The code looks like this on windows:
release_memory(base, size);
attempt_reserve_memory_at(split, base);
attempt_reserve_memory_at(size - split, split_address);
After the release and before the re-reservations, another thread could reserve that memory. Investigate the effects of this.
We've had similar bugs in the past, where memory are temporarily unmapped and non-JVM threads reserves memory in the requested area.
release_memory(base, size);
attempt_reserve_memory_at(split, base);
attempt_reserve_memory_at(size - split, split_address);
After the release and before the re-reservations, another thread could reserve that memory. Investigate the effects of this.
We've had similar bugs in the past, where memory are temporarily unmapped and non-JVM threads reserves memory in the requested area.
- duplicates
-
JDK-8255917 runtime/cds/SharedBaseAddress.java failed "assert(reserved_rgn != 0LL) failed: No reserved region"
-
- Resolved
-
- relates to
-
-
- Closed
-
-
-
- Closed
-
-
JDK-8245035 Clean up os::split_reserved_memory()
-
- Resolved
-
-
-
- Closed
-
-
-
- Closed
-
(1 relates to)