[PATCH v4 1/6] arm64: ftrace: adjust callsite addresses examined by stack tracer

[Mark Rutland]

On Mon, Nov 02, 2015 at 06:41:24PM +0000, Will Deacon wrote:
> On Mon, Nov 02, 2015 at 06:29:28PM +0000, Mark Rutland wrote:
> > On Mon, Nov 02, 2015 at 01:20:43PM -0500, Steven Rostedt wrote:
> > > On Fri, 30 Oct 2015 11:16:19 +0000
> > > Will Deacon <will.deacon at arm.com> wrote:
> > > 
> > > > On Fri, Oct 30, 2015 at 02:25:36PM +0900, AKASHI Takahiro wrote:
> > > > > On arm64, no PC values returned by save_stack_trace() will match to LR
> > > > > values saved in stack frames on a stack after the following commit:
> > > > >     commit e306dfd06fcb ("ARM64: unwind: Fix PC calculation")
> > > > > As a result, the output from stack tracer will be messed up.
> > > > 
> > > > FWIW, we've decided to revert that patch for the moment. We're the only
> > > > architecture making that kind of adjustment, so we should fix that before
> > > > building on top of it.
> > > > 
> > > >
> > > 
> > > What is the status of this patch set. I'm currently pulling in last
> > > minute patches for 4.3 and should the ftrace patch in this series be
> > > applied? (I still need to review it too).
> > 
> > The revert Will mentioned is in v4.3 (see commit 9702970c7bd3e2d6), so
> > this series needs a respin to account for that.
> 
> Right, but we're talking about the core ftrace change in patch 4, which
> is (afaict) independent of the commit you mention above.

Ah; I don't know on that front, sorry for the noise.

Mark.