diff --git a/.gitlab-ci/container/build-deqp.sh b/.gitlab-ci/container/build-deqp.sh
index 4a5a720f309..29e8e0112c6 100644
--- a/.gitlab-ci/container/build-deqp.sh
+++ b/.gitlab-ci/container/build-deqp.sh
@@ -70,6 +70,9 @@ cts_commits_to_backport=(
         # Test alpha-less 10bit formats correctly in wide_color tests
         # Fixes dEQP-EGL.functional.wide_color.*_888_colorspace_*
         dacf752adf388ce4399733ee9b4a3c8a4f317990
+
+        # KHR-GLES3.packed_pixels.*snorm
+        46158c2a1f570aab0dcefba461ddc879323367d5
 )
 
 for commit in "${cts_commits_to_backport[@]}"
diff --git a/.gitlab-ci/image-tags.yml b/.gitlab-ci/image-tags.yml
index 01df0ee7ef7..4b788ba59f4 100644
--- a/.gitlab-ci/image-tags.yml
+++ b/.gitlab-ci/image-tags.yml
@@ -21,14 +21,14 @@ variables:
    DEBIAN_X86_64_TEST_IMAGE_VK_PATH: "debian/x86_64_test-vk"
    DEBIAN_X86_64_TEST_ANDROID_IMAGE_PATH: "debian/x86_64_test-android"
 
-   DEBIAN_X86_64_TEST_ANDROID_TAG: "2023-08-20-deqp-wide_color"
-   DEBIAN_X86_64_TEST_GL_TAG: "2023-08-20-deqp-wide_color"
-   DEBIAN_X86_64_TEST_VK_TAG: "2023-08-30-fossilize"
+   DEBIAN_X86_64_TEST_ANDROID_TAG: "2023-09-01-deqp-snorm"
+   DEBIAN_X86_64_TEST_GL_TAG: "2023-09-01-deqp-snorm"
+   DEBIAN_X86_64_TEST_VK_TAG: "2023-09-01-snorm"
 
    ALPINE_X86_64_BUILD_TAG: "2023-05-01-3.18-bump-1"
    ALPINE_X86_64_LAVA_SSH_TAG: "2023-06-26-first-version"
    FEDORA_X86_64_BUILD_TAG: "2023-07-26-meson-1.2"
-   KERNEL_ROOTFS_TAG: "2023-08-30-fossilize"
+   KERNEL_ROOTFS_TAG: "2023-09-01-snorm"
    KERNEL_TAG: "v6.4.12-for-mesa-ci-f6b4ad45f48d"
 
    WINDOWS_X64_VS_PATH: "windows/x64_vs"
diff --git a/src/amd/ci/angle-radv-stoney-aco-fails.txt b/src/amd/ci/angle-radv-stoney-aco-fails.txt
index 97f7b4ff1ca..396e77ac7d5 100644
--- a/src/amd/ci/angle-radv-stoney-aco-fails.txt
+++ b/src/amd/ci/angle-radv-stoney-aco-fails.txt
@@ -1,17 +1,3 @@
-# Fails on both radv and anv:
-# "Valid format used but glReadPixels failed for input = [GL_RED, GL_BYTE] output = [GL_RGBA, GL_UNSIGNED_BYTE]"
-# Conclusion is it's an ext spec+test bug
-# https://gitlab.khronos.org/opengl/API/-/issues/180
-KHR-GLES3.packed_pixels.pbo_rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.pbo_rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.pbo_rectangle.rgba8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.rgba8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.rgba8_snorm,Fail
-
 # Failing since ANGLE exposed geometry shaders in their commit a4c7abcbc1eb
 # ("Advertise GL_OES_geometry_shader and GL_EXT_tessellation_shader"):
 #
diff --git a/src/intel/ci/angle-anv-tgl-fails.txt b/src/intel/ci/angle-anv-tgl-fails.txt
index edf9f4344d0..82a93eb1ff3 100644
--- a/src/intel/ci/angle-anv-tgl-fails.txt
+++ b/src/intel/ci/angle-anv-tgl-fails.txt
@@ -1,17 +1,3 @@
-# Fails on both radv and anv:
-# "Valid format used but glReadPixels failed for input = [GL_RED, GL_BYTE] output = [GL_RGBA, GL_UNSIGNED_BYTE]"
-# Conclusion is it's an ext spec+test bug
-# https://gitlab.khronos.org/opengl/API/-/issues/180
-KHR-GLES3.packed_pixels.pbo_rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.pbo_rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.pbo_rectangle.rgba8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.rectangle.rgba8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.r8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.rg8_snorm,Fail
-KHR-GLES3.packed_pixels.varied_rectangle.rgba8_snorm,Fail
-
 # ANGLE marks non-shadow linear-filtered textures incomplete (see GLES 3.2 spec
 # 8.17 Texture Completeness), so you get a fallback texture sample, while Mesa
 # incorrectly marks it complete.  The CTS incorrectly expects texture