iwlwifi: pcie: dump error on FW reset handshake failures

If the firmware crashes while we're waiting for the reset
handshake then it cannot possibly make progress anymore,
and we will just time out the wait. That's pointless, so
just stop waiting at that point.

Additionally, if it never acknowledges the reset handshake,
something went wrong.

Dump an error in both of these cases, but we need to do it
synchronously here since the device will be turned off.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20210802170640.8b6a33544b4b.I55f97f70f8efa64db064a9207177a094c60ac8f1@changeidSigned-off-by: Luca Coelho <luciano.coelho@intel.com>

drivers/net/wireless/intel/iwlwifi/pcie/internal.h

@@ -254,6 +254,13 @@ struct cont_rec {
 };
 #endif
 
+enum iwl_pcie_fw_reset_state {
+	FW_RESET_IDLE,
+	FW_RESET_REQUESTED,
+	FW_RESET_OK,
+	FW_RESET_ERROR,
+};
+
 /**
  * struct iwl_trans_pcie - PCIe transport specific data
  * @rxq: all the RX queue data
@@ -405,7 +412,7 @@ struct iwl_trans_pcie {
 	dma_addr_t base_rb_stts_dma;
 
 	bool fw_reset_handshake;
-	bool fw_reset_done;
+	enum iwl_pcie_fw_reset_state fw_reset_state;
 	wait_queue_head_t fw_reset_waitq;
 
 	char rf_name[32];

drivers/net/wireless/intel/iwlwifi/pcie/rx.c

@@ -2228,7 +2228,13 @@ irqreturn_t iwl_pcie_irq_msix_handler(int irq, void *dev_id)
 			"Microcode SW error detected. Restarting 0x%X.\n",
 			inta_fh);
 		isr_stats->sw++;
-		iwl_pcie_irq_handle_error(trans);
+		/* during FW reset flow report errors from there */
+		if (trans_pcie->fw_reset_state == FW_RESET_REQUESTED) {
+			trans_pcie->fw_reset_state = FW_RESET_ERROR;
+			wake_up(&trans_pcie->fw_reset_waitq);
+		} else {
+			iwl_pcie_irq_handle_error(trans);
+		}
 	}
 
 	/* After checking FH register check HW register */
@@ -2296,7 +2302,7 @@ irqreturn_t iwl_pcie_irq_msix_handler(int irq, void *dev_id)
 
 	if (inta_hw & MSIX_HW_INT_CAUSES_REG_RESET_DONE) {
 		IWL_DEBUG_ISR(trans, "Reset flow completed\n");
-		trans_pcie->fw_reset_done = true;
+		trans_pcie->fw_reset_state = FW_RESET_OK;
 		wake_up(&trans_pcie->fw_reset_waitq);
 	}
 

drivers/net/wireless/intel/iwlwifi/pcie/trans-gen2.c

@@ -95,7 +95,7 @@ static void iwl_trans_pcie_fw_reset_handshake(struct iwl_trans *trans)
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	int ret;
 
-	trans_pcie->fw_reset_done = false;
+	trans_pcie->fw_reset_state = FW_RESET_REQUESTED;
 
 	if (trans->trans_cfg->device_family < IWL_DEVICE_FAMILY_AX210)
 		iwl_write_umac_prph(trans, UREG_NIC_SET_NMI_DRIVER,
@@ -106,10 +106,15 @@ static void iwl_trans_pcie_fw_reset_handshake(struct iwl_trans *trans)
 
 	/* wait 200ms */
 	ret = wait_event_timeout(trans_pcie->fw_reset_waitq,
-				 trans_pcie->fw_reset_done, FW_RESET_TIMEOUT);
-	if (!ret)
+				 trans_pcie->fw_reset_state != FW_RESET_REQUESTED,
+				 FW_RESET_TIMEOUT);
+	if (!ret || trans_pcie->fw_reset_state == FW_RESET_ERROR) {
 		IWL_INFO(trans,
 			 "firmware didn't ACK the reset - continue anyway\n");
+		iwl_trans_fw_error(trans, true);
+	}
+
+	trans_pcie->fw_reset_state = FW_RESET_IDLE;
 }
 
 void _iwl_trans_pcie_gen2_stop_device(struct iwl_trans *trans)