Re: linux-rt stall on imx6dl on patch-5.4.22-rt13.patch

[Michael Nazzareno Trimarchi]

Hi

On Fri, May 29, 2020 at 3:31 PM Sebastian Andrzej Siewior
<bigeasy@xxxxxxxxxxxxx> wrote:
>
> On 2020-05-29 10:08:01 [+0200], Michael Nazzareno Trimarchi wrote:
> > Hi all
> Hi,
>
> > I get this stack trace
> > echo 3 > /proc/sys/vm/drop_caches chrt -f 99 sysbench --test=cpu --num-threads=2
> >
> > First of all, I need to know that if it is a valid use case because
> > was raised by the testing team and
> > how can I start and where to take a look at it
>
> Assuming that there is a `;' in front of `chrt' then I wouldn't be
> surprised. You are basically keeping two CPUs busy with sysbench at RT
> priority and then wondering why the system isn't making any progress. If
> it wouldn't be for the RT throttler, you might not even see any output
> and box hardware would appear be dead.

No, I'm not surprised but my answer to them was not enough. Right now
they are executing some bench.

Best
Michael
>
> > Best Regards
> > Michael
> >
> > [  161.709260] 001: sh (480): drop_caches: 3
> > [  180.234358] 001: sched: RT throttling activated
> > [  180.283363] 001: fec 2188000.ethernet eth0: MDIO read timeout
> > [  180.283406] 001: ------------[ cut here ]------------
> > [  180.283416] 001: WARNING: CPU: 1 PID: 133 at
> > drivers/net/phy/phy.c:716 phy_er
> >
> >                                                ror+0x24/0x64
> > [  180.283460] 001: Modules linked in:
> > [  180.283473] 001: CPU: 1 PID: 133 Comm: kworker/1:2 Tainted: G
> >  W
>
> This is hard to read. Especially later in that mail. In future please
> try to not cut the lines.
>
> Sebastian



-- 
| Michael Nazzareno Trimarchi                     Amarula Solutions BV |
| COO  -  Founder                                      Cruquiuskade 47 |
| +31(0)851119172                                 Amsterdam 1018 AM NL |
|                  [`as] http://www.amarulasolutions.com               |