Re: Potential DoS attack on PF due to infinite loop

2017-07-16 Thread Alexander Bluhm
On Tue, Jul 11, 2017 at 12:29:33PM -0700, Jingmin Zhou wrote: > The problem is at line 224. When a LB rule is configured to have 65535 as > the high port, and uint16 variable tmp reaches it, ++(tmp) will wrap around > and hence potentially enter into an infinite loop. Of course, it only > happens

Re: Potential DoS attack on PF due to infinite loop

2017-07-16 Thread Alexander Bluhm
On Tue, Jul 11, 2017 at 12:29:33PM -0700, Jingmin Zhou wrote: > Recently we discovered a potential bug in pf_lb.c. I have commited a fix. Thanks for the report and analysis. bluhm