RFR(XS): JDK-8068655 frame::safe_for_sender() computes incorrect sender_sp value for interpreted frames

[Frederic Parain]

Dan,

Thank you for your review.

My answers are below.

On 29/01/2015 19:24, Daniel D. Daugherty wrote:
> On 1/29/15 8:05 AM, Frederic Parain wrote:
>> Hi Dean,
>>
>> When unextended_sp is not specified, it is set to the sp value.
>>
>> However, after your question, I realized that my fix was not
>> correct but worked because the is_interpreted_frame_valid() had
>> a bug too, it uses sp() where unextended_sp() should be used.
>>
>> Checking the history of these files, is_interpreted_frame_valid() has
>> been written before the distinction between raw sp and unextended sp
>> has been introduced in the frame class. Unfortunately, when
>> unextended_sp was added, is_interpreted_frame_valid() has not been
>> updated.
>>
>> Here's a new webrev where safe_for_sender computes both raw_sp and
>> unextended_sp, and is_interpreted_frame_valid() checks the frame
>> size using unextended_sp instead of raw sp.
>>
>> http://cr.openjdk.java.net/~fparain/8068655/webrev.01/
>
> src/cpu/x86/vm/frame_x86.cpp
>      nit: line 158: saved_fp = (intptr_t*)*(sender_sp -
> frame::sender_sp_offset);
>          Please add a space between the cast and the ptr deref, i.e.:
>
>          line 158: saved_fp = (intptr_t*) *(sender_sp -
> frame::sender_sp_offset);

Done

>      line 214: frame sender(sender_sp, saved_fp, sender_pc);
>          This line is not changed based on the assumption that
>          (sender_sp == sender_unextended_sp). Do you want to
>          assert() that just before line 214?

After writing this assert it looked strange to me.
I removed it and used the 4 arguments constructor to pass the
unextended_sp computed line 154. It makes the code more regular
between interpreted and not interpreted frames.

>      line 580: if (fp() - unextended_sp() > 1024 + ...
>          Previous to this line, the "stack pointer" checks all use sp().
>          A short comment explaining the switch to unextended_sp() would
>          be good here.

Done

> Thumbs up. Your choice on whether to address my (very) minor
> comments on this round.

Thanks Dan,

Fred

>>
>> I also added more comments.
>>
>> Note that I only modify one test in is_interpreted_frame_valid().
>> Use of sp() or unextended_sp() could be discussed for each test,
>> but in their current form they are wrong.
>>
>> Thanks,
>>
>> Fred
>>
>> On 26/01/2015 20:40, Dean Long wrote:
>>> In sender_for_interpreter_frame(), we set the raw sender_sp but also
>>> unextended_sp.  Should
>>> safe_for_sender() be doing the same?
>>>
>>> dl
>>>
>>> On 1/26/2015 6:40 AM, Frederic Parain wrote:
>>>> Greetings,
>>>>
>>>> Please review this small fix in the frame verification code.
>>>> The bug report includes a detailed description of the issue.
>>>>
>>>> CR:
>>>> https://bugs.openjdk.java.net/browse/JDK-8068655
>>>>
>>>> Webrev:
>>>> http://cr.openjdk.java.net/~fparain/8068655/webrev.00/
>>>>
>>>> Thanks,
>>>>
>>>> Fred
>>>>
>>>
>>
>

-- 
Frederic Parain - Oracle
Grenoble Engineering Center - France
Phone: +33 4 76 18 81 17
Email: Frederic.Parain at oracle.com