Re: syslogd suppress message repeated

On 2017/04/17 00:57, Alexander Bluhm wrote: > Hi, > > I have received complaints from a customer that the "last message > repeated" feature of syslogd(8) makes their log auditing difficult. > FreeBSD has -c and -cc to disable it, but our -c is already taken. > > Add -r to suppress the summary

retire icmp6_rip6_input

Hi, As the comment says, icmp6_rip6_input() is mostly duplicated code from rip6_input(). So lets merge these functions together and retire icmp6_rip6_input(). ok? bluhm Index: netinet6/icmp6.c === RCS file:

syslogd suppress message repeated

Hi, I have received complaints from a customer that the "last message repeated" feature of syslogd(8) makes their log auditing difficult. FreeBSD has -c and -cc to disable it, but our -c is already taken. Add -r to suppress the summary line for pipe and remote loghost, as they are most commonly

Re: ld.lld: disable colors/bold/whatever

> Date: Sun, 16 Apr 2017 21:21:24 +0200 > From: Patrick Wildt > > Hi, > > this diff disables the color diagnostics by default so that LLVM's lld > does not throw any ANSI sequences. > > ok? Wouldn't it be better to just change

ld.lld: disable colors/bold/whatever

Hi, this diff disables the color diagnostics by default so that LLVM's lld does not throw any ANSI sequences. ok? Patrick diff --git a/gnu/llvm/tools/lld/ELF/DriverUtils.cpp b/gnu/llvm/tools/lld/ELF/DriverUtils.cpp index 5ba494d791d..36b4052d72a 100644 ---

syslogd multiple tls listen sockets

Hi, This allows syslogd(8) to listen on multiple addresses for incomming TLS connections. ok? bluhm Index: usr.sbin/syslogd/syslogd.8 === RCS file: /data/mirror/openbsd/cvs/src/usr.sbin/syslogd/syslogd.8,v retrieving revision 1.53

Re: arm64 ieee.h

On Sun, Apr 16, 2017 at 05:52:54PM +0200, Mark Kettenis wrote: > Like mips64 and sparc64, the 128-bit long double floating point format > needs the EXT_IMPLICIT_NBIT define. This fixes the lib/libc/printf > regress test. > > While there, remove struct ieee_ldouble, which isn't used and doesn't >

Re: show correct vltime/pltime when setting with ifconfig(8)

On Sat, Apr 08, 2017 at 04:55:23PM +, Florian Obser wrote: > We need to update the ia6_updatetime when changing the vltime/pltime > vio ioctl(2). Otherwise ifconfig shows the wrong times since > vltime/pltime are calculated from this value on export. > > The handling of router

Re: show correct vltime/pltime when setting with ifconfig(8)

anyone? On Sat, Apr 08, 2017 at 04:55:23PM +, Florian Obser wrote: > > We need to update the ia6_updatetime when changing the vltime/pltime > vio ioctl(2). Otherwise ifconfig shows the wrong times since > vltime/pltime are calculated from this value on export. > > The handling of router

check af in rip_input

Hi, The raw ip input functions are called from several places, so I think a check that the address family is correct, is justified. ok? bluhm Index: netinet/raw_ip.c === RCS file:

use af in tcp_input

Hi, Simplify tcp_input() by using the address family passed by pr_input. ok? bluhm Index: netinet/tcp_input.c === RCS file: /data/mirror/openbsd/cvs/src/sys/netinet/tcp_input.c,v retrieving revision 1.339 diff -u -p -r1.339

Re: start building clang alongside gcc on amd64

> Date: Mon, 17 Apr 2017 01:49:18 +1000 > From: Jonathan Gray > > On Sun, Apr 16, 2017 at 11:15:55AM +0200, Mark Kettenis wrote: > > Here is my current diff to build clang alongside gcc. The idea is > > that if you add an architecture to both CLANG_ARCH and GCC4_ARCH, both > >

arm64 ieee.h

Like mips64 and sparc64, the 128-bit long double floating point format needs the EXT_IMPLICIT_NBIT define. This fixes the lib/libc/printf regress test. While there, remove struct ieee_ldouble, which isn't used and doesn't exist on other architectures. ok? Index: arch/arm64/include/ieee.h

Re: start building clang alongside gcc on amd64

On Sun, Apr 16, 2017 at 11:15:55AM +0200, Mark Kettenis wrote: > Here is my current diff to build clang alongside gcc. The idea is > that if you add an architecture to both CLANG_ARCH and GCC4_ARCH, both > compilers get built, but gcc remains the default compiler. It forces > the clang-related

use af in udp_input

Hi, As we have the address family in pr_input now, udp_input can be simplified. ok? bluhm Index: netinet/udp_usrreq.c === RCS file: /data/mirror/openbsd/cvs/src/sys/netinet/udp_usrreq.c,v retrieving revision 1.234 diff -u -p

Re: Noob looking to work on HDMI audio integration

On Sun, Apr 16, 2017 at 5:54 PM, Jonathan Gray wrote: > The hdmi codecs were disabled in azalia as non working audio devices > would attach as audio0 with the real audio device attaching as audio1. > This meant symlinks in /dev had to change or programs had to explicitly > use

Re: Noob looking to work on HDMI audio integration

On Sun, Apr 16, 2017 at 05:04:49PM +0300, Tanel Kriik wrote: > Hi, total newbie here: > > I have intermediate C experience from work and free time, > and use OpenBSD more or less often. > I'm familiar with the OpenBSD coding style and have compiled the > kernel a few times, > but I haven't

Noob looking to work on HDMI audio integration

Hi, total newbie here: I have intermediate C experience from work and free time, and use OpenBSD more or less often. I'm familiar with the OpenBSD coding style and have compiled the kernel a few times, but I haven't dabbled deeper than that. Areas I'm surely lacking in: * Audio processing (only

vi NULL check before free()

Hi, Remove some checks for NULL around free() in vi. vi still sees to work when I build it on i386 and amd64. - Michael Index: cl/cl_screen.c === RCS file: /cvs/src/usr.bin/vi/cl/cl_screen.c,v retrieving revision 1.27 diff -u -p

Re: arm64 _longjmp fix

On Sun, Apr 16, 2017 at 12:27:51PM +0200, Mark Kettenis wrote: > We fixed longjmp(), but left _longjmp() alone. In this case the > return code is still in w1 so the fix is slightly different. Makes > the longjmp regress test pass on arm64. > > ok? ok jsg@ > > > Index:

Fix regress/lib/libc/db

The db-3 subtest fails on 64-bit systems (and has been failing for a long time, maybe forever?). The problem is that (off_t)SIZE_MAX becomes -1, so the if (sb.st_size > (off_t)SIZE_MAX) condition is always true. Since we pass (int)sb.st_size to read(2), I think checking for INT_MAX is

dnsmasq/IP_SENDSRCADDR [st...@openbsd.org: CVS: cvs.openbsd.org: ports]

Anyone have ideas about this? - Forwarded message from Stuart Henderson - From: Stuart Henderson Date: Sun, 16 Apr 2017 04:41:10 -0600 (MDT) To: ports-chan...@openbsd.org Subject: CVS: cvs.openbsd.org: ports CVSROOT:/cvs Module name:

arm64 _longjmp fix

We fixed longjmp(), but left _longjmp() alone. In this case the return code is still in w1 so the fix is slightly different. Makes the longjmp regress test pass on arm64. ok? Index: lib/libc/arch/aarch64/gen/_setjmp.S === RCS

Fix regex regress for clang

clang doesn't like the DEF_WEAK(regfree) that's in the libc code that gets recycled for the test. Diff below defines it away in the same way as we do in ld.so. ok? Index: regress/lib/libc/regex/Makefile === RCS file:

yacc: y.tab.c remains on error

Hi, yacc will produce a y.tab.c file when it is fed input with bad syntax. Maybe this is expected behaviour but is it better for yacc to clean up after itself? When I feed the same badsyntax.y to bison 3.0.4 it doesn't create y.tab.c. yacc creates y.tab.c very early. Unlink the file in done() if

start building clang alongside gcc on amd64

Here is my current diff to build clang alongside gcc. The idea is that if you add an architecture to both CLANG_ARCH and GCC4_ARCH, both compilers get built, but gcc remains the default compiler. It forces the clang-related libraries (libcompiler_rt, libc++abi, libc++) to be built with clang.

Re: umb: aggregate packets on tx

On Sat, Apr 15, 2017 at 11:21:41PM +0200, Alexander Bluhm wrote: > On Mon, Feb 20, 2017 at 04:35:10PM +0100, Gerhard Roth wrote: > > On Mon, 12 Dec 2016 14:50:50 +0100 Gerhard Roth > > wrote: > > > The current umb(4) implementation needs one USB transfer for every packet >