Re: BIOS Shows 4GB memory but OpenBSD 4.7/amd64 SMP detects only 3 GB
From: Siju George sgeorge...@gmail.com but OpenBSD 4.7/amd64 SMP detects only 3 GB. Is there anything more I should do to get the other 1 GB of RAM recognized by the System? This is normal. Large memory support is not yet included in OpenBSD by default for amd64.
Re: BIOS Shows 4GB memory but OpenBSD 4.7/amd64 SMP detects only 3 GB
From: Siju George sgeorge...@gmail.com On Thu, May 27, 2010 at 3:22 PM, Peter Kay (Syllopsium) syllops...@syllopsium.com wrote: From: Siju George sgeorge...@gmail.com but OpenBSD 4.7/amd64 SMP detects only 3 GB. Is there anything more I should do to get the other 1 GB of RAM recognized by the System? This is normal. Large memory support is not yet included in OpenBSD by default for amd64. Is there anything I can do to get this by recompiling the kernel or something? Yes and no; google for 'bigmem' in the misc list for marc.info. You have to edit a source file and recompile the kernel. However, it won't work (to be precise, it will probably crash on boot, or possibly afterwards) unless you have an IOMMU, and most Intel systems don't. OpenBSD doesn't support the AGP/PCI-e GART as an IOMMU, and I'm not sure if it supports VT-d platforms (which you probably aren't running anyway). The only option here is AMD. Until the devs tell us it's working, it's not worth persevering with. PK
Re: Help contacting Richard Stallman
From: Julian Acosta j.acost...@gmail.com Hello! I'm from the Postgraduate Departmen of the ITCC University from Mexico, Really we need to contact with Richard Stallman, just for give us his opinion and answer us some questions about free software, How can I contact him? What's his real email? You'd be better contacting the FSF rather than Stallman directly - don't you think that's overkill? He also may have conducted just one or two interviews and written a couple of articles - just google. Bear in mind that their favoured GPL 'free' software license is not free. It is effectively free as in beer, but not as in free speech[1]. Their definition includes being forced to give away source code, which whilst I understand the viewpoint (of increasing free code), is by any measure a restriction of your freedom. BSD licenses, on the other hand, do not restrict what you can do, although it's good karma to contribute back when using a large amount of free code from others. [1] The GPL allows products to be sold, but seeing as this must include source code, after one sale it only needs someone with a compiler to distribute it freely (as in beer). Peter
Re: Source Overview
From: J.C. Roberts list-...@designtools.org The developers *CONSTANTLY* *ASK* *FOR* *YOUR* *HELP* with testing, but this dull and heavy work is somehow below most people who just talk about wanting to become developers and are looking for shortcuts to becoming one. Since validity is critical, if you cannot test properly and hopefully help in the debugging, then you'll never be any good at writing code. You're not wrong, but that's a rather black and white way of looking at the world. When someone starts a new activity - whether that's coding for OpenBSD, baking cakes or similar, it's usually necessary to have a visible 'quick win' or at least sign of progress that encourages the person to carry on and try a little harder. Testing does not usually fit into that category - it is indeed 'dull and heavy' and usually something people expect to be paid for. I understand and mostly agree with the viewpoint that the best way is to download code, decide on what needs fixing and keep plugging at it until success is achieved. That's also fine if the OpenBSD community wants to perpetuate the type of people that code for it and the size of the community. If (and it is an if) the OpenBSD community wants more resource - both coding and testing, there probably needs to be a degree more flexibility. Or, in short, we need to not deter people straight away, and accept that perhaps sometimes decent programmers start from ones that make lots of mistakes. Perhaps a ports TODO similar to the NetBSD ports TODO might help; it doesn't require quite the same level of kernel or userspace hacking and provides very visible feedback and thanks once completed. Neither would I completely rule out a central TODO list linked off OpenBSD.org. Sure, it might well be ignored, but the possibility remains that someone might take up the task. NetBSD isn't doing too badly with Google's Summer of Code initiatives, either. It might not even be a bad idea to puff up new developers a bit : 'new developer Fred Bloggs decided to solve PR7738 squashing an annoying bug in the ipz(4) driver. John Smith is very grateful for this as it enabled him to use his new ServBladePro NZ20 server' With specific reference to the ISA 486, if there are specific test cases that can be run without taking up hours of interactive time, I have a suitable VLB/ISA 486 that could run them. It's not something I'm interested in using on a regular basis though - I've got other machines that are far easier to work with. PK
Multibooting (was : OpenBSD culture)
OpenBSD does not require a primary partition, nor does NetBSD. Solaris does for the moment, although code to fix that has been committed. I have a Windows 7 x64, OpenBSD, Solaris, NetBSD multiboot. It's not that difficult to arrange. I did most of the partitioning in Windows, setting up a primary partition for Solaris, then logical partitions for OpenBSD and NetBSD. Either the NetBSD or OpenBSD media can then be used to edit the partition types to the recognised ones. Install as normal, then use EasyBCD to edit the Vista/Windows 7 boot menu - modify as appropriate if you're using grub etc or XP..
Re: Multibooting (was : OpenBSD culture)
From: Brad Tilley b...@16systems.com as appropriate if you're using grub etc or XP.. Another Option. Assuming a i386 or amd64 PC: 1. Put another hard drive into the computer. 2. Go into the BIOS and make the new hard drive have higher priority. 3. Boot the computer and install OpenBSD onto the new hard drive (Run dmesg to be sure you're doing the right thing) 4. When you want to go back into the other OS, change the drive priority in the BIOS and reboot. Not pretty, but it works and keeps drives separate and no fooling with grub, partitions, Windows boot loader, etc. If you're going to take /that/ approach, I would suggest a trayless SATA caddy from someone like Icy Dock (be careful - some of their products are garbage, but I can attest that the trayless, fanless SATA caddy is not). You can easily swap the drives in and out without faffing with BIOSes. I use precisely that method for swapping in test systems.
Re: OpenBSD culture?
From: Siju George sgeorge...@gmail.com On Mon, Apr 19, 2010 at 2:59 AM, shweg...@gmail.com wrote: Of course I boot using the Vista bootloader and easybcd to edit the configuration, which saves a lot of headache. The important thing is it can be done. :) How Do you Tell the OpenBSD Installer to install to a logical partition? Same as to install to a primary partition - make sure the partition exists and is of type A6, don't tell the installer you want to use the entire disk and it'll handle most of the rest.
Re: Testing bigmem properly on amd64?
From: Henning Brauer lists-open...@bsws.de * Peter syllops...@syllopsium.com [2010-04-15 03:27]: I know bigmem is still in a state of flux and can be enabled by editing machdep.c and compiling a custom kernel. What's the best way to test and report this? none. bigmem is known broken, otherwise it would be enabled by default. and tests proving the fact, well, you could test that water is wet, too. Fair enough. I'll wait until it's deemed testable again, then. PK
Re: 4.6 patch support
From: Andreas Gerdd kryptos...@gmail.com when 4.8 comes out (a year after 4.6 came out) support for 4.6 will stop. Quite short time. Perhaps, but it /is/ free. There are undoubtedly some people who will backport fixes to earlier versions if you paid them. Our advise is to upgrade to a newer version and plan for that now. It's not magic, in fact it is pretty easy in almost all cases. It is not magic, but it is more than magic if you have only remote ssh access and nothing else. :-( You have multiple options, there's : http://www.openbsd.org/faq/upgrade46.html Which perhaps looks a little scary, but does work. Alternatively try YAIFO http://sourceforge.net/projects/yaifo/ for an ssh enabled install kernel. Of course, you should test both these options on a local machine before attempting it remotely.. PK
Re: OpenBSD Volunteer needed today in Los Angeles - Solved!
On 2/22/2010 9:23 AM, Bret S. Lambert wrote: Unless some benefactor is willing to come forward and deal with the logistical headache of doing the paperwork and keeping it all as up to date as it needs to be, it's not going to happen, even if getting an EAL meant ponies, rainbows, and money trees for everybody. To be severely offtopic, if getting an EAL genuinely meant ponies, rainbows and money trees I'd be quitting my job and working on it right now.. I doubt I'm alone in that What other motivation could you possibly want? Moon on an unobtanium stick? ;) PK
Re: MFM disk geometry
From: Daniel Malament b...@anonix.net Subject: Re: MFM disk geometry Try looking for Total Hardware '99 - your controller might be documented in there. Nice! Thanks. http://th99.dyndns.org/c/C-D/20069.htm Unfortunately, it doesn't look like it's actually all that configurable. Although I don't know what some of those settings actually mean. Does anyone else see ways this can help, or care to explain the settings? :) I think my first course of action would be to use DOS, or possibly OS/2, to override the disk geometry, unless the disk has data on it that can only be accessed from OpenBSD. Yes, I know it's intellectually more fun to get OpenBSD to do it, but for a one off with little practical future use I think I'd use something else. DOS, OS/2 and OpenBSD can of course all be booted from floppy, thus avoiding any early initialisation nastiness. I'm curious as to exactly what the difference between 'driven to interface' and 'intercepted by controller' is. Is this a standard interface vs INT13 thing? I wouldn't call a P3 system with ISA particularly unusual - by the time the P4 came out they were rare (although, readily available especially if needed for industrial applications). PK
Re: MFM disk geometry
From: Daniel Malament b...@anonix.net To: Peter Kay (Syllopsium) syllops...@syllopsium.com I think my first course of action would be to use DOS, or possibly OS/2, to override the disk geometry, unless the disk has data on it that can only be accessed from OpenBSD. Yes, I know it's intellectually more fun to get OpenBSD to do it, but for a one off with little practical future use I think I'd use something else. DOS, OS/2 and OpenBSD can of course all be booted from floppy, thus avoiding any early initialisation nastiness. I'm not sure what you're describing here. Also, accessing the data from DOS still leaves the problem of moving it. Or perhaps I didn't make it sufficiently clear that the goal was to copy the data off the drive... I'm just saying that some operating systems more of the era that the drive was created in are able to specify (override) disk geometries. Obviously moving the data off the drive requires another hard drive, floppies, a network etc exactly the same as under OpenBSD. Alternatively, can disktab be used? The documentation is not entirely transparent on this, but it does appear that disktab might be able to override BIOS parameters. PK
Re: OpenBSD in VirtualBox 3.1.x on non-SMP machine
From: Bayard Bell buf...@mac.com Requires VT-x or AMD-V hardware virtualization support. It would appear they've therefore made VT-x and friends non- configurable. You can file a bug report and see where that goes. Would it be too cynical to suggest using a product which doesn't suck? It really is a pity as I appreciated the work Innotek did for the OS/2 community. The one product that gets sold off to Sun and becomes quite popular also turns out to be the one which continues to be mediocre. Impressive, really, as I thought VirtualBox was derived from Qemu, and OpenBSD works just fine under Qemu on both x86 and sparc platforms, at least. PK
Re: installboot: broken MBR
OpenBSD works just fine in an extended partition, even if the documentation says it requires a primary partition - at least on amd64. However, I seem to remember convention is that extended partitions should be at the end of the disk. In theory this probably shouldn't matter, provided the OS/boot manager is properly written.. I'm using a system that's multibooting Windows 7, Solaris, OpenBSD and NetBSD. All happily co-exist. PK - Original Message - From: Nick Holland n...@holland-consulting.net To: T. Tofus von Blisstein tuffst...@googlemail.com Cc: misc@openbsd.org Sent: Wednesday, January 06, 2010 2:23 PM Subject: Re: installboot: broken MBR Nick Holland wrote: T. Tofus von Blisstein wrote: Hello, I have linux and openbsd installed on a single drive. Linuxy fdisk shows Device Boot Start End Blocks Id System /dev/sda1 1 24017 1929165215 Extended Partition 1 does not end on cylinder boundary. /dev/sda2 * 24018 3040151279480 a6 OpenBSD OpenBSD is in an extended partition...don't know that this works in all (or any) cases, and the fact that it doesn't work in yours doesn't surprise me. ok, I have NO idea what I was seeing here, that's not an extended partition at all. I would have swore that said sda5... *sigh* Nick.
Re: decreasing the size of the distribution
From: Jan Stary h...@stare.cz On Oct 26 00:10:20, Abdullah Sendul wrote: Hi, we are having a couple of openbsd servers, of which, the content is static. I would like to identify all the files needed for this system to run, and then move it to a flash disk to minimise the size of the distribution You can easily move it to a flash disk without minimizing anything. I would tend to agree. OpenBSD is a small operating system and any reasonably sized flash storage will easily contain it. Why on earth make life more difficult for yourself - as soon as you cut the OS down it becomes harder to support! I have a NetBSD system using an IDE-Compact Flash converter and simply use it as normal (at the time, OpenBSD didn't support the hardware fully). I do use flashdist on an OpenBSD firewall, with an extremely cut down subset of OpenBSD. The reason for this is not distribution size, but that the OS remains unchanged every time it boots up (flashdist uses the storage device read only) and the possibility of corruption is removed. Flashdist works well, but from experience I can say it's a bit of a pain in the arse to get working as it's always missing something. The first time I booted it up, I found it didn't include the firmware for my (wired) ethernet card by default, for instance. Save yourself the hassle and just install as normal. PK
Re: 4.6 arriving
From: Lukas Ratajski l.rataj...@h-s-l.de On 09.10.2009, at 08:30, patrick keshishian wrote: arrived in burbank, ca (usa) today. thank you all! tiny little puffy shrine: http://sidster.com/gallery/misc/2009/obsd46-32-21-mugs.jpg Oh man, I'd LOVE to give the 2.1 version a boot opportunity on i386. Just for the sake of curiosity. Anyone offering a copy? Given that 2.1 is just a *tiny* bit pricey, might I suggest : 1) http://ftp.eu.openbsd.org/pub/OpenBSD/2.1/i386/ 2) A donation PK
Re: VirtualBox2.2+OpenBSD4.4 (fail)
Hi guys, (Pardon since lot of people use *BSD and Linux together but if rude, I'll take it off-list) Ok. It installs fine. However, I keep getting segfaults on simple programs (such as xorgconfig). (I don't have exact text/dmesg to dump right now but I can produce it if required) Is it that VirtualBox isn't emulating x86 hardware properly? Or, is it a bug in obsd? (I am thinking the former). Any Ideas/suggestion are entertained (Trying in VMware right now) It's VirtualBox - looks like it's still crap. VMWare works fine, so does qemu. PK
Re: VirtualBox2.2+OpenBSD4.4 (fail)
From: L. V. Lammert l...@omnitec.net On Wed, 7 Oct 2009, Peter Kay - Syllopsium wrote: It's VirtualBox - looks like it's still crap. VMWare works fine, so does qemu. As does VirtualBox with proper hardware support (AMD64 Socket AM2), .. though we do not use X on VMx. Are you seriously saying VirtualBox is a viable option by specifying one hardware platform? I don't see why AMD64 really helps, or AM2 for that matter - some AMD platforms have iommu but that shouldn't be relevant either. Qemu runs OpenBSD without VT. So does VMWare server (although that will transparently switch it on, at least on VMWare server x64) OpenBSD wasn't the only OS VirtualBox had problems with last time I tried, either. VMWare, Qemu and VirtualPC all worked flawlessly. VT may help, but the VMM should run without it. PK
Re: OT: Laptop advice. SSD costs.
From: Henning Brauer lists-open...@bsws.de * Robert rob...@openbsd.pap.st [2009-09-17 16:34]: On Thu, 17 Sep 2009 11:16:58 +0200 Henning Brauer lists-open...@bsws.de wrote: * Aaron Mason simplersolut...@gmail.com [2009-09-17 03:52]: Would these drives by any chance be similar to the 1.8 ZIF drives used in (*shudder*) 5th gen iPods? I have one from my iPod that died a horrible death and it still runs fine. The drive's a 30 gigabyte Toshiba drive, and after the iPod it was used on a PATA to 1.8 ZIF converter and used on a PC running OpenBSD. If this suits your needs, I'm happy to donate it if you can cover postage from Australia. x40 do not use 1.8 zif drives but 1.8 drives with the regular 2.5 pata interface. which is exactly the issue - nothing else does. The iPod might not have suitable drives for the X41, but some other MP3 players do. Possibly the iRiver series, but my memory fails me. I recommend browsing http://forum.thinkpads.com/ for more information. They can give you chapter and verse on both the drives and compact flash options. Just how hard is hacking ACPI to get the X60/X61 working, anyway? Currently I'm trying to fix an old X driver (on an AMD Geode based system : it just turns the screen black) but perhaps I should help to get other bits of hardware I own working. PK
Re: Defending OpenBSD Performance
From: L. V. Lammert l...@omnitec.net On Wed, 16 Sep 2009, Henning Brauer wrote: Building from source is light years more difficult than 'apt-get update apt-get upgrade, or 'yum upgrade' or the like. so don't fucking do it, use releases and packages. *OR* learn how to use environment variables and set your PKG_PATH create an alias for pkg_find and get equivalent functionality. That doesn't help if you're running OpenBSD-STABLE. Updating packages is easy. Snapshots really aren't an option; OpenBSD is a good firewall and networking option but selling the concept of snapshots to management is less than trivial. The example of the BIND fix is a good one. On a server which hadn't built STABLE before it was a bit of a faff to sort out, especially as IIRC the fix wasn't available in CVS until some time after the advisory had been sent. I'll grant that the patch was available direct on the OpenBSD website very quickly, though. At the risk of a flaming, sysmerge is also a pain in the arse. Once you know how to use patch files and diff properly I'm sure it is absolutely wonderful, but it also copes badly with files that have not changed in any significant way. Which is not to say that the enhancements in 4.6's install and beyond are not welcome, of course. Oh, and to belay the predicatable 'well, why don't you fix it' response - I am looking at fixing things, just not the above. PK
Re: Laptop advice. SSD costs.
From: Edd Barrett vex...@gmail.com I have located someone willing to sell me an X41 tablet at a very affordable price, however this is the model with the sucky hitachi disk [..snip..] a) Is there anywhere you can get SSD's for cheaper than 100GBP. I only really need 60GB or so. b) Any other comments? Unless the price is very, very good I wouldn't bother with the X41. If it's not out of warranty it will be so within a month or two. Go for an X60 or X61 instead. With an X61 you'll get a Core2Duo, VT support, 8GB memory support and an SATA disk. Plus it's easily possible to find one still in warranty until 2010 onwards. Only an idiot buys a laptop without a warranty, except when it's staggeringly cheap. If you must persist with the X41, buy a big compact flash and IDE-CF adaptor. It will be faster than the horrid 1.8 PATA hard drive. There are articles on this online. I went down this route last December. The choice was an X31/X32 at 100GBP or less (no warranty), an X41 at 200GBPish or an X60 or X61 from 400 upwards. My rationale was to go for broke. At 200GBPish, you can afford a new netbook which will be as fast (but not as well built) as the X41 and have a warranty. OK, so 400GBP is quite a jump, but for that you can find an X61 on ebay with a warranty and modern support. The only thing that sucks about it is the X3100 graphics, but you're buying a sub notebook, not a gaming laptop. PK
Re: Laptop advice. SSD costs.
- Original Message - From: David Vasek va...@fido.cz It could make sense. However, you won't have working suspend/resume with OpenBSD yet and will have to fight with ACPI and its possible problems. Not speaking about hibernation, which X4x laptops have. Also, a keyboard without Microsoft keys is much more comfortable for some. Each has its own advantages. I'll grant you that suspend is not supposed to be working yet (I haven't actually tried). It will work, eventually, though. Everything else is, as far as I'm aware, absolutely fine. Installation of OpenBSD on an X61 is a breeze, at least for amd64 (That's another advantage, of course : X31/X41 is 32 bit only). The only issue I've found so far is that VLC 0.8.6 is considerably less usable than under Vista x64; the sound lags the video to an unusable degree. I've not yet compiled VLC 1.0 to find out whether the issue is VLC (not unlikely), the sound driver (possible) or X (very unlikely IMO). Accelerated X is fine, and OpenGL on the X3100 appears to be fully functional too. PK
Re: VHS transfer on OpenBSD
From: Jan Stary h...@stare.cz I need to transfer some old VHS tapes into (any) digital video format. On OpenBSD of course. I understand I need (1) a VCR, obviously, to play those tapes (2) a TV card that can input what the VCR outputs (3) a piece of software that can capture the input Before I start shopping 80's style, does anyone have some general DO/DONT advice about (2) and (3), or even a complete working solution? I wouldn't start from there. Video capture isn't exactly my area, but my solution would be : If it's only a couple of tapes - get someone else to do it. If it's old episodes of TV programmes - buy it on DVD/Bluray instead. Otherwise : 1) Find decent hardware (not TV cards) that can capture compressed video in real time (2nd hand ebay may help). 2) Find software that uses that hardware 3) Use the operating system that software runs on. This may include using Windows. I can rarely be arsed now to make things work in one particular OS for a short term one off job, if it's going to be tricky. In particular, my (limited) experience is that video capture on TV cards is A Bit Shit, and capturing uncompressed video is not fun, even if modern hardware is probably adequate to handle it. Also carefully note the limitation of the capture card/software. My ancient SGI O2 will happily capture full frame PAL in real time despite being rather slow. However - it'll only do that for certain types of compression and certain frame sizes - beyond that the custom hardware can't help. Likewise, many of these TV cards that feature 'real time video capture' do so at reduced image sizes and capture speeds. PK
Re: Open Vs Free BSD
From: Anton Parol anton.pa...@sun.com OBSD is the best choice of OS for people who like violent little fish mascots. And it has blue-boot-console-thingy (tm) . Ace. I wasn't going to contribute to this thread, but I have to ask. *What* blue-boot-console-thingy? I'm not sure it's sensible to do direct comparisons of NetBSD/OpenBSD/FreeBSD. At first glance the maxim of 'OpenBSD is more secure, NetBSD is more portable, FreeBSD has better support' applies, however OpenBSD has some platform support that NetBSD does not and NetBSD, for a Unix that on first glance looks a bit hardcore, has quite a large amount of functionality and seems quite willing to experiment with new features. FreeBSD is sometimes promoted as being mostly i386 based, but supports a number of platforms. There's the cultural and administrative differences, too. Custom kernels are pretty much anathema to OpenBSD, encouraged on NetBSD and generally handled by modules on FreeBSD.. PK
Re: Even and Odd numbered OpenBSD versions
From: Mark Romer romes...@gmail.com Hello, just a simple question. We have here at work a old hand at openbsd and he says he only uses openbsd versions that are even numbered. (3.8, 4.0, 4.2, 4.4 etc...) I am not sure why, did not have a chance to ask him. I believe that you should use the latest version available, but what does everything else think? Perhaps he's watching the shit (odd numbered) Star Trek films in between upgrades? There's certainly no automatic suckage of odd numbered OpenBSD releases. (also, various even numbered Star Trek films sucked too). PK
Re: Multiboot OpenBSD with Vista
From: MANI mm.m...@gmail.com Subject: Multiboot OpenBSD with Vista First of all you need to know I am running OpenBSD on my laptop and PC at home happily as sole OS, but unfortunately I need to dual boot my PC at Office because of some proprietary softwares we need at company, the other OS is Microsucks Windows Vista The easiest solution by far is to install Vista first but leave space on the hard disk. Run the OpenBSD install and adjust the partition values so that OpenBSD is using the space instead of 'all the disk'. Once installed, use bcdedit to add an entry for OpenBSD pointing at the partition. It 'just works'. Personally I don't even dislike Vista that much, provided it's given plenty (32 bit : 2GB, 64 bit : 4GB) of RAM. I favour 64bit, even if the driver support is less comprehensive and the memory requirements are higher. PK
Re: Shared IRQ
From: Henry Sieff henry.si...@gmail.com To: Joco Salvatti salva...@gmail.com http://www.openbsd.org/faq/faq12.html 12.7.3 2009/5/14 Joco Salvatti salva...@gmail.com: Hi, I would like to know if a different hardware can shared the same IRQ with another? 12.7.3 is accurate, however there is a difference between 'can it' 'should it' and 'will it' 'should it?' - yes, it should 'can it?' - yes, it can 'will it?' - that's the tricky one. Some devices just don't share interrupts well. Perhaps it's shit hardware, a shit APIC, crappy BIOS, naff driver - whatever. PCI devices can theoretically share interrupts, but that doesn't necessarily mean they will. PK
Re: OT: 10GbE Physical Network Taps
From: J.C. Roberts list-...@designtools.org To: Johan Fredin jo...@spelaroll.se On 09-05-07 05.00, J.C. Roberts wrote: If anyone here mistakenly thinks they can actually run *ANALYSIS* at these speeds with off the shelf components... BAWAHAHAHAHAHAHAHA! Well, depends on what you mean by off the shelf. Procera Networks is doing layer 7 analysis at 40Gbps FD with their PacketLogic PL10k. The hardware used for this is sourced from companies that anyone can by hardware from as far as I know. Of course it's not x86 stuff, but it's off the shelf. :) This is really rather getting off topic, but I would suggest that 'off the shelf' only applies when there are many well known shelves where the kit may easily be obtained, preferably with multiple implementations of the hardware. If you can't drive to a random three decent suppliers and find it in one of them, it is not 'off the shelf'. If the kit can be obtained from a restricted set of sources and features highly up to date technology, yet basically only requires money and a phone call to start the process it is 'leading edge' (if it features old technology it is now over the hill and is 'legacy') If you're calling a company to source FPGA/DSPs or to contract someone to make it for you, you're now into the 'bleeding edge' PK
Re: USB-PS2 converter with KVM?
From: J.C. Roberts list-...@designtools.org Subject: USB-PS2 converter with KVM? I'm attempting to use a USB-to-PS2 converter and running the PS2 through a Belkin KVM. The converts I bought seem to be old USB 1.1 stuff, and they don't play very well with any OS. [..snip..] Can anyone suggest a brand (and model) for a good quality USB-PS2 converter that plays well with Belkin KVM's? Newlink USB-PS/2 convertors are quite good. However, I suspect your problem is the Belkin - not the convertor. My SGI O2 boxes *really* don't like the Belkin Omnicube I still have somewhere, and my pentium OS/2/DOS box won't see the mouse unless the KVM is switched to that box on bootup.. Solution : ebay! I got a very nice Compaq (rebadged Avocent, IIRC) eight way KVM for about thirty quid PK
Re: Build a custom kernel to installation
From: Ricardo Augusto de Souza ricardo.so...@cmtsp.com.br Could you please tell me the steps I must follow? Is it possible enable it at boot -c? I don't wish to be rude, but you're not reading what people are telling you : aac is not enabled in snapshots. you still need to build your own on another machine. It is not possible to enable it using boot -c, as aac support is deliberately disabled due to poor documentation from adaptec. Look at the documentation on how to create a custom kernel. Your best option is to create a custom ramdisk kernel, which you can then netboot off and install over the network, otherwise you're going to have to build your own cd image which is a bit more involved but still not that difficult. If you do do the ramdisk based install, remember to either patch a custom (non ramdisk - i.e. normal) kernel into /mnt following the install, or download all the sets to a local FTP/HTTP/'whatever install method' site and replace bsd/bsd.mp with an aac enabled kernel. Be certain to update your RAID controller firmware to the latest version. I have a Dell PE2400 (Perc2/Si) working fine with an aac enabled kernel here, although I might be a bit dubious of running it on a production basis. PK
Re: Dual-head OpenBSD 4.5 and NVIDIA GeForce 7300 GT
From: Kamil Monticolo k.montic...@gadu-gadu.pl To: misc@openbsd.org Sent: Friday, March 27, 2009 10:38 AM Subject: Dual-head OpenBSD 4.5 and NVIDIA GeForce 7300 GT Hi folks. I have 4.5 GENERIC.MP on my machine and I'd like to have it dual-head. My dmesg is at the bottom, if you want more information or output, I'll give that. I tried to setup new xorg.conf with two screens but without any luck. I also googled for any idea, but it also fails. It works on linux with modified config, but I don't want it on my machines. Every idea or question is welcome. Thanks. I can tell you it works fine on a 7600GT - I only needed to do X -configure. Don't even think I had to explicitly tell it to use xrandr. This was on 4.4/amd64 I do actually have two 7600GTs on the box I'm using; the second isn't recognised automatically though - so if you're looking to use more than two monitors it may be a bit more tricky. PK
Re: Dual-head OpenBSD 4.5 and NVIDIA GeForce 7300 GT
From: Kamil Monticolo k.montic...@gadu-gadu.pl To: Peter Kay - Syllopsium syllops...@syllopsium.com On Fri, 27 Mar 2009 12:29:21 - Peter Kay - Syllopsium syllops...@syllopsium.com wrote: I can tell you it works fine on a 7600GT - I only needed to do X -configure. Don't even think I had to explicitly tell it to use xrandr. This was on 4.4/amd64 I do actually have two 7600GTs on the box I'm using; the second isn't recognised automatically though - so if you're looking to use more than two monitors it may be a bit more tricky. Sounds great:) Are you running X11 out of the box on -CURRENT? Maybe I misconfigured something a bit, but I don't know what . I'm looking for a solution for two screens only. As per my first message - 4.4/amd64. Can't remember if it was a snapshot or not. Literally all I needed to do was X -configure, if I remember correctly. I'll try with an up to date 4.4 snapshot at some point. PK
Re: Dual-head OpenBSD 4.5 and NVIDIA GeForce 7300 GT
From: Kamil Monticolo k.montic...@gadu-gadu.pl On Fri, 27 Mar 2009 12:57:10 - Peter Kay - Syllopsium syllops...@syllopsium.com wrote: Literally all I needed to do was X -configure, if I remember correctly. I'll try with an up to date 4.4 snapshot at some point. Peter, can You post your configuration here? I confgured X server as you did, but it didn't work for me. Thank you. Sure. I may even get around to doing multi adaptor too (I have three monitors connected to the system). May be Sunday before I can get it sorted, but I'll post to the list. PK
Re: Ideas for Getting MATLAB/Mathematica to utilize sparc64 ram that runs openbsd
On Wed, 25 Mar 2009 02:09:08 -0700 Vivek Ayer vivek.a...@gmail.com wrote: Hi guys, I realize openbsd/sparc64 is probably the best port of any OS to the sparc64 architecture, however I work in an environment where matlab/mathematica are greatly needed. I know openbsd/i386 has linux binary emulation, which would do the trick, but I want to use these 2 awesome Sun blade machines that have loads of RAM waiting to be used (8 GB in all just dying to be used). Do you all have any suggestions on how I can matlab/mathematica to possibly use this much memory even though they probably won't be on those systems? Is there some memcached-like solution for this. I have a linux computer on the network which has matlab/mathematica installed. Is there anyway I can generically donate memory from openbsd to linux to make linux think it more? Or some VM solution? I don't know. There is the obvious solution of putting 8GB of memory in the x86 box - even several year old x86 chipsets now support up to 8GB+. I don't know your personal situation - perhaps you're in an educational establishment which has cheap matlab licenses, but can't easily buy hardware. Still, if your place of work can afford 2500$ on a license, it can probably also afford 8GB RAM if you can justify your need for it. Memory can be quite cheap these days. Don't make life difficult for yourself if you can avoid it :). PK
Re: HP Proliant DL385 with Squid at a Gigabit-switch - bad network performance
Whilst I can't comment on the foibles of modern Cisco switches, I can certainly say that Cisco switches I've used somewhat more recently than fifteen years ago (but more than five) refused to autonegotiate to some servers. So far they remain the only switches I've had to manually set the speed and duplex on, AFAICR. Given that experience, I can certainly see why people might enforce a configuration for longer than might be necessary. PK - Original Message - From: Michal mic...@sharescope.co.uk To: misc@openbsd.org Sent: Monday, March 16, 2009 1:51 PM Subject: Re: HP Proliant DL385 with Squid at a Gigabit-switch - bad network performance Sorry but I worked for a very successful company in the UK that didn't use auto neg's on Cisco switches and routers so I wouldn't call it evil AT all, please explain why manual is evil. C -Original Message- From: owner-m...@openbsd.org [mailto:owner-m...@openbsd.org] On Behalf Of Henning Brauer Sent: 16 March 2009 13:29 To: OpenBSD Subject: Re: HP Proliant DL385 with Squid at a Gigabit-switch - bad network performance * Laurent CARON lca...@unix-scripts.info [2009-02-28 21:33]: Steve Shockley wrote: On 2/27/2009 8:43 AM, Laurent CARON wrote: - Forcing speed on switch - Forcing speed on nic Why? This practice made sense when 10baseT gear from different vendors wasn't compatible, but not for the last 15-20 years. This practice still makes sense, at least with broadcom cards. no, it is pure bullshit and the source of many many many errors. just because cisco failed miserably in implementing autoneg for years. even they managed now. so stop spreading this bullshit. autoneg is good. manual is evil. that simple. I always do force the speed on servers. this is extremely stupid. -- Henning Brauer, h...@bsws.de, henn...@openbsd.org BS Web Services, http://bsws.de Full-Service ISP - Secure Hosting, Mail and DNS Services Dedicated Servers, Rootservers, Application Hosting - Hamburg Amsterdam
Re: nfs proxies
From: Jason Dixon ja...@dixongroup.net People always say this but never mean it. I have proof, and I'm sure Theo does too. You have no idea how much a real feature costs to implement. When you present them with the costs they always balk. To be fair, some of them do mean it, but just don't understand the cost of development. Others don't think about it and are caught up in zealotry. Some /think/ they're happy to pay, but when they actually have to put cash on a table suddenly realise they don't want it that badly. Not that this helps a great deal when the end result is almost the same. Peter
Re: Port ZFS to OpenBSD
From: bofh goodb...@gmail.com I have to ask - if you're not copying the code, but only copying the concept/technical requirements over (ie, a rewrite), that new code would be bsd licensed, right? Probably, but this is filesystem code. The last thing you want to do is to replace complex, generally well debugged code with new complex not so well debugged code! That applies for all software, but doubly so for code that can corrupt data and crash the entire system. PK
Re: Best supported arch/workstation
From: Matt KP60 mattk...@gmail.com Hi, I am looking at purchasing a workstation to put OpenBSD on for programming development. What is the best supported arch overall? Or even better what is your most recommended workstation for running OpenBSD? i386, without any doubt whatsoever (imo). amd64 is almost as well supported. I've never personally had any issues with the sparc port (32 bit), but all those machines are old and slow. sgi could frankly do with a bit of work - but hey, it *is* the only commonly available full 64 bit OS available for the O2. PK
Re: laptop choice
IBM's X series are 12.1 only which is outside your 15-17 suggestion. If you want a cheap X series laptop preferably go for the X32 - it's faster and newer than the X31 and has a 2.5 hard drive. The X40/X41 are very OpenBSD compatible but have a 1.8 hard drive (slow, expensive, difficult to replace).. You might find an X32 with at outstanding warranty if you're lucky. Otherwise go for a T series. I wouldn't bother going for an X series unless you need a really portable laptop with lots of battery life. OK, so it's cool, but you get more out of an equivalently priced T series including a much better graphics chipset. If not IBM/Lenovo, I also tend to favour Toshiba laptops, or possibly HP. I presume you've read http://www.openbsd.org/i386-laptop.html and http://jcs.org/laptops/ I think your primary concern should be the warranty. Either get something very cheap with no warranty, or something more expensive with a warranty (which can then be renewed with IBM), plus obviously insurance.. Buying something inbetween is stupid. Personally I've just gone rather over the top and ordered a second hand X61 from ebay (assuming it arrives!). Core2Duo (64 bit, supports VT and NX) and a Santa Rosa chipset (up to 8GB, with painfully priced 4GB modules, should I ever need it) should mean I won't need to change it for a *long* time (hopefully!). PK
Re: laptop choice
From: Mihai Popescu B.S. mihai...@gmail.com I don't need dual core since the support for this is scarce in many operating systems. With respect, scarce in what operating system apart from say, DOS? I can't think of any Unixes or Unix alikes that don't have SMP support (even Plan 9 does). OS/2 has had it as an option since the early 90s. NT right from the start. Same with BeOS. Unless you're using something really obscure I'm struggling to think of anything else. IBM laptops have an excellent record for reliability and support, plus one of the best keyboards you can get. Not cheap though. It's also worth mentioning that the T series have both the p (professional, discrete graphics, possibly other faster stuff) and non p (integrated graphics) series. Not really an issue unless you need OpenGL or games, though. PK
Re: OpenBSD
From: Markus Hennecke [EMAIL PROTECTED] You mean it works great except for newer cards, dual-head setups and a fast X desktop? Yes, but I would not call that great. That's not true. 8800GT support, at the very least, was added back in OpenBSD 4.3 according to the changelog. I'm using two 7600GTs - X works without a hitch in dual head (xrandr) on the first adapter and appears to be accelerated. So far I've not configured it to use the second adapter and run three screens. PK
Re: Virtual Consoles in OpenBSD/macppc
From: Pedro de Oliveira [EMAIL PROTECTED] Hi, Anyone here using OpenBSD/macppc knows if its possible to enable more than one virtual console? I cant seem to find any info about that in the FAQ. http://www.openbsd.org/faq/faq7.html It's not supported. Use 'screen' from packages instead. PK
Hanging X
I'm trying to get X running on my amd64 (Core2Duo) box. It starts but freezes shortly although OpenBSD is still accessible from serial console/ssh. X cannot be killed via kill or kill -9. My configuration is marginally unusual as it contains two 7600GTs. X -configure detects a multihead configuration but only ever displays on one display - the other monitor stops displaying with an out of range error (if it is out of range it must be sending something pretty funky as the monitor can handle up to 2048x1536x75Hz). Also, the second 7600GT appears not to be detected? From dmesg : vga1 at pci1 dev 0 function 0 NVIDIA GeForce 7600 GT rev 0xa1 wsdisplay0 at vga1 mux 1: console (80x25, vt100 emulation) wsdisplay0: screen 1-5 added (80x25, vt100 emulation) ppb1 at pci0 dev 3 function 0 Intel 82975X PCI rev 0x00: apic 2 int 16 (irq 255) pci2 at ppb1 bus 2 NVIDIA GeForce 7600 GT rev 0xa1 at pci2 dev 0 function 0 not configured Any advice how to diagnose this before I submit a bug? Also, how do I activate the second graphics card? Basic hardware should be fine : I run a three monitor configuration in Windows without a problem. PK
Re: 4.4 recently installed
From: Martin Schrvder [EMAIL PROTECTED] 2008/11/10 Douglas A. Tutty [EMAIL PROTECTED]: Put in a couple of big hard drives (I don't know what's there already) and use it as network storage (backup your other computers). And then wonder why it crashes when it does the first fsck. :-( AFAIK 64M will only allow you to fsck 64GB. Seriously: Get yourself a new machine if you can. It will be much faster and consume less power. I'd second that.. Whilst I get the occasional desire for Old Computer Kit and some old computers have a few neat features, generally there really isn't much to compete with a decent Core2Duo/Quad box - it's not that expensive, is low power, fast and supports virtualisation. I can say that considering I also run OpenBSD on a sparcstation 10 and an SGI O2..(*) I do have a 266MHz Pentium thin client running an embedded OpenBSD firewall; I wouldn't necessarily recommend that in your case as yours won't be fanless or consume minimal power. I also have a 486 DX2-66 as a bittorrent box - that actually runs NT 4. Much though I like OpenBSD at times, it's far more effective to add the large disk driver and run utorrent. If you must use your box it'll perform adequately for network based operations - mail, web servers, pf, etc etc. It's a bit underpowered for modern X although you should at least be able to run multiple X terms. Compilation times on older kit tend to be painful. (*) The O2 has some nice video hardware - which is useless in anything other than Irix. The sparcstation has a decent boot monitor, and a cheap multiport 10Mb network card available - which is useless now ADSL goes faster than 10Mb. The O2 uses SCA disks which are still available, if a lot more expensive than SATA. The sparcstation uses narrow SCSI disks which are now defunct, loud and slow. There are workarounds for all of this, but running modern kit really is a lot easier. PK
openbsd sgi - uname -m, packages and mips64
A bit of an oddity. On all other platforms (at least I think so), the output from 'uname -m' matches the name of the directory under packages, except under sgi, for which the directory is 'mips64'. Any chance of this changing for 4.5? I'm presuming no-one is porting to mips32 (netbsd supports the O2, Indigo,Challenge,Indy etc). I realise the O2 port requires a bit of love; I'll see if I can contribute stuff myself. Still, congrats on shipping what appears to be the only 64 bit OS on the O2, regardless of the existing flaws (Irix on the O2 is 32bit, as is netbsd and the workable versions of Linux. 64bit Linux is purely experimental at this stage) PK
Re: openbsd sgi - uname -m, packages and mips64
From: Matthew Weigel [EMAIL PROTECTED] Peter Kay - Syllopsium wrote: A bit of an oddity. On all other platforms (at least I think so), the output from 'uname -m' matches the name of the directory under packages, For all supported platforms, the name of the package directory matches 'machine -a'. Because packages are compiled for a specific processor type, not a platform. For example, the mac68k and mvme68k platforms both have a 'machine -a' output of 'm68k' - ditto with the macppc and socppc platforms. That makes perfect sense, thanks for the response. PK
kernel debugging broken in 4.4-CURRENT?
Kgdb kernel debugging appears to have been broken in 4.4-CURRENT for a week if not longer. The exact same config file works creating a debug kernel under 4.3. I've updated kernel, userland etc. The same thing happens under every virtualised environment I throw it at (vmware, qemu, virtualbox). I get : 'com0: at isa0 port 0x3f8/8 irq 4panic: com_isa_attach: mapping failed ' followed by a kernel panic. In 4.3 it correctly refers to 'pccom' and says that kgdb has been attached. I can't be the only one needing to do kernel debugging with 4.4-CURRENT - am I missing something? kernel config file is : include arch/i386/conf/GENERIC rmoption DDB makeoptions DEBUG=-g option KGDB option KGDB_DEVNAME=\com\,KGDBADDR=0x3f8,KGDBRATE=9600 PK
Re: kernel debugging broken in 4.4-CURRENT?
From: Ted Unangst [EMAIL PROTECTED] To: Peter Kay - Syllopsium [EMAIL PROTECTED] You might be the only one using kgdb. Between 4.3 and 4.4 the pccom device was folded into com, which may not have gone perfectly. OK. A further question, then. What I'm trying to do is debug if_bridge and if_pppoe as I'm stupid enough to try hacking the source to bridge PPPoE(4) to another interface (it doesn't work, and if you hack if_bridge to tell it to add it to the bridge, the bridging doesn't work). if_bridge and if_pppoe are both kernel level files, and I want to set breakpoints to find out where it is/isn't working, so kgdb seemed like the best thing to use. Given that I'm in a minority of kgdb users, what's everyone else using in cases like this? PK
Re: 4.4-current on XenServer 5
From: Stephan A. Rickauer [EMAIL PROTECTED] To: misc misc@openbsd.org Sent: Monday, September 22, 2008 11:24 AM Subject: 4.4-current on XenServer 5 In know virtualization is not one of the primary targets of OpenBSD. However, in case someone is interested, here's a dmesg of 4.4-current booting bsd.rd on latest XenServer 5 (Express, with Intel VT). As you can see, there is no harddisk detected. I am ready to help testing if a developer wants to look at it. Cheers, Stephan pchb0 at pci0 dev 0 function 0 Intel 82441FX rev 0x02 pcib0 at pci0 dev 1 function 0 Intel 82371SB ISA rev 0x00 pciide0 at pci0 dev 1 function 1 Intel 82371SB IDE rev 0x00: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility wd0 at pciide0 channel 0 drive 0: QEMU HARDDISK wd0: 16-sector PIO, LBA48, 5120MB, 10485760 sectors wd0(pciide0:0:0): using PIO mode 0, DMA mode 2 Looks like a harddisk to me... am I missing something? PK
Re: Newbie some problem with OpenBSD
From: Edd Barrett [EMAIL PROTECTED] To: Josh Grosse [EMAIL PROTECTED] Cc: Ling Xiaoheng [EMAIL PROTECTED]; OpenBSD Misc Maillist misc@openbsd.org Sent: Friday, September 12, 2008 4:49 PM Subject: Re: Newbie some problem with OpenBSD On Fri, Sep 12, 2008 at 1:32 PM, Josh Grosse [EMAIL PROTECTED] wrote: Upgrading by compiling source is unsupported. I don't see why that wouldn't work, just so long as you sysmerge. Actually, the issue isn't whether it would work, or whether it's supported. The real question is : Why bother? Compiling by source is vastly slower and more involved than whacking the install cd/disk/netboot in and telling the system to upgrade. As mentioned in response to the original post, it's rarely necessary to compile the source. At this point, there are precisely five fixes included in 4.3-stable, which is what the majority of people running production systems should be running.. The time to compile source is when it brings a tangible benefit. In this case, it doesn't. PK
Bridging pppoe(4) to another NIC - is this even possible, as it appears impossible to change the MTU?
I'm trying to create a transparent bridging firewall with a NIC at one end and PPPoE(4) at the other end. In this case I'm using OpenBSD 4.4-CURRENT sparc (same thing happens on 4.2) on a sparcstation 10 with quad ethernet (qe - 10Mb). The problem is that the bridge cannot be established, probably because the MTUs do not match. The MTU of qe(0 to 3) is 1500. The MTU of pppoe0 (established via pppoe(4)) is 1492 I can't change the MTU of qe0-3. There's an overhead of 8 bytes in PPPoE - does this therefore mean it can never go above 1492? The MTU of pppoe can be modified, but only to 1492 or lower. Additionally I am confused by the OpenBSD 4.4 changelist item : 'Adapt maximum permitted MTU on pppoe(4) to the MTU of the connected Ethernet/VLAN interface.' This, to me, potentially indicates that the MTU of pppoe could be matched to the MTU of the NIC (although, is this perhaps limited by the fact that to do so it would need 1500+8 bytes of overhead, and thus blow the 1500 Ethernet MTU limit?). I tried applying 4.4-CURRENT and the MTU of pppoe stays at 1492. Any solution? Find a NIC which can have its MTU lowered, perhaps? Also, even if I could get the MTUs to match, bridge complains on startup because pppoe0 does not yet exist. Is there a more elegant solution than a shellscript with a delay and a series of brconfig commands to fix this? Cheers! Peter
Re: Bridging pppoe(4) to another NIC - is this even possible, as it appears impossible to change the MTU?
From: ropers [EMAIL PROTECTED] To: Peter Kay - Syllopsium [EMAIL PROTECTED] Cc: misc@openbsd.org Sent: Monday, September 08, 2008 2:05 PM Subject: Re: Bridging pppoe(4) to another NIC - is this even possible, as it appears impossible to change the MTU? 2008/9/8 Peter Kay - Syllopsium [EMAIL PROTECTED]: I'm trying to create a transparent bridging firewall with a NIC at one end and PPPoE(4) at the other end. In this case I'm using OpenBSD 4.4-CURRENT sparc (same thing happens on 4.2) on a sparcstation 10 with quad ethernet (qe - 10Mb). The problem is that the bridge cannot be established, probably because the MTUs do not match. The MTU of qe(0 to 3) is 1500. The MTU of pppoe0 (established via pppoe(4)) is 1492 I can't change the MTU of qe0-3. There's an overhead of 8 bytes in PPPoE - does this therefore mean it can never go above 1492? The MTU of pppoe can be modified, but only to 1492 or lower. Additionally I am confused by the OpenBSD 4.4 changelist item : 'Adapt maximum permitted MTU on pppoe(4) to the MTU of the connected Ethernet/VLAN interface.' This, to me, potentially indicates that the MTU of pppoe could be matched to the MTU of the NIC (although, is this perhaps limited by the fact that to do so it would need 1500+8 bytes of overhead, and thus blow the 1500 Ethernet MTU limit?). I tried applying 4.4-CURRENT and the MTU of pppoe stays at 1492. Any solution? Find a NIC which can have its MTU lowered, perhaps? Also, even if I could get the MTUs to match, bridge complains on startup because pppoe0 does not yet exist. Is there a more elegant solution than a shellscript with a delay and a series of brconfig commands to fix this? Cheers! Peter When you say you want PPPoE at the other end, what exactly do you mean? Is the PPPoE stuff on a separate box that you reach via RJ-45, ie. does your net look loke this: Intranet -- int_if--OpenBSD_bridge--ext_if -- DSL modem w/ PPPoE Or do you want the PPPoE login/dialup stuff to be handled by OpenBSD, ie. does your network look like this: Intranet -- int_if--OBSD_box--ext_if -- DSL modem in dumb as a brick mode If it is the latter then I'm not quite sure where you want to build a transparent bridge. Because IIRC your external interface in this scenario would be a tun interface and you would use NAT. Unless of course... Ok, let me ask you this then: What kind of Internet connectivity do you have / what kind of Internet connection do you have from your ISP? If you are just using an ordinary SOHO user PPPoE offering from a regular ISP, then you more than likely just get ONE IPv4 address, which means you will have to use NAT, not bridging, no two ways about it. Or am I horribly misunderstanding something? A somewhat confused --ropers I have an ADSL connection with 8 IPs, so the topology looks like : Intranet --Int_if--OpenBSD_bridge_with_pf--ext_if--ADSL router bridging PPPoE packets. The router is in bridging mode so 'all' it does is shift the PPPoE packets from the POTS (telephone) connection to the Ethernet port. It passes PPPoE on to the external OpenBSD interface pppoe(4) is bound to. The pppoe(4) should establish the ppp connection and receive the traffic for all 8 IPs. I don't have the space or the inclination to subnet, so I want to transparent bridge with pf for firewalling between pppoe and the internal interface. I realise 1:1 NAT is an option here, but I'd prefer my internal systems to remain with WAN addresses, but protected by a firewall. I'm doing all this because otherwise I have to lose one IP address to my router, and I'd rather not. (What I *actually* want to do eventually is to have three interfaces on the firewall. One internal, one DMZ, one external (pppoe bound) where external-dmz is a transparent firewall and external-internal is NAT) PK
Re: Bridging pppoe(4) to another NIC - is this even possible, as it appears impossible to change the MTU?
From: Vijay Sankar [EMAIL PROTECTED] To: Peter Kay - Syllopsium [EMAIL PROTECTED] Cc: misc@openbsd.org Sent: Monday, September 08, 2008 2:50 PM Subject: Re: Bridging pppoe(4) to another NIC - is this even possible, as it appears impossible to change the MTU? On September 8, 2008 06:43:45 am Peter Kay - Syllopsium wrote: Also, even if I could get the MTUs to match, bridge complains on startup because pppoe0 does not yet exist. Is there a more elegant solution than a shellscript with a delay and a series of brconfig commands to fix this? Not sure whether the following is appropriate under your circumstances but I can try to describe a different solution. We have 8 IP addresses with an ADSL connection (6 with the ISP here calls it a framed route and 2 that are static) and we set pf up as follows: ext_if=pppoe0 int_if=rl0 dmz_if=dc1 scrub out on $ext_if max-mss 1440 One of the 6 addresses is the DMZ interface's IP and I am routing all the other public IP's through this. So I don't have to bridge in my scenario and it has worked very well. Interface fxp0 is connected to the DSL modem and has the Ethernet default MTU of 1500 and pppoe0 has MTU of 1492. fxp0: flags=8843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST mtu 1500 lladdr xx:xx:xx:xx:xx:xx media: Ethernet autoselect (100baseTX full-duplex) status: active pppoe0: flags=8851UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST mtu 1492 dev: fxp0 state: session sid: 0x64e5 PADI retries: 0 PADR retries: 0 time: 36d 04:02:01 sppp: phase network authproto pap authname x groups: pppoe egress inet aaa.bbb.ccc.ddd -- eee.fff.ggg.hhh netmask 0x I am using kernel -mode pppoe. -- Vijay Sankar, M.Eng., P.Eng. ForeTell Technologies Limited 59 Flamingo Avenue, Winnipeg, MB Canada R3J 0X6 Phone: +1 204 885 9535, E-Mail: [EMAIL PROTECTED] OK.. I presume routing is also turned on in your scenario? Unless I'm missing something though, aren't you losing two of your 8 IP addresses - one to PPPoE and one to the DMZ? A main point of me running PPPoE on the firewall is that I only lose one of my 6 available (obviously network and broadcast eat two of my eight) WAN addresses. If I wanted to lose two I could leave it as is, with the router establishing the PPPoE connection, the external interface on the firewall with a WAN IP, and a transparent bridge to the DMZ. PK
