[vm-dev] latest VM crashed on raspberry

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

[vm-dev] latest VM crashed on raspberry

Denis Kudriashov
 
Hi.

I tested latest vm on raspberry

squeak -headless Squeak6.0alpha-16548-32bit.image


it crashed with attached dump with message:

stack page bytes 4096 available headroom 2788 minimum unused headroom 3520


(Segmentation fault)


Details on OS:  Linux raspberrypi 3.12.28+ #709 PREEMPT Mon Sep 8 15:28:00 BST 2014 armv6l GNU/Linux

I have also some old VM which is working fine:
CoInterpreter * VMMaker.oscog-eem.2107 uuid: 19c0fa53-acc2-40f9-9a07-17510e614ae5 Jan 23 2017
StackToRegisterMappingCogit * VMMaker.oscog-eem.2107 uuid: 19c0fa53-acc2-40f9-9a07-17510e614ae5 Jan 23 2017
VM: 201701231021 https://github.com/pharo-project/pharo-vm.git $ Date: Mon Jan 23 11:21:48 2017 +0100 $ Plugins: 201701231021 https://github.com/pharo-project/pharo-vm.git $

Would be nice to get it fixed for upcoming Pharo releaze.

Best regards,
Denis

respberryCrash.dmp (6K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 

> On 25-04-2017, at 10:38 AM, Denis Kudriashov <[hidden email]> wrote:
>
> Details on OS:  Linux raspberrypi 3.12.28+ #709 PREEMPT Mon Sep 8 15:28:00 BST 2014 armv6l GNU/Linux

For starters you need to update that OS. That’s over two years out of date!

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: LD: Lose Device


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Ben Coman
In reply to this post by Denis Kudriashov
 

On Wed, Apr 26, 2017 at 1:38 AM, Denis Kudriashov <[hidden email]> wrote:
 
Hi.

I tested latest vm on raspberry

squeak -headless Squeak6.0alpha-16548-32bit.image


it crashed with attached dump with message:

stack page bytes 4096 available headroom 2788 minimum unused headroom 3520


(Segmentation fault)


Details on OS:  Linux raspberrypi 3.12.28+ #709 PREEMPT Mon Sep 8 15:28:00 BST 2014 armv6l GNU/Linux

I have also some old VM which is working fine:
CoInterpreter * VMMaker.oscog-eem.2107 uuid: 19c0fa53-acc2-40f9-9a07-17510e614ae5 Jan 23 2017
StackToRegisterMappingCogit * VMMaker.oscog-eem.2107 uuid: 19c0fa53-acc2-40f9-9a07-17510e614ae5 Jan 23 2017
VM: 201701231021 https://github.com/pharo-project/pharo-vm.git $ Date: Mon Jan 23 11:21:48 2017 +0100 $ Plugins: 201701231021 https://github.com/pharo-project/pharo-vm.git $

Would be nice to get it fixed for upcoming Pharo releaze.

Best regards,
Denis


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 

On 26-04-2017, at 6:49 AM, Ben Coman <[hidden email]> wrote:


btw, just to check... ARM6 is The Rpi right?

Yup. Although strictly speaking something over 80% of all Pi are v7 cpus the ‘rule’ is to compile v6 to avoid complicated multiple executables. Given that the vm isn’t actually all that big I’d say we could get away with it if we really wanted but so far I’ve not heard any evidence of a major performance difference.

The change I really want to make is going for the AArch64 target. It wouldn’t make a giant difference on a current Pi  (maybe 20%) but it would open up the coming world of 64bit ARM OSs with 3+GHz 48-core CPUs and so on. All I need is 

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Don't document the program; program the document.


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Denis Kudriashov
 
Now I found last working VM:

CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017

StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017

VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $



Also I checked that latest stackVM is working. And I attach newest crash.dmp for today spur vm 

2017-04-26 19:15 GMT+02:00 tim Rowledge <[hidden email]>:
 

On 26-04-2017, at 6:49 AM, Ben Coman <[hidden email]> wrote:


btw, just to check... ARM6 is The Rpi right?

Yup. Although strictly speaking something over 80% of all Pi are v7 cpus the ‘rule’ is to compile v6 to avoid complicated multiple executables. Given that the vm isn’t actually all that big I’d say we could get away with it if we really wanted but so far I’ve not heard any evidence of a major performance difference.

The change I really want to make is going for the AArch64 target. It wouldn’t make a giant difference on a current Pi  (maybe 20%) but it would open up the coming world of 64bit ARM OSs with 3+GHz 48-core CPUs and so on. All I need is 

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Don't document the program; program the document.





raspberryCrash.dmp (6K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Denis Kudriashov
In reply to this post by Ben Coman
 

2017-04-26 15:49 GMT+02:00 Ben Coman <[hidden email]>:

With Pharo result the same. It crashed
Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
In reply to this post by Denis Kudriashov
 

> On 25-04-2017, at 10:38 AM, Denis Kudriashov <[hidden email]> wrote:
>
> Hi.
>
> I tested latest vm on raspberry
> squeak -headless Squeak6.0alpha-16548-32bit.image
>
> it crashed with attached dump with message:
> stack page bytes 4096 available headroom 2788 minimum unused headroom 3520
>

I can now confirm seeing this on my Pi systems, using a built-from-yesterday code. Guess we have some tracking down to do.


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: RCR: Rewind Card Reader


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Nicolas Cellier
 
I had several crash dump on MacOS too, but rather random (only the end of file joined might be relevant)


2017-04-28 21:03 GMT+02:00 tim Rowledge <[hidden email]>:


> On 25-04-2017, at 10:38 AM, Denis Kudriashov <[hidden email]> wrote:
>
> Hi.
>
> I tested latest vm on raspberry
> squeak -headless Squeak6.0alpha-16548-32bit.image
>
> it crashed with attached dump with message:
> stack page bytes 4096 available headroom 2788 minimum unused headroom 3520
>

I can now confirm seeing this on my Pi systems, using a built-from-yesterday code. Guess we have some tracking down to do.


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: RCR: Rewind Card Reader



Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 
I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
In computer science, we stand on each other's feet.


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 

> On 28-04-2017, at 3:34 PM, tim Rowledge <[hidden email]> wrote:
>
> I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
> and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …

And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Fractured Idiom:- VISA LA FRANCE - Don't leave chateau without it


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Denis Kudriashov
 

2017-04-29 21:18 GMT+02:00 tim Rowledge <[hidden email]>:
> I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
> and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …

And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.

Tim, you are probably skipped my mail:

Now I found last working VM:
CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $


(I already checked bintray binaries)

Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 

> On 29-04-2017, at 12:26 PM, Denis Kudriashov <[hidden email]> wrote:
>
>
> 2017-04-29 21:18 GMT+02:00 tim Rowledge <[hidden email]>:
> > I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
> > and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …
>
> And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.
>
> Tim, you are probably skipped my mail:
>
> Now I found last working VM:
> CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
> StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
> VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $
>
> (I already checked bintray binaries)

Ah, I think it more likely that I simply didn’t properly see the date in there. And it’s good to get replication of data too; it makes it less likely that we are hallucinating :-)


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- One sandwich short of a picnic.


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

timrowledge
 
Just to complete the story (hopefully) the latest (as of may 5th 2017) generated code seems to run perfectly well on my Pi. AIUI the proximate cause of the problem was a missed saving of the link register in the new has primitive code.

> On 29-04-2017, at 12:28 PM, tim Rowledge <[hidden email]> wrote:
>
>
>> On 29-04-2017, at 12:26 PM, Denis Kudriashov <[hidden email]> wrote:
>>
>>
>> 2017-04-29 21:18 GMT+02:00 tim Rowledge <[hidden email]>:
>>> I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
>>> and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …
>>
>> And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.
>>
>> Tim, you are probably skipped my mail:
>>
>> Now I found last working VM:
>> CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $
>>
>> (I already checked bintray binaries)
>
> Ah, I think it more likely that I simply didn’t properly see the date in there. And it’s good to get replication of data too; it makes it less likely that we are hallucinating :-)
>
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Useful random insult:- One sandwich short of a picnic.
>
>


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: LJD: Lock Job on Disk


Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Denis Kudriashov
 
So Eliot thank you very much.

I just checked: it works perfectly

2017-05-08 3:15 GMT+02:00 tim Rowledge <[hidden email]>:

Just to complete the story (hopefully) the latest (as of may 5th 2017) generated code seems to run perfectly well on my Pi. AIUI the proximate cause of the problem was a missed saving of the link register in the new has primitive code.

> On 29-04-2017, at 12:28 PM, tim Rowledge <[hidden email]> wrote:
>
>
>> On 29-04-2017, at 12:26 PM, Denis Kudriashov <[hidden email]> wrote:
>>
>>
>> 2017-04-29 21:18 GMT+02:00 tim Rowledge <[hidden email]>:
>>> I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
>>> and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …
>>
>> And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.
>>
>> Tim, you are probably skipped my mail:
>>
>> Now I found last working VM:
>> CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $
>>
>> (I already checked bintray binaries)
>
> Ah, I think it more likely that I simply didn’t properly see the date in there. And it’s good to get replication of data too; it makes it less likely that we are hallucinating :-)
>
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Useful random insult:- One sandwich short of a picnic.
>
>


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: LJD: Lock Job on Disk



Reply | Threaded
Open this post in threaded view
|

Re: [vm-dev] latest VM crashed on raspberry

Eliot Miranda-2
 
Hi Denis,

On May 10, 2017, at 6:32 AM, Denis Kudriashov <[hidden email]> wrote:

So Eliot thank you very much.

I just checked: it works perfectly

you're welcome!  Sorry for breaking it.


2017-05-08 3:15 GMT+02:00 tim Rowledge <[hidden email]>:

Just to complete the story (hopefully) the latest (as of may 5th 2017) generated code seems to run perfectly well on my Pi. AIUI the proximate cause of the problem was a missed saving of the link register in the new has primitive code.

> On 29-04-2017, at 12:28 PM, tim Rowledge <[hidden email]> wrote:
>
>
>> On 29-04-2017, at 12:26 PM, Denis Kudriashov <[hidden email]> wrote:
>>
>>
>> 2017-04-29 21:18 GMT+02:00 tim Rowledge <[hidden email]>:
>>> I managed to find a 2017-01-28 ARM vm (by accident, on https://github.com/OpenSmalltalk/opensmalltalk-vm/releases after hitting the ‘Releases’ button)
>>> and that appears to run OK. So at least we don’t have to go all the way back to 5.1 …
>>
>> And after finally finding a pointer to this bintray thing I can reasonably certainly claim that the break is somewhere between the 20170227 and 20170327 versions.
>>
>> Tim, you are probably skipped my mail:
>>
>> Now I found last working VM:
>> CoInterpreter VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> StackToRegisterMappingCogit VMMaker.oscog-eem.2134 uuid: 4721ad0c-159c-4bf2-9f3e-c9917fcdead5 Feb 27 2017
>> VM: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $ Date: Mon Feb 27 10:43:22 2017 +0100 $ Plugins: 201702270943 https://github.com/OpenSmalltalk/opensmalltalk-vm.git $
>>
>> (I already checked bintray binaries)
>
> Ah, I think it more likely that I simply didn’t properly see the date in there. And it’s good to get replication of data too; it makes it less likely that we are hallucinating :-)
>
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Useful random insult:- One sandwich short of a picnic.
>
>


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: LJD: Lock Job on Disk