this rolls vnetid and parent into a single encap line in ifconfig.
eg:
- vnetid: 7
- parent: ix1
+ encap: vnetid: 7 parent: ix1
after this i would like to get rid of the vlan compat goo.
ok?
Index: ifconfig.c
===
It looks like I've just found a bug in (our) ksh. Not to be brave
enough to fix it right now, but I think it's still worths adding
regression test.
Bash and zsh pass this test without problem.
The idea is using the case...in...esac inside $(...) or `...`.
It starts failing when you add a single c
This is a first stab at HFSC and FQ-CoDel integration via extending
PF queueing operations (pfq_ops) interface. With this FQ-CoDel can
be attached directly to an interface as well as serve as a replacement
for the HFSC queue to improve its characteristics. In essence, in
many setups (router behin
> "servers ntp.ring.nlnog.net \
> local-address 165.254.255.27,2001:728:1808::26"
yes
>
> or allow one to repeat the 'local-address' keyword:
>
> "servers ntp.ring.nlnog.net \
> local-address 165.254.255.27 \
> local-address 2001:728:1808::26"
>
> or use differe
On 28.5.2017. 21:25, Alexander Bluhm wrote:
> On Sun, May 28, 2017 at 07:41:23PM +0200, Hrvoje Popovski wrote:
>> it seems that with
>> $OpenBSD: ip_ipip.c,v 1.81 2017/05/28 13:59:05 bluhm Exp $
>> i can't compile this diff anymore ... with 1.80 everything is fine ...
>> just noticed, nothing else
Hi,
pf blocks IPv4 options, so it should block IPv6 option headers as well.
I am not aware of common use cases where option headers are essential.
For special things allow-opts can be added to pf.conf, but I don't
want to break usual setups.
Did I forget something? Could you test this on real I
On Sun, May 28, 2017 at 10:29:43PM +0200, Sebastian Benoit wrote:
> Peter Hessler(phess...@theapt.org) on 2017.05.28 20:56:39 +0200:
> > looks fine to me, OK
> >
> >
> > On 2017 May 28 (Sun) at 19:12:34 +0200 (+0200), Sebastian Benoit wrote:
> > :Hi,
> > :
> > :testing hennigs diff i found having
Peter Hessler(phess...@theapt.org) on 2017.05.28 20:56:39 +0200:
> looks fine to me, OK
>
>
> On 2017 May 28 (Sun) at 19:12:34 +0200 (+0200), Sebastian Benoit wrote:
> :Hi,
> :
> :testing hennigs diff i found having to specify the socket on the command
> :line annoying:
> :
> : bgpctl -s /var/ru
On Sun, May 28, 2017 at 08:53:16PM +0200, Martin Pieuchot wrote:
> Trying to grab the NET_LOCK() while holding an ifp reference is wrong.
> This creates deadlock as found by Hrvoje Popovski. The reason is that
> if_detach() sleeps holding the NET_LOCK() while waiting for others
> threads to releas
> From: Miod Vallat
> Date: Sun, 28 May 2017 19:22:23 + (UTC)
>
> > This makes MULTIPROCESSOR kernels use copyin32(9) to guarantee
> > atomicity. This will break m88k GENERIC.MP; shouldn't be too
> > difficult to fix for someone whu understands m88k assembly.
Looks plausible; ok kettenis@
Currently when running bgpd -dvv you will see that the SE and the RDE
are crashing during shutdown. Fixing the order of closing the imsg pipes
solves this.
--
:wq Claudio
Index: rde.c
===
RCS file: /cvs/src/usr.sbin/bgpd/rde.c,v
ret
On Sun, May 28, 2017 at 07:41:23PM +0200, Hrvoje Popovski wrote:
> it seems that with
> $OpenBSD: ip_ipip.c,v 1.81 2017/05/28 13:59:05 bluhm Exp $
> i can't compile this diff anymore ... with 1.80 everything is fine ...
> just noticed, nothing else ...
As I am commiting parts of it, the diff does
> This makes MULTIPROCESSOR kernels use copyin32(9) to guarantee
> atomicity. This will break m88k GENERIC.MP; shouldn't be too
> difficult to fix for someone whu understands m88k assembly.
Index: subr.S
===
RCS file: /OpenBSD/src/s
So requested by benno@ print when we receive and send the EOR marker.
This gives an indication when a session is synchronized (esp. the recv
one). Example
neighbor 10.48.0.242 (upstream): sending IPv4 unicast EOR marker
nexthop 10.48.0.242 now valid: via 10.0.0.1
neighbor 10.48.0.242 (upstream): r
Trying to grab the NET_LOCK() while holding an ifp reference is wrong.
This creates deadlock as found by Hrvoje Popovski. The reason is that
if_detach() sleeps holding the NET_LOCK() while waiting for others
threads to release their reference.
Diff below fixes the linkstate vs if_detach() deadloc
looks fine to me, OK
On 2017 May 28 (Sun) at 19:12:34 +0200 (+0200), Sebastian Benoit wrote:
:Hi,
:
:testing hennigs diff i found having to specify the socket on the command
:line annoying:
:
: bgpctl -s /var/run/rdomain123.bgpd.sock sh rib
:
:srly?
:
:I figure when you are debuging things in rd
1 month bump.
CC'd millert@ because he made the switch to getline(3) in r1.18.
Original message below.
--
On Apr 20, 2017, at 8:59 PM, Scott Cheloha wrote:
Check whether or not getline(3) failed before exiting.
It also couldn't hurt to close fp... especially given that we're
freeing line, to
On 28.5.2017. 15:07, Alexander Bluhm wrote:
> On Fri, May 26, 2017 at 03:07:51PM +0200, Martin Pieuchot wrote:
>> This diff get rids of the ipintrq for forwarding. The queue is now used
>> for packets delivered locally which still need the KERNEL_LOCK().
> After committing some cleanup mpi@'s diff
In article <20170528160659.GA46003@lol.local> you wrote:
> Hi Walter,
>
> Thanks for the report, please try out the diff below.
>
The diff works as you explained. It replaces correctly utf-8 with
ascii chars.
Thank you.
Hi,
testing hennigs diff i found having to specify the socket on the command
line annoying:
bgpctl -s /var/run/rdomain123.bgpd.sock sh rib
srly?
I figure when you are debuging things in rdomain , you will want to route
-T exec ksh anyway, so this makes bgpd create /var/run/bgpd.sock. when
r
Theo de Raadt wrote:
> I'm not sure the direction this goes.
>
> Today atalk, and over the next month delete 20 protocols, then anything
> before ARP?
>
> What's the plan here. This code is heavily privsep, pledge, etc. Is
> there a problem with it?
>
> If such a packet showed up on a network
I'm not sure the direction this goes.
Today atalk, and over the next month delete 20 protocols, then anything
before ARP?
What's the plan here. This code is heavily privsep, pledge, etc. Is
there a problem with it?
If such a packet showed up on a network I'd prefer it is *identified*
in some m
Mark Kettenis wrote:
> > * add a DIOCCACHESYNC ioctl that can be used to flush a disk's cache
> >
> > * add code to the file systems that executes this ioctl when a mount
> > is updated from r/w to r/o
>
> Abusing the ioctl codepath in the kernel isn't a good idea. This
> continues to be a maj
* Michal Mazurek [2017-05-28 16:00]:
> Remove atalk support. Significantly shortens the manpage. libpcap still
> supports it. This diff doesn't include the removal of two files:
> appletalk.h and print-atalk.c.
afaict atalk is so dead that the corpse is way beyond the point of
smelling - so yeah,
Hi,
while looking at re(4) on ARM64 I was stumbling upon this sync. I think
the intention was to sync the descriptor where we set the OWN flag, but
we actually sync curidx which is the idx to the last segment instead of
the first.
Opinions?
Patrick
diff --git sys/dev/ic/re.c sys/dev/ic/re.c
in
> Date: Sun, 28 May 2017 17:43:31 +0200 (CEST)
> From: Stefan Fritsch
>
> Hi,
>
> I want to come back to the discussion from Oct. 2016 on tech@ about a
> disk cache flush ioctl.
>
>
> The problem we want to solve is that in case of power loss, there should
> be no data loss on partitions that
Hi Walter,
On Sun, May 28, 2017 at 10:56:19AM +0200, Walter Alejandro Iglesias wrote:
>
> I forgot to Cc this here.
>
>
> - Forwarded message from Walter Alejandro Iglesias
> -
>
> From: Walter Alejandro Iglesias
> To: Anton Lindqvist
> Subject: Re: ksh(1): vi mode UTF-8 bug
> In-R
2017-05-28 18:32 GMT+03:00 Mark Kettenis :
>> From: "Ted Unangst"
>> Date: Sun, 28 May 2017 11:16:21 -0400
>>
>> Vadim Zhukov wrote:
>> > While working on getting Boost & its friends work under Clang,
>> > I've stumbled upon the code that looks like the following:
>> >
>> >decltype(x, y) z = w
2017-05-28 18:35 GMT+03:00 Vadim Zhukov :
> 2017-05-28 18:16 GMT+03:00 Ted Unangst :
>> Vadim Zhukov wrote:
>>> While working on getting Boost & its friends work under Clang,
>>> I've stumbled upon the code that looks like the following:
>>>
>>>decltype(x, y) z = w;
>>
>> why not espie's fix fr
Don't skip the cache flush until the last opening of the device is
closed. Otherwise, when umounting a writable partition while a different
partition is still mounted read-only, the necessary disk flush may be
delayed for a very long time.
diff --git sys/scsi/sd.c sys/scsi/sd.c
index 4cd0e4d13bd.
Other file systems can be changed later.
diff --git sys/msdosfs/msdosfs_vfsops.c sys/msdosfs/msdosfs_vfsops.c
index e13b0b1ea0b..5cd273c4087 100644
--- sys/msdosfs/msdosfs_vfsops.c
+++ sys/msdosfs/msdosfs_vfsops.c
@@ -64,6 +64,7 @@
#include
#include
#include
+#include
#include
#include
Add an ioctl to tell storage devices to flush their internal caches.
Initially ported from netbsd by pedro@
I will implement it for nvme later (if we go ahead with this).
diff --git sys/dev/ata/wd.c sys/dev/ata/wd.c
index cd168223672..859fb454f44 100644
--- sys/dev/ata/wd.c
+++ sys/dev/ata/wd.c
Hi,
I want to come back to the discussion from Oct. 2016 on tech@ about a
disk cache flush ioctl.
The problem we want to solve is that in case of power loss, there should
be no data loss on partitions that are either mounted read-only or are
unmounted. This should also be true if such a partit
2017-05-28 18:16 GMT+03:00 Ted Unangst :
> Vadim Zhukov wrote:
>> While working on getting Boost & its friends work under Clang,
>> I've stumbled upon the code that looks like the following:
>>
>>decltype(x, y) z = w;
>
> why not espie's fix from earlier?
Because I've missed it. :( I have no i
> From: "Ted Unangst"
> Date: Sun, 28 May 2017 11:16:21 -0400
>
> Vadim Zhukov wrote:
> > While working on getting Boost & its friends work under Clang,
> > I've stumbled upon the code that looks like the following:
> >
> >decltype(x, y) z = w;
>
> why not espie's fix from earlier?
Well, I
On 28/05/17(Sun) 17:22, Florian Riehm wrote:
> On 05/28/17 16:09, Martin Pieuchot wrote:
> > On 28/05/17(Sun) 16:02, Florian Riehm wrote:
> >> [...]
> >> Ok, new diff below.
> >
> > I overlooked something! See below:
>
> Good catch!
> I also forgot SRPL_LEAVE.
> fixed.
ok mpi@
> Index: share/
On 05/28/17 16:09, Martin Pieuchot wrote:
> On 28/05/17(Sun) 16:02, Florian Riehm wrote:
>> [...]
>> Ok, new diff below.
>
> I overlooked something! See below:
Good catch!
I also forgot SRPL_LEAVE.
fixed.
Index: share/man/man9/mbuf_tags.9
==
Vadim Zhukov wrote:
> While working on getting Boost & its friends work under Clang,
> I've stumbled upon the code that looks like the following:
>
>decltype(x, y) z = w;
why not espie's fix from earlier?
but clang++ has variadic macros, like C99... the following patch allows
macro dec
While working on getting Boost & its friends work under Clang,
I've stumbled upon the code that looks like the following:
decltype(x, y) z = w;
The "x, y" is a perfectly valid expression, and should work.
But it's not, because decltype is actually defined as a macros
in libcxx:
#if !__is_iden
Hello Martin,
I think you need to do NET_UNLOCK() before return (error);
>
> /* Shutdown and remove trunk ports, return on error */
> + NET_LOCK(s);
> while ((tp = SLIST_FIRST(&tr->tr_ports)) != NULL) {
> if ((error = trunk_port_destroy(tp)) != 0)
>
On 28/05/17(Sun) 16:11, Hrvoje Popovski wrote:
> On 28.5.2017. 15:07, Alexander Bluhm wrote:
> > After committing some cleanup mpi@'s diff looks like this now.
> >
> > bluhm
>
>
> Hi all,
>
> with cvs tree fetched few minutes ago and with this diff i'm getting
> traces in attchment.
Thanks for
Dear team,
I have the following use-case on some of my routers: ntpd will
opportunistically select a source address, regardless of whether that
source address is actually a globally routable IP address. Most of the
time this is great, but not in some deployment scenarios.
For instance, IP address
Hi,
There is a bit spooky code I think in the server_read_http function
after freeing the line buffer and parsing the line :)
Right now I think a double-free bug is prevented because the maxiumum
header length is checked using SERVER_MAXHEADERLENGTH (=8192).
else an integer overflow of clt->clt_
This is a start at pushing locking down into the socket layer.
Use a SRPL list to manage the route PCBs and start running stuff without
the kernel lock where save. Includes some cleanup since rawcb was removed
in along the way.
The goal is to make the block running without the kernel lock larger a
On 28.5.2017. 15:07, Alexander Bluhm wrote:
> After committing some cleanup mpi@'s diff looks like this now.
>
> bluhm
Hi all,
with cvs tree fetched few minutes ago and with this diff i'm getting
traces in attchment.
setup: carp on vlan on trunk (lacp) on 2 x ix
there are so many net diffs, m
On 28/05/17(Sun) 16:02, Florian Riehm wrote:
> [...]
> Ok, new diff below.
I overlooked something! See below:
> Index: sys/netinet/ip_carp.c
> ===
> RCS file: /cvs/src/sys/netinet/ip_carp.c,v
> retrieving revision 1.310
> diff -u -
Our original plan was to unlock the forwarding path without taking any
lock. That's why I added some KERNEL_LOCK()/KERNEL_UNLOCK() around non
MP-safe data structures.
Now we're going to rely on the NET_LOCK() for the interface address and
multicast list. So replace the KERNEL_LOCK()/KERNEL_UNLOC
On 05/28/17 14:04, Martin Pieuchot wrote:
> On 28/05/17(Sun) 13:58, Florian Riehm wrote:
>> On 05/28/17 11:33, Martin Pieuchot wrote:
>>> On 28/05/17(Sun) 10:34, Florian Riehm wrote:
Hi,
after the fix for carp balancing ip-stealth is in, here is the fix for
balancing ip.
>>>
>>>
Remove atalk support. Significantly shortens the manpage. libpcap still
supports it. This diff doesn't include the removal of two files:
appletalk.h and print-atalk.c.
Index: usr.sbin/tcpdump/INSTALL
===
RCS file: /cvs/src/usr.sbin/tc
On Fri, May 26, 2017 at 03:07:51PM +0200, Martin Pieuchot wrote:
> This diff get rids of the ipintrq for forwarding. The queue is now used
> for packets delivered locally which still need the KERNEL_LOCK().
After committing some cleanup mpi@'s diff looks like this now.
bluhm
Index: dev/usb/if_u
OK bluhm@
On Sat, May 27, 2017 at 07:53:30PM +0200, Michal Mazurek wrote:
> Index: sbin/pfctl/parse.y
> ===
> RCS file: /cvs/src/sbin/pfctl/parse.y,v
> retrieving revision 1.659
> diff -u -p -r1.659 parse.y
> --- sbin/pfctl/parse.y
Found with clang. The warning was:
clang -O2 -pipe -Wall -Wmissing-prototypes -Wno-uninitialized
-Wstrict-prototypes -I/usr/src/sbin/pfctl -Werror-implicit-function-declaration
-c parse.c
/usr/src/sbin/pfctl/parse.y:4599:
On Sun, May 28, 2017 at 11:09:39AM +0200, Martin Pieuchot wrote:
> - if_deactivate() needs the NET_LOCK() for carp_ifdetach(). splnet() is
> not need here since if_deactivate() deals with pseudo-interface hooks.
>
> - carp_clone_destroy() needs to grab the NET_LOCK().
>
> In both cases we wan
On 28/05/17(Sun) 13:58, Florian Riehm wrote:
> On 05/28/17 11:33, Martin Pieuchot wrote:
> > On 28/05/17(Sun) 10:34, Florian Riehm wrote:
> >> Hi,
> >>
> >> after the fix for carp balancing ip-stealth is in, here is the fix for
> >> balancing ip.
> >
> > Great!
> >
> >>
> >> Non-stealth balancing
On 05/28/17 11:33, Martin Pieuchot wrote:
> On 28/05/17(Sun) 10:34, Florian Riehm wrote:
>> Hi,
>>
>> after the fix for carp balancing ip-stealth is in, here is the fix for
>> balancing ip.
>
> Great!
>
>>
>> Non-stealth balancing traffic needs some special treatment since it contains
>> layer 3
On Fri, May 26, 2017 at 09:40:49PM +0200, Peter Hessler wrote:
> On 2017 May 26 (Fri) at 20:01:00 +0200 (+0200), Peter Hessler wrote:
> :Apropos of "I found it", I implemented support for RFC 7607. It's a
> :super short RFC, but basically it forbids use of AS 0 anywhere.
> :
> :OK?
>
> Fixed some
with this commit:
On Sun, May 28, 2017 at 03:35:56AM -0600, Florian Obser wrote:
> CVSROOT: /cvs
> Module name: src
> Changes by: flor...@cvs.openbsd.org 2017/05/28 03:35:56
>
> Modified files:
> usr.sbin/slaacd: engine.c engine.h frontend.c slaacd.c slaacd.h
> usr.sbin/slaac
ok benno@
Florian Riehm(m...@friehm.de) on 2017.05.27 10:15:45 +0200:
> Hi,
>
> This patch fixes the carp mode 'balancing ip-stealth'.
>
> Problem:
> System A
> carp1: flags=8843 mtu 1500
> lladdr 00:00:5e:00:01:01
> description: Carp-intern
> index 7 priority 15 llprio 3
2017-05-28 13:14 GMT+03:00 Stefan Sperling :
> If net80211 is not in SCAN state anymore at the point in time where
> the device tells us it is done scanning, do nothing instead of
> proceeding to select an AP anyway and eventually sending some
> commands that will just confuse the firmware.
>
> Thi
thanks for both this and your httpd diff, both commited.
Hiltjo Posthuma(hil...@codemadness.org) on 2017.05.07 20:56:25 +0200:
> Hey,
>
> This replaces the prefix in log messages to __func__. Some older code had
> the prefix still set to another name. Patch below:
>
>
> diff --git a/usr.sbin/re
If net80211 is not in SCAN state anymore at the point in time where
the device tells us it is done scanning, do nothing instead of
proceeding to select an AP anyway and eventually sending some
commands that will just confuse the firmware.
This might silence some error messages people have seen wit
On 28/05/17(Sun) 10:34, Florian Riehm wrote:
> Hi,
>
> after the fix for carp balancing ip-stealth is in, here is the fix for
> balancing ip.
Great!
>
> Non-stealth balancing traffic needs some special treatment since it contains
> layer 3 unicast inside layer 2 multicast.
>
> Now the idea is
On Sun, May 28, 2017 at 10:12:24AM +0200, Michal Mazurek wrote:
> Reduce differences between the two pfctl_osfp.c files.
>
> Apply three commits from pfctl/pfctl_osfp.c, this time with binary change.
>
> Make sure that fingerprint_count gets reset to 0 correctly when we
> flush our list o
On 28/05/17(Sun) 01:39, Alexander Bluhm wrote:
> Hi Martin,
>
> /usr/src/regress/sys/netinet/carp triggers splassert:
>
> splassert: in_delmulti: want 1 have 0
> Starting stack trace...
> in_delmulti(803bc2c0,0,0,0,33ac6118316b76b7,8135a994) at
> in_delmulti+0x54
> carp_multicast
I forgot to Cc this here.
- Forwarded message from Walter Alejandro Iglesias -
From: Walter Alejandro Iglesias
To: Anton Lindqvist
Subject: Re: ksh(1): vi mode UTF-8 bug
In-Reply-To: <20170519121136.GA16623@lol.local>
X-Newsgroups: gmane.os.openbsd.tech
User-Agent: tin/2.4.1-20161224
This adds a side effect free io path to nvme(4) allowing it to be used for
hibernate. I've only tested this in a qemu vm, which successfully hibernates
but falls over weirdly trying to resume.
Index: arch/amd64/amd64/hibernate_machdep.c
Hi,
after the fix for carp balancing ip-stealth is in, here is the fix for
balancing ip.
Non-stealth balancing traffic needs some special treatment since it contains
layer 3 unicast inside layer 2 multicast.
Now the idea is to deal at layer 2 (ether_input()) with the multicast frames
like regula
On Sun, May 28, 2017 at 10:25:14AM +0200, Anton Lindqvist wrote:
> Digging through some old diffs. Any ok's for this one?
ok stsp@
> Tested on amd64.
>
> Index: dsdt.c
> ===
> RCS file: /cvs/src/sys/dev/acpi/dsdt.c,v
> retrieving re
Reduce differences between the two pfctl_osfp.c files.
The pfctl_show_fingerprints() function has changed, the new one from
pfctl uses pfctl_print_title(), which is defined in a different file
(pfctl.c), not available in tcpdump.
tcpdump doesn't use pfctl_show_fingerprints() anyway, so update the
Hi,
On Fri, Apr 07, 2017 at 08:06:47AM +0200, Anton Lindqvist wrote:
> Ping
>
> On Sun, Mar 26, 2017 at 07:10:29PM +1100, Jonathan Gray wrote:
> > On Sun, Mar 26, 2017 at 09:33:44AM +0200, Otto Moerbeek wrote:
> > > On Sun, Mar 26, 2017 at 06:31:41PM +1100, Jonathan Gray wrote:
> > >
> > > > On
Reduce differences between the two pfctl_osfp.c files.
Apply three commits from pfctl/pfctl_osfp.c, this time with binary change.
Make sure that fingerprint_count gets reset to 0 correctly when we
flush our list of fingerprints.
second:
Plug simple memory leak. ``Don't forget to fr
71 matches
Mail list logo