[patch] hrtimer: round up relative start time on low-res arches

[Ingo Molnar <mingo@elte.hu>]

find below a patch with less impact than blanket upwards rounding of 
relatime timeouts. This i think is better for v2.6.16 - it accurately 
documents the problem architectures that have low-res do_gettimeofday(), 
and doesnt impact other architectures. This will go away with John's 
GTOD patchset. I've reviewed every architecture for the (worst-case) 
resolution of their do_gettimeofday() implementations, and this is the 
result of that review.

	Ingo

-----

CONFIG_TIME_LOW_RES is a temporary way for architectures to signal that 
they simply return xtime in do_gettimeoffset(). In this corner-case we 
want to round up by resolution when starting a relative timer, to avoid 
short timeouts. This will go away with the GTOD framework.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

----

 arch/frv/Kconfig       |    4 ++++
 arch/h8300/Kconfig     |    4 ++++
 arch/m68k/Kconfig      |    4 ++++
 arch/m68knommu/Kconfig |    4 ++++
 arch/parisc/Kconfig    |    5 +++++
 arch/v850/Kconfig      |    4 ++++
 kernel/hrtimer.c       |   13 ++++++++++++-
 7 files changed, 37 insertions(+), 1 deletion(-)

Index: linux/arch/frv/Kconfig
===================================================================
--- linux.orig/arch/frv/Kconfig
+++ linux/arch/frv/Kconfig
@@ -25,6 +25,10 @@ config GENERIC_HARDIRQS
 	bool
 	default n
 
+config TIME_LOW_RES
+	bool
+	default y
+
 mainmenu "Fujitsu FR-V Kernel Configuration"
 
 source "init/Kconfig"
Index: linux/arch/h8300/Kconfig
===================================================================
--- linux.orig/arch/h8300/Kconfig
+++ linux/arch/h8300/Kconfig
@@ -33,6 +33,10 @@ config GENERIC_CALIBRATE_DELAY
 	bool
 	default y
 
+config TIME_LOW_RES
+	bool
+	default y
+
 config ISA
 	bool
 	default y
Index: linux/arch/m68k/Kconfig
===================================================================
--- linux.orig/arch/m68k/Kconfig
+++ linux/arch/m68k/Kconfig
@@ -21,6 +21,10 @@ config GENERIC_CALIBRATE_DELAY
 	bool
 	default y
 
+config TIME_LOW_RES
+	bool
+	default y
+
 config ARCH_MAY_HAVE_PC_FDC
 	bool
 	depends on Q40 || (BROKEN && SUN3X)
Index: linux/arch/m68knommu/Kconfig
===================================================================
--- linux.orig/arch/m68knommu/Kconfig
+++ linux/arch/m68knommu/Kconfig
@@ -29,6 +29,10 @@ config GENERIC_CALIBRATE_DELAY
 	bool
 	default y
 
+config TIME_LOW_RES
+	bool
+	default y
+
 source "init/Kconfig"
 
 menu "Processor type and features"
Index: linux/arch/parisc/Kconfig
===================================================================
--- linux.orig/arch/parisc/Kconfig
+++ linux/arch/parisc/Kconfig
@@ -29,6 +29,11 @@ config GENERIC_CALIBRATE_DELAY
 	bool
 	default y
 
+config TIME_LOW_RES
+	bool
+	depends on SMP
+	default y
+
 config GENERIC_ISA_DMA
 	bool
 
Index: linux/arch/v850/Kconfig
===================================================================
--- linux.orig/arch/v850/Kconfig
+++ linux/arch/v850/Kconfig
@@ -28,6 +28,10 @@ config GENERIC_IRQ_PROBE
 	bool
 	default y
 
+config TIME_LOW_RES
+	bool
+	default y
+
 # Turn off some random 386 crap that can affect device config
 config ISA
 	bool
Index: linux/kernel/hrtimer.c
===================================================================
--- linux.orig/kernel/hrtimer.c
+++ linux/kernel/hrtimer.c
@@ -418,8 +418,19 @@ hrtimer_start(struct hrtimer *timer, kti
 	/* Switch the timer base, if necessary: */
 	new_base = switch_hrtimer_base(timer, base);
 
-	if (mode == HRTIMER_REL)
+	if (mode == HRTIMER_REL) {
 		tim = ktime_add(tim, new_base->get_time());
+		/*
+		 * CONFIG_TIME_LOW_RES is a temporary way for architectures
+		 * to signal that they simply return xtime in
+		 * do_gettimeoffset(). In this case we want to round up by
+		 * resolution when starting a relative timer, to avoid short
+		 * timeouts. This will go away with the GTOD framework.
+		 */
+#ifdef CONFIG_TIME_LOW_RES
+		tim = ktime_add(tim, base->resolution);
+#endif
+	}
 	timer->expires = tim;
 
 	enqueue_hrtimer(timer, new_base);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/