Re: Say no to LARVAL

On Wed, May 9, 2018 at 3:31 AM, Martin Pieuchot wrote: > On 08/05/18(Tue) 14:12, Philip Guenther wrote: > > On Tue, 8 May 2018, Martin Pieuchot wrote: > > > The way our kernel allocates and populates new 'struct file *' is > > > currently a complete mess. One of the problems

[patch] adds include statement in dhcpd.conf

Hi, this diff implements the “include” statement, like other daemons. Also the config file can be split between different files (in my case, a big list of client is generated from a script, and I don’t want to modify dhcpd.conf). Although, I am not even sure if this diff is decent and if it’s a

Re: [patch] Use err instead of errx for checking strdup failed

On Thu, May 10 2018, Nan Xiao wrote: > Hi tech@, Hi, > Following is a trivial modification of tcpbench.c, and I think use err > is more better than errx. Apologize if I am wrong, thanks! I think the preferred idiom is err(1, NULL); > Index: tcpbench.c >

Re: adjtime, clock_settime, settimeofday: EINVAL for non-normal time inputs

On Thu, May 10, 2018 at 09:13:27AM -0500, Scott Cheloha wrote: > Hi, > > This sets EINVAL on non-normal input for adjtime(2), clock_settime(2), > and settimeofday(2). > > clock_settime and settimeofday take absolute times as input, and adjtime > is explicitly allowed to be a negative relative

Re: [patch] Set TCP send buffer size only when tcpbench client works in TCP mode

On Thu, May 10, 2018 at 09:13:14PM +0800, Nan Xiao wrote: > int Sflag;/* Socket buffer size (tcp mode) */ > So I think maybe this should work on TCP only. BTW, if it works on both > TCP/UDP, I think differentiate error message will be better: I have changed all comments and

adjtime, clock_settime, settimeofday: EINVAL for non-normal time inputs

Hi, This sets EINVAL on non-normal input for adjtime(2), clock_settime(2), and settimeofday(2). clock_settime and settimeofday take absolute times as input, and adjtime is explicitly allowed to be a negative relative time, so timespecfix() is inapplicable. POSIX specifies that clock_settime(2)

Re: if_indextoname(): Fix typo in reference to RFC

ok ;) Theo Buehler(t...@openbsd.org) on 2018.05.10 14:52:04 +0200: > The text quoted in this comment is section 4.2 from RFC 2553 > ("Basic Socket Interface Extensions for IPv6") > https://tools.ietf.org/html/rfc2553#section-4.2 > > RFC 2533 ("A Syntax for Describing Media Feature Sets") clearly

Re: [patch] Use err instead of errx for checking strdup failed

ok benno@ Nan Xiao(n...@chinadtrace.org) on 2018.05.10 21:27:40 +0800: > Hi tech@, > > Following is a trivial modification of tcpbench.c, and I think use err > is more better than errx. Apologize if I am wrong, thanks! > > Index: tcpbench.c >

[patch] Use err instead of errx for checking strdup failed

Hi tech@, Following is a trivial modification of tcpbench.c, and I think use err is more better than errx. Apologize if I am wrong, thanks! Index: tcpbench.c === RCS file: /cvs/src/usr.bin/tcpbench/tcpbench.c,v retrieving revision

Re: tcp(4) does not mention netinet/tcp.h

On Thu, May 10, 2018 at 03:27:24PM +0300, Vadim Zhukov wrote: > Hi all! > > The tcp(4) page talks about TCP options which are available through > , but doesn't menthion this header. DragonFly BSD, > FreeBSD & Linux do; Solaris mentions this header only; NetBSD is like > us currently. > > Okay?

Re: [patch] Set TCP send buffer size only when tcpbench client works in TCP mode

Very sorry for forgetting to cc reply into tech@. Forwarded Message Subject: Re: [patch] Set TCP send buffer size only when tcpbench client works in TCP mode Date: Thu, 10 May 2018 17:56:25 +0800 From: Nan Xiao To: Alexander Bluhm

if_indextoname(): Fix typo in reference to RFC

The text quoted in this comment is section 4.2 from RFC 2553 ("Basic Socket Interface Extensions for IPv6") https://tools.ietf.org/html/rfc2553#section-4.2 RFC 2533 ("A Syntax for Describing Media Feature Sets") clearly has nothing to do with this function. Index: lib/libc/net/if_indextoname.c

tcp(4) does not mention netinet/tcp.h

Hi all! The tcp(4) page talks about TCP options which are available through , but doesn't menthion this header. DragonFly BSD, FreeBSD & Linux do; Solaris mentions this header only; NetBSD is like us currently. Okay? Or should we remove netinet/in.h from here as well (at least

Re: mpfsafe pfkey socket list

On Wed, May 09, 2018 at 03:45:44PM +0200, Martin Pieuchot wrote: > Diff below turns the list of pfkey sockets into a SRPL. It reuses > the logic introduced by claudio@ for routing sockets. That allows > us to stop grabbing the KERNEL_LOCK() when iterating over the list. > > ok? OK claudio@ >

Re: [patch] Set TCP send buffer size only when tcpbench client works in TCP mode

On Thu, May 10, 2018 at 03:10:53PM +0800, Nan Xiao wrote: > Per my understanding, the tcpbench.Sflag (Socket buffer size) should > only take effect when working in TCP mode. Why? The kernel provides socket buffers for all protocols. Userland can set them and the kernel does the appropriate

Re: pf route to on loopback

Hello, On Wed, May 09, 2018 at 07:05:30PM +0200, Alexander Bluhm wrote: > Hi, > > I while ago I changed pf route-to that it does not send packets > from 127.0.0.1 address to the network. This is necessary for localy > generated icmp packets that would be dropped otherwise. > > Now I found out

[patch] Set TCP send buffer size only when tcpbench client works in TCP mode

Hi tech@, Per my understanding, the tcpbench.Sflag (Socket buffer size) should only take effect when working in TCP mode. But according to my testing, it also works when tcpbench client is in UDP mode: (1) Set buffer size less than DEFAULT_UDP_PKT (1500 - 28): # tcpbench -u -S 1471 127.0.0.1