KiCad segfault.

[Wayne Stambaugh]

I am seeing intermittent segfaults of KiCad. I'm not sure what the specific actions are to make this happen but there doesn't seem to be a pattern. This is as of product branch r5468. Here is the code that causes the segfault:

#0 0x0000000009c9798c in PNS_NODE::isRoot (this=0xfeeefeeefeeefeee)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_node.h:414
#1 0x00000000099743d9 in PNS_NODE::KillChildren (this=0xfeeefeeefeeefeee)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_node.cpp:1141
#2 0x000000000994d6cd in PNS_ROUTER::ClearWorld (this=0x548ff6c0)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_router.cpp:391
#3 0x000000000994d551 in PNS_ROUTER::~PNS_ROUTER (this=0x548ff6c0,
    __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_router.cpp:379
#4 0x00000000099512d3 in PNS_TOOL_BASE::~PNS_TOOL_BASE (this=0x548eaa90,
    __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_tool_base.cpp:72
#5 0x0000000009949fdb in LENGTH_TUNER_TOOL::~LENGTH_TUNER_TOOL (
    this=0x548eaa90, __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/length_tuner_tool.cpp:104
#6 0x000000000994a027 in LENGTH_TUNER_TOOL::~LENGTH_TUNER_TOOL (
    this=0x548eaa90, __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/length_tuner_tool.cpp:107
#7 0x00000000099f1e31 in TOOL_MANAGER::~TOOL_MANAGER (this=0x548d8d40,
    __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/common/tool/tool_manager.cpp:217
#8 0x000000000989da3b in PCB_BASE_FRAME::~PCB_BASE_FRAME (this=0x52d1dc80,
    __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/basepcbframe.cpp:123
#9 0x0000000009b0b686 in PCB_BASE_EDIT_FRAME::~PCB_BASE_EDIT_FRAME (
    this=0x52d1dc80, __in_chrg=<optimized out>)
    at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/pcb_base_edit_frame.h:43

It appears that PNS_NODEs are being used after they have been freed.

[Wayne Stambaugh (stambaughw)]

Here's the build info:

Application: kicad
Version: (after 2015-jan-16 BZR unknown)-product Debug build
wxWidgets: Version 3.0.2 (debug,wchar_t,compiler with C++ ABI 1002,GCC 4.9.2,wx containers,compatible with 2.8)
Platform: Windows 7 (build 7601, Service Pack 1), 64-bit edition, 64 bit, Little endian, wxMSW
Boost version: 1.57.0
         USE_WX_GRAPHICS_CONTEXT=OFF
         USE_WX_OVERLAY=OFF
         KICAD_SCRIPTING=ON
         KICAD_SCRIPTING_MODULES=ON
         KICAD_SCRIPTING_WXPYTHON=ON
         USE_FP_LIB_TABLE=HARD_CODED_ON
         BUILD_GITHUB_PLUGIN=ON
         KICAD_USE_WEBKIT=ON

[Nick Østergaard (nickoe)]

This looks like it is related to the new CERN contributions, so I will tag it as such for their convience.

[Nick Østergaard (nickoe)]

I get something similar, but not quite the same. Sorry for not being a debug build.

This is when having routed a diff pair, then use the single trace length tuner on one trace, and clicking a lot on the second trace after the first one has been tuned.

(gdb) bt
#0 0x00007fffe9062405 in ?? () from /usr/bin/_pcbnew.kiface
#1 0x00007fffe9022244 in ?? () from /usr/bin/_pcbnew.kiface
#2 0x00007fffe905be66 in ?? () from /usr/bin/_pcbnew.kiface
#3 0x00007fffe90226cf in ?? () from /usr/bin/_pcbnew.kiface
#4 0x00007fffe901f55c in LENGTH_TUNER_TOOL::performTuning() () from /usr/bin/_pcbnew.kiface
#5 0x00007fffe902001d in LENGTH_TUNER_TOOL::mainLoop(PNS_ROUTER_MODE) () from /usr/bin/_pcbnew.kiface
#6 0x00007fffe9020ea8 in LENGTH_TUNER_TOOL::TuneSingleTrApplication: pcbnew
Version: (2015-03-03 BZR 5469)-kicad Release build
wxWidgets: Version 3.0.2 (debug,wchar_t,compiler with C++ ABI 1002,GCC 4.9.2,wx containers,compatible with 2.8)
Platform: Linux 3.18.6-1-ARCH x86_64, 64 bit, Little endian, wxGTK
Boost version: 1.57.0
         USE_WX_GRAPHICS_CONTEXT=OFF
         USE_WX_OVERLAY=OFF
         KICAD_SCRIPTING=ON
         KICAD_SCRIPTING_MODULES=ON
         KICAD_SCRIPTING_WXPYTHON=ON
         USE_FP_LIB_TABLE=HARD_CODED_ON
         BUILD_GITHUB_PLUGIN=ON
         KICAD_USE_WEBKIT=OFF
ace(TOOL_EVENT const&) () from /usr/bin/_pcbnew.kiface
#7 0x00007fffe9156371 in ?? () from /usr/bin/_pcbnew.kiface
#8 0x00007fffe6a2c731 in make_fcontext () from /usr/lib/libboost_context.so.1.57.0
#9 0x0000000000000000 in ?? ()

With:
Application: pcbnew
Version: (2015-03-03 BZR 5469)-kicad Release build
wxWidgets: Version 3.0.2 (debug,wchar_t,compiler with C++ ABI 1002,GCC 4.9.2,wx containers,compatible with 2.8)
Platform: Linux 3.18.6-1-ARCH x86_64, 64 bit, Little endian, wxGTK
Boost version: 1.57.0
         USE_WX_GRAPHICS_CONTEXT=OFF
         USE_WX_OVERLAY=OFF
         KICAD_SCRIPTING=ON
         KICAD_SCRIPTING_MODULES=ON
         KICAD_SCRIPTING_WXPYTHON=ON
         USE_FP_LIB_TABLE=HARD_CODED_ON
         BUILD_GITHUB_PLUGIN=ON
         KICAD_USE_WEBKIT=OFF

[Nick Østergaard (nickoe)]

Pretty easy to replicate my use case:

1. Draw track between two pads
2. Select tune track pair
3. Click a few time on a track fast.

Also, I note that I get a gray box in the 0,0 in screencoordinates. Some gray as that hovering info box. Maybe it has something to do with that.

[Wayne Stambaugh (stambaughw)]

I just committed the fix for the problem I was seeing. If this doesn't
fix your problem, please file a separate bug report so it doesn't get lost.

On 3/3/2015 3:08 PM, Nick Østergaard wrote:
> Pretty easy to replicate my use case:
>
> 1. Draw track between two pads
> 2. Select tune track pair
> 3. Click a few time on a track fast.
>
> Also, I note that I get a gray box in the 0,0 in screencoordinates. Some
> gray as that hovering info box. Maybe it has something to do with that.
>

[Nick Østergaard (nickoe)]

Sure. I am just recompiling.

2015-03-03 21:15 GMT+01:00 Wayne Stambaugh <email address hidden>:
> I just committed the fix for the problem I was seeing. If this doesn't
> fix your problem, please file a separate bug report so it doesn't get lost.
>
> On 3/3/2015 3:08 PM, Nick Østergaard wrote:
>> Pretty easy to replicate my use case:
>>
>> 1. Draw track between two pads
>> 2. Select tune track pair
>> 3. Click a few time on a track fast.
>>
>> Also, I note that I get a gray box in the 0,0 in screencoordinates. Some
>> gray as that hovering info box. Maybe it has something to do with that.
>>
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1427803
>
> Title:
> KiCad segfault.
>
> Status in KiCad EDA Software Suite:
> Fix Committed
>
> Bug description:
> I am seeing intermittent segfaults of KiCad. I'm not sure what the
> specific actions are to make this happen but there doesn't seem to be
> a pattern. This is as of product branch r5468. Here is the code that
> causes the segfault:
>
> #0 0x0000000009c9798c in PNS_NODE::isRoot (this=0xfeeefeeefeeefeee)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_node.h:414
> #1 0x00000000099743d9 in PNS_NODE::KillChildren (this=0xfeeefeeefeeefeee)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_node.cpp:1141
> #2 0x000000000994d6cd in PNS_ROUTER::ClearWorld (this=0x548ff6c0)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_router.cpp:391
> #3 0x000000000994d551 in PNS_ROUTER::~PNS_ROUTER (this=0x548ff6c0,
> __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_router.cpp:379
> #4 0x00000000099512d3 in PNS_TOOL_BASE::~PNS_TOOL_BASE (this=0x548eaa90,
> __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/pns_tool_base.cpp:72
> #5 0x0000000009949fdb in LENGTH_TUNER_TOOL::~LENGTH_TUNER_TOOL (
> this=0x548eaa90, __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/length_tuner_tool.cpp:104
> #6 0x000000000994a027 in LENGTH_TUNER_TOOL::~LENGTH_TUNER_TOOL (
> this=0x548eaa90, __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/router/length_tuner_tool.cpp:107
> #7 0x00000000099f1e31 in TOOL_MANAGER::~TOOL_MANAGER (this=0x548d8d40,
> __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/common/tool/tool_manager.cpp:217
> #8 0x000000000989da3b in PCB_BASE_FRAME::~PCB_BASE_FRAME (this=0x52d1dc80,
> __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/basepcbframe.cpp:123
> #9 0x0000000009b0b686 in PCB_BASE_EDIT_FRAME::~PCB_BASE_EDIT_FRAME (
> this=0x52d1dc80, __in_chrg=<optimized out>)
> at C:/msys64/home/wstambaugh/src/kicad/product/pcbnew/pcb_base_edit_frame.h:43
>
> It appears that PNS_NODEs are being used after they have been freed.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/kicad/+bug/1427803/+sub...

[Wayne Stambaugh (stambaughw)]

You may want to stop your build and update your source branch. I just committed another P&S router segfault fix which might be what you are seeing.

[Nick Østergaard (nickoe)]

The commit in 5471 fixed the segfault from my description, but I still have the issue with the gray box appearing on top of every thing on the screencoordinates origin. This happens as soon as I click the track with the tune track lenght tool. It is important not to move the cursor after this is done, otherwise the box will dissapear (and I guess turn into the correct info box near the cursor).

[I have not created a new bug on this, but I can do that tomorrow if not fixed before.]

[Nick Østergaard (nickoe)]

Just FYI regarding the description in #8, it has been reported and it has been fixed. So lets leave it here.