Smalltalk › Instantiations

9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

5 messages Options

Hans-Martin Mosner-3

9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

Hello,

today I found that 9.1 64 bit VM can be crashed by sending #isWeak or #isInFixedSpace to true or false (other primitives in that region might be affected as well).

Most likely these objects have object pointers that are not memory addresses, but the primitives don't check for that.

Since I did not switch to 9.2 yet, I did not test this with the 9.2 VM.

Cheers,

Hans-Martin

--
You received this message because you are subscribed to the Google Groups "VA Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/va-smalltalk/ee27902e-03d9-40bf-ba11-6723c8452083%40googlegroups.com.

Wayne Johnston

Re: 9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

"true isWeak" causes GPF on my 9.2 32-bit image on Windows.

--
You received this message because you are subscribed to the Google Groups "VA Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/va-smalltalk/945e76de-1dfa-4f2e-90d1-33824966ed39%40googlegroups.com.

Seth Berman

Re: 9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

Hi All,

I'll hand this off to the vm team.

I see that things work as expected in 9.2 64-bit.
And the vm code has the appropriate checks in it...so there must be something strange with the codegen for 9.2 32-bit.

Thanks for reporting.

I'll report back when I have an answer.

-Seth

On Tuesday, February 4, 2020 at 7:21:40 AM UTC-5, Wayne Johnston wrote:

"true isWeak" causes GPF on my 9.2 32-bit image on Windows.

--
You received this message because you are subscribed to the Google Groups "VA Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/va-smalltalk/3b5fcb14-397e-49d3-85ea-d94d3bff61d0%40googlegroups.com.

Seth Berman

Re: 9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

Hi All,

Small update. I can repeat in VAST 9.2 64-bit.

As I said, we will figure out what is happening and report back

- Seth

On Tuesday, February 4, 2020 at 8:19:47 AM UTC-5, Seth Berman wrote:

Hi All,

I'll hand this off to the vm team.
I see that things work as expected in 9.2 64-bit.
And the vm code has the appropriate checks in it...so there must be something strange with the codegen for 9.2 32-bit.
Thanks for reporting.
I'll report back when I have an answer.

-Seth

On Tuesday, February 4, 2020 at 7:21:40 AM UTC-5, Wayne Johnston wrote:
"true isWeak" causes GPF on my 9.2 32-bit image on Windows.

--
You received this message because you are subscribed to the Google Groups "VA Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/va-smalltalk/27552f21-5d3c-446c-841a-c6d42b4211f3%40googlegroups.com.

Seth Berman

Re: 9.1: Sending #isWeak or #isInFixedSpace to Booleans crashes 64-bit VM

Hi All,

We have made the appropriate adjustments to primitives VMprObjectIsFixed and VMprObjectIsWeak
and it will be in the forthcoming 9.2.1 patch which is scheduled for this month.

Thank you for reporting.

-Seth

On Tuesday, February 4, 2020 at 9:43:10 AM UTC-5, Seth Berman wrote:

Hi All,

Small update. I can repeat in VAST 9.2 64-bit.
As I said, we will figure out what is happening and report back

- Seth

On Tuesday, February 4, 2020 at 8:19:47 AM UTC-5, Seth Berman wrote:
Hi All,

I'll hand this off to the vm team.
I see that things work as expected in 9.2 64-bit.
And the vm code has the appropriate checks in it...so there must be something strange with the codegen for 9.2 32-bit.
Thanks for reporting.
I'll report back when I have an answer.

-Seth

On Tuesday, February 4, 2020 at 7:21:40 AM UTC-5, Wayne Johnston wrote:
"true isWeak" causes GPF on my 9.2 32-bit image on Windows.

--
You received this message because you are subscribed to the Google Groups "VA Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/va-smalltalk/b71a76c5-44af-4817-a7b7-1af1bd965798%40googlegroups.com.