Re: Towards tcp_input() w/o KERNEL_LOCK()

On Tue, Jun 06, 2017 at 05:15:40PM +0200, Martin Pieuchot wrote: > TCP/UDP are almost ready to run without KERNEL_LOCK() because accesses > to their sockets are serialized via the NET_LOCK(). On the other hand > pfkey and routing sockets accesses still rely on the KERNEL_LOCK(). > > Since we're

Re: tcpdump: drop atalk support

On Thu, Jun 08, 2017 at 09:42:44PM +0200, Michal Mazurek wrote: > Let's start by ignoring the existence of AppleTalk in the manpage, > reducing it by 10%. This leaves mention of atalk in the syntax of libpcap. > > A second diff will remove /etc/atalk.names support reducing the amount > of

Re: radix lookup w/o KERNEL_LOCK()

On Tue, Jun 06, 2017 at 03:36:12PM +0200, Martin Pieuchot wrote: > +#define SALEN(sa)(*(u_char *)sa) Put () around macro arguments. #define SALEN(sa) (*(u_char *)(sa)) > -int > +static int > rn_refines(void *m_arg, void *n_arg) > -int > +static int > rn_inithead0(struct

monop(6): correct number of players in the man page

Hi tech@, The man page says monop monitors a game between 1 to 9 users, but the program enforces a range from 2 to 9. Comments? OK? Index: games/monop/monop.6 === RCS file: /cvs/src/games/monop/monop.6,v retrieving revision 1.15

Re: doas: add confirm to prompt the user on what is to be executed

Thu, 8 Jun 2017 20:17:02 +0200 Adam Wolk > This email is a request for comment, roughly I want to know if others > see this feature as valuable. Hi Adam, Simple shell wrapper functions can replicate the are you sure thing. Only a reminder to focus on implementing just

Re: tcpdump: drop atalk support

Let's start by ignoring the existence of AppleTalk in the manpage, reducing it by 10%. This leaves mention of atalk in the syntax of libpcap. A second diff will remove /etc/atalk.names support reducing the amount of appletalk code significantly. Comments? OK? Index: usr.sbin/tcpdump/tcpdump.8

Re: ifconfig: Fix/improve settimeslot(), simplify get_ts_map() out

On 2017/06/08 19:17, Stuart Henderson wrote: > On 2017/06/08 13:54, Ted Unangst wrote: > > Klemens Nanni wrote: > > > This fixes the primitive parsing route of settimeslot() to allow any > > > possible list of slots and/or ranges, see the update manual section. > > > > There doesn't appear to be

Re: doas: add confirm to prompt the user on what is to be executed

Adam Wolk wrote: > This email is a request for comment, roughly I want to know if others see this > feature as valuable. The diff currently lacks manpage changes, I will work on > those if the general decision is to include this feature. > > I won't cry if we decide to drop this. I would have

Re: ifconfig: Fix/improve settimeslot(), simplify get_ts_map() out

On 2017/06/08 13:54, Ted Unangst wrote: > Klemens Nanni wrote: > > This fixes the primitive parsing route of settimeslot() to allow any > > possible list of slots and/or ranges, see the update manual section. > > There doesn't appear to be any use of timeslot code, is there? > > Better to delete

doas: add confirm to prompt the user on what is to be executed

Hi tech@ This is a feture that came up in a chat I had with Kurt Mosiejczuk. I have been recently reading source daily as a learning experience and decided that implementing the feature we discussed would be a nice exercise. The attached diff extends the configuration syntax with a new option

Re: ifconfig: Fix/improve settimeslot(), simplify get_ts_map() out

Klemens Nanni wrote: > This fixes the primitive parsing route of settimeslot() to allow any > possible list of slots and/or ranges, see the update manual section. There doesn't appear to be any use of timeslot code, is there? Better to delete it entirely. Index: ifconfig.8

ifconfig: Fix/improve settimeslot(), simplify get_ts_map() out

This fixes the primitive parsing route of settimeslot() to allow any possible list of slots and/or ranges, see the update manual section. The old code would happily mask "1,2-" into 0b11, settimeslot() now fails on such broken ranges and also checks for inconsistencies like "4-1", etc.

Re: amd64: EFI boot over network try to load kernel from hd0

On Thu, Jun 08, 2017 at 11:42:44PM +0900, YASUOKA Masahiko wrote: > Hi, > > Tested the diff. It works fine. > > On Wed, 7 Jun 2017 20:23:43 +0200 > Patrick Wildt wrote: > > Do you want it to run diskless with root on NFS? Well, I haven't > > implemented that, but here's a

Re: amd64: EFI boot over network try to load kernel from hd0

Hi, Tested the diff. It works fine. On Wed, 7 Jun 2017 20:23:43 +0200 Patrick Wildt wrote: > Do you want it to run diskless with root on NFS? Well, I haven't > implemented that, but here's a diff that should allow you to load > the kernel from a TFTP server. > > I have a

Re: patch(1) is fucked up

On Thu, 08 Jun 2017 15:33:11 +0200, Marc Espie wrote: > There is zero option that says "assume all those patches are correct and > error out if something untowards happen". > > This is really annoying for ports. > > - prompting for non-existent filenames breaks automated builds... dpb avoids >

patch(1) is fucked up

There is zero option that says "assume all those patches are correct and error out if something untowards happen". This is really annoying for ports. - prompting for non-existent filenames breaks automated builds... dpb avoids that by explicitly zapping stdin. - the new one. In the absence of

isakmpd(8) use-after-free

Michał Koc reported a crash on misc@, turns out it's a use-after-free: http://marc.info/?l=openbsd-misc=149597472223216=2 The trace indicates that argument given to pf_key_v2_stayalive() is no longer valid: #0 conf_get_str (section=0xa8735b03f80 ' , tag=0xa8459272809 "Phase") at

Re: [patch] Avoid system(3) in ikectl

On Fri, May 19, 2017 at 12:32:16AM -0500, Matthew Martin wrote: > ikectl errors in a number of situations where shell special characters > are used. For example: > > % doas ikectl ca test create password \' > [...] > subject=/C=DE/ST=Lower Saxony/L=Hanover/O=OpenBSD/OU=iked/CN=VPN >

Re: [patch] Use readpassphrase in ikectl

On Fri, May 19, 2017 at 12:35:44AM -0500, Matthew Martin wrote: > While making the last patch, I noticed ikectl uses getpass. Use > readpassphrase instead and explicit_bzero the buffers. > > - Matthew Martin What is the goal here? It can't be to use a different buffer size as the same size as

Re: Better handling of short reads

On Wed, Jun 07, 2017 at 23:04 -0500, Amit Kulkarni wrote: > On Wed, 7 Jun 2017 21:27:27 -0500 > Amit Kulkarni wrote: > > > On Thu, 8 Jun 2017 01:57:25 +0200 > > Mike Belopuhov wrote: > > > > > On Wed, Jun 07, 2017 at 18:35 -0500, Amit Kulkarni wrote: >

Re: diff: add missing rtm_send to nd6

Hi Martin, On Thu, Jun 08, 2017 at 09:44:37AM +0200, Martin Pieuchot wrote: > On 08/06/17(Thu) 00:47, Jan Klemkow wrote: > > This diff adds a missing routing message to the neighbor discovery code. > > The message informs the userland about new reachable IPv6 nodes on the > > network. The IPv6

Re: ifconfig.8 doco for vnetid and parent options

Claudio Jeker(cje...@diehard.n-r-g.com) on 2017.06.07 16:05:15 +0200: > Be careful, AFAIK the capitalisation of IEEE standards does matter. > You're right 802.1Q is the correct spelling but not for 802.1ad (where the > lowercase version is the offical standard). IIRC the status of the > standard

Re: diff: add missing rtm_send to nd6

On 08/06/17(Thu) 00:47, Jan Klemkow wrote: > Hi, > > This diff adds a missing routing message to the neighbor discovery code. > The message informs the userland about new reachable IPv6 nodes on the > network. The IPv6 network stack acts more like the IPv4 port by this > diff. Now, the behavior