On 19.06.2017 18:34, Alexey Budankov wrote: > On 19.06.2017 18:24, Andi Kleen wrote: >>> One more: >> >> It may help if you run with KASAN enabled. That often >> finds problems earlier. >> >> Also could use ftrace and use ftrace_dump_on_oops to >> see the sequence (but that has a lot of overhead >> and sometimes hides problems) > > Ok. Will try. Thanks.
Tested patch v5 with KASAN enabled. No memory corruptions were revealed so far. > >> >> -Andi >> >> > >
