RE: [PATCH v3] Added perf functionality to mmdc driver

From: Zhengyu Shen <zhengyu.shen@nxp.com>
To: Mark Rutland <mark.rutland@arm.com>, Shawn Guo <shawnguo@kernel.org>
Cc: Zhi Li <lznuaa@gmail.com>, Frank Li <frank.li@nxp.com>,
	"linux-arm-kernel@lists.infradead.org" 
	<linux-arm-kernel@lists.infradead.org>,
	kernel list <linux-kernel@vger.kernel.org>,
	"peterz@infradead.org" <peterz@infradead.org>,
	"mingo@redhat.com" <mingo@redhat.com>,
	"acme@kernel.org" <acme@kernel.org>,
	"alexander.shishkin@linux.intel.com" 
	<alexander.shishkin@linux.intel.com>
Subject: RE: [PATCH v3] Added perf functionality to mmdc driver
Date: Tue, 30 Aug 2016 20:01:18 +0000	[thread overview]
Message-ID: <DB5PR04MB1430D8A0284123B5C475800E9FE00@DB5PR04MB1430.eurprd04.prod.outlook.com> (raw)
In-Reply-To: <20160830125420.GG1223@leverpostej>

> Hi,
> 
> On Tue, Aug 30, 2016 at 07:43:29PM +0800, Shawn Guo wrote:
> > On Mon, Aug 29, 2016 at 11:06:44AM -0500, Zhi Li wrote:
> > > On Wed, Aug 17, 2016 at 2:42 PM, Zhengyu Shen
> <zhengyu.shen@nxp.com> wrote:
> > > > MMDC is a multi-mode DDR controller that supports DDR3/DDR3L
> > > > x16/x32/x64 and LPDDR2 two channel x16/x32 memory types. MMDC is
> > > > configurable, high performance, and optimized. MMDC is present on
> > > > i.MX6 Quad and i.MX6 QuadPlus devices, but this driver only supports
> i.MX6 Quad at the moment.
> > > > MMDC provides registers for performance counters which read via
> > > > this driver to help debug memory throughput and similar issues.
> > > >
> > > > $ perf stat -a -e
> > > > mmdc/busy-cycles/,mmdc/read-accesses/,mmdc/read-
> bytes/,mmdc/total-cycles/,mmdc/write-accesses/,mmdc/write-bytes/ dd
> if=/dev/zero of=/dev/null bs=1M count=5000 Performance counter stats for
> 'dd if=/dev/zero of=/dev/null bs=1M count=5000':
> > > >
> > > >          898021787      mmdc/busy-cycles/
> > > >           14819600      mmdc/read-accesses/
> > > >             471.30 MB   mmdc/read-bytes/
> > > >         2815419216      mmdc/total-cycles/
> > > >           13367354      mmdc/write-accesses/
> > > >             427.76 MB   mmdc/write-bytes/
> > > >
> > > >        5.334757334 seconds time elapsed
> > > >
> > > > Signed-off-by: Zhengyu Shen <zhengyu.shen@nxp.com>
> > > > Signed-off-by: Frank Li <frank.li@nxp.com>
> > >
> > > Shawn Guo:
> > >
> > >          No new comments got more than 1 weeks.
> > >          Did you plan accept it?
> >
> > @Mark, how do you think of this version?
> 
> Sorry, I've been away for the last week and haven't had the chance to look at
> this yet. I will try to get round to it in the next few days.
> 
> In the meantime, could you please try attacking this with Vince's perf fuzzer
> [1] (as root, or with perf_event_paranoid dropped to -1)? It's rather good at
> finding (subtle) issues in drivers.
> 
> Thanks,
> Mark.
> 
> [1] https://github.com/deater/perf_event_tests

Hi, I've done some testing with the fuzzer. Mmdc was only responsible 
for one crash which I fixed (had to remove the event from the pmu properly). 
Other drivers also cause crashes and the program reports that events are
Throttling. Is this normal?