pf half-open tcp in state table

Hi, I've recently setup a new pair of OpenBSD 6.2 pf firewalls (with carp) in my lab, and that's not performing very well. tcp-based NFS v3 and v4 traffic (between Linux clients and a NetApp server) through it is struggling, and some SSH or HTTPS transfers are stalling, with their states

Re: CVS: cvs.openbsd.org: src

On Thu, Feb 8, 2018, at 9:01 PM, Theo de Raadt wrote: > CVSROOT: /cvs > Module name: src > Changes by: dera...@cvs.openbsd.org 2018/02/08 20:01:24 > > Modified files: > sys/dev: rnd.c > > Log message: > Situation occur where bootloader cannot supply kernel with early >

Re: simplify kqueue timers

On Fri, Feb 09, 2018 at 03:43:52PM +1300, Jonathan Matthew wrote: > kqueue timers are specified in milliseconds, and timeout_add_msec can do the > conversion to ticks for us, so we can trim a few lines here. Using > timeout_add_msec preserves the 'at least one tick' behaviour that was added > in

Kill old x86 ASM mutex code

i386 and amd64 have been using the MI mutex code for some time now without problem. Time to remove the old assembly code, ok? Index: arch/amd64/amd64/genassym.cf === RCS file: /cvs/src/sys/arch/amd64/amd64/genassym.cf,v retrieving

Re: em: minimum ethernet frame size

On 02/02/18(Fri) 12:18, Michele Curti wrote: > Hi, > in sys/dev/pci/if_em.c at line 469 there is: > sc->hw.min_frame_size = > ETHER_MIN_LEN + ETHER_CRC_LEN; > > But ETHER_MIN_LEN already includes the CRC size: > #define ETHER_MIN_LEN 64 /* Minimum frame length, CRC included

Re: CVS: cvs.openbsd.org: src

> On Thu, Feb 8, 2018, at 9:01 PM, Theo de Raadt wrote: > > CVSROOT:/cvs > > Module name:src > > Changes by: dera...@cvs.openbsd.org 2018/02/08 20:01:24 > > > > Modified files: > > sys/dev: rnd.c > > > > Log message: > > Situation occur where bootloader cannot supply

replace the hash in bridge with a red-black tree

siphash is slower than an rbt lookup. we also save a bit of memory by not allocating 1024 hash buckets. tests? ok? Index: if_bridge.c === RCS file: /cvs/src/sys/net/if_bridge.c,v retrieving revision 1.304 diff -u -p -r1.304

Introduce killjobc()

Diff below moves some job control bits out of exit1(). It is extracted from guenther@'s proctreelk diff. It's currently a noop so I'd like to get it in to reduce the locking diff. ok? Index: kern/kern_exit.c === RCS file:

Re: Introduce killjobc()

Martin Pieuchot(m...@openbsd.org) on 2018.02.09 21:31:55 +0100: > Diff below moves some job control bits out of exit1(). It is extracted > from guenther@'s proctreelk diff. It's currently a noop so I'd like to > get it in to reduce the locking diff. > > ok? ok > Index: kern/kern_exit.c >

iostat: cap wait/interval at 100 million seconds

Hi, nanosleep(2) won't take more than 100 million seconds at a time, so we ought to cap interval when we read it in. Otherwise, an oversized interval argument causes iostat(8) to print as fast as it can. While here, "interval" is called "wait" in the documentation when the new syntax is in use,

Re: Kill old x86 ASM mutex code

On Fri, Feb 09, 2018 at 09:06:21PM +0100, Martin Pieuchot wrote: > i386 and amd64 have been using the MI mutex code for some time now > without problem. Time to remove the old assembly code, ok? > no objection here. ok mlarkin > Index: arch/amd64/amd64/genassym.cf >

Re: ospfd getting confused about who is DR

On 2018/02/09 03:39, Claudio Jeker wrote: > On netsplits it can happen that on join multiple ospfd end up as DR. > In my case with 3 routers the one cut off stays DR even though the rest of > the network already has a DR and BDR. Very likely this is what I've seen. My layout has been roughly like

Re: Introduce killjobc()

Martin Pieuchot wrote: > Diff below moves some job control bits out of exit1(). It is extracted > from guenther@'s proctreelk diff. It's currently a noop so I'd like to > get it in to reduce the locking diff. > > ok? i misread subject as kill objc, but ok anyway.

Re: CVS: cvs.openbsd.org: src

On Fri, Feb 9, 2018, at 3:46 PM, Theo de Raadt wrote: > > On Thu, Feb 8, 2018, at 9:01 PM, Theo de Raadt wrote: > > > CVSROOT: /cvs > > > Module name: src > > > Changes by: dera...@cvs.openbsd.org 2018/02/08 20:01:24 > > > > > > Modified files: > > > sys/dev: rnd.c > > >

tidy up get_hibernate_io_function

It's hard to see exactly what the dv->dv_parent->etc. checks are doing. Before adding sdmmc I'd like to tidy up a bit. ok? Index: arch/amd64/amd64/hibernate_machdep.c === RCS file: /cvs/src/sys/arch/amd64/amd64/hibernate_machdep.c,v

Re: iostat: cap wait/interval at 100 million seconds

On Sat, Feb 10, 2018 at 01:12:59PM +1300, Theo Buehler wrote: > On Fri, Feb 09, 2018 at 05:41:34PM -0600, Scott Cheloha wrote: > > Hi, > > > > nanosleep(2) won't take more than 100 million seconds at a time, > > so we ought to cap interval when we read it in. Otherwise, an > > oversized interval

Re: iostat: cap wait/interval at 100 million seconds

On Fri, Feb 09, 2018 at 05:41:34PM -0600, Scott Cheloha wrote: > Hi, > > nanosleep(2) won't take more than 100 million seconds at a time, > so we ought to cap interval when we read it in. Otherwise, an > oversized interval argument causes iostat(8) to print as fast as > it can. > > While here,

Re: CVS: cvs.openbsd.org: src

> On Fri, Feb 9, 2018, at 3:46 PM, Theo de Raadt wrote: > > > On Thu, Feb 8, 2018, at 9:01 PM, Theo de Raadt wrote: > > > > CVSROOT:/cvs > > > > Module name:src > > > > Changes by: dera...@cvs.openbsd.org 2018/02/08 20:01:24 > > > > > > > > Modified files: > > > > sys/dev

remove lo0 handing from etc/netstart

Since today lo(4) will create 127.0.0.1 and the 127/8 reject route automatically. This means /etc/netstart no longer needs to do that. I will commit this in a few days so that people can make sure to have -current kernels. -- :wq Claudio Index: etc/netstart