radeondrm(4) display port output number assignation

2013-08-27 Thread Alexis de BRUYN
Hi tech@, On an ATI Radeon HD 7870 display card with 6 mini-display-ports, I have plugged screens on the first 3 DP (identified DP1 - DP2 - DP3 on the bracket). In Xorg.0.log, the connected output are listed as : [14.513] (II) RADEON(0): Output DisplayPort-0 connected [14.513] (II)

Re: openbsd ioctl fix (in6.c)

2013-08-27 Thread Martin Pieuchot
On 22/08/13(Thu) 23:31, Claudio Jeker wrote: On Wed, Aug 21, 2013 at 09:59:56AM -0700, Loganaden Velvindron wrote: I'm not sure if applies to OpenBSD as well, but NetBSD also disallowed SIOCSIFDSTADDR for ioctl. [...] Index: sys/netinet6/in6.c

Re: openbsd ioctl fix (in6.c)

2013-08-27 Thread Loganaden Velvindron
On Tue, Aug 27, 2013 at 10:37:30AM +0200, Martin Pieuchot wrote: On 22/08/13(Thu) 23:31, Claudio Jeker wrote: On Wed, Aug 21, 2013 at 09:59:56AM -0700, Loganaden Velvindron wrote: I'm not sure if applies to OpenBSD as well, but NetBSD also disallowed SIOCSIFDSTADDR for ioctl. [...]

Re: defer routing table updates on link state changes

2013-08-27 Thread Martin Pieuchot
On 26/08/13(Mon) 13:36, Mike Belopuhov wrote: hi, in order to make our life a bit easier and prevent rogue accesses to the routing table from the hardware interrupt context violating all kinds of spl assumptions we would like if_link_state_change that is called by network device drivers in

Re: defer routing table updates on link state changes

2013-08-27 Thread Mike Belopuhov
On 27 August 2013 13:39, Martin Pieuchot mpieuc...@nolizard.org wrote: I think that's the right approach but the current code generating interfaces indexes is too clever from my point of view, it tries to reuse the last index if possible. This could lead to some funny races if we detach and

Split rtinit()

2013-08-27 Thread Martin Pieuchot
So I started to play with the routine table and I'm slowly trying to unify the various code paths to add and delete route entries. The diff below is a first step, it splits rtinit() into rt_add() and rt_delete() there should be no functional change. ok? Index: net/if.c

Remove unused argument from *rtrequest()

2013-08-27 Thread Martin Pieuchot
In order to define a proper API for our routine table, I'd like to turn the struct rt_addrinfo into a private type (ie: only used in route.c and rtsock.c). This type is used by a lost of code in our network stack to add or delete a route but also in the various *rtrequest() functions. However in

Re: Remove unused argument from *rtrequest()

2013-08-27 Thread Kenneth R Westerback
On Tue, Aug 27, 2013 at 03:58:51PM +0200, Martin Pieuchot wrote: In order to define a proper API for our routine table, I'd like to turn the struct rt_addrinfo into a private type (ie: only used in route.c and rtsock.c). This type is used by a lost of code in our network stack to add or

Re: defer routing table updates on link state changes

2013-08-27 Thread Kenneth R Westerback
On Tue, Aug 27, 2013 at 01:54:34PM +0200, Mike Belopuhov wrote: On 27 August 2013 13:39, Martin Pieuchot mpieuc...@nolizard.org wrote: I think that's the right approach but the current code generating interfaces indexes is too clever from my point of view, it tries to reuse the last index

Re: Split rtinit()

2013-08-27 Thread Kenneth R Westerback
On Tue, Aug 27, 2013 at 03:38:49PM +0200, Martin Pieuchot wrote: So I started to play with the routine table and I'm slowly trying to unify the various code paths to add and delete route entries. The diff below is a first step, it splits rtinit() into rt_add() and rt_delete() there should be

Re: Remove unused argument from *rtrequest()

2013-08-27 Thread Mike Belopuhov
On 27 August 2013 15:58, Martin Pieuchot mpieuc...@nolizard.org wrote: In order to define a proper API for our routine table, I'd like to turn the struct rt_addrinfo into a private type (ie: only used in route.c and rtsock.c). This type is used by a lost of code in our network stack to add or