Creates the patchfile which lowers the "priority" of clocksource tsc below acpi
on x86 systems (File 997-tsc_on_vm.patch, which I send as standalone in 2/2).
I'm doing this, because tsc is not stable on virtualized machines nor on smp
systems, and so acpi gets selected instead of tsc and then jiffies (if acpi is
built in, which I would recommend for x86).
Signed-off-by: Alexander Stadler <sa.maillists at univie.ac.at>
------
diff -urN a/8.09/target/linux/generic-2.6/patches-2.6.25/997-tsc_on_vm.patch
b/8.09/target/linux/generic-2.6/patches-2.6.25/997-tsc_on_vm.patch
--- a/8.09/target/linux/generic-2.6/patches-2.6.25/997-tsc_on_vm.patch
1970-01-01 01:00:00.000000000 +0100
+++ b/8.09/target/linux/generic-2.6/patches-2.6.25/997-tsc_on_vm.patch
2009-04-17 12:01:48.000000000 +0200
@@ -0,0 +1,24 @@
+diff -urN openwrt-8.09/arch/x86/kernel/tsc_32.c
openwrt/arch/x86/kernel/tsc_32.c
+--- a/arch/x86/kernel/tsc_32.c 2008-09-08 12:20:51.000000000 +0200
++++ b/arch/x86/kernel/tsc_32.c 2009-04-17 11:55:40.000000000 +0200
+@@ -297,7 +297,7 @@
+
+ static struct clocksource clocksource_tsc = {
+ .name = "tsc",
+- .rating = 300,
++ .rating = 199,
+ .read = read_tsc,
+ .mask = CLOCKSOURCE_MASK(64),
+ .mult = 0, /* to be set */
+diff -urN openwrt-8.09/arch/x86/kernel/tsc_64.c
openwrt/arch/x86/kernel/tsc_64.c
+--- a/arch/x86/kernel/tsc_64.c 2008-09-08 12:20:51.000000000 +0200
++++ b/arch/x86/kernel/tsc_64.c 2009-04-17 11:55:42.000000000 +0200
+@@ -305,7 +305,7 @@
+
+ static struct clocksource clocksource_tsc = {
+ .name = "tsc",
+- .rating = 300,
++ .rating = 199,
+ .read = read_tsc,
+ .mask = CLOCKSOURCE_MASK(64),
+ .shift = 22,
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
