Re: [PATCH v3] tracing: add cond_resched to ftrace_replace_code()

From: Anders Roxell
Date: Wed Dec 05 2018 - 05:50:53 EST


On Tue, 4 Dec 2018 at 23:33, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
>
> On Tue, 4 Dec 2018 20:40:44 +0100
> Anders Roxell <anders.roxell@xxxxxxxxxx> wrote:
>
> > When running in qemu on an kernel built with allmodconfig and debug
> > options (in particular kcov and ubsan) enabled, ftrace_replace_code
> > function call take minutes. The ftrace selftest calls
> > ftrace_replace_code to look >40000 through
> > ftrace_make_call/ftrace_make_nop, and these end up calling
> > __aarch64_insn_write/aarch64_insn_patch_text_nosync.
> >
> > Microseconds add up because this is called in a loop for each dyn_ftrace
> > record, and this triggers the softlockup watchdog unless we let it sleep
> > occasionally.
> >
> > Rework so that we call cond_resched() if !irqs_disabled() && !preempt_count().
>
> This isn't urgent is it?

No, it is not urgent. I'm trying to get allmodconfig to boot on arm64.

> That is, it doesn't need a stable tag?

Also I don't think its a regression, since clearly nobody has ever been
able to boot an arm64 allmodconfig kernel or anything close to that.

Cheers,
Anders