Ivan Maidanski
Ivan Maidanski
Latest build: https://app.travis-ci.com/github/ivmai/bdwgc/jobs/573526977 Source: b75f1aa (master)
Build: https://app.travis-ci.com/github/ivmai/bdwgc/jobs/576835719 Source: release-8_2 (4919305150)
Also, sometimes SIGSEGV happens in sigsuspend (on world restart).
It is reported that a couple of bugs in signal-related code of kernel have been fixed and the issue is not observed on kernel 5.4.0-5.4.
Also, probably this should fix #419
> It is reported that a couple of bugs in signal-related code of kernel have been fixed and the issue is not observed on kernel 5.4.0-5.4. I confirm the issue...
For some reason the following config hangs: -D GC_THREADS -D THREAD_LOCAL_ALLOC -D GC_ENABLE_SUSPEND_THREAD
Current workaround for kFreeBSD: * release-7_6: do not define THREAD_LOCAL_ALLOC if configure --enable-gcj-support (because configure does not have --disable-thread-local-alloc option) * master: do bot define GC_ENABLE_SUSPEND_THREAD if configure --disable-thread-local-alloc
According to the [man page](https://man7.org/linux/man-pages/man2/userfaultfd.2.html): Since Linux 5.7, userfaultfd supports write-protect mode. The user needs to first check availability of this feature using UFFDIO_API ioctl against the feature bit UFFD_FEATURE_PAGEFAULT_FLAG_WP...
I wonder if UFFDIO_WRITEPROTECT affects map count (i.e. can exceed vm.max_map_count? (related to #426).