### System details

Local validation baseline; this is not the environment from the third-party C103/C106 report:

- Dell XPS 14 DA14260 (Panther Lake); XPS 16 uses the same Omarchy `linux-ptl` kernel path
- Intel BE211 / Panther Lake CNVi Wi-Fi (`8086:e440`), `iwlwifi` with `iwlmld`
- Omarchy `4.0.0rc2-1`
- `linux-ptl 7.1.8.arch1-3`
- Installed `linux-firmware-intel 20260622-1`
- Currently loaded firmware: `102.07fca168.0 sc-a0-wh-b0-c102.ucode`

### What's wrong?

#### Summary

**Tracking item:** this is not a reproduced Omarchy 7.1 defect. It tracks the future stable Linux 7.2 `linux-ptl` rebase and C106 validation for the XPS 14/16 kernel path.

The XPS 14 and XPS 16 use Omarchy's `linux-ptl` kernel. The current 7.1.8 driver supports Panther Lake SC firmware only through Core 102, so it cannot load the newer C106 firmware now shipped by Arch.

A BE211 user testing mainline Linux 7.2-rc5 reported a separate throughput degradation with C103: uplink performance fell over the lifetime of the association without kernel errors or an SKB runaway, while changing only the firmware from C103 to C106 restored performance. That report is currently a comment on #5827, but it is distinct from the original OOM tracked there.

The C103 degradation has not been reproduced on Omarchy and is structurally unreachable on Linux 7.1 because that driver never selects firmware newer than C102. Omarchy XPS users therefore are not exposed to the reported C103 regression today.

#### Why this is separate from #5827

The original #5827 failure was a rapid `skbuff_small_head` / `skbuff_head_cache` allocation storm followed by OOM kills. It strongly matches the upstream iwlmld TSO segmentation bug fixed by:

- https://github.com/torvalds/linux/commit/92cee08dc4f00e77fd1317e4343c5d458b0abab7
- Stable 7.0.y backport: https://github.com/gregkh/linux/commit/cbe1c8245e4469d1aa6e12e5d913611376d23788

That fix is present in Linux 7.0.11 and in the current `linux-ptl` 7.1.8 kernel. Any recurrence of the same SKB runaway on 7.1.8 or newer should return to #5827 with the booted-kernel banner and fresh logs.

The C106 report explicitly states that there was **no SKB runaway**. Its symptom was progressive throughput collapse, making it a separate firmware/driver tracking item:

- https://github.com/basecamp/omarchy/issues/5827#issuecomment-5263292247

#### Upstream and Arch status

- Linux 7.1 limits Panther Lake SC firmware to C102:
  https://github.com/torvalds/linux/blob/v7.1/drivers/net/wireless/intel/iwlwifi/cfg/sc.c#L13
- Linux 7.2 supports Panther Lake SC firmware through C106:
  https://github.com/torvalds/linux/blob/v7.2-rc5/drivers/net/wireless/intel/iwlwifi/cfg/sc.c#L13
- Linux 7.2 also sets the minimum supported SC core to C102:
  https://github.com/torvalds/linux/blob/v7.2-rc5/drivers/net/wireless/intel/iwlwifi/cfg/sc.c#L16
- Upstream C106 support:
  https://github.com/torvalds/linux/commit/a6136ca2dd9773d6bcd45e8290403536c9c71054
- Arch now ships `linux-firmware-intel 20260810-2` in Core, including `iwlwifi-sc-a0-wh-b0-c106.ucode`:
  https://archlinux.org/packages/core/any/linux-firmware-intel/
  https://archlinux.org/packages/core/any/linux-firmware-intel/files/

Installing the newer firmware package on Linux 7.1 is not sufficient: the 7.1 driver requests at most C102 and therefore will not select C106.

A ceiling-only C106 patch on Linux 7.1 would also retain a different accepted firmware window from upstream Linux 7.2. Taking the complete 7.2 driver through the normal rebase is safer than forcing only the final C106 constant onto 7.1.

#### Recommended implementation

1. Wait for the stable Linux 7.2 release rather than forcing C106 onto Linux 7.1.
2. Rebase `omacom-io/omarchy-pkgs` `linux-ptl` to the stable 7.2 Arch kernel base.
3. Use Arch's current `linux-firmware-intel` package; Omarchy does not need to redistribute the firmware.
4. Verify on an XPS 14, and an XPS 16 when available, that the kernel journal names `sc-a0-wh-b0-c106.ucode`. Loading C103 is a failed C106 validation, not a successful fallback result.
5. Run controlled uplink and downlink tests across at least 30 minutes of one continuous association with no reconnect. Take three 60-second runs in each direction immediately after association, then repeat that set at the end of the 30-minute window. Record association uptime, every result, median, spread, and sampled RX/TX MCS and NSS. With stable link conditions, the final-set median should remain at least 75% of the initial-set median and must not collapse toward the reported 2–22 Mbit/s range.
6. Pin and record `disable_11be` for every run. Validate the shipped Omarchy state (`disable_11be=Y`) first. Any EHT-enabled comparison should be a separate controlled arm and should not change the shipped default as part of this issue.
7. Check the journal for iwlwifi errors, firmware API rejection, SKB growth, or other regressions.

#### Completion criteria

- A stable Linux 7.2-based `linux-ptl` package is published for the XPS 14/16 path.
- The journal explicitly reports `sc-a0-wh-b0-c106.ucode` after reboot with `linux-firmware-intel >= 20260810`; a C103 load does not pass.
- The journal contains no `no suitable firmware` or `unable to support your firmware API` error.
- Initial and final sets of three 60-second runs in both directions span at least 30 minutes without reconnect. Association uptime, medians, spread, MCS/NSS samples, and pinned `disable_11be` state are recorded; under stable link conditions, the final-set median remains at least 75% of the initial-set median.
- No iwlwifi firmware errors or `skbuff_*` runaway are observed during validation.

No change in the main `basecamp/omarchy` repository is expected for the firmware support itself. The implementation belongs to the `linux-ptl` package in `omarchy-pkgs`.