Re: [PATCH v1 1/2] perf auxtrace: Change to use SMP memory barriers

[Leo Yan]

On Mon, May 31, 2021 at 05:48:03PM +0200, Peter Zijlstra wrote:
> On Mon, May 31, 2021 at 10:53:02PM +0800, Leo Yan wrote:
> > Hi Peter, Adrian,
> > 
> > On Thu, May 27, 2021 at 11:57:37AM +0200, Peter Zijlstra wrote:
> > > On Thu, May 27, 2021 at 12:24:15PM +0300, Adrian Hunter wrote:
> > > 
> > > > > If all we want is a compiler barrier, then shouldn't that be what we use?
> > > > > i.e. barrier()
> > 
> > Sorry for a bit late.  Just bring up one question before I respin
> > this patch set.
> > 
> > > > I guess you are saying we still need to stop potential re-ordering across
> > > > CPUs, so please ignore my comments.
> > > 
> > > Right; so the ordering issue is real, consider:
> > > 
> > > 	CPU0 (kernel)		CPU1 (user)
> > > 
> > > 	write data		read head
> > > 	smp_wmb()		smp_rmb()
> > > 	write head		read data
> > 
> > One thing should be mentioned is the Linux kernel has _not_ used an
> > explict "smb_wmb()" between writing AUX trace data and updating header
> > "aux_head".  Please see the function perf_aux_output_end():
> 
> I think we pushed that into the driver. There is nothing the generic
> code can do here.
> 
> It is the drivers responsibility of ensuring the data is stable before
> calling perf_aux_output_end() or something along those lines.

Thanks for explaination.  I reviewed the drivers, some of them have
used memory barriers (e.g. Intel-PT, Arm SPE), but some drivers miss
to use memory barriers before calling perf_aux_output_end() (Like Arm
CoreSight, Intel-bts).

Will address this issue in next patch set.

Thanks,
Leo