strict alignment fix for gif bridging

2010-10-28 Thread Claudio Jeker
In most cases etherip encapsulation works but it fails horribly when the ethernet header added by ether_output was prepended into a new mbuf. Add magic code to fixup the alignment of the packet so that the IP header starts again on a word boundary. With this my sparc64 is able to vether - bridge

mbuf print function for DDB

2010-10-28 Thread Claudio Jeker
Many network related crashes need some deeper inspection at an mbuf. I got sick and tired of x /x-ing an mbuf and disect the headers. I wrote a simple function that prints all important fields in human readable form. Just call m_print(0xdeadbeef) to print the mbuf at 0xdeadbeef. Example output:

Informative commerciali

2010-10-28 Thread Anna-29
Informazioni commerciali Per la tutela del credito, sia nella fase prae-contrattuale che in quella prae-contenziosa, fornendo accurate informative volte, da un lato alla minimizzazione dei rischi nell'assunzione di nuova clientela, e dall'altro ad ottimizzare la fase del recupero dei crediti,

do not forward 802.1D reserved addresses in bridge(4)

2010-10-28 Thread Paul de Weerd
Currently, our bridge(4) code forwards ethernet frames with destination MAC address 01:80:C2:00:00:0X, except for X=0 (STP BPDUs). This is not allowed according to 802.1D-2004, section 7.12.6: these addresses are reserved (and currently used by eg. LLDP). In turn, this causes issues for ports

libstdc++ diff: using tr1/memory should work with -fno-rtti

2010-10-28 Thread Vladimir Kirillov
Hello, t...@! Currently, the c++ tr1 implementation in OpenBSD tree implicitly uses typeid() in shared_ptr, thus using -fno-rtti is not possible here. It turns out that gcc guys came onto the issue only a year ago: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42019 The diff: Index:

Re: yield in pf_table

2010-10-28 Thread Ted Unangst
On Fri, 29 Oct 2010, Ted Unangst wrote: now that the atomic flag is gone, the yield diff is simpler. once again, the idea is that unbounded (or of unknown bounds) loops in the kernel are bad because you hog the cpu. so be polite and yield from time to time. anybody use tables heavily

Re: yield in pf_table

2010-10-28 Thread Damien Miller
On Fri, 29 Oct 2010, Ted Unangst wrote: On Fri, 29 Oct 2010, Ted Unangst wrote: now that the atomic flag is gone, the yield diff is simpler. once again, the idea is that unbounded (or of unknown bounds) loops in the kernel are bad because you hog the cpu. so be polite and yield from

Se aCERca el verano, tu cuerpo Necesita CER

2010-10-28 Thread Estetica Spa
Your email client cannot read this email. To view it online, please go here: http://my.newsxmail.com/display.php?M=19918804C=971ee1c8133e35124a52eb58fe86f807S=1347L=386N=227 http://my.newsxmail.com/link.php?M=19918804N=1347L=11753F=T PROMOCIONES 2010 QUERES TENER UN CUERPO PERFECTO? PLAN CER