dm_test_han..._vmax_update(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… NAME πŸ’…

okay so bestie, this page is all about dm_test_handle_vmin_vmax_update β€” and yes, the truncated title up top (dm_test_han..._vmax_update(9)) is just the page name being ✨dramatic✨. We're in the Kernel API Manual today, and the whole thing is literally "Test the deferred vmin/vmax worker." That's the vibe. πŸŽ€


✨ SYNOPSIS ✨

void dm_test_handle_vmin_vmax_update (struct kunit *test );

That's it. That's the whole call. One argument, no return value, no cap. 🫢


πŸ¦‹ ARGUMENTS πŸ¦‹

  • test β€” the KUnit test context. The main character of this function signature ngl, it's what gives the test its little identity. πŸ’«

🌸 DESCRIPTION 🌸

So here's the ✨lore✨, bestie. This worker does three things, and it does them in order, and we stan a queen with a routine πŸ‘‘:

  • it applies the cached timing adjust to the stream via dc_stream_adjust_vmin_vmax
  • it drops the stream reference taken when the work was scheduled
  • it frees the work and its adjust copy

Now, the trick that makes this testable without dragging real hardware into the group chat: a fake dc with a current_state lets the adjust walk an empty pipe list β€” like, no pipes, nothing to trip over, and crucially without touching hardware. Slay. Hardware stays unbothered. πŸ”₯

And one more detail that is lowkey the whole point: the work and the adjust are kmalloc'd because the worker frees them. Like, it will literally free them, so they cannot live on the stack. Not me stressing this β€” the manpage stresses it. It's giving "whoever allocates it, the worker deallocates it." πŸ’€


🌈 SEE ALSO 🌈

πŸŽ€ Kernel file πŸŽ€

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

🧚 The rest of the squad 🧚

like, the full friend group, no one left out, fr fr:


k that's the whole page bestie, nothing skipped, every single test name accounted for πŸ’–

and like, a lil sign-off before you close the tab: this whole thing ships out of ./drivers/gpu/drm/amd/display/..., dated October 2026, and it's the dm_test_han..._vmax_update(9) page β€” periodt. πŸ¦‹βœ¨πŸŽ€

πŸ“¦ where this page comes from