Re: iked.conf.5, ipsec.conf.5: Quote $domain in tag string

2020-02-16 Thread Jason McIntyre
On Sun, Feb 16, 2020 at 12:23:40AM +0100, Klemens Nanni wrote: > On Sat, Feb 15, 2020 at 10:30:52PM +, Jason McIntyre wrote: > > from a practical point of view, is there a reason to say when expansion > > happens? by this i mean, what (if any) difference does it have for the > > user - they

dumpfs: don't pick alternate superblock

2020-02-16 Thread Otto Moerbeek
Hi, If the block size is 64k, the first alternate ffs1 superblock ends up in a location first looked at by dumpfs. fsck_ffs(8) (see setup.c) and ffs_mountfs() in sys/ufs/ffs/ffs_vfsops.c have protection against that case, since we really want the primary superblock, that's the one that is

arm64 syscall ABI change

2020-02-16 Thread Mark Kettenis
In order to fix a speculative execution issue on various ARM CPUs, the OpenBSD/arm64 system call ABI has been changed. System calls now skip the two instructions immediately following the system call instruction. This allows us to insert a barrier that blocks the CPU from speculating further

Set UF_EXCLOSE inside finishdup()

2020-02-16 Thread Visa Hankala
The closing of a file can be a complex operation, and hence it would be good to hold as few locks as possible when calling closef(). In particular, the code behind close(2) already releases the file descriptor table lock before closef(). This might not be strictly necessary when the type of the

Re: iked.conf.5: Add BUGS section

2020-02-16 Thread Klemens Nanni
On Fri, Feb 14, 2020 at 08:56:25PM +0100, Klemens Nanni wrote: > I'm experiencing some issues with iked(8) and the first one took > me much longer than appreciated: order of commands matters. iked/iked.conf is not special, ipsecctl/ipsec.conf obviously has it, too. So while the problem of

Re: vmctl.8, vm.conf.5: DHCP is configured on the first interface only

2020-02-16 Thread Klemens Nanni
On Sun, Feb 16, 2020 at 06:35:10AM +0100, Theo Buehler wrote: > I don't think this patch is correct. > > The vmctl part contradicts the "LOCAL INTERFACES" section, which > explains how the addresses are calculated and also states: > > Multiple -L options can be provided to the 'vmctl start'

examples/bgpd.conf: Remove trailing backslashes

2020-02-16 Thread Klemens Nanni
Reading it for the first time, I thought `mynetworks' is special and that's why it has these line continuations, but it's not and the example looks much nicer without. The `bogons' block directly following doesn't use it, either. Feedback? OK? Index: etc/examples/bgpd.conf

Re: bgpd.conf.5: Tag groups

2020-02-16 Thread Klemens Nanni
On Sun, Feb 16, 2020 at 09:25:51PM +0100, Klemens Nanni wrote: > Going through the example with bgpd.conf(5) side by side, jumping to the > "group" tag shows > > group descr Neighbors in this group will be matched. > AS as-number >Neighbors with this AS will be

Re: bgpd.conf.5: Tag groups

2020-02-16 Thread Klemens Nanni
On Sun, Feb 16, 2020 at 09:47:43PM +0100, Ingo Schwarze wrote: > Sure. I'm not convinced you even need OKs for such simple tagging > corrections. > > By the way, just > > .Tg > .Ic neighbor > > .Tg > .Ic group > > is sufficient, the .Tg macro defaults to the first argument of the

bgpd.conf.5: Tag groups

2020-02-16 Thread Klemens Nanni
Going through the example with bgpd.conf(5) side by side, jumping to the "group" tag shows group descr Neighbors in this group will be matched. AS as-number Neighbors with this AS will be matched. ... which is somewhere in the manual that assumes

Re: bgpd.conf.5: Tag groups

2020-02-16 Thread Ingo Schwarze
Hi Klemens, Klemens Nanni wrote on Sun, Feb 16, 2020 at 09:32:52PM +0100: > On Sun, Feb 16, 2020 at 09:25:51PM +0100, Klemens Nanni wrote: >> Going through the example with bgpd.conf(5) side by side, jumping to the >> "group" tag shows >> >> group descr Neighbors in this group will be

Re: snmp(1) fix parse agent

2020-02-16 Thread Martijn van Duren
jan@ found some issues with this diff during u2k20. Here's an updated diff. Changes since previous diff are: - If we don't know the protocol yet, we can't assume a port is specified. - If we can create a ip6? socket don't wait to connect until we're out of the loop. It passes everything I've

Re: Raise spl for updating kn_status

2020-02-16 Thread Visa Hankala
On Sat, Feb 15, 2020 at 09:42:53PM +0100, Martin Pieuchot wrote: > On 15/02/20(Sat) 16:56, Visa Hankala wrote: > > When I added the knote_acquire() call in kqueue_register(), I overlooked > > the fact that the knote could be modified by a (soft) interrupt. > > Interrupts have to be blocked when

Re: em(4) towards multiqueues

2020-02-16 Thread Hrvoje Popovski
On 14.2.2020. 18:28, Martin Pieuchot wrote: > I'm running this on: > > em0 at pci1 dev 0 function 0 "Intel I210" rev 0x03: msi > em0 at pci0 dev 20 function 0 "Intel I354 SGMII" rev 0x03: msi > > More tests are always welcome ;) em0 at pci0 dev 25 function 0 "Intel 82579LM" rev

Re: snmp(1) fix parse agent

2020-02-16 Thread Martijn van Duren
On 2/17/20 8:12 AM, Martijn van Duren wrote: > jan@ found some issues with this diff during u2k20. > Here's an updated diff. Changes since previous diff are: > - If we don't know the protocol yet, we can't assume a port is > specified. > - If we can create a ip6? socket don't wait to connect

Re: ospf6d: rework rde_lsdb.c

2020-02-16 Thread Remi Locherer
On Sat, Feb 15, 2020 at 11:37:12AM +0100, Denis Fondras wrote: > 3 changes in rde_lsdb.c > - lsa_find_lsid() has redondant parameters > - call to lsa_self() can be simplified (== ospfd) > - update debug messages to be more suitable > ok remi@ > Index: rde.c >