src/renderer/mod.rs
+++ b/gfx/wr/webrender/src/renderer/mod.rs
@@ -2483,6 +2483,17 @@ impl Renderer {
         {
             let _timer = self.gpu_profiler.start_timer(GPU_TAG_SETUP_TARGET);
             self.device.bind_draw_target(draw_target);
+
+            if self.device.get_capabilities().supports_qcom_tiled_rendering {
+                self.device.gl().start_tiling_qcom(
+                    target.dirty_rect.min.x.max(0) as _,
+                    target.dirty_rect.min.y.max(0) as _,
+                    target.dirty_rect.width() as _,
+                    target.dirty_rect.height() as _,
+                    0,
+                );
+            }
+
             self.device.enable_depth_write();
             self.set_blend(false, framebuffer_kind);
 
@@ -2583,6 +2594,9 @@ impl Renderer {
         }
 
         self.device.invalidate_depth_target();
+        if self.device.get_capabilities().supports_qcom_tiled_rendering {
+            self.device.gl().end_tiling_qcom(gl::COLOR_BUFFER_BIT0_QCOM);
+        }
     }
 
     /// Draw an alpha batch container into a given draw target. This is used
@@ -3349,6 +3363,21 @@ impl Renderer {
         {
             let _timer = self.gpu_profiler.start_timer(GPU_TAG_SETUP_TARGET);
             self.device.bind_draw_target(draw_target);
+
+            if self.device.get_capabilities().supports_qcom_tiled_rendering {
+                let preserve_mask = match target.clear_color {
+                    Some(_) => 0,
+                    None => gl::COLOR_BUFFER_BIT0_QCOM,
+                };
+                self.device.gl().start_tiling_qcom(
+                    target.used_rect.min.x.max(0) as _,
+                    target.used_rect.min.y.max(0) as _,
+                    target.used_rect.width() as _,
+                    target.used_rect.height() as _,
+                    preserve_mask,
+                );
+            }
+
             self.device.disable_depth();
             self.set_blend(false, framebuffer_kind);
 
@@ -3479,6 +3508,9 @@ impl Renderer {
         if clear_depth.is_some() {
             self.device.invalidate_depth_target();
         }
+        if self.device.get_capabilities().supports_qcom_tiled_rendering {
+            self.device.gl().end_tiling_qcom(gl::COLOR_BUFFER_BIT0_QCOM);
+        }
     }
 
     fn draw_blurs(