linux-kernel - Re: [PATCH 10/14] perf test: Add thread loop test shell scripts

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJ9a7VjQ4B8yudKqHZ_f8pGRGcM9t=6M0zfQK9-LmyTf4pZbZw@mail.gmail.com>
Date:   Fri, 8 Jul 2022 11:27:42 +0100
From:   Mike Leach <mike.leach@...aro.org>
To:     Carsten Haitzler <carsten.haitzler@...s.arm.com>
Cc:     James Clark <james.clark@....com>, linux-kernel@...r.kernel.org,
        coresight@...ts.linaro.org, mathieu.poirier@...aro.org,
        linux-perf-users@...r.kernel.org, acme@...nel.org,
        Suzuki K Poulose <suzuki.poulose@....com>
Subject: Re: [PATCH 10/14] perf test: Add thread loop test shell scripts

Hi,

On Fri, 8 Jul 2022 at 10:22, Carsten Haitzler
<carsten.haitzler@...s.arm.com> wrote:
>
>
>
> On 7/5/22 14:53, James Clark wrote:
> >
> >
> > On 01/07/2022 13:07, carsten.haitzler@...s.arm.com wrote:
> >> From: "Carsten Haitzler (Rasterman)" <raster@...terman.com>
> >>
> >> Add a script to drive the thread loop test that gathers data so
> >> it passes a minimum bar (in this case do we get any perf context data
> >> for every thread).
> >>
> >> Signed-off-by: Carsten Haitzler <carsten.haitzler@....com>
> >
> > Hi Carsten,
> >
> > I checked this on N1SDP and I get failures in both threads tests. This is
> > because it's looking for "CID=..." when in my output threads are shown as
> > "VMID=...":
> >
> >      Idx:628048; ID:10;       I_ADDR_CTXT_L_64IS0 : Address & Context, Long, 64 bit, IS0.; Addr=0x0000AAAAE3BF0B18; Ctxt: AArch64,EL0, NS; VMID=0xa588c;
> >
> > I think with a change to the grep it should work.
>
> Errrr... I get no VMID= ... it's all
>
> Idx:563008; ID:12;      I_ADDR_CTXT_L_64IS0 : Address & Context, Long, 64
> bit, IS0.; Addr=0x0000AAAAE4B00A60; Ctxt: AArch64,EL0, NS; CID=0x00004aff;
>
> are you using containers or something? because:
>
>              if(context.updated_c)
>              {
>                  oss << "CID=0x" << std::hex << std::setfill('0') <<
> std::setw(8) << context.ctxtID << "; ";
>              }
>              if(context.updated_v)
>              {
>                  oss << "VMID=0x" << std::hex << std::setfill('0') <<
> std::setw(4) << context.VMID << "; ";
>              }
>
> I'm running without any containers etc. - bare metal. Haven't bothered
> with any VM stuff.
>
> In OpenOCD the CID should be the the pid/thread id. It seems to not be
> the same thing as VMID. I haven't traced this beyond here as to exactly
> what this represents though my first reaction is "This is extra VM info
> and not the PID/TID being looked for". OpenOCD is full of tests with log
> dumps that produce CID and VMID:
>
> Idx:1676; ID:10;        I_ADDR_CTXT_L_64IS0 : Address & Context, Long,
> 64 bit, IS0.; Addr=0xFFFFFFC000096A00; Ctxt: AArch64,EL1, NS;
> CID=0x00000000; VMID=0x0000;
>
> A quick git grep CID= in OpenCD will show them all. My understanding is
> CID is the thread/process ID and thus the test/check "Do we get reported
> data from all threads? - anything?".
>
> I don't think using VMID is right. The fact you are missing a CID is an
> issue though...
>

The register used for linux TID trace is dependent on the EL of the kernel.
EL1 => CONTEXT_IDR_EL1
EL2 => CONTEXT_IDR_EL2.

By design, the trace hardware traces CONTEXT_IDR_EL2 as the VMID packet.

So, depending on your kernel build, TID can validly be traced as CID or VMID

Regards

Mike

> > Thanks
> > James
> >
> >> ---
> >>   .../coresight/thread_loop_check_tid_10.sh     | 19 +++++++++++++++++++
> >>   .../coresight/thread_loop_check_tid_2.sh      | 19 +++++++++++++++++++
> >>   2 files changed, 38 insertions(+)
> >>   create mode 100755 tools/perf/tests/shell/coresight/thread_loop_check_tid_10.sh
> >>   create mode 100755 tools/perf/tests/shell/coresight/thread_loop_check_tid_2.sh
> >>
> >> diff --git a/tools/perf/tests/shell/coresight/thread_loop_check_tid_10.sh b/tools/perf/tests/shell/coresight/thread_loop_check_tid_10.sh
> >> new file mode 100755
> >> index 000000000000..7c13636fc778
> >> --- /dev/null
> >> +++ b/tools/perf/tests/shell/coresight/thread_loop_check_tid_10.sh
> >> @@ -0,0 +1,19 @@
> >> +#!/bin/sh -e
> >> +# CoreSight / Thread Loop 10 Threads - Check TID
> >> +
> >> +# SPDX-License-Identifier: GPL-2.0
> >> +# Carsten Haitzler <carsten.haitzler@....com>, 2021
> >> +
> >> +TEST="thread_loop"
> >> +. $(dirname $0)/../lib/coresight.sh
> >> +ARGS="10 1"
> >> +DATV="check-tid-10th"
> >> +DATA="$DATD/perf-$TEST-$DATV.data"
> >> +STDO="$DATD/perf-$TEST-$DATV.stdout"
> >> +
> >> +SHOW_TID=1 perf record -s $PERFRECOPT -o "$DATA" "$BIN" $ARGS > $STDO
> >> +
> >> +perf_dump_aux_tid_verify "$DATA" "$STDO"
> >> +
> >> +err=$?
> >> +exit $err
> >> diff --git a/tools/perf/tests/shell/coresight/thread_loop_check_tid_2.sh b/tools/perf/tests/shell/coresight/thread_loop_check_tid_2.sh
> >> new file mode 100755
> >> index 000000000000..a067145af43c
> >> --- /dev/null
> >> +++ b/tools/perf/tests/shell/coresight/thread_loop_check_tid_2.sh
> >> @@ -0,0 +1,19 @@
> >> +#!/bin/sh -e
> >> +# CoreSight / Thread Loop 2 Threads - Check TID
> >> +
> >> +# SPDX-License-Identifier: GPL-2.0
> >> +# Carsten Haitzler <carsten.haitzler@....com>, 2021
> >> +
> >> +TEST="thread_loop"
> >> +. $(dirname $0)/../lib/coresight.sh
> >> +ARGS="2 20"
> >> +DATV="check-tid-2th"
> >> +DATA="$DATD/perf-$TEST-$DATV.data"
> >> +STDO="$DATD/perf-$TEST-$DATV.stdout"
> >> +
> >> +SHOW_TID=1 perf record -s $PERFRECOPT -o "$DATA" "$BIN" $ARGS > $STDO
> >> +
> >> +perf_dump_aux_tid_verify "$DATA" "$STDO"
> >> +
> >> +err=$?
> >> +exit $err



-- 
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK