RFR(XS): 8245051: c1 is broken if it is compiled by gcc without -fno-lifetime-dse

Liu, Xin xxinliu at amazon.com
Tue May 19 01:44:27 UTC 2020

On 5/18/20, 5:56 PM, "Kim Barrett" <kim.barrett at oracle.com> wrote:

    CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.

    > On May 17, 2020, at 5:34 PM, Liu, Xin <xxinliu at amazon.com> wrote:
    > Hi, Kim,
    > Thank you to review my patch. I have removed the friend class BlockBegin.
    > Here is the new revision:  http://cr.openjdk.java.net/~xliu/8245051/01/webrev/

    Looks good.

    > About --with-extra-cflags, I completely agree. However,  how to configure OpenJDK is not under control. There're so many linux distributions.

    I’m not sure what you mean by that.  The given configure options simply aren’t valid.
    And if some linux distribution is patching the OpenJDK to allow such a configuration,
    that’ s their lookout.
Recently, I found different linux distros have their convenient ways to configure OpenJDK.  
centos7: https://git.centos.org/rpms/java-11-openjdk/blob/c7/f/SPECS/java-11-openjdk.spec#_1327
ubuntu: https://git.launchpad.net/~openjdk/ubuntu/+source/openjdk/+git/openjdk/tree/debian/rules?h=openjdk-11#n223
yes, they should take care of cflags.  I will file a bug in centos Bugzilla.

    > In addition, it might cause subtle bugs if the toolchain is not gcc. From my side, I'd like to get rid of undefine behaviors as many as we can.

    I agree that it’s desirable to eliminate unnecessary or unintentional UB.

    > Yes, I still need a reviewer and a sponsor.

    Looks like Paul Hohense has volunteered.

Thank you!

More information about the hotspot-compiler-dev mailing list