Re: hints static wiring
In message [EMAIL PROTECTED] Greg Lehey writes: : 2. You must have a /boot/device.hints file, and it must contain at : least some entries Add "unless you compile the hints statically into the kernel, in which case it may be empty." and you will be correct. There's plenty of potential for foot shooting, but it will generally only impact older, legacy non-pnp systems. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Brooks Davis writes: : Which begs the question: when should the hints file be updated? : : When you add new devices which require hints to work. Since GENERIC.hints : covers the defaults for most of those devices and I don't see too many : new isa drivers going in to the kernel I think that roughly translates to : "never" on modern system and "rarely" on strange systems made of scavenged : parts. Heck, on most legacy free systems, it really does translate : to never because you can't add anything that would effect drivers that : need hints (device wiring being the only exception I can think of there). Yes. Usually you can just copy GENERIC.hints and forget it. However, if you traffic in older hardware in non-plug and play mode, you'll have to cope with updating/creating/modifying a device.hints file. As this hardware becomes rarer, you'll have to deal less and less. Note, however, that until the isapnpbios code is corrected to probe in the right order (before all other isa devices rather than after like it does now with the isapnp hardware) you'll need hints on even the most modern machines. The snippet that I posted earlier was one that seemed to work on my vaio. I believe there's work on a userconfig in boot-loader in the works as well, which is a large part of the motivation here. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
First, I want to thank everyone for taking the time to explain what was going on. It's now clear that I was confused, and things aren't as bad as I thought. I'd like to see the Makefile changes so that if there wasn't an empty /boot/device.hints, one was created, but that's relatively minor. Brooks Davis writes: This is more correct. The new world order says that hints are not in the kernel, instead they are loaded by the loader at boot time. By default they are loaded from /boot/device.hints. Without hints in some form old, stupid devices don't work. Unfortunatly, PC consoles are old, stupid devices for compatability reasons so it's best to have a working hints file around. Just curious - are there no defaults at all if you don't have a hints file somewhere? That seems a bit strange. mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Tue, Aug 29, 2000 at 04:15:09AM -0500, Mike Meyer wrote: Brooks Davis writes: This is more correct. The new world order says that hints are not in the kernel, instead they are loaded by the loader at boot time. By default they are loaded from /boot/device.hints. Without hints in some form old, stupid devices don't work. Unfortunatly, PC consoles are old, stupid devices for compatability reasons so it's best to have a working hints file around. Just curious - are there no defaults at all if you don't have a hints file somewhere? That seems a bit strange. If you don't have a hints file in /boot and you didn't suck one in as a static config during boot then, yes there are no defaults. I think the idea is that we should eventualy get to a state where (on legacy free systems at least) we really can run with no hints because everything is PnP. This also means you should be able to boot generic on some really strange systems just by modifying the hints file. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Mon, Aug 28, 2000 at 06:29:21PM -0700, Brooks Davis wrote: On Tue, Aug 29, 2000 at 10:25:26AM +0930, Greg Lehey wrote: At the very least, there appears to be confusion about how to use the hints. I can see two conflicting views here: 1. You must have a /boot/device.hints file, but it may be empty. This is minimally correct. I.e. that's what the build system requires. This works if you build static hints into your kernel. Please stop saying FUD. The *INSTALL* system requires /boot/device.hints. The *BUILD* process does not. 2. You must have a /boot/device.hints file, and it must contain at least some entries. This is more correct. The new world order says that hints are not in the kernel, instead they are loaded by the loader at boot time. I would quite to so far. I know many that will continue to compile static hints into their kernels until that ability is removed. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Tue, Aug 29, 2000 at 10:30:24AM -0700, David O'Brien wrote: On Mon, Aug 28, 2000 at 06:29:21PM -0700, Brooks Davis wrote: This is minimally correct. I.e. that's what the build system requires. This works if you build static hints into your kernel. Please stop saying FUD. The *INSTALL* system requires /boot/device.hints. The *BUILD* process does not. That's what I ment. I ment build as opposed to runtime. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Brooks Davis writes: : On Tue, Aug 29, 2000 at 10:30:24AM -0700, David O'Brien wrote: : On Mon, Aug 28, 2000 at 06:29:21PM -0700, Brooks Davis wrote: : This is minimally correct. I.e. that's what the build system requires. : This works if you build static hints into your kernel. : : Please stop saying FUD. The *INSTALL* system requires : /boot/device.hints. The *BUILD* process does not. : : That's what I ment. I ment build as opposed to runtime. I committed some changes to UPDATING that I think are complete, fair and accurate. Please look at the new UPDATING file and let me know what you think. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
David O'Brien writes: On Mon, Aug 28, 2000 at 06:29:21PM -0700, Brooks Davis wrote: On Tue, Aug 29, 2000 at 10:25:26AM +0930, Greg Lehey wrote: At the very least, there appears to be confusion about how to use the 2. You must have a /boot/device.hints file, and it must contain at least some entries. This is more correct. The new world order says that hints are not in the kernel, instead they are loaded by the loader at boot time. I would quite to so far. I know many that will continue to compile static hints into their kernels until that ability is removed. I suspect I'm going to be one of those people, but let's ask the next question. How do the two approaches compare from a performance standpoint? I figure having them wired in means you avoid parsing them at boot time, which should be faster. How smart is the code about devices with hints that aren't configured in the kernel? Do both methods throw out such data, or do both save it? Basically, which will provide the smallest running kernel, and if it's the same, which will boot fastest? Thanx, mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Mike Meyer writes: : I suspect I'm going to be one of those people, but let's ask the next : question. How do the two approaches compare from a performance : standpoint? Completely identical. One loads the table from a file, the other reads it from the kernel. : Basically, which will provide the smallest running kernel, and if it's : the same, which will boot fastest? They should be the same size. I'd be surprised if you could measure a difference. You'll get more savings by not compiling in just one driver. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Brooks Davis writes: On Sun, Aug 27, 2000 at 09:33:21PM +0900, Motomichi Matsuzaki wrote: Doing 'make install' without /boot/device.hints is failed, saying "You must set up a /boot/device.hints file first." Is this right? You should read cvs-all. There was a commit by Peter which forces you to install a /boot/device.hints file to install a kernel as an anti-foot shooting measure. An empty file (ie touch /boot/device.hints) is acceptable for those who don't want to use a hints file. I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Thanx, mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? cvs-all is not appropriate. I am noticing a 3-7 day lag on UPDATING. Bad. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Actually, device.hints isn't used in the build process. Your KERNEL.hints file is hard-coded into the kernel when your kernel is built (assuming you use one). /boot/device.hints is used to override the "hardcoded" values of hints, KERNEL.hints, at boot time. Sometimes, people can make a mistake in KERNEL.hints, and it's necessary to override those hints with /boot/device.hints. So, device.hints is created after-the-fact, and not part of the kernel build. Of course, if you don't have any hints to override, then just install an empty device.hints file. device.hints is there to save you from rebuilding your kernel every time you want to change a hint value. -Donn To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Donn Miller writes: Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Actually, device.hints isn't used in the build process. In that case, why does the kernel build process fail if it doesn't exist? KERNEL.hints file is hard-coded into the kernel when your kernel is built (assuming you use one). /boot/device.hints is used to override the "hardcoded" values of hints, KERNEL.hints, at boot time. Sometimes, people can make a mistake in KERNEL.hints, and it's necessary to override those hints with /boot/device.hints. So, device.hints is created after-the-fact, and not part of the kernel build. Of course, if you don't have any hints to override, then just install an empty device.hints file. Will the system fail to boot if there isn't an empty device.hints file? mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Mike Meyer wrote: Donn Miller writes: Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Actually, device.hints isn't used in the build process. In that case, why does the kernel build process fail if it doesn't exist? Probably because you have `hints "BLABLA.hints"' line in your kernel config file. KERNEL.hints file is hard-coded into the kernel when your kernel is built (assuming you use one). /boot/device.hints is used to override the "hardcoded" values of hints, KERNEL.hints, at boot time. Sometimes, people can make a mistake in KERNEL.hints, and it's necessary to override those hints with /boot/device.hints. So, device.hints is created after-the-fact, and not part of the kernel build. Of course, if you don't have any hints to override, then just install an empty device.hints file. Will the system fail to boot if there isn't an empty device.hints file? No, it will boot, but some devices (like keyboard, console etc) would not work. -Maxim To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Maxim Sobolev writes: Mike Meyer wrote: Donn Miller writes: Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Actually, device.hints isn't used in the build process. In that case, why does the kernel build process fail if it doesn't exist? Probably because you have `hints "BLABLA.hints"' line in your kernel config file. That doesn't really answer the question. Yup, I use GENERIC.hints. That exists. I can see why that not existing would cause problems, but not /boot/device.hints? *Especially* when I'm building a kernel for a different machine? KERNEL.hints file is hard-coded into the kernel when your kernel is built (assuming you use one). /boot/device.hints is used to override the "hardcoded" values of hints, KERNEL.hints, at boot time. Sometimes, people can make a mistake in KERNEL.hints, and it's necessary to override those hints with /boot/device.hints. So, device.hints is created after-the-fact, and not part of the kernel build. Of course, if you don't have any hints to override, then just install an empty device.hints file. Will the system fail to boot if there isn't an empty device.hints file? No, it will boot, but some devices (like keyboard, console etc) would not work. That's clearly not true - I just removed an empty /boot/device.hints and rebooted, and all those things work fine. I can believe that such things won't work if they aren't specified in some hints file, but an empty /boot/device.hints doesn't do anything more to specify them than one that isn't there. mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Mike Meyer wrote: Will the system fail to boot if there isn't an empty device.hints file? No, it will boot, but some devices (like keyboard, console etc) would not work. That's clearly not true - I just removed an empty /boot/device.hints and rebooted, and all those things work fine. I can believe that such things won't work if they aren't specified in some hints file, but an empty /boot/device.hints doesn't do anything more to specify them than one that isn't there. That's probably because you have hints compiled into your kernel. Try to compile kernel w/o hints and use it with empty/unexistent /boot/device.hints and you will see what I mean. -Maxim To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Maxim Sobolev writes: Mike Meyer wrote: Will the system fail to boot if there isn't an empty device.hints file? No, it will boot, but some devices (like keyboard, console etc) would not work. That's clearly not true - I just removed an empty /boot/device.hints and rebooted, and all those things work fine. I can believe that such things won't work if they aren't specified in some hints file, but an empty /boot/device.hints doesn't do anything more to specify them than one that isn't there. That's probably because you have hints compiled into your kernel. Try to compile kernel w/o hints and use it with empty/unexistent /boot/device.hints and you will see what I mean. Well, yeah, I'd expect that. I'm still trying to figure out what *good* failing to compile unless there's an empty /boot/device.hints does. I mean, if I didn't provide kernel hints, it would make some sense if the build process could determine that it was building on the machine it was running on. mike To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Ahh..., I tried to summarize my opinion. If you find any misunderstandings of me, please correct them. *** What's happen if there's no /boot/device.hints? case A kernel has no built-in hints ... some devices would not work, system would stall! You can tell whole hints to the kernel interactively on /boot/loader, however it's a tiresome task. YES, this is the abominable situation. To avoid it, you should be warned at kernel-install-time. case B kernel has built-in hints (i.e. config has "hints" line) B-1: wrong hints ... some devices would not work, system could stall. You can correct hints interactively on /boot/loader. You can override hints by making /boot/device.hints also. B-2: suitable hints ... everything goes OK You can override hints by /boot/device.hints. Currently, 'make install' will be aborted in every case above, but this treatment is suitable only for case A. And it would be technically possible to limit this treatment to case A. This treatment will do accoring to Makefile, which is controled by config(8) (src/usr.sbin/config/mkmakefile.c). -- Motomichi Matsuzaki [EMAIL PROTECTED] Dept. of Biological Sciences, Grad. School of Science, Univ. of Tokyo, Japan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Mon, Aug 28, 2000 at 03:19:15AM -0500, Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Having an empty one will not help you, but installing a post hints change GENERIC without a hints file will results in a system that does not boot. Thus you are required to have a hints file. The ability to install a bogus (empty) one is for those who want static hints. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Mon, Aug 28, 2000 at 08:24:50AM -0500, Mike Meyer wrote: Well, yeah, I'd expect that. I'm still trying to figure out what *good* failing to compile unless there's an empty /boot/device.hints The kernel does not fail to *BUILD*. ``make install'' is what fails. I agree that the requirement is somewhat anonying. But, it is better to have this slight anonyance than for many to to install and boot a kernel that would be broken. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Tony Fleisher writes: : Just a suggestion, but isn't this the type of thing that : should be added to UPDATING? Quoting from the UPDATING file: ... 2825: /boot/device.hints is now required for installkernel to succeed. ... 2612: Peter took an axe to config(8). Besure that you read his mail on the topic before even thinking about updating. You will need to create a /boot/device.hints or add a hints directive to your config file to compile them in statically. The format of the config file has changed as well. Please see GENERIC or NEWCARD for examples of the new format. ... Granted, I did commit the first entry only a few hours before you posted... Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Matthew Jacob writes: : I do read cvs-all, and I missed it. Not did I find device.hints in the : relevant Makefiles. Can you provide a pointer to details on how : /boot/device.hints is used in the build process, or how having an : empty one keeps you from shooting yourself in the foot? : : cvs-all is not appropriate. I am noticing a 3-7 day lag on UPDATING. : Bad. cvs-all *IS*REQUIRED* for all people running -current. UPDATING tries to cull things from there on an as needed basis. It is a service that gets done when I have time. If someone wants to pay me a stipend to drop everything the instant something is committed to the tree and update UPDATING, then the lag will improve. Otherwise, 3-7 days really isn't that bad and will continue to be the case. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Matthew Jacob writes: : I do read cvs-all, and I missed it. Not did I find device.hints in the : relevant Makefiles. Can you provide a pointer to details on how : /boot/device.hints is used in the build process, or how having an : empty one keeps you from shooting yourself in the foot? : : cvs-all is not appropriate. I am noticing a 3-7 day lag on UPDATING. : Bad. cvs-all *IS*REQUIRED* for all people running -current. UPDATING tries to cull things from there on an as needed basis. It is a service that gets done when I have time. If someone wants to pay me a stipend to drop everything the instant something is committed to the tree and update UPDATING, then the lag will improve. Otherwise, 3-7 days really isn't that bad and will continue to be the case. Warner Oops- I realize that what I said might have been construed as criticism- not meant at all! What I meant is that while cvs-all can be read by everyone, it's not always obvious from the flood of mail there, or if you're not a developer, what needs to change. In my opinion, people making major changes that require something in UPDATING, should coordinate with you *before* the commit. Only 5 or 6 brain cells are needed for this- I sure wish some of my fellow committers weren't such skinflints in this area. -matt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Mike Meyer writes: : Will the system fail to boot if there isn't an empty device.hints : file? If the kernel doesn't have a hints file compiled into it, then you will have problems. However, you may not have a video console. I've been able to boot my laptop with a kernel that had no hints and this was the result. All the PCI based things worked, as did those things that had a PnP ID, except the video console. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Matthew Jacob writes: : In my opinion, people making major changes that require something in : UPDATING, should coordinate with you *before* the commit. Only 5 or : 6 brain cells are needed for this- I sure wish some of my fellow : committers weren't such skinflints in this area. About 10% of the commmits that are known to need an UPDATING entry get sent to me first. About 50%-70% are posted to -current, -stable or committers, which is less well because I get busy.. The remaining entries are not well communicated and could be near the "oops" category. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
In message [EMAIL PROTECTED] Mike Meyer writes: : Maxim Sobolev writes: : Mike Meyer wrote: : : Donn Miller writes: :Mike Meyer wrote: : I do read cvs-all, and I missed it. Not did I find device.hints in the : relevant Makefiles. Can you provide a pointer to details on how : /boot/device.hints is used in the build process, or how having an : empty one keeps you from shooting yourself in the foot? :Actually, device.hints isn't used in the build process. : In that case, why does the kernel build process fail if it doesn't : exist? : Probably because you have `hints "BLABLA.hints"' line in your kernel config : file. : : That doesn't really answer the question. Yup, I use : GENERIC.hints. That exists. I can see why that not existing would : cause problems, but not /boot/device.hints? *Especially* when I'm : building a kernel for a different machine? The build of the kernel isn't forbidden by not having /boot/device.hints, just the install. I just copied my GENERIC.hints to /boot/device.hints and things were happy. : That's clearly not true - I just removed an empty /boot/device.hints : and rebooted, and all those things work fine. I can believe that such : things won't work if they aren't specified in some hints file, but an : empty /boot/device.hints doesn't do anything more to specify them than : one that isn't there. Specifically, the console will not work without hints. These hints can be compiled in or in /boot/device.hints. You need to have hint.atkbdc.0.at="isa" hint.atkbdc.0.port="0x060" hint.atkbd.0.at="atkbdc" hint.atkbd.0.irq="1" hint.atkbd.0.flags="0x1" hint.vga.0.at="isa" hint.sc.0.at="isa" hint.sc.0.flags="0x100" At a minimum, but I might be mistaken about that. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Monday, 28 August 2000 at 8:24:50 -0500, Mike Meyer wrote: Maxim Sobolev writes: Mike Meyer wrote: Will the system fail to boot if there isn't an empty device.hints file? No, it will boot, but some devices (like keyboard, console etc) would not work. That's clearly not true - I just removed an empty /boot/device.hints and rebooted, and all those things work fine. I can believe that such things won't work if they aren't specified in some hints file, but an empty /boot/device.hints doesn't do anything more to specify them than one that isn't there. That's probably because you have hints compiled into your kernel. Try to compile kernel w/o hints and use it with empty/unexistent /boot/device.hints and you will see what I mean. Well, yeah, I'd expect that. I'm still trying to figure out what *good* failing to compile unless there's an empty /boot/device.hints does. I mean, if I didn't provide kernel hints, it would make some sense if the build process could determine that it was building on the machine it was running on. On Monday, 28 August 2000 at 14:45:23 -0600, Warner Losh wrote: In message [EMAIL PROTECTED] Mike Meyer writes: Maxim Sobolev writes: Mike Meyer wrote: Donn Miller writes: Mike Meyer wrote: I do read cvs-all, and I missed it. Not did I find device.hints in the relevant Makefiles. Can you provide a pointer to details on how /boot/device.hints is used in the build process, or how having an empty one keeps you from shooting yourself in the foot? Actually, device.hints isn't used in the build process. In that case, why does the kernel build process fail if it doesn't exist? Probably because you have `hints "BLABLA.hints"' line in your kernel config file. That doesn't really answer the question. Yup, I use GENERIC.hints. That exists. I can see why that not existing would cause problems, but not /boot/device.hints? *Especially* when I'm building a kernel for a different machine? The build of the kernel isn't forbidden by not having /boot/device.hints, just the install. I just copied my GENERIC.hints to /boot/device.hints and things were happy. That's clearly not true - I just removed an empty /boot/device.hints and rebooted, and all those things work fine. I can believe that such things won't work if they aren't specified in some hints file, but an empty /boot/device.hints doesn't do anything more to specify them than one that isn't there. Specifically, the console will not work without hints. These hints can be compiled in or in /boot/device.hints. You need to have hint.atkbdc.0.at="isa" hint.atkbdc.0.port="0x060" hint.atkbd.0.at="atkbdc" hint.atkbd.0.irq="1" hint.atkbd.0.flags="0x1" hint.vga.0.at="isa" hint.sc.0.at="isa" hint.sc.0.flags="0x100" At a minimum, but I might be mistaken about that. At the very least, there appears to be confusion about how to use the hints. I can see two conflicting views here: 1. You must have a /boot/device.hints file, but it may be empty. 2. You must have a /boot/device.hints file, and it must contain at least some entries. I ran into this same problem yesterday. I had noticed it in the cvs mailing list, and I found the first entry in UPDATING. But it didn't say what to put in, and I found no other documentation. Finally John Baldwin told me to copy my GENERIC.hints, so I did that, and it worked. But it seems that we should have some documentation here. On the face of it, (1) above seems the most obvious solution. In that case, 'make install' shouldn't fail if there's no device.hints file, it should make one. If it's (2), it can still copy the MYKERNEL.hints file. Which begs the question: when should the hints file be updated? Greg -- Finger [EMAIL PROTECTED] for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Tue, Aug 29, 2000 at 10:25:26AM +0930, Greg Lehey wrote: At the very least, there appears to be confusion about how to use the hints. I can see two conflicting views here: 1. You must have a /boot/device.hints file, but it may be empty. This is minimally correct. I.e. that's what the build system requires. This works if you build static hints into your kernel. 2. You must have a /boot/device.hints file, and it must contain at least some entries. This is more correct. The new world order says that hints are not in the kernel, instead they are loaded by the loader at boot time. By default they are loaded from /boot/device.hints. Without hints in some form old, stupid devices don't work. Unfortunatly, PC consoles are old, stupid devices for compatability reasons so it's best to have a working hints file around. I ran into this same problem yesterday. I had noticed it in the cvs mailing list, and I found the first entry in UPDATING. But it didn't say what to put in, and I found no other documentation. Finally John Baldwin told me to copy my GENERIC.hints, so I did that, and it worked. But it seems that we should have some documentation here. On the face of it, (1) above seems the most obvious solution. In that case, 'make install' shouldn't fail if there's no device.hints file, it should make one. If it's (2), it can still copy the MYKERNEL.hints file. Installing GENERIC.hints might run the risk of turning the current anti-footshooting mechinism into a foot shooting device on non-standard hardware, but I'm kinda inclined to say, who cares since that's a very far out edge case. The only problem is that it might annoy those who insist on static hints. Which begs the question: when should the hints file be updated? When you add new devices which require hints to work. Since GENERIC.hints covers the defaults for most of those devices and I don't see too many new isa drivers going in to the kernel I think that roughly translates to "never" on modern system and "rarely" on strange systems made of scavenged parts. Heck, on most legacy free systems, it really does translate to never because you can't add anything that would effect drivers that need hints (device wiring being the only exception I can think of there). -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
On Sun, Aug 27, 2000 at 09:33:21PM +0900, Motomichi Matsuzaki wrote: When kernel is built with static device wiring (i.e. 'hints' line is enabled in the config file), is /boot/device.hints required? Doing 'make install' without /boot/device.hints is failed, saying "You must set up a /boot/device.hints file first." Is this right? You should read cvs-all. There was a commit by Peter which forces you to install a /boot/device.hints file to install a kernel as an anti-foot shooting measure. An empty file (ie touch /boot/device.hints) is acceptable for those who don't want to use a hints file. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: hints static wiring
Just a suggestion, but isn't this the type of thing that should be added to UPDATING? "This file contains a list, in reverse chronologocal order, of major breakages in tracking -current." TOny. On Sun, 27 Aug 2000, Brooks Davis wrote: On Sun, Aug 27, 2000 at 09:33:21PM +0900, Motomichi Matsuzaki wrote: When kernel is built with static device wiring (i.e. 'hints' line is enabled in the config file), is /boot/device.hints required? Doing 'make install' without /boot/device.hints is failed, saying "You must set up a /boot/device.hints file first." Is this right? You should read cvs-all. There was a commit by Peter which forces you to install a /boot/device.hints file to install a kernel as an anti-foot shooting measure. An empty file (ie touch /boot/device.hints) is acceptable for those who don't want to use a hints file. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message