Re: RFR (XXS): JDK-8138610 add assert to ThreadLocalAllocBuffer::clear_before_allocation to check the storage above top

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: RFR (XXS): JDK-8138610 add assert to ThreadLocalAllocBuffer::clear_before_allocation to check the storage above top

Joseph Provino
Hi, could I get reviews for this trivial change?

Thanks.

joe

>> On 14/02/2017 4:57 AM, Joseph Provino wrote:
>>> Please review this one line change:
>>>
>>> CR: JDK-8138610 <https://bugs.openjdk.java.net/browse/JDK-8138610> add
>>> assert to ThreadLocalAllocBuffer::clear_before_allocation to check the
>>> storage above top
>>>
>>> Webrev: http://cr.openjdk.java.net/~jprovino/8138610/webrev.00
>>>
>>> Tests:  Passes JPRT.
>>>
>>> thanks.
>>>
>>> joe
>>>
>

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: RFR (XXS): JDK-8138610 add assert to ThreadLocalAllocBuffer::clear_before_allocation to check the storage above top

Kim Barrett
> On Mar 2, 2017, at 2:24 PM, Joseph Provino <[hidden email]> wrote:
>
> Hi, could I get reviews for this trivial change?
>
> Thanks.
>
> joe
>
>>> On 14/02/2017 4:57 AM, Joseph Provino wrote:
>>>> Please review this one line change:
>>>>
>>>> CR: JDK-8138610 <https://bugs.openjdk.java.net/browse/JDK-8138610> add
>>>> assert to ThreadLocalAllocBuffer::clear_before_allocation to check the
>>>> storage above top
>>>>
>>>> Webrev: http://cr.openjdk.java.net/~jprovino/8138610/webrev.00
>>>>
>>>> Tests:  Passes JPRT.
>>>>
>>>> thanks.
>>>>
>>>> joe

Calling this a trivial change is a bit of a stretch, as the analysis
needed to understand it is extensive.

And having done some of that, I'm not sure the proposed assertion
is correct.

I *think* we can get here with a completely full TLAB, in which case I
think *top() may be looking at who knows what, possibly even
inaccessible memory.

The CR doesn't provide much motivation for this assertion.  It looks
to me like somebody's leftover debugging code, that had been buried in
and extracted from the large JVMCI addition.

Loading...