Re: Major Surprise with xdm on 3.7
> I now think that there is a security advantage to using xdm to bring up > KDE - namely, that there is no unprotected console session which can be > hijacked by someone sitting down at the computer, finding the session > from which KDE was started, and putting startkde in the background. > Well, what happens when they reboot the box, go in with boot -s, get the root shell and cause havoc? If you're that concerned about people sitting down in front of your machine to cause problems, then protect it with something physical (e.g. caged rack with lock).
Re: Random lladdr to ifconfig for ath0 - cannot connect
> etc and works fine. When we change the MAC to something random with ifconfig > and then associate, we get nothing. No surprise here. When you change the lladdr you're only changing what's sent as the source ethernet address on ethernet-style frames. This will not affect the 802.11 level station's address. The chip will still be sending frames with the original MAC as the SA, while the ethernet frames will be sent with what you specified to ifconfig. At any rate, any sane AP will drop said packets. If you want to really do this, then you need to change the MAC that's programmed onto the card. This will propagate down to the OBSD generic ethernet code. This differs based on wifi chip make, so you'll have to figure that one out. I'd be willing to bet it's a bit of a pain in the butt. Regards, Justin
Cisco Aironet MPI350
Hello All, I recently purchased a Thinkpad T40 which contains the Cisco Aironet MPI350 minipci adapter. After an afternoon of research and hacking the various aironet driver files, I'm hoping someone can offer some assistance. If you read this far, I'm sure you already know the MPI350 is different from the other Aironet miniPCI cards in that it does DMA in addition to PIO and also advertises has two seperate BARs to be mapped for "mem" and "aux mem". I've used the FreeBSD (primarily) and Linux drivers for reference. So far, I can get the card to attach and can read some basic data from the card. It gets through the generic an_attach without a hitch, an_read_record seems fine so far (e.g. can read the genconfig, capabilities, etc...). Mac address is printed properly, so I'm receiving data from the card. I also printf'd some other things in the capabilities and genconfig structure and verified the results appear valid. The fun comes inside of an_init, where we set the SSID list. This fails due to the check (reply.an_status & AM_CMD_QUAL_MASK) (from the FBSD code). So I'm curious to know if anyone knows anybody who might have access to this documentation and could help answer a few questions as and when they come? For instance, what does checking the an_status against the bits in AN_CMD_QUAL_MASK reveal? Does it indicate the type of failure? I've also seen that some records read with an_read_record have a length mismatch, such that the "record length mismatch" warning is printed. Is this a fatal problem? Is there some difference between how records are passed with the other MiniPCI aironets or have I screwed something up? Also, is pci_mapreg_map and the related functions documented anywhere other than /usr/src/sys? Their implementations seem straightforward, but I'm not sure if there are any underlying subtle semantics I'm missing... Any insight would be most helpful. I know this probably isn't the most relevent thing to post to misc, but I'm hoping I can solicit some assistance from an interested party. - Justin
Re: 3.7CDs arrived today...
> heh, i got my cd's today too. which is awsome. also i finally got > something special. I received my CDs yesterday too. Excellent :). Installed two machines, upgraded one, rock and roll. However, the jewel case is screwed up (2 of three mounts are disintigrated), just like every single time I've ever ordered (since 2.7). I wonder if it's possible to use a higher quality triple jewel case? Oh well... - Justin
Re: Postfix (IBM) license question
> Do you really want to be in the situation where you have to indemnify > someone who has the desire and the resources to sue IBM? Why would you indemnify the person suing you? Furthermore, why make a warranted commercial distribution if you can't live up to your claims? What the OpenBSD project produces is not a comercial distribution since it's given with no warranty whatsoever and you don't have to pay for it. Anyone a lawyer or have a lawyer friend who might help decipher this one? - Justin
Re: Postfix (IBM) license question
> I've been reading that section as meaning "If you sell something with > with Postfix on it, and what you sell ends up in a lawsuit involving > part of the Postfix written by others that you changed, YOU carry the > burden of defending the portion written by others as well as your own > portion of the defense." IANAL also If you read #5 and #6 it should be obvious that there is absolutely no warranty of any kind whatsoever attached to the piece of software as offered by the author and contributors. However, if someone is making a commercial product that guarantees feature-x or some performance metric and the product fails to live up to it's warranted claim, then #4 applies to you. In this case, I read #4 to mean that if you distribute with no warranty or liability claimed (e.g. like all of OpenBSD) #4 does not apply. However, if some schmuck takes the OpenBSD-derived stuff and creates a commercial product including it, warranting some feature, then #4 requires that individual to take full responsibility for all shortcomings of their warranted claims and also has to indemnify upstream developers from getting screwed. The net result isn't really too different from someone taking BSD licensed code written by a 3rd party, and selling it with some commercial terms and guaranteed warranty that they can't meet. I think the only real purpose of #4 is to explicity protect upstream developers. Anyway, that's my interpretation :). While it's probably benign, I can see why such a verbose license is rejected by the OpenBSD team. - Justin
Re: Memory management
> Where can i find information about Memory management in OpenBSD ? > (Virtual memory, swap, etc) http://www.netbsd.org/Documentation/kernel/uvm.html I'm sure many differences exist by now, but this was the starting point for UVM. - Justin