diff --git a/src/freedreno/common/freedreno_dev_info.h b/src/freedreno/common/freedreno_dev_info.h
index 7d9631d9b62..e2170df697c 100644
--- a/src/freedreno/common/freedreno_dev_info.h
+++ b/src/freedreno/common/freedreno_dev_info.h
@@ -247,6 +247,13 @@ struct fd_dev_info {
        * fully compatible.
        */
       bool ubwc_unorm_snorm_int_compatible;
+
+      /* Blob doesn't use hw binning with GS on all a6xx and a7xx, however
+       * in Turnip it worked without issues until a750. On a750 there are CTS
+       * failures when e.g. dEQP-VK.subgroups.arithmetic.framebuffer.* in
+       * parallel with "forcebin". It is exacerbated by using "syncdraw".
+       */
+      bool no_gs_hw_binning_quirk;
    } a7xx;
 };
 
diff --git a/src/freedreno/common/freedreno_devices.py b/src/freedreno/common/freedreno_devices.py
index 2f6aa02573a..a213cc32b1d 100644
--- a/src/freedreno/common/freedreno_devices.py
+++ b/src/freedreno/common/freedreno_devices.py
@@ -825,6 +825,7 @@ a7xx_750 = A7XXProps(
         # example dEQP-VK.image.load_store.with_format.2d.*. Disable this for
         # now.
         #supports_ibo_ubwc = True,
+        no_gs_hw_binning_quirk = True,
     )
 
 a730_magic_regs = dict(
diff --git a/src/freedreno/vulkan/tu_cmd_buffer.cc b/src/freedreno/vulkan/tu_cmd_buffer.cc
index 850603a5e7b..1edfec6a383 100644
--- a/src/freedreno/vulkan/tu_cmd_buffer.cc
+++ b/src/freedreno/vulkan/tu_cmd_buffer.cc
@@ -858,6 +858,10 @@ use_hw_binning(struct tu_cmd_buffer *cmd)
    const struct tu_framebuffer *fb = cmd->state.framebuffer;
    const struct tu_tiling_config *tiling = &fb->tiling[cmd->state.gmem_layout];
 
+   if (cmd->state.rp.has_gs &&
+       cmd->device->physical_device->info->a7xx.no_gs_hw_binning_quirk)
+      return false;
+
    /* XFB commands are emitted for BINNING || SYSMEM, which makes it
     * incompatible with non-hw binning GMEM rendering. this is required because
     * some of the XFB commands need to only be executed once.
@@ -908,17 +912,21 @@ use_sysmem_rendering(struct tu_cmd_buffer *cmd,
    if (cmd->state.rp.disable_gmem)
       return true;
 
-   /* XFB is incompatible with non-hw binning GMEM rendering, see use_hw_binning */
-   if (cmd->state.rp.xfb_used && !cmd->state.tiling->binning_possible)
-      return true;
+   if (!cmd->state.tiling->binning_possible ||
+       (cmd->state.rp.has_gs &&
+        cmd->device->physical_device->info->a7xx.no_gs_hw_binning_quirk)) {
+      /* XFB is incompatible with non-hw binning GMEM rendering, see
+       * use_hw_binning */
+      if (cmd->state.rp.xfb_used)
+         return true;
 
-   /* QUERY_TYPE_PRIMITIVES_GENERATED is incompatible with non-hw binning
-    * GMEM rendering, see use_hw_binning.
-    */
-   if ((cmd->state.rp.has_prim_generated_query_in_rp ||
-        cmd->state.prim_generated_query_running_before_rp) &&
-       !cmd->state.tiling->binning_possible)
-      return true;
+      /* QUERY_TYPE_PRIMITIVES_GENERATED is incompatible with non-hw binning
+       * GMEM rendering, see use_hw_binning.
+       */
+      if (cmd->state.rp.has_prim_generated_query_in_rp ||
+          cmd->state.prim_generated_query_running_before_rp)
+         return true;
+   }
 
    if (TU_DEBUG(GMEM))
       return false;
@@ -3183,6 +3191,9 @@ tu_pipeline_update_rp_state(struct tu_cmd_state *cmd_state)
    if (cmd_state->pipeline_has_tess) {
       cmd_state->rp.has_tess = true;
    }
+   if (cmd_state->pipeline_has_gs) {
+      cmd_state->rp.has_gs = true;
+   }
 }
 
 VKAPI_ATTR void VKAPI_CALL
@@ -3234,6 +3245,7 @@ tu_CmdBindPipeline(VkCommandBuffer commandBuffer,
    cmd->state.prim_order_gmem = pipeline->prim_order.state_gmem;
    cmd->state.pipeline_sysmem_single_prim_mode = pipeline->prim_order.sysmem_single_prim_mode;
    cmd->state.pipeline_has_tess = pipeline->active_stages & VK_SHADER_STAGE_TESSELLATION_CONTROL_BIT;
+   cmd->state.pipeline_has_gs = pipeline->active_stages & VK_SHADER_STAGE_GEOMETRY_BIT;
    cmd->state.pipeline_disable_gmem = gfx_pipeline->feedback_loop_may_involve_textures;
 
    tu_pipeline_update_rp_state(&cmd->state);
@@ -3723,6 +3735,7 @@ tu_render_pass_state_merge(struct tu_render_pass_state *dst,
 {
    dst->xfb_used |= src->xfb_used;
    dst->has_tess |= src->has_tess;
+   dst->has_gs |= src->has_gs;
    dst->has_prim_generated_query_in_rp |= src->has_prim_generated_query_in_rp;
    dst->disable_gmem |= src->disable_gmem;
    dst->sysmem_single_prim_mode |= src->sysmem_single_prim_mode;
diff --git a/src/freedreno/vulkan/tu_cmd_buffer.h b/src/freedreno/vulkan/tu_cmd_buffer.h
index 0bc3db8b92f..ab171e1ab8c 100644
--- a/src/freedreno/vulkan/tu_cmd_buffer.h
+++ b/src/freedreno/vulkan/tu_cmd_buffer.h
@@ -276,6 +276,7 @@ struct tu_render_pass_state
 {
    bool xfb_used;
    bool has_tess;
+   bool has_gs;
    bool has_prim_generated_query_in_rp;
    bool disable_gmem;
    bool sysmem_single_prim_mode;
@@ -504,6 +505,7 @@ struct tu_cmd_state
    bool pipeline_feedback_loop_ds;
    bool pipeline_sysmem_single_prim_mode;
    bool pipeline_has_tess;
+   bool pipeline_has_gs;
    bool pipeline_disable_gmem;
 
    bool pipeline_blend_lrz, pipeline_bandwidth;