Request for review: 8014138: Add VM option to facilitate the writing of CDS tests

harold seigel harold.seigel at oracle.com
Thu May 9 12:16:25 PDT 2013


I'll clean them up along with the strcpy().

Thanks, Harold

On 5/9/2013 3:14 PM, Daniel D. Daugherty wrote:
> On 5/9/13 1:10 PM, harold seigel wrote:
>> I guess Parfait can't detect that strcat is copying a literal string?
>
> The problem is overflow of the destination parameter.It presumes
> if you are using the strncat() version that you're actually passing
> a valid "remaining available space" value.
>
> Dan
>
>
>>
>> On 5/9/2013 3:06 PM, Daniel D. Daugherty wrote:
>>> On 5/9/13 12:56 PM, harold seigel wrote:
>>>> Hi Zhengyu,
>>>>
>>>> Thanks for looking at this.  Comments inline.
>>>>
>>>> Harold
>>>>
>>>> On 5/9/2013 2:35 PM, Zhengyu Gu wrote:
>>>>> Hi Harold,
>>>>>
>>>>> arguments.cpp
>>>>>
>>>>> 3186 - 3187: why not  strcat(shared_archive_path, "classes.jsa")
>>>> I'll clean that up.
>>>
>>> I think strcat() will set off a Parfait warning. You should use
>>> strncat() with the appropriate length parameters...
>>>
>>> Dan
>>>
>>>
>>>>>
>>>>> Also, I don't see where SharedArchiveFile is used.
>>>> In function set_shared_archive_path(), if SharedArchiveFile is null 
>>>> then the original archive file path is used.
>>>> Else the value of SharedArchiveFile is copied to 
>>>> shared_archive_path and returned as the function value.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Zhengyu
>>>>>
>>>>>
>>>>>
>>>>> On 5/9/2013 1:44 PM, harold seigel wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Please review this fix for bug 8014138.
>>>>>>
>>>>>> This enhancement adds a new diagnostic option, 
>>>>>> "-XX:SharedArchiveFile=<file-path>" that takes a string 
>>>>>> specifying the path of the shared archive file. Note that if the 
>>>>>> option is specified when the shared archive is written, then it 
>>>>>> must be used when the shared archive is being read (i.e. 
>>>>>> -Xshare:on).
>>>>>>
>>>>>> This change was tested with JCK lang and vm tests, UTE 
>>>>>> vm.quick.testlist and vm.mlvm.testlist tests, and jtreg tests. In 
>>>>>> addition, the test included in this change was run on Linux 
>>>>>> 32/64, Solaris 32/64, Windows 7, and Mac OS.
>>>>>>
>>>>>> Open webrev at http://cr.openjdk.java.net/~hseigel/bug_8014138/ 
>>>>>> <http://cr.openjdk.java.net/%7Ehseigel/bug_8014138/>
>>>>>>
>>>>>> Bug link at http://bugs.sun.com/view_bug.do?bug_id=8014138
>>>>>>
>>>>>> Thanks, Harold
>>>>>
>>>>
>>>
>>
>



More information about the hotspot-runtime-dev mailing list