[PATCH] wifi: mt76: mt7921: fix kernel panic by accessing invalid 6GHz channel info

AngeloGioacchino Del Regno angelogioacchino.delregno at collabora.com
Tue Oct 17 07:25:24 PDT 2023


Il 17/10/23 15:04, Deren Wu ha scritto:
> From: Ming Yen Hsieh <mingyen.hsieh at mediatek.com>
> 
> When the chip not support 6GHz capability, the channels of 6GHz information
> should not be updated.
> 
> May get the stacktrace without this patch.
> <1>[   19.442078] Unable to handle kernel NULL pointer dereference at
> virtual address 0000000000000014
> <1>[   19.457535] Mem abort info:
> <1>[   19.465329]   ESR = 0x0000000096000004
> <1>[   19.473295]   EC = 0x25: DABT (current EL), IL = 32 bits
> <1>[   19.482354]   SET = 0, FnV = 0
> <1>[   19.489143]   EA = 0, S1PTW = 0
> <1>[   19.495991]   FSC = 0x04: level 0 translation fault
> <1>[   19.504554] Data abort info:
> <1>[   19.511111]   ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
> <1>[   19.520269]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> <1>[   19.528988]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> <1>[   19.537960] user pgtable: 4k pages, 48-bit VAs, pgdp=00000001027a9000
> <1>[   19.548014] [0000000000000014] pgd=0000000000000000, p4d=000000000000
> <0>[   19.558429] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
> <4>[   19.568270] Modules linked in: mt7921e mt7921_common mt792x_lib
> mt76_connac_lib mt76 mac80211 btusb btintel cfg80211 btmtk snd_sof_ipc_msg_
> btrtl snd_sof_ipc_flood_test btbcm bluetooth snd_sof_mt8195 uvcvideo
> mtk_adsp_common snd_sof_xtensa_dsp uvc snd_sof_of snd_sof videobuf2_vmalloc
> ecdh_generic ecc snd_sof_utils cros_ec_lid_angle cros_ec_sensors crct10dif_
> cros_ec_sensors_core cros_usbpd_logger crypto_user fuse ip_tables ipv6
> <4>[   19.614237] CPU: 1 PID: 105 Comm: kworker/1:1 Not tainted
> 6.6.0-rc6-next-20231017+ #324
> <4>[   19.625957] Hardware name: Acer Tomato (rev2) board (DT)
> <4>[   19.634970] Workqueue: events mt7921_init_work [mt7921_common]
> <4>[   19.644522] pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTY
> <4>[   19.655182] pc : mt7921_regd_notifier+0x180/0x290 [mt7921_common]
> <4>[   19.664983] lr : mt7921_regd_notifier+0xd4/0x290 [mt7921_common]
> <4>[   19.674679] sp : ffff800080acba80
> <4>[   19.681649] x29: ffff800080acba80 x28: 0000000000000000 x27: ffff4faf
> <4>[   19.692483] x26: 0000000000000000 x25: 0000000000000000 x24: ffff4faf
> <4>[   19.703294] x23: 00000000ffffe926 x22: ffff4faf16031fa0 x21: 00000000
> <4>[   19.714108] x20: 000000000000001c x19: ffff4faf16ba6f40 x18: 00000000
> <4>[   19.724928] x17: 0000000000000000 x16: ffffac6b891c2750 x15: ffff8000
> <4>[   19.735722] x14: 0000000000000180 x13: 0000000000000000 x12: 00000000
> <4>[   19.746478] x11: 0000000000000002 x10: ffff4faf01c21780 x9 : ffffac6b
> <4>[   19.757214] x8 : 00000000006c0000 x7 : ffffac6b6b020cf0 x6 : ffffac6b
> <4>[   19.767945] x5 : ffffac6b6b020d00 x4 : ffffac6b6b020cf8 x3 : ffff4faf
> <4>[   19.778648] x2 : 0000000000000000 x1 : 000000000000001c x0 : 00000000
> <4>[   19.789366] Call trace:
> <4>[   19.795381]  mt7921_regd_notifier+0x180/0x290 [mt7921_common]
> <4>[   19.804675]  wiphy_update_regulatory+0x2bc/0xa08 [cfg80211]
> <4>[   19.813864]  wiphy_regulatory_register+0x4c/0x88 [cfg80211]
> <4>[   19.823029]  wiphy_register+0x75c/0x8d0 [cfg80211]
> <4>[   19.831446]  ieee80211_register_hw+0x70c/0xc10 [mac80211]
> <4>[   19.840479]  mt76_register_device+0x168/0x2e8 [mt76]
> <4>[   19.849008]  mt7921_init_work+0xdc/0x250 [mt7921_common]
> <4>[   19.857817]  process_one_work+0x148/0x3c0
> <4>[   19.865292]  worker_thread+0x32c/0x450
> <4>[   19.872489]  kthread+0x11c/0x128
> <4>[   19.879173]  ret_from_fork+0x10/0x20
> <0>[   19.886153] Code: f0000041 9100a021 94000aef aa0003f9 (b9401780)
> <4>[   19.895634] ---[ end trace 0000000000000000 ]---
> 
> Reported-by: AngeloGioacchino Del Regno <angelogioacchino.delregno at collabora.com>
> Closes: https://lore.kernel.org/all/927e7d50-826d-4c92-9931-3c59b18c6945@collabora.com/
> Fixes: 09382d8f8641 ("mt7921: update the channel usage when the regd domain changed")
> Signed-off-by: Ming Yen Hsieh <mingyen.hsieh at mediatek.com>
> Signed-off-by: Deren Wu <deren.wu at mediatek.com>

I can confirm the regression is fixed with this patch.

# Acer Chromebook Spin 513 CP513-2H (MT8195)
Tested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno at collabora.com>

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno at collabora.com>

Thanks,
Angelo



More information about the Linux-mediatek mailing list