<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    Mark,<br>
    <br>
    1) maybe I missed posting my later version of a2l, which
    automatically strips the stack pointers from the ":" address syntax
    so you can copy&paste the USB output. Here it is:<br>
    <br>
    <tt>#!/bin/bash</tt><tt><br>
    </tt><tt>elf=~/esp/Open-Vehicle-Monitoring-System-3/vehicle/OVMS.V3/build/ovms3.elf</tt><tt><br>
    </tt><tt>for adr in $* ; do</tt><tt><br>
    </tt><tt>  if [[ "$adr" =~ "elf" ]] ; then</tt><tt><br>
    </tt><tt>    elf="$adr"</tt><tt><br>
    </tt><tt>  else</tt><tt><br>
    </tt><tt>    cmd+=" -ex 'l *${adr/:*/}'"</tt><tt><br>
    </tt><tt>  fi</tt><tt><br>
    </tt><tt>done</tt><tt><br>
    </tt><tt>cmd+=" -ex 'q'"</tt><tt><br>
    </tt><tt>echo "Using elf file: $elf"</tt><tt><br>
    </tt><br>
    <br>
    2) You're right, the bug is tx_frame with null origin overwriting
    body.bus in the union. I didn't notice that when checking Marko's
    submission.<br>
    <br>
    tx_frame is a copy of the last frame given to the bus for
    transmission. The queue msg is gone when the TX done IRQ comes in,
    and Marko needed a copy of the frame the TX IRQ relates to. I asked
    him (see PR discussion), he checked and confirmed that all TX is
    done sequentially, so a single buffer is sufficient.<br>
    <br>
    Swapping the lines would work. The frame.origin also shouldn't be
    null, but the handler should tolerate that. …oops, tx_frame also
    doesn't get initialized in the canbus constructor, so there's also
    potentially garbage in tx_frame if due to some bug a TX IRQ is
    generated or processed without a previous tx.<br>
    <br>
    I'll do the fixes… and also rename tx_frame to m_tx_frame for
    consistency.<br>
    <br>
    Regards,<br>
    Michael<br>
    <br>
    <br>
    <div class="moz-cite-prefix">Am 12.12.19 um 01:20 schrieb Mark
      Webb-Johnson:<br>
    </div>
    <blockquote type="cite"
      cite="mid:DDE51878-283B-45E6-8423-FF095EF621A6@webb-johnson.net">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div class="">Two issues:</div>
      <div class=""><br class="">
      </div>
      <div class=""><u class=""><b class="">1] A2L</b></u></div>
      <div class=""><br class="">
      </div>
      My a2l is this:
      <div class=""><br class="">
      </div>
      <blockquote style="margin: 0 0 0 40px; border: none; padding:
        0px;" class="">
        <div class="">
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">#!/bin/bash</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">elf=3.2.007.ovms3.elf</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">for
                adr in $* ; do</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                if [[ "$adr" =~ "elf" ]] ; then</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  elf="$adr"</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                else</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  cmd+=" -ex 'l *$adr'"</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                fi</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">done</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">cmd+="
                -ex 'q'"</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">echo
                "Using elf file: $elf"</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">echo
                "xtensa-esp32-elf-gdb -batch $elf $cmd"</span></font></div>
          <div class=""><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">eval
                xtensa-esp32-elf-gdb -batch $elf $cmd 2>/dev/null #|
                grep "^0x.* is in "</span></font></div>
        </div>
      </blockquote>
      <div class="">
        <div><br class="">
        </div>
        <div>When I run it, I get:</div>
        <div><br class="">
        </div>
      </div>
      <blockquote style="margin: 0 0 0 40px; border: none; padding:
        0px;" class="">
        <div><font class="" face="Andale Mono"><span style="font-style:
              normal; font-size: 14px;" class="">$ a2l 3.2.007.ovms3.elf
              0x400d5e4c:0x3ffc5c40 0x7ffffffd:0x3ffc5c90</span></font></div>
        <div><font class="" face="Andale Mono"><span style="font-style:
              normal; font-size: 14px;" class="">Using elf file:
              3.2.007.ovms3.elf</span></font></div>
        <div><font class="" face="Andale Mono"><span style="font-size:
              14px;" class="">xtensa-esp32-elf-gdb -batch
              3.2.007.ovms3.elf  -ex 'l *0x400d5e4c:0x3ffc5c40' -ex 'l
              *0x7ffffffd:0x3ffc5c90' -ex ‘q'</span></font></div>
      </blockquote>
      <div class="">
        <div><br class="">
        </div>
        <div>And a manual run gives:</div>
        <div><br class="">
        </div>
      </div>
      <blockquote style="margin: 0 0 0 40px; border: none; padding:
        0px;" class="">
        <div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">$
                xtensa-esp32-elf-gdb -batch 3.2.007.ovms3.elf  -ex 'l
                *0x400d5e4c:0x3ffc5c40' -ex 'l *0x7ffffffd:0x3ffc5c90'
                -ex 'q'</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">Junk
                at end of line specification.</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""><br
                  class="">
              </span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">$ </span><span
                style="font-size: 14px;" class="">xtensa-esp32-elf-gdb
                3.2.007.ovms3.elf</span></font></div>
          <div><font class="" face="Andale Mono"><span style="font-size:
                14px;" class="">GNU gdb (crosstool-NG
                crosstool-ng-1.22.0-80-g6c4433a) 7.10</span></font></div>
          <div><span style="font-size: 14px; font-family: "Andale
              Mono";" class="">(gdb) l *0x400d5e4c:0x3ffc5c40</span></div>
          <div><font class="" face="Andale Mono"><span style="font-size:
                14px;" class="">
                <div>Junk at end of line specification.</div>
                <div>(gdb) l *0x7ffffffd:0x3ffc5c90</div>
                <div>(gdb) quit</div>
              </span></font></div>
        </div>
      </blockquote>
      <div class="">
        <div><br class="">
        </div>
        <div><u class=""><b class="">2] Frame origin</b></u></div>
        <div><br class="">
        </div>
        <div>Regarding the CAN_txcallback, I think you are correct. And
          both generators of that message set the frame correctly. So my
          initial thought was that it is either a memory corruption, or
          somewhere else sending a frame with garbage data.</div>
        <div><br class="">
        </div>
        <div>I do see this technique used in both the mcp2515 and
          esp32can drivers:</div>
        <div><br class="">
        </div>
      </div>
      <blockquote style="margin: 0 0 0 40px; border: none; padding:
        0px;" class="">
        <div class="">
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">msg.body.bus
                = me;</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">msg.body.frame
                = me->tx_frame;</span></font></div>
        </div>
      </blockquote>
      <div class="">
        <div><br class="">
        </div>
        <div>I don’t normally just copy structures over like that. I
          memcpy() them, but I guess it must work. However, as our
          CAN_queue_msg_t is a union of CAN_frame_t (frame, with first
          member of the structure a canbus*) and 'canbus* bus’, that is
          a little worrying. It seems that the msg.body.bus will get
          overwritten with whatever is in msg.body.frame.origin. I can’t
          see anywhere that tx_frame.origin is set - which is scary
          because that would mean it is always random junk.</div>
        <div><br class="">
        </div>
        <div>Maybe it works if tx_frame.origin is set to the bus before
          anything else, but not in Greg’s circumstances where something
          else arrives first. Perhaps related to obd2ecu? I only see
          tx_frame set in can.cpp canbus::Write. I don’t really
          understand the tx_frame approach at all, and why the frame is
          just not passed on the queue.</div>
        <div><br class="">
        </div>
      </div>
      <blockquote style="margin: 0 0 0 40px; border: none; padding:
        0px;" class="">
        <div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">//
                CAN Frame</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">//
                Note: Take care changing this structure, as it is a
                union with</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">//
                CAN_log_message_t and position of 'origin' is fixed.</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class="">struct
                CAN_frame_t</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                {</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                canbus*     origin;                   // Origin of the
                frame</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                CanFrameCallback * callback;          // Frame-specific
                callback. Is called when this frame is successfully sent
                (or sending failed)</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                CAN_FIR_t   FIR;                      // Frame
                information record</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                uint32_t    MsgID;                    // Message ID</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                union</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  {</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  uint8_t   u8[8];                    // Payload byte
                access</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  uint32_t  u32[2];                   // Payload u32
                access (Att: little endian!)</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  uint64_t  u64;                      // Payload u64
                access (Att: little endian!)</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                  } data;</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""><br
                  class="">
              </span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                esp_err_t Write(canbus* bus=NULL, TickType_t
                maxqueuewait=0);  // bus: NULL=origin</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""> 
                };</span></font></div>
          <div><font class="" face="Andale Mono"><span
                style="font-style: normal; font-size: 14px;" class=""><br
                  class="">
              </span></font></div>
          <div>
            <div><font class="" face="Andale Mono"><span
                  style="font-size: 14px;" class="">// CAN message<br
                    class="">
                  typedef struct<br class="">
                    {<br class="">
                    CAN_queue_type_t type;<br class="">
                    union<br class="">
                      {<br class="">
                      CAN_frame_t frame;  // CAN_frame<br class="">
                      canbus* bus;<br class="">
                      } body;<br class="">
                    } CAN_queue_msg_t;</span></font></div>
          </div>
        </div>
      </blockquote>
      <div class="">
        <div><br class="">
        </div>
        <div>This approach seems to date back to the swcan support merge
          (f94ae5a1b).</div>
        <div><br class="">
        </div>
        <div>Should we swap around, and set the msg.body.bus after
          msg.body.frame? Or am I missing something…</div>
        <div><br class="">
        </div>
        <div>Regards, Mark.</div>
        <div><br class="">
          <blockquote type="cite" class="">
            <div class="">On 12 Dec 2019, at 2:29 AM, Michael Balzer
              <<a href="mailto:dexter@expeedo.de" class=""
                moz-do-not-send="true">dexter@expeedo.de</a>> wrote:</div>
            <br class="Apple-interchange-newline">
            <div class="">
              <meta http-equiv="Content-Type" content="text/html;
                charset=UTF-8" class="">
              <div class=""> Mark,<br class="">
                <br class="">
                good example why not to use addr2line: I think that
                result is wrong. a2l uses gdb which gives:<br class="">
                <br class="">
                <tt class=""><a class="moz-txt-link-abbreviated"
href="mailto:balzer@leela:~/esp/Open-Vehicle-Monitoring-System-3/vehicle/OVMS.V3"
                    moz-do-not-send="true">balzer@leela:~/esp/Open-Vehicle-Monitoring-System-3/vehicle/OVMS.V3</a>>
                  a2l tmp/3.2.007.ovms3.elf 0x400d5e4c:0x3ffc5c40
                  0x7ffffffd:0x3ffc5c90<br class="">
                  Using elf file: tmp/3.2.007.ovms3.elf<br class="">
                  0x400d5e4c is in CAN_rxtask(void*)
(/home/openvehicles/build/Open-Vehicle-Monitoring-System-3.1/vehicle/OVMS.V3/components/can/src/can.cpp:730).<br
                    class="">
                  725                 
                  me->IncomingFrame(&msg.body.frame);<br class="">
                  726                } while (loop);<br class="">
                  727              break;<br class="">
                  728              }<br class="">
                  729            case CAN_txcallback:<br class="">
                  730             
                  msg.body.bus->TxCallback(&msg.body.frame,
                  true);<br class="">
                  731              break;<br class="">
                  732            case CAN_txfailedcallback:<br class="">
                  733             
                  msg.body.bus->TxCallback(&msg.body.frame,
                  false);<br class="">
                  734             
                  msg.body.bus->LogStatus(CAN_LogStatus_Error);<br
                    class="">
                </tt><tt class=""><br class="">
                </tt>…and that actually makes sense and matches the
                register dump.<br class="">
                <br class="">
                If I read the gdb disassembly correctly, A10 =
                msg.body.bus, so Greg's got a CAN_txcallback msg without
                a bus.<br class="">
                <br class="">
                Hardening the rxtask against null here would probably
                avoid the crash, but I don't see yet how that could be
                possible.<br class="">
                Both esp32can and mcp2515 set the bus field to their
                object addresses, which cannot be null.<br class="">
                <br class="">
                Regards,<br class="">
                Michael<br class="">
                <br class="">
                <br class="">
                <div class="moz-cite-prefix">Am 11.12.19 um 13:45
                  schrieb Mark Webb-Johnson:<br class="">
                </div>
                <blockquote type="cite"
                  cite="mid:1BF79214-A026-4AD0-97D4-5C845FBB0328@webb-johnson.net"
                  class="">
                  <meta http-equiv="Content-Type" content="text/html;
                    charset=UTF-8" class="">
                  Can’t get a2l working at the moment. The addr2line
                  gives:
                  <div class=""><br class="">
                  </div>
                  <blockquote style="margin: 0 0 0 40px; border: none;
                    padding: 0px;" class="">
                    <div class="">
                      <div class="">addr2line -e 3.2.007.ovms3.elf
                        0x400d5e4c:0x3ffc5c40 0x7ffffffd:0x3ffc5c90</div>
                      <div class="">/home/openvehicles/build/Open-Vehicle-Monitoring-System-3.1/vehicle/OVMS.V3/components/can/src/can.cpp:551</div>
                    </div>
                  </blockquote>
                  <div class="">
                    <div class="">
                      <div class=""><br class="">
                      </div>
                      <div class="">That is:</div>
                      <div class=""><br class="">
                      </div>
                    </div>
                  </div>
                  <blockquote style="margin: 0 0 0 40px; border: none;
                    padding: 0px;" class="">
                    <div class="">
                      <div class="">
                        <div class="">
                          <div class="">void
                            canbus::LogInfo(CAN_log_type_t type, const
                            char* text)</div>
                          <div class="">  {</div>
                          <div class="">  MyCan.LogInfo(this, type,
                            text);    <—— HERE</div>
                          <div class="">  }</div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                  <div class="">
                    <div class="">
                      <div class=""><br class="">
                      </div>
                      <div class="">ELF is at:</div>
                      <div class=""><br class="">
                      </div>
                    </div>
                  </div>
                  <blockquote style="margin: 0 0 0 40px; border: none;
                    padding: 0px;" class="">
                    <div class="">
                      <div class="">
                        <div class=""><a
href="http://api.openvehicles.com/firmware/ota/v3.2/main/3.2.007.ovms3.elf"
                            class="" moz-do-not-send="true">http://api.openvehicles.com/firmware/ota/v3.2/main/3.2.007.ovms3.elf</a></div>
                      </div>
                    </div>
                  </blockquote>
                  <div class="">
                    <div class="">
                      <div class=""><br class="">
                      </div>
                      <div class="">Regards, Mark.</div>
                      <div class=""><br class="">
                        <blockquote type="cite" class="">
                          <div class="">On 11 Dec 2019, at 5:20 PM,
                            Michael Balzer <<a
                              href="mailto:dexter@expeedo.de" class=""
                              moz-do-not-send="true">dexter@expeedo.de</a>>
                            wrote:</div>
                          <br class="Apple-interchange-newline">
                          <div class="">
                            <meta http-equiv="Content-Type"
                              content="text/html; charset=UTF-8"
                              class="">
                            <div class="">
                              <div class="moz-cite-prefix">Mark,<br
                                  class="">
                                <br class="">
                                Greg uses your build, the crash point
                                seems to be consistent, can you post the
                                a2l on this?<br class="">
                                <br class="">
                                Regards,<br class="">
                                Michael<br class="">
                                <br class="">
                                PS: Greg, would you mind switching to
                                EAP to beta test future releases?<br
                                  class="">
                                <br class="">
                                <br class="">
                                Am 11.12.19 um 04:33 schrieb Greg D.:<br
                                  class="">
                              </div>
                              <blockquote type="cite"
                                cite="mid:48cd3f20-409f-d201-640a-31b6e7788e92@gmail.com"
                                class="">
                                <pre class="moz-quote-pre" wrap="">Hi folks,

Well, the module updated to 3.2.007 last night.  I just checked on it,
and it appears that it didn't exactly survive.  Crashed while running
the autoconfig script.  Log with two cycles attached.

I tried renaming the /store/events directory to /store/was_events since
it seems like Duktape was getting in the way, but that didn't resolve
the crash.  I manually enabled wifi so I could manage the module, and
moved it back to 3.2.005 from the other partition.  Seems stable again.

The car is going into the Service Center tomorrow (the perennial issue
with 1146 alerts), so I need to have the module stable so that I can
keep an eye on it.  Going to leave it on 3.2.005 for now, unless someone
has a quick fix in the next few hours...

Otherwise, any ideas for troubleshooting after I get the car back back
(hopefully end of day)?

Greg

</pre>
                                <br class="">
                                <fieldset class="mimeAttachmentHeader"></fieldset>
                                <pre class="moz-quote-pre" wrap="">_______________________________________________
OvmsDev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:OvmsDev@lists.openvehicles.com" moz-do-not-send="true">OvmsDev@lists.openvehicles.com</a>
<a class="moz-txt-link-freetext" href="http://lists.openvehicles.com/mailman/listinfo/ovmsdev" moz-do-not-send="true">http://lists.openvehicles.com/mailman/listinfo/ovmsdev</a>
</pre>
                              </blockquote>
                              <br class="">
                              <br class="">
                              <pre class="moz-signature" cols="144">-- 
Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
Fon 02333 / 833 5735 * Handy 0176 / 206 989 26
</pre>
                            </div>
_______________________________________________<br class="">
                            OvmsDev mailing list<br class="">
                            <a
                              href="mailto:OvmsDev@lists.openvehicles.com"
                              class="" moz-do-not-send="true">OvmsDev@lists.openvehicles.com</a><br
                              class="">
                            <a class="moz-txt-link-freetext"
                              href="http://lists.openvehicles.com/mailman/listinfo/ovmsdev"
                              moz-do-not-send="true">http://lists.openvehicles.com/mailman/listinfo/ovmsdev</a><br
                              class="">
                          </div>
                        </blockquote>
                      </div>
                      <br class="">
                    </div>
                  </div>
                  <br class="">
                  <fieldset class="mimeAttachmentHeader"></fieldset>
                  <pre class="moz-quote-pre" wrap="">_______________________________________________
OvmsDev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:OvmsDev@lists.openvehicles.com" moz-do-not-send="true">OvmsDev@lists.openvehicles.com</a>
<a class="moz-txt-link-freetext" href="http://lists.openvehicles.com/mailman/listinfo/ovmsdev" moz-do-not-send="true">http://lists.openvehicles.com/mailman/listinfo/ovmsdev</a>
</pre>
                </blockquote>
                <br class="">
                <pre class="moz-signature" cols="160">-- 
Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
Fon 02333 / 833 5735 * Handy 0176 / 206 989 26
</pre>
              </div>
              _______________________________________________<br
                class="">
              OvmsDev mailing list<br class="">
              <a href="mailto:OvmsDev@lists.openvehicles.com" class=""
                moz-do-not-send="true">OvmsDev@lists.openvehicles.com</a><br
                class="">
              <a class="moz-txt-link-freetext" href="http://lists.openvehicles.com/mailman/listinfo/ovmsdev">http://lists.openvehicles.com/mailman/listinfo/ovmsdev</a><br
                class="">
            </div>
          </blockquote>
        </div>
        <br class="">
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <pre class="moz-quote-pre" wrap="">_______________________________________________
OvmsDev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:OvmsDev@lists.openvehicles.com">OvmsDev@lists.openvehicles.com</a>
<a class="moz-txt-link-freetext" href="http://lists.openvehicles.com/mailman/listinfo/ovmsdev">http://lists.openvehicles.com/mailman/listinfo/ovmsdev</a>
</pre>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="160">-- 
Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
Fon 02333 / 833 5735 * Handy 0176 / 206 989 26
</pre>
  </body>
</html>