mariabackup hangs on prepare

mariabackup --prepare --export --use-memory=1G --target-dir=/backupdb/2020-04-23-04-29-52

The prepare gets stuck here for hours:
.....
.....
2020-04-23 4:29:58 0 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
2020-04-23 4:29:58 0 [Note] InnoDB: Uses event mutexes
2020-04-23 4:29:58 0 [Note] InnoDB: Compressed tables use zlib 1.2.11
2020-04-23 4:29:58 0 [Note] InnoDB: Number of pools: 1
2020-04-23 4:29:58 0 [Note] InnoDB: Using SSE2 crc32 instructions
2020-04-23 4:29:58 0 [Note] InnoDB: Initializing buffer pool, total size = 1G, instances = 1, chunk size = 1G
2020-04-23 4:29:59 0 [Note] InnoDB: Completed initialization of buffer pool
2020-04-23 4:29:59 0 [Note] InnoDB: page_cleaner coordinator priority: -20

Suspecting:
https://jira.mariadb.org/browse/MDEV-19401

I asked to remove the --export, same result.

We tried to collect info with 'gdb' and 'perf' and they don't return anything, in only one case gdb returned this snapshot (one single sample) attached as backtrace-txt.log

One of the three threads apparently is in xtrabackup.cc:4284:

/* Wait for threads to exit */
while (1) {
os_thread_sleep(1000000);
pthread_mutex_lock(&count_mutex);
bool stop = count == 0;
pthread_mutex_unlock(&count_mutex);
if (stop)

}

So wondering if it's stuck in a sleep loop, not sure if it's compatible with the observations.