diff mbox series

[3/7] dmaengine: ste_dma40: Add dev helper variable

Message ID 20230417-ux500-dma40-cleanup-v1-3-b26324956e47@linaro.org
State Superseded
Headers show
Series DMA40 SRAM refactoring and cleanup | expand

Commit Message

Linus Walleij April 17, 2023, 7:55 a.m. UTC
The &pdev->dev device pointer is used so many times in the
probe() and d40_hw_detect_init() functions that a local *dev
variable makes the code way easier to read.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 drivers/dma/ste_dma40.c | 50 +++++++++++++++++++++++++------------------------
 1 file changed, 26 insertions(+), 24 deletions(-)
diff mbox series

Patch

diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 236269d35a53..ef2a2fdaa82e 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -3104,6 +3104,7 @@  static int __init d40_phy_res_init(struct d40_base *base)
 static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 {
 	struct stedma40_platform_data *plat_data = dev_get_platdata(&pdev->dev);
+	struct device *dev = &pdev->dev;
 	struct clk *clk;
 	void __iomem *virtbase;
 	struct resource *res;
@@ -3117,15 +3118,15 @@  static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 	u32 cid;
 	u8 rev;
 
-	clk = clk_get(&pdev->dev, NULL);
+	clk = clk_get(dev, NULL);
 	if (IS_ERR(clk)) {
-		d40_err(&pdev->dev, "No matching clock found\n");
+		d40_err(dev, "No matching clock found\n");
 		goto check_prepare_enabled;
 	}
 
 	clk_ret = clk_prepare_enable(clk);
 	if (clk_ret) {
-		d40_err(&pdev->dev, "Failed to prepare/enable clock\n");
+		d40_err(dev, "Failed to prepare/enable clock\n");
 		goto disable_unprepare;
 	}
 
@@ -3151,11 +3152,11 @@  static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 			& 255) << (i * 8);
 
 	if (cid != AMBA_CID) {
-		d40_err(&pdev->dev, "Unknown hardware! No PrimeCell ID\n");
+		d40_err(dev, "Unknown hardware! No PrimeCell ID\n");
 		goto unmap_io;
 	}
 	if (AMBA_MANF_BITS(pid) != AMBA_VENDOR_ST) {
-		d40_err(&pdev->dev, "Unknown designer! Got %x wanted %x\n",
+		d40_err(dev, "Unknown designer! Got %x wanted %x\n",
 			AMBA_MANF_BITS(pid),
 			AMBA_VENDOR_ST);
 		goto unmap_io;
@@ -3171,7 +3172,7 @@  static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 	 */
 	rev = AMBA_REV_BITS(pid);
 	if (rev < 2) {
-		d40_err(&pdev->dev, "hardware revision: %d is not supported", rev);
+		d40_err(dev, "hardware revision: %d is not supported", rev);
 		goto unmap_io;
 	}
 
@@ -3189,7 +3190,7 @@  static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 
 	num_log_chans = num_phy_chans * D40_MAX_LOG_CHAN_PER_PHY;
 
-	dev_info(&pdev->dev,
+	dev_info(dev,
 		 "hardware rev: %d @ %pa with %d physical and %d logical channels\n",
 		 rev, &res->start, num_phy_chans, num_log_chans);
 
@@ -3209,7 +3210,7 @@  static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 	base->phy_size = resource_size(res);
 	base->virtbase = virtbase;
 	base->plat_data = plat_data;
-	base->dev = &pdev->dev;
+	base->dev = dev;
 	base->phy_chans = ((void *)base) + ALIGN(sizeof(struct d40_base), 4);
 	base->log_chans = &base->phy_chans[num_phy_chans];
 
@@ -3505,7 +3506,8 @@  static int __init d40_of_probe(struct platform_device *pdev,
 
 static int __init d40_probe(struct platform_device *pdev)
 {
-	struct stedma40_platform_data *plat_data = dev_get_platdata(&pdev->dev);
+	struct device *dev = &pdev->dev;
+	struct stedma40_platform_data *plat_data = dev_get_platdata(dev);
 	struct device_node *np = pdev->dev.of_node;
 	struct device_node *np_lcpa;
 	int ret = -ENOENT;
@@ -3522,7 +3524,7 @@  static int __init d40_probe(struct platform_device *pdev)
 				goto report_failure;
 			}
 		} else {
-			d40_err(&pdev->dev, "No pdata or Device Tree provided\n");
+			d40_err(dev, "No pdata or Device Tree provided\n");
 			goto report_failure;
 		}
 	}
@@ -3541,24 +3543,24 @@  static int __init d40_probe(struct platform_device *pdev)
 	/* Get IO for logical channel parameter address (LCPA) */
 	np_lcpa = of_parse_phandle(np, "sram", 0);
 	if (!np_lcpa) {
-		dev_err(&pdev->dev, "no LCPA SRAM node\n");
+		dev_err(dev, "no LCPA SRAM node\n");
 		goto report_failure;
 	}
 	/* This is no device so read the address directly from the node */
 	ret = of_address_to_resource(np_lcpa, 0, &res_lcpa);
 	if (ret) {
-		dev_err(&pdev->dev, "no LCPA SRAM resource\n");
+		dev_err(dev, "no LCPA SRAM resource\n");
 		goto report_failure;
 	}
 	base->lcpa_size = resource_size(&res_lcpa);
 	base->phy_lcpa = res_lcpa.start;
-	dev_info(&pdev->dev, "found LCPA SRAM at 0x%08x, size 0x%08x\n",
+	dev_info(dev, "found LCPA SRAM at 0x%08x, size 0x%08x\n",
 		 base->phy_lcpa, base->lcpa_size);
 
 	/* We make use of ESRAM memory for this. */
 	val = readl(base->virtbase + D40_DREG_LCPA);
 	if (base->phy_lcpa != val && val != 0) {
-		dev_warn(&pdev->dev,
+		dev_warn(dev,
 			 "[%s] Mismatch LCPA dma 0x%x, def %08x\n",
 			 __func__, val, base->phy_lcpa);
 	} else
@@ -3567,7 +3569,7 @@  static int __init d40_probe(struct platform_device *pdev)
 	base->lcpa_base = ioremap(base->phy_lcpa, base->lcpa_size);
 	if (!base->lcpa_base) {
 		ret = -ENOMEM;
-		d40_err(&pdev->dev, "Failed to ioremap LCPA region\n");
+		d40_err(dev, "Failed to ioremap LCPA region\n");
 		goto release_base;
 	}
 	/* If lcla has to be located in ESRAM we don't need to allocate */
@@ -3576,7 +3578,7 @@  static int __init d40_probe(struct platform_device *pdev)
 							"lcla_esram");
 		if (!res) {
 			ret = -ENOENT;
-			d40_err(&pdev->dev,
+			d40_err(dev,
 				"No \"lcla_esram\" memory resource\n");
 			goto destroy_cache;
 		}
@@ -3584,7 +3586,7 @@  static int __init d40_probe(struct platform_device *pdev)
 						resource_size(res));
 		if (!base->lcla_pool.base) {
 			ret = -ENOMEM;
-			d40_err(&pdev->dev, "Failed to ioremap LCLA region\n");
+			d40_err(dev, "Failed to ioremap LCLA region\n");
 			goto destroy_cache;
 		}
 		writel(res->start, base->virtbase + D40_DREG_LCLA);
@@ -3592,7 +3594,7 @@  static int __init d40_probe(struct platform_device *pdev)
 	} else {
 		ret = d40_lcla_allocate(base);
 		if (ret) {
-			d40_err(&pdev->dev, "Failed to allocate LCLA area\n");
+			d40_err(dev, "Failed to allocate LCLA area\n");
 			goto destroy_cache;
 		}
 	}
@@ -3603,7 +3605,7 @@  static int __init d40_probe(struct platform_device *pdev)
 
 	ret = request_irq(base->irq, d40_handle_interrupt, 0, D40_NAME, base);
 	if (ret) {
-		d40_err(&pdev->dev, "No IRQ defined\n");
+		d40_err(dev, "No IRQ defined\n");
 		goto destroy_cache;
 	}
 
@@ -3611,7 +3613,7 @@  static int __init d40_probe(struct platform_device *pdev)
 
 		base->lcpa_regulator = regulator_get(base->dev, "lcla_esram");
 		if (IS_ERR(base->lcpa_regulator)) {
-			d40_err(&pdev->dev, "Failed to get lcpa_regulator\n");
+			d40_err(dev, "Failed to get lcpa_regulator\n");
 			ret = PTR_ERR(base->lcpa_regulator);
 			base->lcpa_regulator = NULL;
 			goto destroy_cache;
@@ -3619,7 +3621,7 @@  static int __init d40_probe(struct platform_device *pdev)
 
 		ret = regulator_enable(base->lcpa_regulator);
 		if (ret) {
-			d40_err(&pdev->dev,
+			d40_err(dev,
 				"Failed to enable lcpa_regulator\n");
 			regulator_put(base->lcpa_regulator);
 			base->lcpa_regulator = NULL;
@@ -3642,7 +3644,7 @@  static int __init d40_probe(struct platform_device *pdev)
 
 	ret = dma_set_max_seg_size(base->dev, STEDMA40_MAX_SEG_SIZE);
 	if (ret) {
-		d40_err(&pdev->dev, "Failed to set dma max seg size\n");
+		d40_err(dev, "Failed to set dma max seg size\n");
 		goto destroy_cache;
 	}
 
@@ -3651,7 +3653,7 @@  static int __init d40_probe(struct platform_device *pdev)
 	if (np) {
 		ret = of_dma_controller_register(np, d40_xlate, NULL);
 		if (ret)
-			dev_err(&pdev->dev,
+			dev_err(dev,
 				"could not register of_dma_controller\n");
 	}
 
@@ -3701,7 +3703,7 @@  static int __init d40_probe(struct platform_device *pdev)
 	kfree(base->phy_res);
 	kfree(base);
  report_failure:
-	d40_err(&pdev->dev, "probe failed\n");
+	d40_err(dev, "probe failed\n");
 	return ret;
 }