diff --git a/.pick_status.json b/.pick_status.json
index 62417940e14..792cd6d0919 100644
--- a/.pick_status.json
+++ b/.pick_status.json
@@ -3484,7 +3484,7 @@
         "description": "pan/kmod: Allow mmap() on foreign buffers",
         "nominated": true,
         "nomination_type": 2,
-        "resolution": 0,
+        "resolution": 1,
         "main_sha": null,
         "because_sha": "b5e47ba8941467a4e2f53907a6645a0703852e62",
         "notes": null
diff --git a/src/panfrost/lib/kmod/panfrost_kmod.c b/src/panfrost/lib/kmod/panfrost_kmod.c
index e834b18e14d..b00c5f2853a 100644
--- a/src/panfrost/lib/kmod/panfrost_kmod.c
+++ b/src/panfrost/lib/kmod/panfrost_kmod.c
@@ -350,11 +350,16 @@ panfrost_kmod_bo_import(struct pan_kmod_dev *dev, uint32_t handle, uint64_t size
          goto err_free_bo;
       }
 
-      /* If the BO comes from a different subsystem, we don't allow
-       * mmap() to avoid the CPU-sync churn.
+      /* FIXME: If the BO comes from a different subsystem
+       * (args.extra_flags & DRM_PANTHOR_BO_IS_IMPORTED), we should normally
+       * add extra DMA_BUF_IOCTL_SYNC calls around CPU accesses to ensure the
+       * CPU mapping consistency, but this is something we never worried about
+       * (we've always assumed exporters were exposing uncached mappings with
+       * NOP {begin,end}_cpu_access() implementations), and it worked fine until
+       * now.
+       * The long term plan is to hook up DMA_BUF_IOCTL_SYNC, but this requires
+       * more work.
        */
-      if (args.extra_flags & DRM_PANFROST_BO_IS_IMPORTED)
-         flags |= PAN_KMOD_BO_FLAG_NO_MMAP;
    }
 
    pan_kmod_bo_init(&panfrost_bo->base, dev, NULL, size,
diff --git a/src/panfrost/lib/kmod/panthor_kmod.c b/src/panfrost/lib/kmod/panthor_kmod.c
index 37013dea836..c958de188eb 100644
--- a/src/panfrost/lib/kmod/panthor_kmod.c
+++ b/src/panfrost/lib/kmod/panthor_kmod.c
@@ -445,11 +445,16 @@ panthor_kmod_bo_import(struct pan_kmod_dev *dev, uint32_t handle, uint64_t size,
          goto err_free_bo;
       }
 
-      /* If the BO comes from a different subsystem, we don't allow
-       * mmap() to avoid the CPU-sync churn.
+      /* FIXME: If the BO comes from a different subsystem
+       * (args.extra_flags & DRM_PANTHOR_BO_IS_IMPORTED), we should normally
+       * add extra DMA_BUF_IOCTL_SYNC calls around CPU accesses to ensure the
+       * CPU mapping consistency, but this is something we never worried about
+       * (we've always assumed exporters were exposing uncached mappings with
+       * NOP {begin,end}_cpu_access() implementations), and it worked fine until
+       * now.
+       * The long term plan is to hook up DMA_BUF_IOCTL_SYNC, but this requires
+       * more work.
        */
-      if (args.extra_flags & DRM_PANTHOR_BO_IS_IMPORTED)
-         flags |= PAN_KMOD_BO_FLAG_NO_MMAP;
    }
 
    /* Create a unsignalled syncobj on import. Will serve as a