diff options
| author | Neeraj Sanjay Kale <[email protected]> | 2025-07-14 07:30:15 +0000 |
|---|---|---|
| committer | Luiz Augusto von Dentz <[email protected]> | 2025-07-23 14:32:06 +0000 |
| commit | 634fd53a63be4798da356dbc7251046b713d992a (patch) | |
| tree | 09707238a9fe32ffeae59b7ddf0fd250b2953ade /net/core/dev_api.c | |
| parent | Bluetooth: ISO: Support SCM_TIMESTAMPING for ISO TS (diff) | |
| download | kernel-634fd53a63be4798da356dbc7251046b713d992a.tar.gz kernel-634fd53a63be4798da356dbc7251046b713d992a.zip | |
Bluetooth: btnxpuart: Correct the Independent Reset handling after FW dump
This adds proper handling for the independent reset command sent by the
driver after FW dump is complete.
In normal scenario, the independent reset vendor command gives a success
response before jumping to bootcode.
However, when FW goes in a bad state, and sends out FW dump packets, the
independent reset command does not get any response from the controller.
[ 159.807732] Bluetooth: hci0: ==== Start FW dump ===
[ 180.759060] Bluetooth: hci0: ==== FW dump complete ===
[ 182.779208] Bluetooth: hci0: command 0xfcfc tx timeout
[ 183.364974] Bluetooth: hci0: ChipID: 7601, Version: 0
[ 183.368490] Bluetooth: hci0: Request Firmware: nxp/uartspi_n61x_v1.bin.se
[ 184.679977] Bluetooth: hci0: FW Download Complete: 417064 bytes
[ 187.963102] Bluetooth: hci0: Opcode 0x0c03 failed: -110
As a fix for such scenario, the independent reset vendor command is sent
using the __hci_cmd_send() API, which does not expect any response for
vendor commands.
__hci_cmd_send is non blocking, so before the tx_work is scheduled, it
sometimes gets canceled and 3F|FC command is never sent. Adding a small
delay after __hci_cmd_send allows the command to be sent to the
controller.
Signed-off-by: Neeraj Sanjay Kale <[email protected]>
Tested-by: Jean-Yves Salaün <[email protected]>
Signed-off-by: Luiz Augusto von Dentz <[email protected]>
Diffstat (limited to 'net/core/dev_api.c')
0 files changed, 0 insertions, 0 deletions