Attached another old case of the same assert, this time it looks before shutdown, during rollback of uncommitted transaction

I somewhat reliably can reproduce this issue by steps below :

mkdir /bkup
mysql < /test/sakila-schema.sql
mysql < /test/sakila-data.sql
mysql -e "START TRANSACTION; DELETE FROM payment;SELECT SLEEP(10000);" sakila &
sleep 15
/test/xtrabackup --backup --target-dir=/bkup
/test/xtrabackup --prepare --target-dir=/bkup

InnoDB: Starting in background the rollback of uncommitted transactions
2016-12-11 01:51:35 7f5b17bff700 InnoDB: Rolling back trx with id 2443, 16049 rows to undo

InnoDB: Progress in percents: 1InnoDB: Read page 0 from tablespace for space 2 name mysql/innodb_index_stats key_id 0 encryption 0 handle 9.
InnoDB: Read page 0 from tablespace for space 1 name mysql/innodb_table_stats key_id 0 encryption 0 handle 11.
InnoDB: Read page 0 from tablespace for space 27 name sakila/payment key_id 0 encryption 0 handle 12.
2InnoDB: Percona XtraDB (http://www.percona.com) 5.6.32-79.0 started; log sequence number 10442629
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 442016-12-11 01:51:38 7f5b14bf9700 InnoDB: Assertion failure in thread 140029166851840 in file buf0flu
.cc line 2431
InnoDB: Failing assertion: srv_max_dirty_pages_pct_lwm <= srv_max_buf_pool_modified_pct

my.cnf used (all other is default):
[xtrabackup]
user=root
socket=/dt/m.sock
[client]
socket=/dt/m.sock
[mysqld]
innodb-buffer-pool-size=128M
socket=/dt/m.sock
log_error=/dt/error.log

This looks fixed with linked patches, I will close it in few days, unless any objections.

I cannot reproduce the problem in recent builds - Closing as fixed by linked commits