RFR(XS): 8043264: hsdis library not picked up correctly on expected paths
serguei.spitsyn at oracle.com
serguei.spitsyn at oracle.com
Sat May 17 06:10:12 UTC 2014
Kris,
I've submitted the jprt job:
2014-05-17-060826.sspitsyn.hotspot
Thanks,
Serguei
On 5/16/14 3:38 PM, Krystal Mok wrote:
> Thank you, Serguei!
>
> BTW, Could you or Staffan sponsor this change and help me push it?
>
> Best regards,
> Kris
>
>
> On Fri, May 16, 2014 at 3:20 PM, serguei.spitsyn at oracle.com
> <mailto:serguei.spitsyn at oracle.com> <serguei.spitsyn at oracle.com
> <mailto:serguei.spitsyn at oracle.com>> wrote:
>
> On 5/15/14 1:19 PM, Krystal Mok wrote:
>> Hi everyone,
>>
>> May I have a couple of review for this small patch, please?
>>
>> *Bug*: https://bugs.openjdk.java.net/browse/JDK-8043264
>>
>> *Patch*: (against jdk9/hs)
>>
>> diff -r 466b58fa837b src/share/vm/compiler/disassembler.cpp
>> --- a/src/share/vm/compiler/disassembler.cppThu May 15 11:35:26
>> 2014 -0700
>> +++ b/src/share/vm/compiler/disassembler.cppThu May 15 13:14:58
>> 2014 -0700
>> @@ -86,7 +86,7 @@
>> {
>> // Match "jvm[^/]*" in jvm_path.
>> const char* base = buf;
>> - const char* p = strrchr(buf, '/');
>> + const char* p = strrchr(buf, *os::file_separator());
>> if (p != NULL) lib_offset = p - base + 1;
>> p = strstr(p ? p : base, "jvm");
>> if (p != NULL) jvm_offset = p - base;
>> @@ -111,7 +111,7 @@
>> if (_library == NULL) {
>> // 3. <home>/jre/lib/<arch>/hsdis-<arch>.so
>> buf[lib_offset - 1] = '\0';
>> - const char* p = strrchr(buf, '/');
>> + const char* p = strrchr(buf, *os::file_separator());
>> if (p != NULL) {
>> lib_offset = p - buf + 1;
>> strcpy(&buf[lib_offset], hsdis_library_name);
>
> The fix looks good.
>
> Thanks,
> Serguei
>
>
>>
>> *Description*:
>>
>> (Copied from the bug)
>>
>> On Windows, the hsdis library isn't picked up correctly in all
>> expected paths described in Disassembler::load_library():
>>
>> // Find the disassembler shared library.
>> // Search for several paths derived from libjvm, in this order:
>> // 1. <home>/jre/lib/<arch>/<vm>/libhsdis-<arch>.so (for
>> compatibility)
>> // 2. <home>/jre/lib/<arch>/<vm>/hsdis-<arch>.so
>> // 3. <home>/jre/lib/<arch>/hsdis-<arch>.so
>> // 4. hsdis-<arch>.so (using LD_LIBRARY_PATH)
>>
>> The reason is that the code that concatenates the paths doesn't
>> take os::file_separator() into account, and always uses '/'
>> instead, like:
>>
>> const char* p = strrchr(buf, '/');
>>
>> The fix is to change the use of '/' into *os::file_separator() on
>> two lines in Disassembler::load_library().
>>
>> *Testing*
>>
>> The change was tested by a friend of mine on Windows and he said
>> hsdis.dll could be picked up from expected paths after the
>> change. I don't have a Windows box to test it myself, so I'd like
>> someone from Oracle to help me test it.
>>
>> P.S. I wanted to upload a webrev, but somehow I couldn't get the
>> the connection to work. Could it be that my publickey is missing
>> on cr.openjdk.java.net <http://cr.openjdk.java.net>?
>>
>> The authenticity of host 'cr.openjdk.java.net
>> <http://cr.openjdk.java.net> (137.254.56.61)' can't be established.
>> RSA key fingerprint is
>> d2:d1:ee:03:4b:11:45:49:ec:46:99:8c:e4:c3:00:c4.
>> Are you sure you want to continue connecting (yes/no)? yes
>> Warning: Permanently added 'cr.openjdk.java.net
>> <http://cr.openjdk.java.net>,137.254.56.61' (RSA) to the list of
>> known hosts.
>> Permission denied (publickey).
>> lost connection
>>
>> Best regards,
>> Kris (OpenJDK username: kmo)
>
>
More information about the hotspot-runtime-dev
mailing list