<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">Now the change passed JPRT successfully.  This looks good.<div class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Nov 11, 2015, at 10:11 AM, Christian Thalinger <<a href="mailto:christian.thalinger@oracle.com" class="">christian.thalinger@oracle.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><meta http-equiv="Content-Type" content="text/html charset=windows-1252" class=""><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">There was a bug in 32-bit x86 which Gilles fixed:<div class=""><br class=""></div><div class=""><span style="font-family: Courier;" class="">diff -r 24fd08e99b35 -r 372cb6e59b0d src/cpu/x86/vm/sharedRuntime_x86_32.cpp</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">--- a/src/cpu/x86/vm/sharedRuntime_x86_32.cpp</span><span class="Apple-tab-span" style="font-family: Courier; white-space: pre;">    </span><span style="font-family: Courier;" class="">Wed Nov 11 10:33:08 2015 +0100</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">+++ b/src/cpu/x86/vm/sharedRuntime_x86_32.cpp</span><span class="Apple-tab-span" style="font-family: Courier; white-space: pre;">    </span><span style="font-family: Courier;" class="">Wed Nov 11 18:34:15 2015 +0100</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">@@ -2946,7 +2946,8 @@</span><br style="font-family: Courier;" class=""><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">  oop_maps->add_gc_map( __ pc()-start, map);</span><br style="font-family: Courier;" class=""><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">-  // Discard arg to fetch_unroll_info</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">+  // Discard args to fetch_unroll_info</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">+  __ pop(rcx);</span><br style="font-family: Courier;" class=""><span style="font-family: Courier;" class="">  __ pop(rcx);</span><br style="font-family: Courier;" class=""><br style="font-family: Courier;" class=""><font face="Courier" class="">  __ get_thread(dcx);</font></div><div class=""><font face="Courier" class=""><br class=""></font><div class=""><blockquote type="cite" class=""><div class="">On Nov 9, 2015, at 6:50 AM, Gilles Duboscq <<a href="mailto:gilles.m.duboscq@oracle.com" class="">gilles.m.duboscq@oracle.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class="">I updated the webrev to export the UnrollBlock::_unpack_kind field through vmStruct and JVMCI's HotSpotVMConfig:<br class=""><br class=""><a href="http://cr.openjdk.java.net/~gdub/webrev-8141133.1" class="">http://cr.openjdk.java.net/~gdub/webrev-8141133.1</a><br class=""><br class="">The only difference to the previous webrev are the changes in vmStruct.cpp  and HotSpotVMConfig.java<br class=""><br class=""> Gilles<br class=""><br class="">On 05/11/15 16:28, Gilles Duboscq wrote:<br class=""><blockquote type="cite" class="">Hi,<br class=""><br class="">Please review the following change:<br class=""><br class="">Webrev:<br class=""><a href="http://cr.openjdk.java.net/~gdub/webrev-8141133/" class="">http://cr.openjdk.java.net/~gdub/webrev-8141133/</a><br class="">JIRA:<br class="">https://bugs.openjdk.java.net/browse/JDK-8141133<br class=""><br class="">Scopes with rethrow_exception set to true need to be handled for all types of deoptimization, not just for uncommon traps.<br class="">This change moves their handling from uncommon_trap_inner to fetch_unroll_info_helper and uses exception_oop and exec_mode = Unpack_exception rather than setting a pending exception.<br class="">fetch_unroll_info_helper can now override the exec_mode by passing it through the UnrollBlock using the existing unpack_kind field.<br class=""><br class=""> Gilles<br class=""><br class=""></blockquote></div></div></blockquote></div><br class=""></div></div></div></blockquote></div><br class=""></div></body></html>