On Omarchy, when windows are floating, moving focus with
SUPER + ←/→/↑/↓ leaves the newly focused window behind the others.
ALT + TAB doesn't have this problem, because Omarchy binds that key twice:
once to cycle focus and once to raise the window
(/usr/share/omarchy/default/hypr/bindings/tiling.lua):
o.bind("ALT + TAB", "Focus on next window", hl.dsp.window.cycle_next())
On a workspace using Hyprland's scrolling layout, the keyboard resize keys
(SUPER + =, SUPER + ALT + =, SUPER + CTRL + = in Omarchy) cannot make the
last window wider. You first have to open another window to its right.
This gist fixes that with a small script and a few keybinding overrides.
Tested on Omarchy 4.0.4 with Hyprland 0.56.2 (Lua config).
| --- a/drivers/tty/serial/8250/8250_lpss.c | |
| +++ b/drivers/tty/serial/8250/8250_lpss.c | |
| @@ -45,6 +45,10 @@ | |
| #define BYT_TX_OVF_INT 0x820 | |
| #define BYT_TX_OVF_INT_MASK BIT(1) | |
| +#define BYT_PRV_RESETS 0x804 | |
| +#define BYT_PRV_RESETS_FUNC 0x3 | |
| +#define BYT_PRV_RESETS_IDMA BIT(2) | |
| + |
Codec: Cirrus Logic CS4208 (codec SSID 106b:6400) · Verified on: MacBook8,1 running
Omarchy (Arch Linux), kernel 7.2.3-arch1-3, PipeWire 1.6.8 + WirePlumber 0.5.17
Headphones work. Internal speakers are completely silent. PipeWire shows an active sink,
speaker-test exits 0, the PCM is RUNNING — and nothing comes out.
MacBook9,1 / MacBook10,1 (2016–2017)? You want the other guide: Silent internal speakers on a 12-inch MacBook (MacBook9,1 / MacBook10,1).
The MacBook8,1's Broadcom Bluetooth controller is attached to an Intel LPSS UART, not USB, and on a stock kernel it never appears: there is no hci0, bluetooth.service stays inactive and bluetoothctl reports No default controller available. This gist contains the investigation and a working fix: a small out-of-tree module plus two small patches to hci_bcm and btbcm, packaged with DKMS.
Tested on Arch Linux (Omarchy), kernel 7.2.3-arch1-3, BlueZ 5.87. After the fix the controller initialises (Broadcom, HCI 4.2, LMP subversion 0x6152, build 1288), is powered on by BlueZ, and LE scanning works.
Three separate problems stack up:
Codec: Cirrus Logic CS4208 · Verified on: MacBook10,1 (2017) running Omarchy (Arch Linux), kernel 7.2.3, PipeWire + WirePlumber 0.5.17
Headphones work. Internal speakers are completely silent. Every mixer looks fine,
PipeWire shows an active sink, speaker-test exits 0 — and nothing comes out.
MacBook8,1 (12-inch, Early 2015)? This guide does not work on your model. The leifliddy driver below is built from a MacBook9,1 capture and stays silent on the 8,1 even with the chime un-muted. See the model-specific guide instead:
macOS adjusts the Touch Bar and display brightness from the ambient light sensor. Linux doesn't, but everything needed is already exposed — this is a ~250-line userspace daemon, no root, no udev rules, no kernel patches.
Out of the box on Linux, the speakers on a T2 MacBook sound thin and harsh next to macOS. It isn't your imagination and it isn't just EQ.
These machines have four drivers — two woofers and two tweeters — and Linux plays the same full-range signal into all of them. The tweeters get bass they physically cannot reproduce, the woofers get treble, and everything is louder and nastier than it should be. macOS instead runs a real DSP stack: an FIR crossover that sends each driver only its own band, psychoacoustic bass synthesis,
Result: idle draw 23 W → ~10 W, suspend/resume still working.
Update 2026-09-19 — kernel 7.2.x needs a second patch. The amdgpu patch alone is no longer enough: on 7.2 a real suspend hangs the machine hard. See Kernel 7.2: the second patch. 7.1.8 is unaffected.
| #import "SenTestCase+MethodSwizzling.m" | |
| @interface ExampleTest : SenTestCase {} | |
| + (BOOL)trueMethod; | |
| + (BOOL)falseMethod; | |
| @end | |
| @implementation ExampleTest | |
| + (BOOL)trueMethod { return YES; } |