drivers: qcom: rpmh-rsc: output debug information from RSC

Output the TCS state when the debug api is invoked. The state of the
TCS, the contents and the IRQ status is presented. Additionally, crash
the system if any TCS is busy to help with the debug.

Change-Id: I7ef5480cc6da7208d5f82351d863561fcc1abf68
Signed-off-by: Raju P.L.S.S.S.N <rplsssn@codeaurora.org>
Signed-off-by: Lina Iyer <ilina@codeaurora.org>

drivers/soc/qcom/rpmh-internal.h

@@ -93,6 +93,7 @@ struct rpmh_ctrlr {
  * @tcs_in_use: s/w state of the TCS
  * @lock:       synchronize state of the controller
  * @client:     handle to the DRV's client.
+ * @irq:        IRQ at gic
  */
 struct rsc_drv {
 	const char *name;
@@ -105,6 +106,7 @@ struct rsc_drv {
 	DECLARE_BITMAP(tcs_in_use, MAX_TCS_NR);
 	spinlock_t lock;
 	struct rpmh_ctrlr client;
+	int irq;
 };
 
 int rpmh_rsc_send_data(struct rsc_drv *drv, const struct tcs_request *msg);
@@ -117,4 +119,5 @@ int rpmh_rsc_write_pdc_data(struct rsc_drv *drv, const struct tcs_request *msg);
 
 void rpmh_tx_done(const struct tcs_request *msg, int r);
 
+void rpmh_rsc_debug(struct rsc_drv *drv);
 #endif /* __RPM_INTERNAL_H__ */

drivers/soc/qcom/rpmh-rsc.c

@@ -65,6 +65,16 @@
 #define RSC_PDC_DRV_DATA		0x38
 #define RSC_PDC_DATA_OFFSET		0x08
 
+#define ACCL_TYPE(addr)			((addr >> 16) & 0xF)
+#define NR_ACCL_TYPES			3
+
+static const char * const accl_str[] = {
+	"", "", "", "CLK", "VREG", "BUS",
+};
+
+static struct rsc_drv *__rsc_drv[2];
+static int __rsc_count;
+
 static u32 read_tcs_reg(struct rsc_drv *drv, int reg, int tcs_id, int cmd_id)
 {
 	return readl_relaxed(drv->tcs_base + reg + RSC_DRV_TCS_OFFSET * tcs_id +
@@ -450,8 +460,8 @@ int rpmh_rsc_send_data(struct rsc_drv *drv, const struct tcs_request *msg)
 	do {
 		ret = tcs_write(drv, msg);
 		if (ret == -EBUSY) {
-			pr_info_ratelimited("TCS Busy, retrying RPMH message send: addr=%#x\n",
-					    msg->cmds[0].addr);
+			pr_info_ratelimited("DRV:%s TCS Busy, retrying RPMH message send: addr=%#x\n",
+					    drv->name, msg->cmds[0].addr);
 			udelay(10);
 		}
 	} while (ret == -EBUSY);
@@ -624,6 +634,106 @@ int rpmh_rsc_write_pdc_data(struct rsc_drv *drv, const struct tcs_request *msg)
 	return 0;
 }
 
+static struct tcs_group *get_tcs_from_index(struct rsc_drv *drv, int tcs_id)
+{
+	unsigned int i;
+
+	for (i = 0; i < TCS_TYPE_NR; i++) {
+		if (drv->tcs[i].mask & BIT(tcs_id))
+			return &drv->tcs[i];
+	}
+
+	return NULL;
+}
+
+static void print_tcs_info(struct rsc_drv *drv, int tcs_id, unsigned long *accl)
+{
+	struct tcs_group *tcs_grp = get_tcs_from_index(drv, tcs_id);
+	const struct tcs_request *req = get_req_from_tcs(drv, tcs_id);
+	unsigned long cmds_enabled;
+	u32 addr, data, msgid, sts, irq_sts;
+	bool in_use = test_bit(tcs_id, drv->tcs_in_use);
+	int i;
+
+	if (!tcs_grp || !req)
+		return;
+
+	sts = read_tcs_reg(drv, RSC_DRV_STATUS, tcs_id, 0);
+	cmds_enabled = read_tcs_reg(drv, RSC_DRV_CMD_ENABLE, tcs_id, 0);
+	if (!cmds_enabled)
+		return;
+
+	data = read_tcs_reg(drv, RSC_DRV_CONTROL, tcs_id, 0);
+	irq_sts = read_tcs_reg(drv, RSC_DRV_IRQ_STATUS, 0, 0);
+	pr_warn("Request: tcs-in-use:%s active_tcs=%s(%d) state=%d wait_for_compl=%u]\n",
+		(in_use ? "YES" : "NO"),
+		((tcs_grp->type == ACTIVE_TCS) ? "YES" : "NO"),
+		tcs_grp->type, req->state, req->wait_for_compl);
+	pr_warn("TCS=%d [ctrlr-sts:%s amc-mode:0x%x irq-sts:%s]\n",
+		tcs_id, sts ? "IDLE" : "BUSY", data,
+		(irq_sts & BIT(tcs_id)) ? "COMPLETED" : "PENDING");
+
+	for_each_set_bit(i, &cmds_enabled, MAX_CMDS_PER_TCS) {
+		addr = read_tcs_reg(drv, RSC_DRV_CMD_ADDR, tcs_id, i);
+		data = read_tcs_reg(drv, RSC_DRV_CMD_DATA, tcs_id, i);
+		msgid = read_tcs_reg(drv, RSC_DRV_CMD_MSGID, tcs_id, i);
+		sts = read_tcs_reg(drv, RSC_DRV_CMD_STATUS, tcs_id, i);
+		pr_warn("\tCMD=%d [addr=0x%x data=0x%x hdr=0x%x sts=0x%x enabled=1]\n",
+			i, addr, data, msgid, sts);
+		if (!(sts & CMD_STATUS_ISSUED))
+			continue;
+		if (!(sts & CMD_STATUS_COMPL))
+			*accl |= BIT(ACCL_TYPE(addr));
+	}
+}
+
+void rpmh_rsc_debug(struct rsc_drv *drv)
+{
+	struct irq_data *rsc_irq_data = irq_get_irq_data(drv->irq);
+	bool irq_sts;
+	int i;
+	int busy = 0;
+	unsigned long accl = 0;
+	char str[20] = "";
+
+	pr_warn("RSC:%s\n", drv->name);
+
+	for (i = 0; i < drv->num_tcs; i++) {
+		if (!test_bit(i, drv->tcs_in_use))
+			continue;
+		busy++;
+		print_tcs_info(drv, i, &accl);
+	}
+
+	if (!rsc_irq_data) {
+		pr_err("No IRQ data for RSC:%s\n", drv->name);
+		return;
+	}
+
+	irq_get_irqchip_state(drv->irq, IRQCHIP_STATE_PENDING, &irq_sts);
+	pr_warn("HW IRQ %lu is %s at GIC\n", rsc_irq_data->hwirq,
+		irq_sts ? "PENDING" : "NOT PENDING");
+
+	for_each_set_bit(i, &accl, ARRAY_SIZE(accl_str)) {
+		strlcat(str, accl_str[i], sizeof(str));
+		strlcat(str, " ", sizeof(str));
+	}
+
+	if (busy && !irq_sts)
+		pr_warn("ERROR:Accelerator(s) { %s } at AOSS did not respond\n",
+			str);
+	else if (irq_sts)
+		pr_warn("ERROR:Possible lockup in Linux\n");
+
+	/*
+	 * The TCS(s) are busy waiting, we have no way to recover from this.
+	 * If this debug function is called, we assume it's because timeout
+	 * has happened.
+	 * Crash and report.
+	 */
+	BUG_ON(busy);
+}
+
 static int rpmh_probe_tcs_config(struct platform_device *pdev,
 				 struct rsc_drv *drv)
 {
@@ -760,6 +870,8 @@ static int rpmh_rsc_probe(struct platform_device *pdev)
 	if (irq < 0)
 		return irq;
 
+	drv->irq = irq;
+
 	ret = devm_request_irq(&pdev->dev, irq, tcs_tx_done,
 			       IRQF_TRIGGER_HIGH | IRQF_NO_SUSPEND,
 			       drv->name, drv);
@@ -774,6 +886,7 @@ static int rpmh_rsc_probe(struct platform_device *pdev)
 	INIT_LIST_HEAD(&drv->client.batch_cache);
 
 	dev_set_drvdata(&pdev->dev, drv);
+	__rsc_drv[__rsc_count++] = drv;
 
 	return devm_of_platform_populate(&pdev->dev);
 }

drivers/soc/qcom/rpmh.c

@@ -322,8 +322,12 @@ int rpmh_write(const struct device *dev, enum rpmh_state state,
 		return ret;
 
 	ret = wait_for_completion_timeout(&compl, RPMH_TIMEOUT_MS);
-	WARN_ON(!ret);
-	return (ret > 0) ? 0 : -ETIMEDOUT;
+	if (!ret) {
+		rpmh_rsc_debug(ctrlr_to_drv(ctrlr));
+		return -ETIMEDOUT;
+	}
+
+	return 0;
 }
 EXPORT_SYMBOL(rpmh_write);
 
@@ -457,7 +461,7 @@ int rpmh_write_batch(const struct device *dev, enum rpmh_state state,
 			 * the completion that we're going to free once
 			 * we've returned from this function.
 			 */
-			WARN_ON(1);
+			rpmh_rsc_debug(ctrlr_to_drv(ctrlr));
 			ret = -ETIMEDOUT;
 			goto exit;
 		}