https://bugzilla.kernel.org/show_bug.cgi?id=214823 Bug ID: 214823 Summary: RTL8821CE Bluetooth adapter randomly stopped working, only suspend then resume can reset it Product: Drivers Version: 2.5 Kernel Version: 5.14.14 Hardware: All OS: Linux Tree: Mainline Status: NEW Severity: high Priority: P1 Component: Bluetooth Assignee: linux-bluetooth@xxxxxxxxxxxxxxx Reporter: thanhdatwarriorok@xxxxxxxxx Regression: No Tried to downgrade rtl8821c version, bluetooth adapter still randomly stop working The only fix I found was to suspend then resume, restarting doesn't help ~ > sudo dmesg | grep 'Bluetooth: hci' [sudo] password for nellowly: [ 32.269597] Bluetooth: hci0: RTL: examining hci_ver=08 hci_rev=000c lmp_ver=08 lmp_subver=8821 [ 32.272590] Bluetooth: hci0: RTL: rom_version status=0 version=1 [ 32.272595] Bluetooth: hci0: RTL: loading rtl_bt/rtl8821c_fw.bin [ 32.272611] Bluetooth: hci0: RTL: loading rtl_bt/rtl8821c_config.bin [ 32.272626] Bluetooth: hci0: RTL: cfg_sz 10, total sz 21678 [ 32.858653] Bluetooth: hci0: RTL: fw version 0x826ca99e [11324.989363] Bluetooth: hci0: command 0x2005 tx timeout [11327.038375] Bluetooth: hci0: command 0x200b tx timeout [11329.085367] Bluetooth: hci0: command 0x200c tx timeout [11331.133356] Bluetooth: hci0: command 0x0401 tx timeout [11337.968450] Bluetooth: hci0: link tx timeout [11337.968455] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11340.030351] Bluetooth: hci0: command 0x041f tx timeout [11340.030362] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11342.078353] Bluetooth: hci0: command 0x0406 tx timeout [11342.078362] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11348.001581] Bluetooth: hci0: link tx timeout [11348.001591] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11348.003695] Bluetooth: hci0: link tx timeout [11348.003703] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11348.004799] Bluetooth: hci0: link tx timeout [11348.004807] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11348.017537] Bluetooth: hci0: link tx timeout [11348.017545] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11349.001714] Bluetooth: hci0: link tx timeout [11349.001719] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11350.077371] Bluetooth: hci0: command 0x0c24 tx timeout [11350.077383] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11352.125366] Bluetooth: hci0: command 0x0c52 tx timeout [11352.125371] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11354.173357] Bluetooth: hci0: command 0x0405 tx timeout [11354.173365] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11356.221372] Bluetooth: hci0: command 0x0408 tx timeout [11356.221377] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [11370.557366] Bluetooth: hci0: link tx timeout [11370.557376] Bluetooth: hci0: killing stalled connection 94:db:56:17:9e:c2 [11372.605355] Bluetooth: hci0: command 0x0408 tx timeout [11372.605359] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [12701.885371] Bluetooth: hci0: Timed out waiting for suspend events [12701.885380] Bluetooth: hci0: Suspend timeout bit: 6 [12701.885388] Bluetooth: hci0: command 0x0408 tx timeout [12701.885394] Bluetooth: hci0: No gpio to reset Realtek device, ignoring [12701.885406] Bluetooth: hci0: Suspend notifier action (3) failed: -110 I am not familiar with kernel's source but there is this part I found that seems revelant for at least a work around (can fast reset bluetooth adapter when it stops working). I just don't know how to patch kernel so... drivers/bluetooth/hci_h5.c /* Suspend/resume support. On many devices the RTL BT device loses power during * suspend/resume, causing it to lose its firmware and all state. So we simply * turn it off on suspend and reprobe on resume. This mirrors how RTL devices * are handled in the USB driver, where the USB_QUIRK_RESET_RESUME is used which * also causes a reprobe on resume. */ static int h5_btrtl_suspend(struct h5 *h5) { serdev_device_set_flow_control(h5->hu->serdev, false); gpiod_set_value_cansleep(h5->device_wake_gpio, 0); gpiod_set_value_cansleep(h5->enable_gpio, 0); return 0; } -- You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
