<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>