Re: Correct error number in crypto(9)

Michael McConville wrote: > Michael McConville wrote: > > Does this make sense? > > I just realized that the allocation failure checks earlier in the > function return ENOBUFS. This probably makes more sense for the sake of > consistency. The best I can tell, the only use of this function is in

Re: ddb: Do not ask for a symbol table

On Fri, Feb 26, 2016 at 7:33 AM, Martin Pieuchot wrote: > This diff removes the "symbol table" argument of the functions dealing > with symbols. While here keep the definition of ``db_symtab'' private > because we only have one global symbol table. I'm a fairly active ddb

Re: arm: store curcpu pointer in thread id register

On Sat, Feb 20, 2016 at 1:49 PM, Patrick Wildt wrote: > since ARMv6 the coprocessor provides special registers to store software > defined values. Those registers are: > > * TPIDRURW -> kernel RW, user RW > * TPIDRURO -> kernel RW, user RO > * TPIDRPRW -> kernel RW > >

Re: move mnt_maxsymlinklen to struct ufsmount

Sorry, the second message was not supposed to go to the list...

move mnt_maxsymlinklen to struct ufsmount

The concept of differentiating between "short" and "long" symlinks is specific to ufs/, but the relevant variable is contained in struct mount, which is used by all filesystems. I think it would be a worthwile effort to clear this up by moving maxsymlinklen to struct ufsmount instead. Filesystem

Re: move mnt_maxsymlinklen to struct ufsmount

Hello Todd, Stefan, this is the logical continuation of the previous maxsymlinklen diff. At least in my mind it is logical. ;) I have a dream that one day filesystem code will be actually readable and changeable without being a UFS lawyer. I would appreciate if you could have a look at this.

kdump.1: fix description of -e

There is no emulation "bsd". Our default is "native". Index: kdump.1 === RCS file: /cvs/src/usr.bin/kdump/kdump.1,v retrieving revision 1.27 diff -u -p -r1.27 kdump.1 --- kdump.1 2 Oct 2015 19:12:45 - 1.27 +++ kdump.1

Re: properly inject UDP header for udpencap over IPv6

On Fri, Feb 26, 2016 at 09:55:35PM +0100, Patrick Wildt wrote: > bluhm proposed to add M_UDP_CSUM_OUT to the mbuf csum flags. I just > tried his proposal and it looks like the following diff works. Yes, also OK bluhm@ for this. > diff --git sys/netinet/ipsec_output.c sys/netinet/ipsec_output.c >

Re: properly inject UDP header for udpencap over IPv6

On Fri, Feb 26, 2016 at 08:25:02PM +0100, Alexander Bluhm wrote: > On Fri, Feb 26, 2016 at 04:56:49PM +0100, Mike Belopuhov wrote: > > I think your revised diff should be committed. Any objections? > > OK bluhm@ > > > > uh->uh_sum = 0; > > > +#ifdef INET6 > > > + if

Re: properly inject UDP header for udpencap over IPv6

On Fri, Feb 26, 2016 at 04:56:49PM +0100, Mike Belopuhov wrote: > I think your revised diff should be committed. Any objections? OK bluhm@ > > uh->uh_sum = 0; > > +#ifdef INET6 > > + if (tdb->tdb_dst.sa.sa_family == AF_INET6) > > + uh->uh_sum =

Re: rtadvd: remove more dead code

On Tue, Feb 09, 2016 at 02:17:18AM +0100, J??r??mie Courr??ges-Anglas wrote: > > - a few *cnt members of struct rainfo aren't used for anything > - the SIOCGIFPREFIX_IN6 ioctl has been deprecated since June 2002 > - prefix_match() and in6a_site_allrouters are remnants from the > Renumbering

Re: Correct error number in crypto(9)

Michael McConville wrote: > Does this make sense? I just realized that the allocation failure checks earlier in the function return ENOBUFS. This probably makes more sense for the sake of consistency. > Index: sys/crypto/cryptosoft.c >

Correct error number in crypto(9)

Does this make sense? Index: sys/crypto/cryptosoft.c === RCS file: /cvs/src/sys/crypto/cryptosoft.c,v retrieving revision 1.80 diff -u -p -r1.80 cryptosoft.c --- sys/crypto/cryptosoft.c 10 Dec 2015 21:00:51 - 1.80 +++

Re: /32 routes don't stick

On 26/02/2016 16:53, Sevan Janiyan wrote: > in my hostname.if I have > !/sbin/route add -net 192.168.0.1 -netmask 255.255.255.255 -link > -cloning -iface vio0 > !/sbin/route add default -ifa 10.0.0.1 192.168.0.1 > > Which works. To clarify 192.168.0.1 is the address assigned to my interface

Re: /32 routes don't stick

On 26/02/2016 16:38, Matthieu Herrb wrote: > hostname.em0: > > inet 92.224.149.DDD/32 > ! route add 91.224.148.0/32 -link -iface em0 in my hostname.if I have !/sbin/route add -net 192.168.0.1 -netmask 255.255.255.255 -link -cloning -iface vio0 !/sbin/route add default -ifa 10.0.0.1

/32 routes don't stick

Hi, My hosting ISP is using a routing model where every machine has a IP address with a /32 mask and uses a default route specificed at the link level. For Linux machines it means something like this with iproute2: ip link set eth0 up ip addr add 91.224.149.DDD/32 dev eth0 ip route add

Re: libedit audit

Ingo Schwarze said: > I will post a selection of the most important diffs here, but in order > to not overwhelm the list, not all of them. If you like to see all of > them, please tell me privately and you'll be in the loop. FWIW if individual diffs are not going to be reviewed one-by-one in the

Re: diff to /etc/examples/printcap and sysmerge

On 2016/02/26 16:31, Antoine Jacoutot wrote: > On Fri, Feb 26, 2016 at 09:17:16AM -0600, Chris Bennett wrote: > > On Thu, Feb 25, 2016 at 07:56:54PM -0700, Theo de Raadt wrote: > > > Chris, you continue to amaze me. > > > > > > Upon running sysmerge, that will break everyone's setup. > > > > > >

Re: properly inject UDP header for udpencap over IPv6

Hi Patrick, I think your revised diff should be committed. Any objections? On Fri, Feb 26, 2016 at 15:12 +0100, Patrick Wildt wrote: > Bump with updated diff, now that tree is unlocked again. > > Patrick > > diff --git sys/netinet/ipsec_output.c sys/netinet/ipsec_output.c > index

Rename kbd_trap()

I'm currently working on a kernel subsystem to dynamically instrument kernel functions. For that I'm heavily using ddb(4) code. Now I would like to prevent ddb(4) functions to be instrumented or depend on instrumented functions. So I'm slowly making an audit to reduce the amount of external

Re: diff to /etc/examples/printcap and sysmerge

On Fri, Feb 26, 2016 at 09:17:16AM -0600, Chris Bennett wrote: > On Thu, Feb 25, 2016 at 07:56:54PM -0700, Theo de Raadt wrote: > > Chris, you continue to amaze me. > > > > Upon running sysmerge, that will break everyone's setup. > > > > Like, can you try stuff before you send it out? > > > >

ddb: Do not ask for a symbol table

This diff removes the "symbol table" argument of the functions dealing with symbols. While here keep the definition of ``db_symtab'' private because we only have one global symbol table. ok? Index: ddb/db_elf.c === RCS file:

Re: diff to /etc/examples/printcap and sysmerge

On Thu, Feb 25, 2016 at 07:56:54PM -0700, Theo de Raadt wrote: > Chris, you continue to amaze me. > > Upon running sysmerge, that will break everyone's setup. > > Like, can you try stuff before you send it out? > > I'm done. > > > /etc/examples/printcap doesn't match > > #define

Re: arm: store curcpu pointer in thread id register

Bump, now that the tree is unlocked again. Patrick On Sat, Feb 20, 2016 at 10:49:36PM +0100, Patrick Wildt wrote: > Hi, > > since ARMv6 the coprocessor provides special registers to store software > defined values. Those registers are: > > * TPIDRURW -> kernel RW, user RW > * TPIDRURO ->

Re: properly inject UDP header for udpencap over IPv6

Bump with updated diff, now that tree is unlocked again. Patrick diff --git sys/netinet/ipsec_output.c sys/netinet/ipsec_output.c index 91c319f..18a3f57 100644 --- sys/netinet/ipsec_output.c +++ sys/netinet/ipsec_output.c @@ -375,13 +375,30 @@ ipsp_process_done(struct mbuf *m, struct tdb *tdb)

a pool_setipl for every pool_init

i am going to make setting the IPL mandantory for pools. eventually, the IPL will be set in pool_init, but the first step toward that is to actually set it on every pool. i just finished this. it compiles and boots on sparc64, but more eyes and tests would be appreciated. dlg Index:

libedit audit

Hi, i have begun an audit of libedit. I will post a selection of the most important diffs here, but in order to not overwhelm the list, not all of them. If you like to see all of them, please tell me privately and you'll be in the loop. Actually, i'd love it if one or two additional people

Re: ddb: remove support for multiple symbol tables

On 25/02/16(Thu) 11:39, Mike Larkin wrote: > On Thu, Feb 25, 2016 at 03:05:21PM +0100, Martin Pieuchot wrote: > [...] > > - return 0; > > + return db_elf_sym_lookup(_symtab, symstr); > > } > > From what I can tell, db_lookup is only called one time in the tree. > Rather than replace