On 17 February 2014 03:34, Xishi Qiu <qiuxishi@xxxxxxxxxx> wrote: > If we want to debug the kernel memory, we should turn on CONFIG_KMEMCHECK > and rebuild the kernel. This always takes a long time and sometimes > impossible, e.g. users don't have the kernel source code or the code > is different from "www.kernel.org" (private features may be added to the > kernel, and usually users can not get the whole code). > > This patch adds a new command-line "kmemcheck=3", then the kernel will run > as the same as CONFIG_KMEMCHECK=off even CONFIG_KMEMCHECK is turn on. > "kmemcheck=0/1/2" is the same as originally. This means we can always turn > on CONFIG_KMEMCHECK, and use "kmemcheck=3" to control it on/off with out > rebuild the kernel. > > In another word, "kmemcheck=3" is equivalent: > 1) turn off CONFIG_KMEMCHECK > 2) rebuild the kernel > 3) reboot > > The different between kmemcheck=0 and 3 is the used memory and nr_cpus. > Also kmemcheck=0 can used in runtime, and kmemcheck=3 is only used in boot. > boottime: kmemcheck=0/1/2/3 (command-line) > runtime: kmemcheck=0/1/2 (/proc/sys/kernel/kmemcheck) This is not the right way to do what you want. The behaviour that we want is: - CONFIG_KMEMCHECK=y + kmemcheck=0 (boot parameter) should have a minimal runtime impact and not limit the number of CPUs - CONFIG_KMEMCHECK=y + kmemcheck=1 should limit the number of CPUs during boot - setting kmemcheck to 1 via /proc/sys/kernel/kmemcheck should probably return an error if more than 1 CPU is online Vegard -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>