Hi,
I notice a regression report on Bugzilla [1]. Quoting from it:
> In linux 6.5.2, loading firmware version 83 on AX210 device appears to fail (repeatedly) with a timeout. Reverting to linux 6.4.12 successfully loads firmware 78.
>
> My nic:
> $ lspci -kd::280
> 08:00.0 Network controller: Intel Corporation Wi-Fi 6 AX210/AX211/AX411 160MHz (rev 1a)
> Subsystem: Rivet Networks Wi-Fi 6 AX210/AX211/AX411 160MHz
> Kernel driver in use: iwlwifi
> Kernel modules: iwlwifi
>
> The failing kernel:
> $ pacman -Qp /var/cache/pacman/pkg/linux-6.5.2.arch1-1-x86_64.pkg.tar.zst
> linux 6.5.2.arch1-1
>
> The error:
> $ journalctl -b -1 _KERNEL_DEVICE=+pci:${$(lspci -Dd::280)[(w)1]} + _TRANSPORT=kernel > dmesg65.log
>
> [..attached..]
>
> The interesting bit:
> Sep 09 20:30:28 kernel: iwlwifi 0000:08:00.0: WRT: Invalid buffer destination
> Sep 09 20:30:29 kernel: ------------[ cut here ]------------
> Sep 09 20:30:29 kernel: Timeout waiting for hardware access (CSR_GP_CNTRL 0xffffffff)
> Sep 09 20:30:29 kernel: WARNING: CPU: 13 PID: 679 at drivers/net/wireless/intel/iwlwifi/pcie/trans.c:2190 __iwl_trans_pcie_gr>
> Sep 09 20:30:29 kernel: Modules linked in: iwlmvm(+) snd_sof_pci_intel_tgl snd_sof_intel_hda_common soundwire_intel snd_sof_i>
> Sep 09 20:30:29 kernel: snd_hwdep intel_rapl_msr dell_smm_hwmon processor_thermal_rfim i2c_i801 realtek btmtk alienware_wmi >
> Sep 09 20:30:29 kernel: CPU: 13 PID: 679 Comm: modprobe Not tainted 6.5.2-arch1-1 #1 d2912f929551bc8e9b95af790b8285a77c25fa29
> Sep 09 20:30:29 kernel: Hardware name: Dell Inc. XPS 8950/0R6PCT, BIOS 1.2.1 03/25/2022
> [...]
> Sep 09 20:30:29 kernel: Call Trace:
> Sep 09 20:30:29 kernel: <TASK>
> Sep 09 20:30:29 kernel: ? __iwl_trans_pcie_grab_nic_access+0x14a/0x150 [iwlwifi 25a8da985d322177fdc2dbc451d4271c449a7a6f]
> Sep 09 20:30:29 kernel: ? __warn+0x81/0x130
> Sep 09 20:30:29 kernel: ? __iwl_trans_pcie_grab_nic_access+0x14a/0x150 [iwlwifi 25a8da985d322177fdc2dbc451d4271c449a7a6f]
> Sep 09 20:30:29 kernel: ? report_bug+0x171/0x1a0
> Sep 09 20:30:29 kernel: ? prb_read_valid+0x1b/0x30
> Sep 09 20:30:29 kernel: ? handle_bug+0x3c/0x80
> Sep 09 20:30:29 kernel: ? exc_invalid_op+0x17/0x70
> Sep 09 20:30:29 kernel: ? asm_exc_invalid_op+0x1a/0x20
> Sep 09 20:30:29 kernel: ? __iwl_trans_pcie_grab_nic_access+0x14a/0x150 [iwlwifi 25a8da985d322177fdc2dbc451d4271c449a7a6f]
> Sep 09 20:30:29 kernel: iwl_trans_pcie_grab_nic_access+0x1a/0x40 [iwlwifi 25a8da985d322177fdc2dbc451d4271c449a7a6f]
> Sep 09 20:30:29 kernel: iwl_read_prph+0x1d/0x60 [iwlwifi 25a8da985d322177fdc2dbc451d4271c449a7a6f]
> Sep 09 20:30:29 kernel: iwl_mvm_load_ucode_wait_alive+0x2d9/0x620 [iwlmvm 7d9113127caff2df016f1a19aad637aa20200412]
> [...]
> Sep 09 20:30:29 kernel: ---[ end trace 0000000000000000 ]---
> Sep 09 20:30:29 kernel: iwlwifi 0000:08:00.0: iwlwifi transaction failed, dumping registers
> [...]
>
> See attachment for full log.
>
> Boot after revert to 6.4.12 with working firmware:
> $ journalctl --no-hostname -b _KERNEL_DEVICE=+pci:${$(lspci -Dd::280)[(w)1]} + _TRANSPORT=kernel > dmesg64.log
>
> [..attached..]
>
> The interesting bit:
> Sep 09 20:48:23 kernel: iwlwifi 0000:08:00.0: loaded firmware version 78.3bfdc55f.0 ty-a0-gf-a0-78.ucode op_mode iwlmvm
>
>
> $ pacman -Ql linux-firmware | grep ty.a0.gf.a0
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-59.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-66.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-72.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-73.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-74.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-77.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-78.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-79.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-81.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0-83.ucode.zst
> linux-firmware /usr/lib/firmware/iwlwifi-ty-a0-gf-a0.pnvm.zst
>
See Bugzilla for the full thread and attached journal logs.
On BZ, I have asked the reporter (Cc'ed) to test the mainline.
Anyway, I'm adding this regression to regzbot:
#regzbot introduced: 399762de769c4e https://bugzilla.kernel.org/show_bug.cgi?id=217894
#regzbot title: CSR_GP_CNTRL hardware access timeout due to firmware API bump
Thanks.
[1]: https://bugzilla.kernel.org/show_bug.cgi?id=217894
--
An old man doll... just what I always wanted! - Clara
