Don't hardcode the register name maximum length (used for alignment
purposes) to 14, but compute it dynamically. The small performance
impact isn't an issue as this is debugging code.
Signed-off-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
---
.../platform/rockchip/rkisp1/rkisp1-debug.c | 21 +++++++++++++------
1 file changed, 15 insertions(+), 6 deletions(-)
diff --git a/drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c b/drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c
index 2c226f20f525..28a69323cb38 100644
--- a/drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c
+++ b/drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c
@@ -11,8 +11,10 @@
#include <linux/debugfs.h>
#include <linux/delay.h>
#include <linux/device.h>
+#include <linux/minmax.h>
#include <linux/pm_runtime.h>
#include <linux/seq_file.h>
+#include <linux/string.h>
#include "rkisp1-common.h"
#include "rkisp1-regs.h"
@@ -32,22 +34,29 @@ static int rkisp1_debug_dump_regs(struct rkisp1_device *rkisp1,
struct seq_file *m, unsigned int offset,
const struct rkisp1_debug_register *regs)
{
+ const struct rkisp1_debug_register *reg;
+ int width = 0;
u32 val, shd;
int ret;
+ for (reg = regs; reg->name; ++reg)
+ width = max_t(int, width, strlen(reg->name));
+
+ width += 1;
+
ret = pm_runtime_get_if_in_use(rkisp1->dev);
if (ret <= 0)
return ret ? : -ENODATA;
- for ( ; regs->name; ++regs) {
- val = rkisp1_read(rkisp1, offset + regs->reg);
+ for (reg = regs; reg->name; ++reg) {
+ val = rkisp1_read(rkisp1, offset + reg->reg);
- if (regs->shd) {
- shd = rkisp1_read(rkisp1, offset + regs->shd);
- seq_printf(m, "%14s: 0x%08x/0x%08x\n", regs->name,
+ if (reg->shd) {
+ shd = rkisp1_read(rkisp1, offset + reg->shd);
+ seq_printf(m, "%*s: 0x%08x/0x%08x\n", width, reg->name,
val, shd);
} else {
- seq_printf(m, "%14s: 0x%08x\n", regs->name, val);
+ seq_printf(m, "%*s: 0x%08x\n", width, reg->name, val);
}
}
--
Regards,
Laurent Pinchart
