Some time points are reached from atomic context, e.g. FW_RSP_OR_NOTIF, TX_FAILED, FAKE_TX, and more. The debug commands of a time point are always sent synchronously, which means that we sleep in atomic contexts. The reason we didn't have issues so far is probably because we never had to send hcmds in those timepoints. Simply send the hcmd asynchrounously. Assisted-by: LLM Fixes: 3ed34fbf9d3b ("iwlwifi: dbg_ini: support FW response/notification region type") Reviewed-by: Johannes Berg Reviewed-by: Emmanuel Grumbach Signed-off-by: Miri Korenblit --- drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c b/drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c index a64f958bbe5f..6636c21f3704 100644 --- a/drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c +++ b/drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c @@ -881,6 +881,7 @@ static void iwl_dbg_tlv_send_hcmds(struct iwl_fw_runtime *fwrt, u16 hcmd_len = le32_to_cpu(node->tlv.length) - sizeof(*hcmd); struct iwl_host_cmd cmd = { .id = WIDE_ID(hcmd_data->group, hcmd_data->id), + .flags = CMD_ASYNC, .len = { hcmd_len, }, .data = { hcmd_data->data, }, }; -- 2.34.1