On 05.03.2012 [15:40:56 +0100], Michal Marek wrote: > On 29.2.2012 20:01, Nishanth Aravamudan wrote: > > Noticed something on a fresh FC16 based install on a test server in our > > lab. Autotest by default runs `make dep` and it seems the clock is > > pretty badly skewed on this server, resulting in an infinite loop of: > > > > 01:09:37 DEBUG| Running 'make dep' > > 01:09:38 DEBUG| [stderr] make: Warning: File `/usr/local/autotest/tmp/build/linux/arch/powerpc/Makefile' has modification time 112143 s in the future > > 01:09:38 DEBUG| [stderr] make[1]: Warning: File `/usr/local/autotest/tmp/build/linux/arch/powerpc/Makefile' has modification time 112143 s in the future > > 01:09:38 DEBUG| [stderr] make[2]: Warning: File `scripts/Makefile.host' has modification time 112143 s in the future > > 01:09:38 DEBUG| [stdout] HOSTCC scripts/basic/fixdep > > 01:09:38 DEBUG| [stderr] make[2]: warning: Clock skew detected. Your build may be incomplete. > > 01:09:38 DEBUG| [stderr] make[2]: Warning: File `scripts/Makefile.host' has modification time 112143 s in the future > [...] > > > Anyone know why that happens? > > Because the system time time is several hours, if not days, behind, so > generated files are older than source files and thus dependencies cannot > work. You need to fix the system time to be at least ahead of the mtime > of the sources. That makes sense. > > Obvious fixes are to either not run `make dep` > > I don't think this is limited to make dep. Well, killing `make dep`, which was just looping, and executing `make -j16 vmlinux` did seem to complete. But it was a while ago and maybe I'm mis-remembering. Ah well, I've fixed up my testing scripts to ensure an ntp run is performed before trying to build. Thanks, Nish -- Nishanth Aravamudan <nacc@xxxxxxxxxx> IBM Linux Technology Center -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
