On Wed, 30 Sep 2015 11:46:31 +0200 Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote: > On Wed, Sep 30, 2015 at 10:56 AM, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote: > > On Wed, 30 Sep 2015 09:35:23 +0200 (CEST) > > Jiri Kosina <jikos@xxxxxxxxxx> wrote: > > > > > >> [ ... snip a lot of stuff ... ] > >> > What shall I do... play with lockdep (print_irqtrace_events) in > >> > del_timer_sync()? > > > > Also, do you have PARAVIRT enabled by chance? > > > > Yes. > > CONFIG_PARAVIRT=y > > Is that the corresponding indicator in dmesg-log? > > [ 0.000000] Booting paravirtualized kernel on bare hardware > > What is your suspicion, Lt. Columbo? > Heh, I loved that show. Can you see if you still have the issue if you disable that. PARAVIRT adds a bit of magic when run on baremetal. See arch/x86/kernel/alterative.c. On boot up, things like disabling interrupts, which are originally compiled to perform a function call, that on a virtual machine would do hypervisor calls into the host, and on baremetal does the real code. But that alterative.c file will do live patching of the code to not do the function call but instead do the real work in place (if it can). Something tells me that with clang, things are not as they're expected, and perhaps the modification isn't working. -- Steve -- To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
