Crash with the new compactor on 64-bit linux

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Crash with the new compactor on 64-bit linux

Levente Uzonyi
 
Hi All,

I've been seeing random crashes with the VM having the new compactor. It
takes anywhere between 5 mintues to a few hours for the crash to happen.
The crash dumps didn't seem to contain relevant information and sometimes
had no C call stack at all.
So I decided to run the VM with gdb to catch the bug, and got the
following:

Program received signal SIGSEGV, Segmentation fault.
lookupOrdinaryNoMNUEtcInClass (class=0) at
/home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
68119 /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c: No such file or directory.
(gdb) bt
#0  lookupOrdinaryNoMNUEtcInClass (class=0) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
#1  0x000000000043fd97 in lookupOrdinaryreceiver (selector=21887312, rcvr=<optimized out>) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:19195
#2  0x000000000048f128 in ceSICMiss (receiver=14637768) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/cogitX64.c:3232
#3  0x0000000000900461 in ?? ()
#4  0x00007ffffff98c51 in ?? ()
#5  0x000000000045af68 in interpret () at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:2650
#6  0x0000000000901010 in ?? ()
#7  0x0000000000df5ac8 in ?? ()
#8  0x0000000000df5780 in ?? ()
#9  0x0000000000df6da8 in ?? ()
#10 0x0000000000df6fe8 in ?? ()
#11 0x0000000000000000 in ?? ()
(gdb) info registers
rax            0x0 0
rbx            0x2e78e0 3045600
rcx            0x0 0
rdx            0x12e78e0 19822816
rsi            0x2e78e0 3045600
rdi            0x0 0
rbp            0x14df950 0x14df950
rsp            0x7ffffff98b60 0x7ffffff98b60
r8             0xffffffffffffffff -1
r9             0x2 2
r10            0x7ffffffaac00 140737488006144
r11            0xa087df 10520543
r12            0x14df950 21887312
r13            0xa087ef 10520559
r14            0x15903f0 22610928
r15            0x12e78e0 19822816
rip            0x43f9b6 0x43f9b6 <lookupOrdinaryNoMNUEtcInClass+38>
eflags         0x10283 [ CF SF IF RF ]
cs             0x33 51
ss             0x2b 43
ds             0x0 0
es             0x0 0
fs             0x0 0
gs             0x0 0

I obviously didn't have any source files at those paths, but the line
numbers may still be of help.
The VM was cog_linux64x64_squeak.cog.spur_201702211732 from Travis with
the plugins copied over from some previous version, because, as I
mentioned it in another mail, newer builds come with no plugins at all.

Levente
Reply | Threaded
Open this post in threaded view
|

Re: Crash with the new compactor on 64-bit linux

Eliot Miranda-2
 
Hi Levente,

    any clues on how to reproduce this?  I don't mind reproducibility, but being able to kick off a press that has a good chance of crashing eventually would be progress :-)

On Mon, Feb 27, 2017 at 7:52 AM, Levente Uzonyi <[hidden email]> wrote:

Hi All,

I've been seeing random crashes with the VM having the new compactor. It takes anywhere between 5 mintues to a few hours for the crash to happen.
The crash dumps didn't seem to contain relevant information and sometimes had no C call stack at all.
So I decided to run the VM with gdb to catch the bug, and got the following:

Program received signal SIGSEGV, Segmentation fault.
lookupOrdinaryNoMNUEtcInClass (class=0) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
68119   /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c: No such file or directory.
(gdb) bt
#0  lookupOrdinaryNoMNUEtcInClass (class=0) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
#1  0x000000000043fd97 in lookupOrdinaryreceiver (selector=21887312, rcvr=<optimized out>) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:19195
#2  0x000000000048f128 in ceSICMiss (receiver=14637768) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/cogitX64.c:3232
#3  0x0000000000900461 in ?? ()
#4  0x00007ffffff98c51 in ?? ()
#5  0x000000000045af68 in interpret () at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:2650
#6  0x0000000000901010 in ?? ()
#7  0x0000000000df5ac8 in ?? ()
#8  0x0000000000df5780 in ?? ()
#9  0x0000000000df6da8 in ?? ()
#10 0x0000000000df6fe8 in ?? ()
#11 0x0000000000000000 in ?? ()
(gdb) info registers
rax            0x0      0
rbx            0x2e78e0 3045600
rcx            0x0      0
rdx            0x12e78e0        19822816
rsi            0x2e78e0 3045600
rdi            0x0      0
rbp            0x14df950        0x14df950
rsp            0x7ffffff98b60   0x7ffffff98b60
r8             0xffffffffffffffff       -1
r9             0x2      2
r10            0x7ffffffaac00   140737488006144
r11            0xa087df 10520543
r12            0x14df950        21887312
r13            0xa087ef 10520559
r14            0x15903f0        22610928
r15            0x12e78e0        19822816
rip            0x43f9b6 0x43f9b6 <lookupOrdinaryNoMNUEtcInClass+38>
eflags         0x10283  [ CF SF IF RF ]
cs             0x33     51
ss             0x2b     43
ds             0x0      0
es             0x0      0
fs             0x0      0
gs             0x0      0

I obviously didn't have any source files at those paths, but the line numbers may still be of help.
The VM was cog_linux64x64_squeak.cog.spur_201702211732 from Travis with the plugins copied over from some previous version, because, as I mentioned it in another mail, newer builds come with no plugins at all.

Levente



--
_,,,^..^,,,_
best, Eliot
Reply | Threaded
Open this post in threaded view
|

Re: Crash with the new compactor on 64-bit linux

Levente Uzonyi
 
IIRC it mostly happened when I tried to save code with Monticello by
pressing the Save or the Changes button, but most of the time that just
works.

Levente
Reply | Threaded
Open this post in threaded view
|

Re: Crash with the new compactor on 64-bit linux

Eliot Miranda-2
 
Hi Levente,

On Mon, Feb 27, 2017 at 1:25 PM, Levente Uzonyi <[hidden email]> wrote:

IIRC it mostly happened when I tried to save code with Monticello by pressing the Save or the Changes button, but most of the time that just works.

That's weird. I've been using the new compactor for over a week now under my Spur work image and have never seen a crash, let along in Monticello, and I'm saving VMMaker.oscog which is large.  What's the version info for th eVM you're using?


_,,,^..^,,,_
best, Eliot
Reply | Threaded
Open this post in threaded view
|

Re: Crash with the new compactor on 64-bit linux

Eliot Miranda-2
In reply to this post by Levente Uzonyi
 
Hi Levente, Hi All,

On Mon, Feb 27, 2017 at 7:52 AM, Levente Uzonyi <[hidden email]> wrote:

Hi All,

I've been seeing random crashes with the VM having the new compactor. It takes anywhere between 5 mintues to a few hours for the crash to happen.
The crash dumps didn't seem to contain relevant information and sometimes had no C call stack at all.
So I decided to run the VM with gdb to catch the bug, and got the following:

Program received signal SIGSEGV, Segmentation fault.
lookupOrdinaryNoMNUEtcInClass (class=0) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
68119   /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c: No such file or directory.
(gdb) bt
#0  lookupOrdinaryNoMNUEtcInClass (class=0) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:68119
#1  0x000000000043fd97 in lookupOrdinaryreceiver (selector=21887312, rcvr=<optimized out>) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:19195
#2  0x000000000048f128 in ceSICMiss (receiver=14637768) at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/cogitX64.c:3232
#3  0x0000000000900461 in ?? ()
#4  0x00007ffffff98c51 in ?? ()
#5  0x000000000045af68 in interpret () at /home/travis/build/OpenSmalltalk/opensmalltalk-vm/spur64src/vm/gcc3x-cointerp.c:2650
#6  0x0000000000901010 in ?? ()
#7  0x0000000000df5ac8 in ?? ()
#8  0x0000000000df5780 in ?? ()
#9  0x0000000000df6da8 in ?? ()
#10 0x0000000000df6fe8 in ?? ()
#11 0x0000000000000000 in ?? ()
(gdb) info registers
rax            0x0      0
rbx            0x2e78e0 3045600
rcx            0x0      0
rdx            0x12e78e0        19822816
rsi            0x2e78e0 3045600
rdi            0x0      0
rbp            0x14df950        0x14df950
rsp            0x7ffffff98b60   0x7ffffff98b60
r8             0xffffffffffffffff       -1
r9             0x2      2
r10            0x7ffffffaac00   140737488006144
r11            0xa087df 10520543
r12            0x14df950        21887312
r13            0xa087ef 10520559
r14            0x15903f0        22610928
r15            0x12e78e0        19822816
rip            0x43f9b6 0x43f9b6 <lookupOrdinaryNoMNUEtcInClass+38>
eflags         0x10283  [ CF SF IF RF ]
cs             0x33     51
ss             0x2b     43
ds             0x0      0
es             0x0      0
fs             0x0      0
gs             0x0      0

I obviously didn't have any source files at those paths, but the line numbers may still be of help.
The VM was cog_linux64x64_squeak.cog.spur_201702211732 from Travis with the plugins copied over from some previous version, because, as I mentioned it in another mail, newer builds come with no plugins at all.

Next time can you include the info in the crash dump or the output of -version?  It's much more informative.  It looks like this:

spur64cfvm -version
/Users/eliot/oscogvm/build.macos64x64/squeak.cog.spur/CocoaFast.app/Contents/MacOS/Squeak
5.0 5.0.201702250108 Mac OS X built on Feb 24 2017 17:17:00 PST Compiler: 4.2.1 Compatible Apple LLVM 7.0.0 (clang-700.1.76) [Production Spur 64-bit VM]
CoInterpreter VMMaker.oscog-eem.2138 uuid: 8b98184c-9061-45e1-b4ce-bd9b8bc8c802 Feb 24 2017
StackToRegisterMappingCogit VMMaker.oscog-eem.2139 uuid: a905d032-fb67-4132-a5bf-5d59e668f2e5 Feb 24 2017
VM: 201702250108 eliot@Sisyphus:oscogvm $ Date: Fri Feb 24 17:08:29 2017 -0800 $
Plugins: 201702250108 eliot@Sisyphus:oscogvm $

excerpt from crash.mp:
Segmentation fault Tue Jan 17 10:39:18 2017


/scratch/bwesterg/vmwork/vm/products/nscogspurlinux/lib/nsvm/5.0-201701142054/nsvm
Newspeak VM version: 5.0-201701142054  Tue Jan 17 10:31:43 PST 2017 gcc 4.8.3 [Production Spur ITHB VM]
Built from: CoInterpreter VMMaker.oscogSPP-eem.2091 uuid: 8df2d889-5c88-4a52-b3f4-c94b29b0ef80 Jan 17 2017
With: StackToRegisterMappingCogit VMMaker.oscog-eem.2082 uuid: f987e1db-f4c0-40ed-86ec-7a478cb89067 Jan 17 2017
Revision: VM: 201701142054 bwesterg@epoxy:/scratch/bwesterg/vmwork/vm $ Date: Sat Jan 14 12:54:19 2017 -0800 $
Plugins: 201701142054 bwesterg@epoxy:/scratch/bwesterg/vmwork/vm $
Build host: Linux epoxy 2.6.18-194.el5 #1 SMP Tue Mar 16 21:52:39 EDT 2010 i686 i686 i386 GNU/Linux
plugin path: ../vm/products/nscogspurlinux/lib/nsvm/5.0-201701142054 [default: /scratch/bwesterg/vmwork/vm/products/nscogspurlinux/lib/nsvm/5.0-201701142054/]

Thanks!
_,,,^..^,,,_
best, Eliot