I found some records of ciMethodData may contain mismatched oops. eg. in the attachment line 14130, it declares that there will be 2 non-null oops followed(see '2' after tag 'oops'). however, one only is recorded.

ciMethodData kotlin/coroutines/jvm/internal/ContinuationImpl <init>
(Lkotlin/coroutines/Continuation;)V 2 21538 orig 80 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 data
26 0x40007 0x402 0x70 0x4e9c 0x70005 0x4d55 0x0 0x7f6a5841c3c0 0xa3
0x7f6a5841c470 0xa4 0xc0003 0x4e9c 0x18 0x110002 0x529e 0x0 0x0 0x0 0x0
0x0 0x0 0x9 0x2 0x6 0x0 oops 2 7
com/example/ProductAttRouter$withRequestLoggingContext$1 methods 0

I don't understand how this happens. ciReplay can't load this record correctly.

One workaround is to truncate oops if ReplayIgnoreInitErrors is true. This allows ciReplay to proceed.

diff --git a/src/hotspot/share/ci/ciReplay.cpp b/src/hotspot/share/ci/ciReplay.cpp
index a9731d4bfb7..9744bbab290 100644
--- a/src/hotspot/share/ci/ciReplay.cpp
+++ b/src/hotspot/share/ci/ciReplay.cpp
@@ -870,7 +870,13 @@ class CompileReplay : public StackObj {
     for (int i = 0; i < rec->_classes_length; i++) {
       int offset = parse_int("offset");
       if (had_error()) {
- return;
+ if (ReplayIgnoreInitErrors) {
+ _error_message = NULL;
+ rec->_classes_length = i;
+ break;
+ } else {
+ return;
+ }
       }
       Klass* k = parse_klass(CHECK);
       rec->_classes_offsets[i] = offset;


I wrote an email to ask why this happens in the first place.
https://mail.openjdk.java.net/pipermail/hotspot-dev/2022-May/060459.html