dm_test_poll...turns_cached(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… NAME πŸ’…

okay bestie, welcome to the manual page for dm_test_poll...turns_cached(9) β€” yes, that is literally the title, no cap. The actual function here is dm_test_poll_dac_load_returns_cached, and it tests the DAC load detection shortcut. Iconic behavior, honestly.


✨ SYNOPSIS ✨

void dm_test_poll_dac_load_returns_cached (struct kunit *test );

πŸ§ͺ ARGUMENTS πŸ§ͺ

  • test β€” the KUnit test context. That's the whole argument, bestie. One (1) singular test context, serving looks.

🌸 DESCRIPTION 🌸

okay so picture this: the previous connection was already established by analog DAC load detection, and polling is not forced. In that case? The connector is not re-detected, and its cached status is returned unchanged. Literally untouched. The vibe check passes, the cache stays winning.

The connector is embedded in an amdgpu_device so that drm_to_adev resolves. That's the setup, that's the lore, that's the whole situation. 🫢

🌈 the kernel file (start here bestie)

  • ./drivers/gpu/drm/amd/display/amdgpu_dm/tests/amdgpu_dm_connector_test.c

like, that's the actual file the whole page is about β€” everything under it is the family tree πŸ’…


🧚 subconnector + content type squad

πŸŽ€ colour depth adjusting (the clock math girlies)

πŸ¦‹ output color space (so many, and honestly iconic for it)

πŸ’œ color depth from info + bpc conversions

🫢 connector type mapping era

🌸 duplicate modes, crtc masks, aspect ratios

πŸŽ€ crtc timing copy + decide

⭐ funcs reset, atomic dup state, hdr

πŸ’… set_property (putting values in)

πŸ’… get_property (getting them back out)

πŸ”₯ refresh rates + freesync video mode

🧚 cacp caps, round one

πŸ“ panel type detective work, round one

πŸ¦‹ subconnector updating, fbc init, mst detect

πŸ‘‘ finding the first crtc

🌸 panel type detective work, round two

🧚 cacp caps, round two

πŸ’– fill_stream (she fills streams, that's the whole thing)

🌈 create_stream

πŸ” detect

πŸŽ€ register / unregister / destroy / atomic check / cec

πŸ¦‹ edid handling + validation + encoders + native modes

✨ common modes + ddc + freesync modes

🧚 i2c + amd vsdb + displayid vrr

πŸ“ mode valid + cec + s3 suspend/resume

πŸ’œ the final boss: update_* and scaling


bestie real talk: that's genuinely every single one, no skips, no "etc." notes, i checked twice 😭 if a name isn't here i promise it wasn't in the original either πŸ’–

okay and like, that's the whole page!! you made it to the end, proud of you fr fr 🫢 the source is ./drivers/gpu/drm/amd/display/..., it's dated October 2026, and the whole thing is filed under the dm_test_poll...turns_cached(9) header β€” go be the main character of your own connector tests, periodt ✨🌸

πŸ“¦ where this page comes from