Hi Bastian, Welp... it was edited by an LLM (Gmail auto-polish and OpenCode for the coding test script). I was being lazy, but the bug is real. Since NetBird is essential for my usage, I'm migrating to another distro soon (undecided which), partly for an A/B test, on a new SSD; I will preserve the Debian 13 OS SSD for a while.
Also let me know what you want to test. I'd also be happy to share what I prompted OpenCode, the test script, and the logs, if you like this my report with help from opencode: Follow-up on the AC 9560 CNVi hard-freeze (Caps Lock blink, zero crash logs — PCH bus hang starving NVMe DMA). Hardware: Advan WorkPro (DN21C-140P3-CS), i3-1220P Alder Lake-P, 8086:51f0/8086:0034, BIOS DN21CV203, Debian Trixie, firmware iwlwifi-so-a0-jf-b0-77.ucode (v77.f92b5fed.0). Boot-time signatures unchanged on every boot: RFIm deactivated reason=4, NMI_INTERRUPT_UNKNOWN (0x84), FSEQ_ERROR_CODE 0x60000000, 0x0BADCAFE VDIG_MIRROR. Controlled isolation results (kernel 6.12.94+deb13-amd64, iwlwifi fw_restart=0, swcrypto=1 bt_coex_active=0 power_save=0 uapsd_disable=1): - NetBird (WireGuard wt0) enabled → hard freeze in ~20h (Aug 31→Sep 1), then again in ~3h16m pure idle (Sep 2). Both left zero crash logs. - NetBird disabled → 72h clean, then 5d7h clean, then 11 days (Sep 11→Sep 22) with zero freezes; every reboot in that window shows a clean shutdown sequence in the journal. Firmware counters frozen (NMI=2 boot-only, no new events). - 5-min heavy iperf3 via wlo1 (3.3 GB) and via wt0 (2.9 GB), plus 72 min burst-idle cycles, all clean — the trigger is idle/power-state timing, not throughput. Also confirmed on this hardware: enable_ini=0 is unusable (fails with -110 INIT ucode, wlo1 never registers), and BIOS C-state limit / intel_idle.max_cstate=1 had no effect. Happy to test any debug patch or provide journalctl -k boot logs / tlp-stat on request. Just reporting it, i will stop if you don't want me to continue. This is the first time in 15 years I've found a bug on a Debian OS that I couldn't solve by Googling and reading forums. I'll also make sure not to overly use LLM to polish my report features going forward. Best regards, Yongky On Tue, 8 Sept 2026 at 23:35, Debian Bug Tracking System < [email protected]> wrote: > This is an automatic notification regarding your Bug report > which was filed against the src:linux package: > > #1145628: src:linux: iwlwifi AC 9560 CNVi causes PCH bus deadlock on > concurrent TLS traffic with Kernel 6.12 > > It has been closed by Bastian Blank <[email protected]>. > > Their explanation is attached below along with your original report. > If this explanation is unsatisfactory and you have not received a > better one in a separate message then please contact Bastian Blank < > [email protected]> by > replying to this email. > > > -- > 1145628: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1145628 > Debian Bug Tracking System > Contact [email protected] with problems > > > > ---------- Forwarded message ---------- > From: Bastian Blank <[email protected]> > To: [email protected] > Cc: > Bcc: > Date: Tue, 8 Sep 2026 18:32:42 +0200 > Subject: Re: Bug#1145628: src:linux: iwlwifi AC 9560 CNVi causes PCH bus > deadlock on concurrent TLS traffic with Kernel 6.12 > Hi > > This bug report was most likely written by a LLM. It lacks all the > information we collect via the preferred way of using reportbug. We > consider this unactionable and will just close it without looking > further. > > For the future: please use "reportbug kernel". > > Bastian > > -- > Landru! Guide us! > -- A Beta 3-oid, "The Return of the Archons", stardate > 3157.4 > > > ---------- Forwarded message ---------- > From: Milk Mon <[email protected]> > To: [email protected] > Cc: > Bcc: > Date: Wed, 26 Aug 2026 09:30:54 +0700 > Subject: src:linux: iwlwifi AC 9560 CNVi causes PCH bus deadlock on > concurrent TLS traffic with Kernel 6.12 > Package: src:linux > Version: 6.12.101-1 > Severity: important > Tags: upstream > > Dear Maintainer, > > *** Summary *** > Under high-concurrency HTTPS/TLS traffic (such as multi-threaded web > scraping or parallel asset downloads), the Intel AC 9560 CNVi wireless > interface triggers a hardware microcode stall that deadlocks the internal > PCH chipset ring bus on Intel 12th Gen (Alder Lake-P) platforms. > > *** Symptoms & Absence of Crash Logs *** > - Sudden, instantaneous hard system lockup accompanied by a rhythmically > blinking Caps Lock LED. > - IMPORTANT: No kernel panic backtrace, oops, or kdump vmcore is written > to disk because the sudden freeze deadlocks the internal motherboard PCH > bus, freezing NVMe DMA write transactions before the kernel can flush the > crash logs. > - Both software watchdogs (nmi_watchdog=1, hardlockup_panic=1) and > hardware watchdogs (iTCO_wdt nowayout=1) fail to trigger an automatic > platform reboot due to the PCH bus lockup. > - Hardware integrity was thoroughly ruled out: Memtest86+ v7.20 completed > 7 full passes (0 errors across 16GB RAM) and sustained 100% CPU load across > all 12 cores held steady at 62°C without any throttling; the issue > reproduces strictly under multi-stream network traffic over iwlwifi. > > *** Hardware & Environment *** > System: Advan WorkPro Laptop > CPU: 12th Gen Intel Core i3-1220P (Alder Lake-P hybrid 2P+8E, 12T) > Kernel: Linux workpro 6.12.101+deb13-amd64 #1 SMP PREEMPT_DYNAMIC Debian > 6.12.101-1 (2026-08-05) x86_64 GNU/Linux > OS: Debian GNU/Linux 13 (Trixie) [13.0] > Installation: Fresh, clean Debian 13 installation (a few days old, > official Debian repositories only) > Wireless Device: > 00:14.3 Network controller [0280]: Intel Corporation Alder Lake-P PCH CNVi > WiFi [8086:51f0] (rev 01) > DeviceName: Onboard - Ethernet > Subsystem: Intel Corporation Dual Band Wi-Fi 5(802.11ac) Wireless-AC 9560 > 160MHz 2x2 [Jefferson Peak] [8086:0034] > Kernel driver in use: iwlwifi > Kernel modules: iwlwifi > > *** Pre-Freeze Diagnostic Log Snippets (Captured from Prior Boot > Initializations) *** > iwlwifi 0000:00:14.3: Detected Intel(R) Wireless-AC 9560 160MHz > iwlwifi 0000:00:14.3: firmware: failed to load > iwlwifi-so-a0-jf-b0-89.ucode down to 80 (-2 Not Found) > iwlwifi 0000:00:14.3: loaded fallback firmware iwlwifi-so-a0-jf-b0-77.ucode > iwlwifi 0000:00:14.3: 0x00000084 | NMI_INTERRUPT_UNKNOWN > iwlwifi 0000:00:14.3: timeout waiting for FW reset ACK (inta_hw=0x0) > iwlwifi 0000:00:14.3: Failed to start RT ucode: -110 > > *** Root Cause & Verified Workaround *** > 1. Upgraded to firmware-iwlwifi 20260622-1~bpo13+1 from trixie-backports. > 2. Configured /etc/modprobe.d/iwlwifi.conf: > options iwlwifi power_save=0 11n_disable=1 > options iwlmvm power_scheme=1 > > With 11n_disable=1, the system successfully completed 200 concurrent TLS > streams at 100% throughput with 0 freezes and rock-solid stability. > > *** Proposed Action for Debian *** > 1. Consider backporting firmware-iwlwifi 20260622 into trixie point > releases. > 2. Document or consider module parameter adjustments (11n_disable=1) for > Intel Alder Lake CNVi AC 9560 chipsets to prevent bus deadlocks under high > TLS concurrency. > > -- System Information: > Debian Release: 13.0 > Architecture: x86_64 > Kernel: 6.12.101+deb13-amd64 >
