wifi: iwlwifi: pcie: point invalid TFDs to invalid data

There are occasionally bugs which cause the device to try
to use a TFD that it wasn't supposed to, and these are
very hard to diagnose. Fill all unused TFDs with a debug
command that immediately causes an error to be detected
in these cases.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230816104355.10a9af1ca91f.Ifc790d62c52b4bc9a74c9581610af498509f5759@changeidSigned-off-by: Johannes Berg <johannes.berg@intel.com>

drivers/net/wireless/intel/iwlwifi/fw/api/debug.h

@@ -29,6 +29,11 @@ enum iwl_debug_cmds {
 	 * &struct iwl_dbg_host_event_cfg_cmd
 	 */
 	HOST_EVENT_CFG = 0x3,
+	/**
+	 * @INVALID_WR_PTR_CMD: invalid write pointer, set in the TFD
+	 *	when it's not in use
+	 */
+	INVALID_WR_PTR_CMD = 0x6,
 	/**
 	 * @DBGC_SUSPEND_RESUME:
 	 * DBGC suspend/resume commad. Uses a single dword as data:

drivers/net/wireless/intel/iwlwifi/iwl-trans.h

@@ -1069,6 +1069,7 @@ struct iwl_trans_txqs {
  * @mbx_addr_1_step: step address data 1
  * @pcie_link_speed: current PCIe link speed (%PCI_EXP_LNKSTA_CLS_*),
  *	only valid for discrete (not integrated) NICs
+ * @invalid_tx_cmd: invalid TX command buffer
  */
 struct iwl_trans {
 	bool csme_own;
@@ -1133,6 +1134,8 @@ struct iwl_trans {
 
 	u8 pcie_link_speed;
 
+	struct iwl_dma_ptr invalid_tx_cmd;
+
 	/* pointer to trans specific struct */
 	/*Ensure that this pointer will always be aligned to sizeof pointer */
 	char trans_specific[] __aligned(sizeof(void *));

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

@@ -2018,6 +2018,30 @@ void iwl_trans_pcie_free_pnvm_dram_regions(struct iwl_dram_regions *dram_regions
 	memset(desc_dram, 0, sizeof(*desc_dram));
 }
 
+static void iwl_pcie_free_invalid_tx_cmd(struct iwl_trans *trans)
+{
+	iwl_pcie_free_dma_ptr(trans, &trans->invalid_tx_cmd);
+}
+
+static int iwl_pcie_alloc_invalid_tx_cmd(struct iwl_trans *trans)
+{
+	struct iwl_cmd_header_wide bad_cmd = {
+		.cmd = INVALID_WR_PTR_CMD,
+		.group_id = DEBUG_GROUP,
+		.sequence = cpu_to_le16(0xffff),
+		.length = cpu_to_le16(0),
+		.version = 0,
+	};
+	int ret;
+
+	ret = iwl_pcie_alloc_dma_ptr(trans, &trans->invalid_tx_cmd,
+				     sizeof(bad_cmd));
+	if (ret)
+		return ret;
+	memcpy(trans->invalid_tx_cmd.addr, &bad_cmd, sizeof(bad_cmd));
+	return 0;
+}
+
 void iwl_trans_pcie_free(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
@@ -2048,6 +2072,8 @@ void iwl_trans_pcie_free(struct iwl_trans *trans)
 		iwl_pcie_free_ict(trans);
 	}
 
+	iwl_pcie_free_invalid_tx_cmd(trans);
+
 	iwl_pcie_free_fw_monitor(trans);
 
 	iwl_trans_pcie_free_pnvm_dram_regions(&trans_pcie->pnvm_data,
@@ -3684,6 +3710,9 @@ struct iwl_trans *iwl_trans_pcie_alloc(struct pci_dev *pdev,
 
 	init_waitqueue_head(&trans_pcie->sx_waitq);
 
+	ret = iwl_pcie_alloc_invalid_tx_cmd(trans);
+	if (ret)
+		goto out_no_pci;
 
 	if (trans_pcie->msix_enabled) {
 		ret = iwl_pcie_init_msix_handler(pdev, trans_pcie);

drivers/net/wireless/intel/iwlwifi/queue/tx.c

@@ -10,6 +10,7 @@
 #include "fw/api/commands.h"
 #include "fw/api/tx.h"
 #include "fw/api/datapath.h"
+#include "fw/api/debug.h"
 #include "queue/tx.h"
 #include "iwl-fh.h"
 #include "iwl-scd.h"
@@ -119,6 +120,15 @@ int iwl_txq_gen2_set_tb(struct iwl_trans *trans, struct iwl_tfh_tfd *tfd,
 	return idx;
 }
 
+static void iwl_txq_set_tfd_invalid_gen2(struct iwl_trans *trans,
+					 struct iwl_tfh_tfd *tfd)
+{
+	tfd->num_tbs = 0;
+
+	iwl_txq_gen2_set_tb(trans, tfd, trans->invalid_tx_cmd.dma,
+			    trans->invalid_tx_cmd.size);
+}
+
 void iwl_txq_gen2_tfd_unmap(struct iwl_trans *trans, struct iwl_cmd_meta *meta,
 			    struct iwl_tfh_tfd *tfd)
 {
@@ -146,7 +156,7 @@ void iwl_txq_gen2_tfd_unmap(struct iwl_trans *trans, struct iwl_cmd_meta *meta,
 					 DMA_TO_DEVICE);
 	}
 
-	tfd->num_tbs = 0;
+	iwl_txq_set_tfd_invalid_gen2(trans, tfd);
 }
 
 void iwl_txq_gen2_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq)
@@ -1025,11 +1035,21 @@ static void iwl_txq_stuck_timer(struct timer_list *t)
 	iwl_force_nmi(trans);
 }
 
+static void iwl_txq_set_tfd_invalid_gen1(struct iwl_trans *trans,
+					 struct iwl_tfd *tfd)
+{
+	tfd->num_tbs = 0;
+
+	iwl_pcie_gen1_tfd_set_tb(trans, tfd, 0, trans->invalid_tx_cmd.dma,
+				 trans->invalid_tx_cmd.size);
+}
+
 int iwl_txq_alloc(struct iwl_trans *trans, struct iwl_txq *txq, int slots_num,
 		  bool cmd_queue)
 {
-	size_t tfd_sz = trans->txqs.tfd.size *
-		trans->trans_cfg->base_params->max_tfd_queue_size;
+	size_t num_entries = trans->trans_cfg->gen2 ?
+		slots_num : trans->trans_cfg->base_params->max_tfd_queue_size;
+	size_t tfd_sz;
 	size_t tb0_buf_sz;
 	int i;
 
@@ -1039,8 +1059,7 @@ int iwl_txq_alloc(struct iwl_trans *trans, struct iwl_txq *txq, int slots_num,
 	if (WARN_ON(txq->entries || txq->tfds))
 		return -EINVAL;
 
-	if (trans->trans_cfg->gen2)
-		tfd_sz = trans->txqs.tfd.size * slots_num;
+	tfd_sz = trans->txqs.tfd.size * num_entries;
 
 	timer_setup(&txq->stuck_timer, iwl_txq_stuck_timer, 0);
 	txq->trans = trans;
@@ -1080,6 +1099,15 @@ int iwl_txq_alloc(struct iwl_trans *trans, struct iwl_txq *txq, int slots_num,
 	if (!txq->first_tb_bufs)
 		goto err_free_tfds;
 
+	for (i = 0; i < num_entries; i++) {
+		void *tfd = iwl_txq_get_tfd(trans, txq, i);
+
+		if (trans->trans_cfg->gen2)
+			iwl_txq_set_tfd_invalid_gen2(trans, tfd);
+		else
+			iwl_txq_set_tfd_invalid_gen1(trans, tfd);
+	}
+
 	return 0;
 err_free_tfds:
 	dma_free_coherent(trans->dev, tfd_sz, txq->tfds, txq->dma_addr);
@@ -1397,7 +1425,7 @@ void iwl_txq_gen1_tfd_unmap(struct iwl_trans *trans,
 
 	meta->tbs = 0;
 
-	tfd->num_tbs = 0;
+	iwl_txq_set_tfd_invalid_gen1(trans, tfd);
 }
 
 #define IWL_TX_CRC_SIZE 4