Tone Equalizer does not utilize GPU

Describe the bug
I just noticed that the Tone Equalizer module does not utilize the capability of the GPU if OpenCL is activated. Instead of using the GPU it is using the CPU and, thus, slowing down the pixelpipe. I am not sure if this by intention or a limitation of this module. Below you can see which module took the longest for rendering.

To Reproduce
run darktable with darktable -d perf
edit any image and note the output in the terminal

Expected behavior

Screenshots
163,913255 [dev_process_preview] pixel pipeline processing took 0,112 secs (0,135 CPU)
164,266260 [dev_pixelpipe] took 0,351 secs (1,786 CPU) initing base buffer [full]
164,366978 [dev_pixelpipe] took 0,101 secs (0,185 CPU) processed lens correction' on GPU, blended on GPU [full] 164,369589 [dev_pixelpipe] took 0,003 secs (0,000 CPU) processed crop and rotate' on GPU, blended on GPU [full]
164,628985 [dev_pixelpipe] took 0,259 secs (1,248 CPU) processed tone equalizer' on CPU, blended on CPU [full] 164,642569 [dev_pixelpipe] took 0,014 secs (0,009 CPU) processed input color profile' on GPU, blended on GPU [full]
164,783572 [dev_pixelpipe] took 0,141 secs (0,126 CPU) processed contrast equalizer' on GPU, blended on GPU [full] 164,800374 [dev_pixelpipe] took 0,017 secs (0,008 CPU) processed sharpen' on GPU, blended on GPU [full]
164,937148 [dev_pixelpipe] took 0,137 secs (0,116 CPU) processed local contrast' on GPU, blended on GPU [full] 164,940186 [dev_pixelpipe] took 0,003 secs (0,000 CPU) processed output color profile' on GPU, blended on GPU [full]
164,958459 [dev_pixelpipe] took 0,018 secs (0,054 CPU) processed `display encoding' on CPU, blended on CPU [full]
164,960713 [dev_process_image] pixel pipeline processing took 1,159 secs (3,669 CPU)

Platform (please complete the following information):
Ubuntu 20.04
darktable 3.1.0-git1300.75abcb81c

Start by running darktable with darktable -d perf, edit an image, and inspect the pixel-pipeline timing for the tone equalizer. Trace the tone equalizer processing path and its OpenCL handling to determine whether GPU processing is supported; done means the module uses the GPU when OpenCL is enabled, with performance output confirming GPU processing.