aboutsummaryrefslogtreecommitdiffstats
path: root/arch/arm/mach-at91/Kconfig
diff options
context:
space:
mode:
authorDavid Brownell <david-b@pacbell.net>2007-11-12 17:59:10 +0100
committerRussell King <rmk+kernel@arm.linux.org.uk>2008-01-26 15:00:31 +0000
commit5248c657898c018bcd23ef77759fa1d6c690bdf4 (patch)
tree48204e22be9aa73f5e8dba4ed79df6517a1f7b21 /arch/arm/mach-at91/Kconfig
parent[ARM] 4688/1: at91: speed-up irq processing (diff)
downloadlinux-dev-5248c657898c018bcd23ef77759fa1d6c690bdf4.tar.xz
linux-dev-5248c657898c018bcd23ef77759fa1d6c690bdf4.zip
[ARM] 4646/1: AT91: configurable HZ, default to 128
This makes HZ configurable on AT91, following the model used on OMAP. It defaults to a power of two on AT91rm9200 chips, avoiding rounding errors which come from dividing a 32 KiHz clock to generate scheduler irqs; and uses 100 on AT91sam926x chips, using MCK/16 (multi-MHZ). Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Acked-by: Remy Bhmer <linux@bohmer.net> Acked-by: Andrew Victor <andrew@sanpeople.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'arch/arm/mach-at91/Kconfig')
-rw-r--r--arch/arm/mach-at91/Kconfig16
1 files changed, 16 insertions, 0 deletions
diff --git a/arch/arm/mach-at91/Kconfig b/arch/arm/mach-at91/Kconfig
index 05a9f8a1b45e..214733e897e5 100644
--- a/arch/arm/mach-at91/Kconfig
+++ b/arch/arm/mach-at91/Kconfig
@@ -219,6 +219,22 @@ config AT91_PROGRAMMABLE_CLOCKS
Select this if you need to program one or more of the PCK0..PCK3
programmable clock outputs.
+config AT91_TIMER_HZ
+ int "Kernel HZ (jiffies per second)"
+ range 32 1024
+ depends on ARCH_AT91
+ default "128" if ARCH_AT91RM9200
+ default "100"
+ help
+ On AT91rm9200 chips where you're using a system clock derived
+ from the 32768 Hz hardware clock, this tick rate should divide
+ it exactly: use a power-of-two value, such as 128 or 256, to
+ reduce timing errors caused by rounding.
+
+ On AT91sam926x chips, or otherwise when using a higher precision
+ system clock (of at least several MHz), rounding is less of a
+ problem so it can be safer to use a decimal values like 100.
+
endmenu
endif