fbdev: pxa3xx: use ktime_get_ts64 for time stamps

Started by Arnd Bergmann on Nov 7, 2017 · 4 messages from 2 authors · Last reply Jan 4, 2018

Subsystems (2)

Maintainers (2)

STALE3232d

open the first message on its own page

  1. [PATCH 2/2] fbdev: radeon: stop use ktime_get() for HZ calibration
    Related patches
Message details and related content
do_gettimeofday() is deprecated because it is not y2038 safe, so I'm
changing the calculation for the diagnostic output over to using
'timespec64'.

We really only print time deltas here, so changing it to monotonic
time makes this more robust, the correct accessor for this is
ktime_get_ts64().

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 drivers/video/fbdev/pxa3xx-gcu.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/video/fbdev/pxa3xx-gcu.c b/drivers/video/fbdev/pxa3xx-gcu.c
index 3366076f9e0f..ddca4425adc8 100644
--- a/drivers/video/fbdev/pxa3xx-gcu.c
+++ b/drivers/video/fbdev/pxa3xx-gcu.c
@@ -104,7 +104,7 @@ struct pxa3xx_gcu_priv {
 	wait_queue_head_t	  wait_idle;
 	wait_queue_head_t	  wait_free;
 	spinlock_t		  spinlock;
-	struct timeval 		  base_time;
+	struct timespec64	  base_time;
 
 	struct pxa3xx_gcu_batch *free;
 	struct pxa3xx_gcu_batch *ready;
@@ -126,18 +126,20 @@ gc_writel(struct pxa3xx_gcu_priv *priv, unsigned int off, unsigned long val)
 
 #define QPRINT(priv, level, msg)					\
 	do {								\
-		struct timeval tv;					\
+		struct timespec64 ts;					\
 		struct pxa3xx_gcu_shared *shared = priv->shared;	\
 		u32 base = gc_readl(priv, REG_GCRBBR);			\
 									\
-		do_gettimeofday(&tv);					\
+		ktime_get_ts64(&ts);					\
+		ts = timespec64_sub(ts, priv->base_time);		\
 									\
-		printk(level "%ld.%03ld.%03ld - %-17s: %-21s (%s, "	\
+		printk(level "%lld.%03ld.%03ld - %-17s: %-21s (%s, "	\
 			"STATUS "					\
 			"0x%02lx, B 0x%08lx [%ld], E %5ld, H %5ld, "	\
 			"T %5ld)\n",					\
-			tv.tv_sec - priv->base_time.tv_sec,		\
-			tv.tv_usec / 1000, tv.tv_usec % 1000,		\
+			(s64)(tv.tv_sec),				\
+			tv.tv_usec / NSEC_PER_MSEC,			\
+			(tv.tv_usec % NSEC_PER_MSEC) / USEC_PER_MSEC,	\
 			__func__, msg,					\
 			shared->hw_running ? "running" : "   idle",	\
 			gc_readl(priv, REG_GCISCR),			\
@@ -164,7 +166,7 @@ pxa3xx_gcu_reset(struct pxa3xx_gcu_priv *priv)
 	priv->shared->buffer_phys = priv->shared_phys;
 	priv->shared->magic = PXA3XX_GCU_SHARED_MAGIC;
 
-	do_gettimeofday(&priv->base_time);
+	ktime_get_ts64(&priv->base_time);
 
 	/* set up the ring buffer pointers */
 	gc_writel(priv, REG_GCRBLR, 0);
-- 
2.9.0

[PATCH 2/2] fbdev: radeon: stop use ktime_get() for HZ calibration

Message details and related content
do_gettimeofday() is deprecated and a bit clumsy. This changes
radeon_probe_pll_params() over to using ktime_get() with monotonic
times. There is no need to check for negative values any more
since the monotonic clocksource cannot go backwards, but I'm
adding a check for zero-division in case of a bad clocksource.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 drivers/video/fbdev/aty/radeon_base.c | 18 +++++++-----------
 1 file changed, 7 insertions(+), 11 deletions(-)
diff --git a/drivers/video/fbdev/aty/radeon_base.c b/drivers/video/fbdev/aty/radeon_base.c
index 8ad1643e7d1c..8b7048fbf6dc 100644
--- a/drivers/video/fbdev/aty/radeon_base.c
+++ b/drivers/video/fbdev/aty/radeon_base.c
@@ -583,8 +583,8 @@ static int radeon_probe_pll_params(struct radeonfb_info *rinfo)
 	int hTotal, vTotal, num, denom, m, n;
 	unsigned long long hz, vclk;
 	long xtal;
-	struct timeval start_tv, stop_tv;
-	long total_secs, total_usecs;
+	ktime_t start_time, stop_time;
+	u64 total_usecs;
 	int i;
 
 	/* Ugh, we cut interrupts, bad bad bad, but we want some precision
@@ -600,7 +600,7 @@ static int radeon_probe_pll_params(struct radeonfb_info *rinfo)
 		if (((INREG(CRTC_VLINE_CRNT_VLINE) >> 16) & 0x3ff) = 0)
 			break;
 
-	do_gettimeofday(&start_tv);
+	start_time = ktime_get();
 
 	for(i=0; i<1000000; i++)
 		if (((INREG(CRTC_VLINE_CRNT_VLINE) >> 16) & 0x3ff) != 0)
@@ -610,18 +610,14 @@ static int radeon_probe_pll_params(struct radeonfb_info *rinfo)
 		if (((INREG(CRTC_VLINE_CRNT_VLINE) >> 16) & 0x3ff) = 0)
 			break;
 	
-	do_gettimeofday(&stop_tv);
+	stop_time = ktime_get();
 	
 	local_irq_enable();
 
-	total_secs = stop_tv.tv_sec - start_tv.tv_sec;
-	if (total_secs > 10)
+	total_usecs = ktime_us_delta(stop_time, start_time);
+	if (total_usecs >= 10 * USEC_PER_SEC || total_usecs = 0)
 		return -1;
-	total_usecs = stop_tv.tv_usec - start_tv.tv_usec;
-	total_usecs += total_secs * 1000000;
-	if (total_usecs < 0)
-		total_usecs = -total_usecs;
-	hz = 1000000/total_usecs;
+	hz = USEC_PER_SEC/(u32)total_usecs;
  
 	hTotal = ((INREG(CRTC_H_TOTAL_DISP) & 0x1ff) + 1) * 8;
 	vTotal = ((INREG(CRTC_V_TOTAL_DISP) & 0x3ff) + 1);
-- 
2.9.0
Message details and related content
On Tue, Nov 7, 2017 at 3:13 PM, Arnd Bergmann [off-list ref] wrote:
do_gettimeofday() is deprecated because it is not y2038 safe, so I'm
changing the calculation for the diagnostic output over to using
'timespec64'.

We really only print time deltas here, so changing it to monotonic
time makes this more robust, the correct accessor for this is
ktime_get_ts64().

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Please ignore this patch, my randconfig builder found a typo that
I had not noticed before.

         Arnd

Re: [PATCH 2/2] fbdev: radeon: stop use ktime_get() for HZ calibration

Message details and related content
On Tuesday, November 07, 2017 03:13:06 PM Arnd Bergmann wrote:
do_gettimeofday() is deprecated and a bit clumsy. This changes
radeon_probe_pll_params() over to using ktime_get() with monotonic
times. There is no need to check for negative values any more
since the monotonic clocksource cannot go backwards, but I'm
adding a check for zero-division in case of a bad clocksource.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Patch queued for 4.16, thanks.

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help