Re: ospfd: add IMSG_IFADDRADD to deal with "sh /etc/netstart if"

On 08/21/17 18:57, Remi Locherer wrote: On Mon, Jul 24, 2017 at 04:59:46PM +0200, Remi Locherer wrote: On Fri, Jul 21, 2017 at 06:24:06PM +0200, Remi Locherer wrote: On Fri, Jul 21, 2017 at 02:45:03PM +0200, Florian Riehm wrote: On 06/25/17 23:47, Remi Locherer wrote: Hi, ospfd does not

Re: partial compat checking w/OF_is_compatible() ?

> Date: Wed, 23 Aug 2017 00:17:01 +0300 > From: Artturi Alm > > On Tue, Aug 22, 2017 at 09:47:27PM +0200, Mark Kettenis wrote: > > > Date: Tue, 22 Aug 2017 22:12:51 +0300 > > > From: Artturi Alm > > > > > > Hi, > > > > > > i came across dwc2 usb

unbound 1.6.5, fix trust anchor tracking

Update to 1.6.5 to fix trust anchor installation when two anchors are present. Concretely, fixes root.key creation if unbound is installed between sep11 and oct11 2017. Only one real file changed, I placed it at the top of the diff, then there's a bunch of autoconf/manpage vomit below. ok?

relayd interrupted transfers

Hi! relay_error() sets se_done even if write buffer is not drained and relay_write() will close the connection if se_done is set I have a case with 76k json payload where relay_error() detects EOF on read socket, sets so_done but write socket is not drained yet and socket is closed before all

Re: vi tilde expands without shell

That will only work for ~/foo and not ~other/bar. Wouldn't it make sense to use glob(3) in argv_exp2()? - todd

Re: partial compat checking w/OF_is_compatible() ?

> Date: Tue, 22 Aug 2017 22:12:51 +0300 > From: Artturi Alm > > Hi, > > i came across dwc2 usb node like this: > > usb@101c { > compatible = "snps,dwc2"; > reg = <0x101c 0x4>; > interrupts = <0x0 0x11

partial compat checking w/OF_is_compatible() ?

Hi, i came across dwc2 usb node like this: usb@101c { compatible = "snps,dwc2"; reg = <0x101c 0x4>; interrupts = <0x0 0x11 0x4>; clocks = <0x2 0x1c9>; clock-names = "otg"; dr_mode

vi tilde expands without shell

Hi, vi(1) currently has support for filename tab expanding, but to expand the tilde '~' character to $HOME it has to vfork/exec ksh. Yikes. 770 vi CALL execve(0x2cf0b7bce80,0x7f7bf960,0x2cef006cc00) 770 vi NAMI "/bin/ksh" 770 ksh RET execve 0 90346 vi RET

cu(1) aborted by pledge on a non-TTY device

Hi, Today while trying to connect to my ERL via serial with cu(1) instead of cuaU0 I used usb0 (yes, I keep forgetting the correct one). But this is actually the first time I typed it wrong and ran cu(1) with it and strangely enough saw it being aborted by pledge saying that ioctl(TIOCEXCL) was

Re: Race against ipsec_in_use

On Tue, Aug 22, 2017 at 01:13:40PM +0200, Martin Pieuchot wrote: > ipsec_in_use is protected by the NET_LOCK(), so it has to be checked > after grabbing it. > > ok? OK bluhm@

Re: Blocking mem alloc & sosetopt()

On Tue, Aug 22, 2017 at 10:13:31AM +0200, Martin Pieuchot wrote: > ok? OK bluhm@ > @@ -1310,7 +1308,6 @@ ip_pcbopts(struct mbuf **pcbopt, struct > return (0); > > bad: > - (void)m_free(m); > return (EINVAL); > } You could replace all goto bad with return (EINVAL).

Race against ipsec_in_use

Diff below fixes a race reported by Hrvoje Popovski. If `ipsec_in_use' is enabled after the softnettq checked for it but before it grabbed the NET_LOCK(), IPsec code will be executed without KERNEL_LOCK(). ipsec_in_use is protected by the NET_LOCK(), so it has to be checked after grabbing it.

Re: Negative array index read (ac97.c)

On Tue, Aug 22, 2017 at 09:33:56AM +0100, Ricardo Mestre wrote: > Updated version for readability ok jsg@ > > Index: ac97.c > === > RCS file: /cvs/src/sys/dev/ic/ac97.c,v > retrieving revision 1.82 > diff -u -p -u -r1.82 ac97.c >

Re: avoid scan_dmesg in armv7 md_installboot

> Date: Tue, 22 Aug 2017 13:04:22 +1000 > From: Jonathan Gray > > If there is whitespace in the pattern sed returns scan_dmesg will sort > all of the words. As hw.product is available on ramdisks use that > instead to avoid the problem. Makes sense to me. ok kettenis@ In the

isakmpd leak on error

By reviewing my last isakmpd(8) diff to fix a use-after-free, hshoexer@ pointed out that if exchange_establish() fails, `arg' is leaked. This is not a new issue. However it generally happens under memory pressure, and when you're under memory pressure leaking memory is not a clever thing to do.

Re: Negative array index read (ac97.c)

Updated version for readability Index: ac97.c === RCS file: /cvs/src/sys/dev/ic/ac97.c,v retrieving revision 1.82 diff -u -p -u -r1.82 ac97.c --- ac97.c 14 Sep 2016 06:12:19 - 1.82 +++ ac97.c 22 Aug 2017 08:31:43

Re: Blocking mem alloc & sosetopt()

On 21/08/17(Mon) 16:21, Martin Pieuchot wrote: > In order to stop allocating multiple mbufs and possibly waiting with > the socket lock held in NFS, I'd like sosetopt() to no longer free the > mbuf it receives. > > This diff makes sosetopt() similar to sogetopt(), the caller is > responsible for

Negative array index read (ac97.c)

Hi tech, Check if cp->dev value is invalid prior to using it as an array index and only then make the assignment si = >source_info[cp->dev]. This is mentioned in Coverity CID 1453243 and 1453334. OK? Index: ac97.c === RCS file:

Re: avoid scan_dmesg in armv7 md_installboot

hi seems to make sense and reads fine but I'm not able to check it before tonight. cheers Robert Am 22. August 2017 05:04:22 MESZ schrieb Jonathan Gray : >If there is whitespace in the pattern sed returns scan_dmesg will sort >all of the words. As hw.product is available on

Re: xenodm.1: Fix Ic markup

On Tue, Aug 22, 2017 at 08:14:31AM +0200, Klemens Nanni wrote: > Missing dot breaks semantic markup. Committed. Thanks. > > Index: man/xenodm.man > === > RCS file: /cvs/xenocara/app/xenodm/man/xenodm.man,v > retrieving revision 1.7

xenodm.1: Fix Ic markup

Missing dot breaks semantic markup. Index: man/xenodm.man === RCS file: /cvs/xenocara/app/xenodm/man/xenodm.man,v retrieving revision 1.7 diff -u -p -r1.7 xenodm.man --- man/xenodm.man 4 Mar 2017 15:48:32 - 1.7 +++