remove vlan(4) specific ifconfig output

the vlan specific output is redundant because vlan supports the generic vnetid and parent ioctls now. this leaves the code for setting vlan specific stuff because there's a slight semantic change i want to discuss separately. ifconfig output on a vlan interface changes a little like this:

Re: tweak msk to avoid ifq_deq_begin/commit/rollback

> On 31 May 2017, at 10:58, Stuart Henderson wrote: > > I'll try it tomorrow.. 9 times out of 10 I hit the vp problem and have to > reboot though, so it'll be tricky to say whether it fixes what I sometimes > see with msk. > > IIRC installing on the overdrive 1000 was

Re: shorten ifconfig output for vnetids and parent interfaces

On Mon, May 29, 2017 at 03:34:51PM +1000, David Gwynne wrote: > 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. how about

Re: tcpdump: drop atalk support

EtherTalk (Appletalk over Ethernet) was removed in Mac OS X v10.6 in 2009. You never know what might be flying across your network Ian McWilliam From: owner-t...@openbsd.org on behalf of Henning Brauer

lex scanopt_destroy()

Hi, In lex the return value of scanopt_destroy() is never used. This function is a wrapper for free(), and free() has no value to return. - Michael Index: scanopt.c === RCS file: /cvs/src/usr.bin/lex/scanopt.c,v retrieving

[patch] rebound.c

Be more consistent with logerr usage. Index: rebound.c === RCS file: /cvs/src/usr.sbin/rebound/rebound.c,v retrieving revision 1.83 diff -u -p -u -r1.83 rebound.c --- rebound.c 27 Apr 2017 16:09:32 - 1.83 +++ rebound.c 31 May

Re: tweak msk to avoid ifq_deq_begin/commit/rollback

I'll try it tomorrow.. 9 times out of 10 I hit the vp problem and have to reboot though, so it'll be tricky to say whether it fixes what I sometimes see with msk. IIRC installing on the overdrive 1000 was just "dd miniroot to a USB stick and boot from it". On 31 May 2017 01:29:06 David

tweak msk to avoid ifq_deq_begin/commit/rollback

ie, do the space check before trying to dequeue and mbuf. this also moves it to using m_defrag. i dont have an msk plugged in and i dont know how to use the overdrive 1000 i have here. if someone could test and ok this, it would be great. Index: if_msk.c

Re: [PATCH] ntpd: allow to specify a source IP address for outgoing queries

Dear team, Henning Brauer (off-list) made a few suggestions, which I summerized in the following four points: 1) poor initialization style, instead of: struct xxx yyy = { .property = zzz }; use: struct xxx yyy; yyy.property = zzz; 2) avoid

rework bgpd ext community and support origin validation state

After looking at what IETF and IANA did to extended communities I decided to refactor the code a major bit and hopefully that will work a bit better than what we had before. When I implemented it first I foolishliy assumed that the transitive flag is a flag but it seems that people decided

[patch] typo smtpd/dns.c

fix typo Index: dns.c === RCS file: /cvs/src/usr.sbin/smtpd/dns.c,v retrieving revision 1.83 diff -u -p -u -r1.83 dns.c --- dns.c 28 Oct 2015 07:28:13 - 1.83 +++ dns.c 30 May 2017 22:09:15 - @@ -246,7 +246,7 @@

Re: Unlock IP forwarding paths

On Tue, May 30, 2017 at 10:12:39PM +0200, Alexander Bluhm wrote: > On Tue, May 30, 2017 at 10:04:14PM +0200, Alexander Bluhm wrote: > > On Tue, May 30, 2017 at 11:48:50AM +0200, Martin Pieuchot wrote: > > > Hrvoje Popovski found that ip{,6}_send_dispatch() also need the IPsec > > > dance. > > > >

Re: Properly serialize pflow's sc_outputqueue

this is basically the code i had before mpi told me that mq_ was not needed. *sigh* ok. Visa Hankala(v...@openbsd.org) on 2017.05.30 16:19:46 +: > On Wed, May 31, 2017 at 01:52:31AM +1000, Jonathan Matthew wrote: > > On Tue, May 30, 2017 at 01:04:07PM +, Visa Hankala wrote: > > > Index:

Re: Unlock IP forwarding paths

On Tue, May 30, 2017 at 10:04:14PM +0200, Alexander Bluhm wrote: > On Tue, May 30, 2017 at 11:48:50AM +0200, Martin Pieuchot wrote: > > Hrvoje Popovski found that ip{,6}_send_dispatch() also need the IPsec > > dance. > > > > Updated diff below. > > I have tests this diff with my ipsec regress

Re: [PATCH] ntpd: allow to specify a source IP address for outgoing queries

Works for me, verified both v4 and v6 selection actually uses that address. Suggested a small improvement for ntpd.conf(5) to Job privately. Would love to see this feature get imported :) Thanks Job! Paul On Tue, May 30, 2017 at 06:10:03PM +0200, Job Snijders wrote: | On Sun, May 28, 2017 at

Re: Unlock IP forwarding paths

On Tue, May 30, 2017 at 11:48:50AM +0200, Martin Pieuchot wrote: > Hrvoje Popovski found that ip{,6}_send_dispatch() also need the IPsec > dance. > > Updated diff below. I have tests this diff with my ipsec regress and a non-MP kernel. It crashed. bluhm panic: kernel diagnostic assertion

vmm: exception injection plumbing

This diff adds the ability to inject exceptions into the guest. It is needed for a variety of things coming soon (XO kernel .text, #UD on unsupported instructions, routing RDRAND/RDSEED to arc4random, etc). ok? Index: amd64/vmm.c

Re: sys/socket.h: make sstosa() available to everyone

On Tue, May 30, 2017 at 01:29:07PM -0600, Theo de Raadt wrote: > I don't think this trivial thing should be pushed into the public > namespace. > > Personally I think this construct is really contrived. ok. Another downside might be that it can negatively impact portability. Thanks, Job

Re: sys/socket.h: make sstosa() available to everyone

I don't think this trivial thing should be pushed into the public namespace. Personally I think this construct is really contrived. > Might be out of my depth here, but would be nice if the sstosa() is > available to everyone, not just _KERNEL > > If accepted, 'define sstosa' can to be removed

sys/socket.h: make sstosa() available to everyone

Hi, Might be out of my depth here, but would be nice if the sstosa() is available to everyone, not just _KERNEL If accepted, 'define sstosa' can to be removed from usr.sbin/ftp-proxy/ftp-proxy.c. Kind regards, Job --- sys/sys/socket.h | 16 ++-- 1 file changed, 10 insertions(+),

Re: Remove extra brackets from MINIMUM definition ber.c

On 05/30/17 20:15, Ted Unangst wrote: Fred wrote: I was greping the tree for MINIMUM and this one stood out as it was (((a) rather than ((a). not sure why this one stood out. the 16 definitions of MAXIMUM in bin and sbin are all the same. i don't think we really need churn changing either

Re: Remove extra brackets from MINIMUM definition ber.c

Fred wrote: > I was greping the tree for MINIMUM and this one stood out as it was > (((a) rather than ((a). not sure why this one stood out. the 16 definitions of MAXIMUM in bin and sbin are all the same. i don't think we really need churn changing either way.

Re: Remove extra brackets from MINIMUM definition ber.c

On 05/30/17 19:57, Theo de Raadt wrote: No. This is consistant with a couple hundred others in the tree, which were originally copied from sys/param.h before we cleaned userland from using the file. The brackets were good enough for CSRG, and they are good enough for me... This patch removes

Re: Remove extra brackets from MINIMUM definition ber.c

No. This is consistant with a couple hundred others in the tree, which were originally copied from sys/param.h before we cleaned userland from using the file. The brackets were good enough for CSRG, and they are good enough for me... > This patch removes the redundant brackets form the MINIMUM

Remove extra brackets from MINIMUM definition ber.c

This patch removes the redundant brackets form the MINIMUM macro in ber.c Cheers Fred Index: ber.c === RCS file: /cvs/src/usr.sbin/ldapd/ber.c,v retrieving revision 1.12 diff -u -p -u -p -r1.12 ber.c --- ber.c 11 Feb 2017

Re: let's add PF_LOCK()

Hello Martin, > > rw_exit_write(); > > export_pflow(cur); > > rw_enter_write(); > > + rw_enter_write(_lock); > > } > > This is not needed, you're not diffing against the latest version of > net/pf.c. indeed my tree is old by couple hours.

iwm: add support for missed beacons notification

This adds support for iwm firmware's missed beacon notification. With this, the driver will notice "dead air" when a laptop moves too far away from the AP. It will automatically try to connect to another AP that is closer, or again to the same AP when it moves back in range. ok? Index: if_iwm.c

Re: airport.7: mandoc -Tlint fixes

Hi Michael, Michael Reed wrote on Tue, May 30, 2017 at 01:20:08PM -0400: > The new airport.7 has a few small formatting issues as determined > by `mandoc -Tlint`. This diff fixes them. Done, thanks. I also fixed .Nd and .Lk usage while there. Yours, Ingo > Index: airport.7 >

Re: operator.7: Remove ref to nonexistant FILES

On Tue, May 30, 2017 at 01:18:15PM -0400, Michael Reed wrote: > Hi, > > Per [1], /usr/share/misc/operator no longer exists, so > there's no need to reference it. Fixed, thanks

Re: let's add PF_LOCK()

On 30/05/17(Tue) 19:11, Alexandr Nedvedicky wrote: > @@ -1322,9 +1330,18 @@ > #if NPFLOW > 0 > if (cur->state_flags & PFSTATE_PFLOW) { > /* XXXSMP breaks atomicity */ > + /* > + * The only way how state gets freed from memory is when it > +

Re: detect post threaded condition

patrick keshishian wrote: > On Tue, May 30, 2017 at 03:48:06AM -0400, Ted Unangst wrote: > > + write(2, GREATSCOTT, sizeof(GREATSCOTT)); >^^ > maybe sizeof(GREATSCOTT)-1 indeed. (and typo fixed, thanks all)

Re: let's add PF_LOCK()

Hello Mike, thanks for looking at my stuff. > > It's great stuff, I think you should move forward with this. > A couple of nits: > > - pfvar*.h doesn't have tabs after #define, just one space fixed > > - do you really want to put MTX in PF_FRAG_MTX_* defines? >why not PF_FRAG_LOCK?

airport.7: mandoc -Tlint fixes

Hi, The new airport.7 has a few small formatting issues as determined by `mandoc -Tlint`. This diff fixes them. Index: airport.7 === RCS file: /cvs/src/share/man/man7/airport.7,v retrieving revision 1.1 diff -u -p -r1.1

operator.7: Remove ref to nonexistant FILES

Hi, Per [1], /usr/share/misc/operator no longer exists, so there's no need to reference it. [1]: https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share/misc/Attic/operator Index: operator.7 === RCS file:

Re: detect post threaded condition

On Tue, May 30, 2017 at 03:48:06AM -0400, Ted Unangst wrote: > talking to stsp, he reminded me of a problematic bug that took some time to > track down in some desktop software that shall not be named. after a program > calls fork(), the child has only a single thread. per the standard, it needs >

amd64 clang: fix bootstrap build

On amd64 we can now build the bootstrap with clang if we * skip the integrated assembler for assym.h; * use as(1) to build biosboot.S and the various versions of srt0.S. I successfully tested a clang-built pxeboot(8) and boot(8). Index: Makefile.inc

Re: Randomize link of kernel, and unmap startup code

> This change relinks kernel objects randomly, and unmaps the bootcode > component of locore during boot. This makes gadgets harder to find. > > The current linker layout is: > > locore.o [bring-up code + asm runtime] > rest of kernel .o, in order supplied by config(8) > > The new

Re: Unlock IP forwarding paths

On Tue, May 30, 2017 at 11:48:50AM +0200, Martin Pieuchot wrote: > Updated diff below. OK bluhm@ > Index: net/if.c > === > RCS file: /cvs/src/sys/net/if.c,v > retrieving revision 1.502 > diff -u -p -r1.502 if.c > --- net/if.c 30

Re: Properly serialize pflow's sc_outputqueue

On Wed, May 31, 2017 at 01:52:31AM +1000, Jonathan Matthew wrote: > On Tue, May 30, 2017 at 01:04:07PM +, Visa Hankala wrote: > > Index: net/if_pflow.c > > === > > RCS file: src/sys/net/if_pflow.c,v > > retrieving revision 1.78 >

Re: [PATCH] ntpd: allow to specify a source IP address for outgoing queries

On Sun, May 28, 2017 at 10:52:24PM +0200, Sebastian Benoit wrote: > which makes me think: > would a global local-address be good enough? Attached is a patch that allows you to specify the source for outgoing queries, as a global option. Example ntpd.conf: query from 165.254.255.33 query

Re: Properly serialize pflow's sc_outputqueue

On Tue, May 30, 2017 at 01:04:07PM +, Visa Hankala wrote: > On Mon, May 29, 2017 at 03:33:35PM +, Visa Hankala wrote: > > Currently, access to pflow's sc_outputqueue is not serialized properly. > > The producer has the NET_LOCK(), while the consumer does not. > > mpi@ suggested using

Re: Properly serialize pflow's sc_outputqueue

On 30/05/17(Tue) 13:04, Visa Hankala wrote: > On Mon, May 29, 2017 at 03:33:35PM +, Visa Hankala wrote: > > Currently, access to pflow's sc_outputqueue is not serialized properly. > > The producer has the NET_LOCK(), while the consumer does not. > > mpi@ suggested using mbuf_queue to solve the

id -R: getrtable

add option -R for printing the routing table id is running under ok? diff --git usr.bin/id/id.1 usr.bin/id/id.1 index 35728ee53ee..c5727927bc6 100644 --- usr.bin/id/id.1 +++ usr.bin/id/id.1 @@ -55,6 +55,8 @@ .Fl p .Op Ar user .Nm id +.Fl R +.Nm id .Fl u Op Fl nr .Op Ar user .Sh

Re: id -R routing domain

Looks good to me, after kern_pledge.c allows getrtable for "stdio" Then the sneaky 0 can go away. Index: kern_pledge.c === RCS file: /cvs/src/sys/kern/kern_pledge.c,v retrieving revision 1.209 diff -u -p -u -r1.209 kern_pledge.c ---

id -R routing domain

make it easy to find out which rdomain a shell is in. id -R. one small problem is that pledge doesn't permit getrtable(). Index: id.1 === RCS file: /cvs/src/usr.bin/id/id.1,v retrieving revision 1.18 diff -u -p -r1.18 id.1 --- id.1

Re: Fix some w(1) warnings, remove bin/ps/fmt.c

Ingo Schwarze wrote: > Hi, > > Brian Callahan wrote on Mon, May 29, 2017 at 04:47:42PM -0400: > > > [...] functions fmt_putc and fmt_puts. > > w(1) is doing a reacharound to ps(1) for these functions. > > [...] > > A grep of the tree shows that the only place these two functions > > are used is

Re: Fix some w(1) warnings, remove bin/ps/fmt.c

Hi, Brian Callahan wrote on Mon, May 29, 2017 at 04:47:42PM -0400: > [...] functions fmt_putc and fmt_puts. > w(1) is doing a reacharound to ps(1) for these functions. > [...] > A grep of the tree shows that the only place these two functions > are used is in usr.bin/w/w.c so I put them in there

Re: Properly serialize pflow's sc_outputqueue

On Mon, May 29, 2017 at 03:33:35PM +, Visa Hankala wrote: > Currently, access to pflow's sc_outputqueue is not serialized properly. > The producer has the NET_LOCK(), while the consumer does not. > mpi@ suggested using mbuf_queue to solve the issue. mpi@ commented that the pflow output task

Re: let's add PF_LOCK()

oh, not again... I'm sorry for not attaching patch to the first email On Tue, May 30, 2017 at 02:34:32PM +0200, Alexandr Nedvedicky wrote: > Hello, > > patch delivers two changes to PF: > > it adds PF_LOCK() et. al. At the moment the PF_LOCK() sort of > duplicates the current

let's add PF_LOCK()

Hello, patch delivers two changes to PF: it adds PF_LOCK() et. al. At the moment the PF_LOCK() sort of duplicates the current NET_LOCK(). It essentially synchronizes packets with ioctl(2) and timer thread, which purges states. The future work is going to break PF_LOCK into

Re: efiboot serial console support

Thanks, Let me update the diff. On Tue, 30 May 2017 13:58:29 +0200 Patrick Wildt wrote: > On Tue, May 30, 2017 at 12:58:48PM +0200, YASUOKA Masahiko wrote: >> The following diff is to support serial console on efiboot. > > Nice. >> >> It uses ACPI UID to identify the port

Re: Kill db_sym_t

> X-Virus-Scanned: by XS4ALL Virus Scanner > X-Originating-IP: 87.142.3.46 > Date: Tue, 30 May 2017 12:58:21 +0200 > From: Martin Pieuchot > > We are only supporting ELF so this abstraction type is no longer needed. > > I'll make my life easier with CTF if I don't have to cast

Re: ksh(1): vi mode UTF-8 bug

Hi Ingo, On Tue, May 30, 2017 at 2:28 AM, Ingo Schwarze wrote: > Walter Alejandro Iglesias wrote on Mon, May 29, 2017 at 06:44:40PM +0200: > >> Are those wide char versions of C functions consistent enough to write >> a separate implementation to be loaded when LC_TYPE is set

Merge pfkeyv2_socket and rawcb into one struct

This is more or less the same thing for PF_KEY that we now do in PF_ROUTE. Use one PCB LIST on the keycb and embedd the rawcb in that PF_KEY cb. Diff also has a few variable renames in it to make this code less alien regarding the rest of our kernel. Mainly use so instead of socket and

Re: vmd vioblk start/finish

On Tue, May 30, 2017 at 07:00:08AM -0400, Ted Unangst wrote: > This splits the read/write functions into top and bottom halves. It doesn't > change much yet, but this is a requirement for async IO. The start funtion > turns the request into an ioinfo (to be completed eventually by a thread) and >

move pbc list from rawcb to routecb (and later pfkey cb)

This is a step I need to do to make progress on the PF_KEY cleanup I'm doing. Both PF_ROUTE and PF_KEY need to start to take care of their own PCB list and so move the LIST_ENTRY out of rawcb into routecb. This allows me to do the same in PF_KEY which will be sent as the next diff. -- :wq

Re: efiboot serial console support

On Tue, May 30, 2017 at 12:58:48PM +0200, YASUOKA Masahiko wrote: > Hi, > > The following diff is to support serial console on efiboot. Nice. > > It uses ACPI UID to identify the port number (com0, com1 and so on) of > probed serial interface. But I'm not sure wether com0-com3 are always >

Re: Makefile.cross tweaks

On Mon, May 29, 2017 at 08:18:08PM +, Miod Vallat wrote: > The following diff attempts to cross-build more things, in particular > gnu/lib (except for libiberty). It also passes the proper optimization > flags so that libstdc++-v3 gets built with optimization. Doesn't build for arm64,

Re: KSH bug: case inside command substitution

Hi Vadim, On Mon, May 29, 2017 at 05:09:11AM +0300, Vadim Zhukov wrote: > 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

vmd vioblk start/finish

This splits the read/write functions into top and bottom halves. It doesn't change much yet, but this is a requirement for async IO. The start funtion turns the request into an ioinfo (to be completed eventually by a thread) and the finish function retrives the result. (for now, we just do the

efiboot serial console support

Hi, The following diff is to support serial console on efiboot. It uses ACPI UID to identify the port number (com0, com1 and so on) of probed serial interface. But I'm not sure wether com0-com3 are always mapped UID 0-3 as expected. Though I think this is good enough. Comment? diff --git

Kill db_sym_t

We are only supporting ELF so this abstraction type is no longer needed. I'll make my life easier with CTF if I don't have to cast types when looking for type information. Ok to kill this abstraction? Index: arch/amd64/amd64/db_trace.c

wsfont: remove qvss8x15, vt220l8x8, and vt220l8x10?

Hi, This would be the last round of console fonts removal. Those fonts are not compiled in by default, and are smaller than the smallest font (bold8x16_iso1) we currently include. Comments? OK to remove them?

Re: Unlock IP forwarding paths

On 30.5.2017. 11:48, Martin Pieuchot wrote: > On 30/05/17(Tue) 10:45, Martin Pieuchot wrote: >> Diff below moves IPv4 & IPv6 incoming/forwarding path, PIPEX ppp >> processing and IPv4 & IPv6 dispatch functions outside the KERNEL_LOCK(). >> >> We currently rely on the NET_LOCK() serializing access

Re: tcpdump: print-bgp.c: update error code 7

On Tue, May 30, 2017 at 11:50:57AM +0200, Michal Mazurek wrote: > On 11:27:28, 30.05.17, Claudio Jeker wrote: > > On a side note. The notification error code 7 seems to be wrong. > > The capability error codes made it never into a standard and now > > error code 7 if for enhanced route refresh. >

ext2: flush cache on mount -ur

The same as is already implemented for ffs & msdosfs. ok? Cheers, Stefan diff --git a/sys/ufs/ext2fs/ext2fs_vfsops.c b/sys/ufs/ext2fs/ext2fs_vfsops.c index 53eaa05a32a..98d5536418c 100644 --- a/sys/ufs/ext2fs/ext2fs_vfsops.c +++ b/sys/ufs/ext2fs/ext2fs_vfsops.c @@ -181,6 +181,7 @@

ext2fs: Mark superblock as not modified when written

I have seen spurious "file system not clean; please fsck(8)" warnings during "mount -ur". Set e2fs_fmod = 0 when writing the superblock (as ffs does). ok? Cheers, Stefan diff --git a/sys/ufs/ext2fs/ext2fs_vfsops.c b/sys/ufs/ext2fs/ext2fs_vfsops.c index 98d5536418c..372c4d6f1fc 100644 ---

Re: tcpdump: drop atalk support

> * Theo de Raadt [2017-05-30 10:56]: > > > How about just dropping support for /etc/appletalk.names, which as far > > > as I can tell was never used, and drop the manpage bit, reducing it by > > > 10%. Most of the text in the manpage is outdated anyway, talking about > > >

Re: tcpdump: drop atalk support

* Theo de Raadt [2017-05-30 10:56]: > > How about just dropping support for /etc/appletalk.names, which as far > > as I can tell was never used, and drop the manpage bit, reducing it by > > 10%. Most of the text in the manpage is outdated anyway, talking about > >

Re: ip_input() in pr_input style

On 30/05/17(Tue) 11:40, Alexander Bluhm wrote: > On Tue, May 30, 2017 at 08:45:53AM +0200, Martin Pieuchot wrote: > > On 29/05/17(Mon) 23:45, Alexander Bluhm wrote: > > > Hi, > > > > > > Convert ip_input(), ip_our(), ip_deliver() functions to pr_input > > > parameter passing and protocol return

tcpdump: print-bgp.c: update error code 7

On 11:27:28, 30.05.17, Claudio Jeker wrote: > On a side note. The notification error code 7 seems to be wrong. > The capability error codes made it never into a standard and now > error code 7 if for enhanced route refresh. > > So I would replace bgpnotify_minor_cap with > static const char

Re: detect post threaded condition

Il 30 mag 2017 09:48, "Ted Unangst" ha scritto: talking to stsp, he reminded me of a problematic bug that took some time to track down in some desktop software that shall not be named. after a program calls fork(), the child has only a single thread. per the standard, it

Re: Unlock IP forwarding paths

On 30/05/17(Tue) 10:45, Martin Pieuchot wrote: > Diff below moves IPv4 & IPv6 incoming/forwarding path, PIPEX ppp > processing and IPv4 & IPv6 dispatch functions outside the KERNEL_LOCK(). > > We currently rely on the NET_LOCK() serializing access to most global > data structures for that. IP

Re: ip_input() in pr_input style

On Tue, May 30, 2017 at 08:45:53AM +0200, Martin Pieuchot wrote: > On 29/05/17(Mon) 23:45, Alexander Bluhm wrote: > > Hi, > > > > Convert ip_input(), ip_our(), ip_deliver() functions to pr_input > > parameter passing and protocol return style. Reset mp to NULL in > > a few places to fail at mbuf

Re: 802.11 multicast frames and the basic rate set

On Tue, May 30, 2017 at 11:03:09AM +0200, Stefan Sperling wrote: > ok? Always wait for the second diff! In the previous version I missed an indirection via wn->ridx[]. wrong: rinfo = _rates[ieee80211_min_basic_rate(ic)] correct: rinfo = _rates[wn->ridx[ieee80211_min_basic_rate(ic)]]; Because of

Re: tcpdump: enable some more bgp info

On Tue, May 30, 2017 at 11:06:10AM +0200, Michal Mazurek wrote: > On 10:43:30, 30.05.17, Job Snijders wrote: > > In the registry created by RFC 6608, the value "0" is the BGP Finite > > State Machine Error subcode meaning "Unspecified Error". I think that > > when a name is assigned to a value,

pcdisplay(4): properly disable hardware cursor

Hi tech@, Use the CD bit (Cursor Disable) in the cursor start register to properly disable hardware cursor. >From NetBSD. See commit message [1] and diff [2] for more information. Comments? OK? [1]

Re: tcpdump: enable some more bgp info

On 10:43:30, 30.05.17, Job Snijders wrote: > In the registry created by RFC 6608, the value "0" is the BGP Finite > State Machine Error subcode meaning "Unspecified Error". I think that > when a name is assigned to a value, the name should be printed (like > your patch does for subcode values 1,

802.11 multicast frames and the basic rate set

Many of our drivers hardcode multicast frames to 1 Mbit/s on 2 GHz. This assumes 11b interop, which is mandatory for all 2GHz PHYs and thus generally works. However, access points may modify the networks' basic rate set such that it differs from the mandatory PHY rate set. (Standard laywers may

Re: tcpdump: drop atalk support

> How about just dropping support for /etc/appletalk.names, which as far > as I can tell was never used, and drop the manpage bit, reducing it by > 10%. Most of the text in the manpage is outdated anyway, talking about > /etc/atalk.names - support for which was removed in 2004 with the > privsep

Re: tcpdump: drop atalk support

How about just dropping support for /etc/appletalk.names, which as far as I can tell was never used, and drop the manpage bit, reducing it by 10%. Most of the text in the manpage is outdated anyway, talking about /etc/atalk.names - support for which was removed in 2004 with the privsep work.

Unlock IP forwarding paths

Diff below moves IPv4 & IPv6 incoming/forwarding path, PIPEX ppp processing and IPv4 & IPv6 dispatch functions outside the KERNEL_LOCK(). We currently rely on the NET_LOCK() serializing access to most global data structures for that. IP input queues are no longer used in the forwarding case.

Re: tcpdump: enable some more bgp info

On Tue, May 30, 2017 at 10:21:17AM +0200, Michal Mazurek wrote: > On 12:15:06, 29.05.17, Job Snijders wrote: > > perhaps add a comment like /* RFC 6608 */ above the below: > > Right, it will make it more consistent. > > > > +static const char *bgpnotify_minor_fsm[] = { > > > + NULL, "In OpenSent

Re: tcpdump: enable some more bgp info

On Tue, May 30, 2017 at 10:21:17AM +0200, Michal Mazurek wrote: > On 12:15:06, 29.05.17, Job Snijders wrote: > > perhaps add a comment like /* RFC 6608 */ above the below: > > Right, it will make it more consistent. > > > > +static const char *bgpnotify_minor_fsm[] = { > > > + NULL, "In OpenSent

Re: tcpdump: enable some more bgp info

On 2017 May 30 (Tue) at 10:21:17 +0200 (+0200), Michal Mazurek wrote: :On 12:15:06, 29.05.17, Job Snijders wrote: :> perhaps add a comment like /* RFC 6608 */ above the below: : :Right, it will make it more consistent. : :> > +static const char *bgpnotify_minor_fsm[] = { :> > + NULL, "In OpenSent

Re: Randomize link of kernel, and unmap startup code

On Tue, May 30, 2017 at 02:16:45AM -0600, Theo de Raadt wrote: > This change relinks kernel objects randomly, and unmaps the bootcode > component of locore during boot. This makes gadgets harder to find. > > The current linker layout is: > > locore.o [bring-up code + asm runtime] >

Re: tcpdump: enable some more bgp info

> > > +static const char *bgpnotify_minor_fsm[] = { > > > + NULL, "In OpenSent State", "In OpenConfirm State", > > > + "In Established State", > > > +}; > > > > and maybe s/NULL/"Unspecified Error"/ > > If it's NULL, then tcpdump will print out the number: > > if (p == NULL) { >

Re: tcpdump: enable some more bgp info

On 12:15:06, 29.05.17, Job Snijders wrote: > perhaps add a comment like /* RFC 6608 */ above the below: Right, it will make it more consistent. > > +static const char *bgpnotify_minor_fsm[] = { > > + NULL, "In OpenSent State", "In OpenConfirm State", > > + "In Established State", > > +}; >

Randomize link of kernel, and unmap startup code

This change relinks kernel objects randomly, and unmaps the bootcode component of locore during boot. This makes gadgets harder to find. The current linker layout is: locore.o [bring-up code + asm runtime] rest of kernel .o, in order supplied by config(8) The new linker layout

Re: ksh(1): vi mode UTF-8 bug

Hi, On Mon, May 29, 2017 at 04:16:06PM +0200, Ingo Schwarze wrote: > Hi, > > Anton Lindqvist wrote on Sun, May 28, 2017 at 06:07:00PM +0200: > > On Sun, May 28, 2017 at 10:56:19AM +0200, Walter Alejandro Iglesias wrote: > > >> There is still a similar issue when you try to "replace" a utf-8 >

detect post threaded condition

talking to stsp, he reminded me of a problematic bug that took some time to track down in some desktop software that shall not be named. after a program calls fork(), the child has only a single thread. per the standard, it needs to get to exec() as quickly as possible. per the quality standards

Re: Introduce ipv{4,6}_input()

On Tue, May 30, 2017 at 09:27:28AM +0200, Martin Pieuchot wrote: > Here's a simple refactoring diff to simplify the next one. The upcoming > diff will be unlocking IP forwarding paths, so I want to keep it as > small as possible. > > The idea of the refactoring below is to introduce two wrappers

Introduce ipv{4,6}_input()

Here's a simple refactoring diff to simplify the next one. The upcoming diff will be unlocking IP forwarding paths, so I want to keep it as small as possible. The idea of the refactoring below is to introduce two wrappers functions around IP queues. In the next iteration they will call the

Re: Fix some mg(1) warnings

commited, thanks! On Mon, May 29, 2017 at 03:15:21PM -0400, Brian Callahan wrote: > Hi tech -- > > This diff cleans up some warnings found with WARNINGS=yes and > compiling with both cc and clang. I ignored the -Wsign-compare > and -Wmissing-prototypes warnings. > > ~Brian > > Index: dir.c >

Re: ip_input() in pr_input style

On 29/05/17(Mon) 23:45, Alexander Bluhm wrote: > Hi, > > Convert ip_input(), ip_our(), ip_deliver() functions to pr_input > parameter passing and protocol return style. Reset mp to NULL in > a few places to fail at mbuf use after free. Rename ipv4_input() > to ip_input(). > > Goal is to

Re: Silence some warnings in tftp(1)

> Silence some warnings in tftp(1) found by WARNINGS=yes. I'm getting tired of these emails which say there are warnings, then FAILS TO SHOW THEM, and proceeds to do huge variable renaming refactorings of the source files. Thereby requiring whoever gets this in to redo it all the way from the

Re: Silence some warnings in calendar(1)

Brian Callahan wrote: > Hi tech -- > > Fix some warnings found with WARNINGS=yes. I believe setting > ev1 = NULL in io.c is a false positive by cc but I fixed it anyway. > > ~Brian > > Index: day.c > === > RCS file: