Re: HEADS UP: bdevsw has changed its character
Greg Lehey wrote: A day or so ago, bdevsw changed from being an array of struct cdevsws to an inline function. That's not a problem in itself: do a make world and all will be well. It does, however, mean that klds which were compiled before the change will no longer load. I had a report today of one person who forgot to do this, and his source tree is on a vinum volume, and he can't load vinum. Beware! That isn't all.. If you are using kld modules, it is a Really Good Idea to keep your kld modules compiled and in sync with your kernel while internal interfaces are changing. kget(8) also will need a recompile. People who are running -current for stability had better be damn careful and be very selective about what they choose to run as a stable snapshot. Cheers, -Peter To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
On Sunday, 9 May 1999 at 16:15:51 +0800, Peter Wemm wrote: Greg Lehey wrote: A day or so ago, bdevsw changed from being an array of struct cdevsws to an inline function. That's not a problem in itself: do a make world and all will be well. It does, however, mean that klds which were compiled before the change will no longer load. I had a report today of one person who forgot to do this, and his source tree is on a vinum volume, and he can't load vinum. Beware! That isn't all.. If you are using kld modules, it is a Really Good Idea to keep your kld modules compiled and in sync with your kernel while internal interfaces are changing. Right, it was just pretty unfortunate in this case (and I heard of another one) that the person had his source tree mounted on a vinum volume, so he could no longer access it :-) If this happens to anybody else, there's a version of the vinum kld compiled this morning on ftp://ftp.lemis.com/pub/vinum/vinum.ko. kget(8) also will need a recompile. People who are running -current for stability had better be damn careful and be very selective about what they choose to run as a stable snapshot. I think the real problem is that the klds get built with make world and not with a kernel build. How about changing that? I've got the opposite problem on another machine: I did a make world, but not a reboot, and now my Linux emulation is broken. Greg -- See complete headers for address, home page and phone numbers finger g...@lemis.com for PGP public key To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
In message 19990509081553.6ab261...@spinner.netplex.com.au, Peter Wemm write s: Greg Lehey wrote: A day or so ago, bdevsw changed from being an array of struct cdevsws to an inline function. That's not a problem in itself: do a make world and all will be well. It does, however, mean that klds which were compiled before the change will no longer load. I had a report today of one person who forgot to do this, and his source tree is on a vinum volume, and he can't load vinum. Beware! That isn't all.. If you are using kld modules, it is a Really Good Idea to keep your kld modules compiled and in sync with your kernel while internal interfaces are changing. kget(8) also will need a recompile. People who are running -current for stability had better be damn careful and be very selective about what they choose to run as a stable snapshot. Amen brother. In general any commit to sys/kern/* or sys/sys/* should make you pay a LOT of attention. -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
In message 19990509174914.a22...@freebie.lemis.com, Greg Lehey writes: People who are running -current for stability had better be damn careful and be very selective about what they choose to run as a stable snapshot. I think the real problem is that the klds get built with make world and not with a kernel build. How about changing that? I've got the opposite problem on another machine: I did a make world, but not a reboot, and now my Linux emulation is broken. Well, it is a problem you cannot solve in general, in particular as we start to see 3rd parth KLDs... -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
Greg Lehey wrote: On Sunday, 9 May 1999 at 16:15:51 +0800, Peter Wemm wrote: Greg Lehey wrote: A day or so ago, bdevsw changed from being an array of struct cdevsws to an inline function. That's not a problem in itself: do a make world and all will be well. It does, however, mean that klds which were compiled before the change will no longer load. I had a report today of one person who forgot to do this, and his source tree is on a vinum volume, and he can't load vinum. Beware! That isn't all.. If you are using kld modules, it is a Really Good Idea to keep your kld modules compiled and in sync with your kernel while internal interfaces are changing. Right, it was just pretty unfortunate in this case (and I heard of another one) that the person had his source tree mounted on a vinum volume, so he could no longer access it :-) If this happens to anybody else, there's a version of the vinum kld compiled this morning on ftp://ftp.lemis.com/pub/vinum/vinum.ko. Which is why I use pseudo-device vinum :-) But the risk there is that the ioctl's change as /sbin/vinum is compiled seperately. kget(8) also will need a recompile. People who are running -current for stability had better be damn careful and be very selective about what they choose to run as a stable snapshot. I think the real problem is that the klds get built with make world and not with a kernel build. How about changing that? I've got the opposite problem on another machine: I did a make world, but not a reboot, and now my Linux emulation is broken. I have a system in mind to replace config(8) entirely and will allow flexible kernel and module building, but it's barely even on the drawing board. It will make sys/modules obsolete, which is the main reason why I don't want code in there. (The screensavers are misplaced). I also have half thought about moving some particular kernel-specific tools under this system, but I'm not at all sure about that. On one hand, libkvm etc could benefit, along with ps etc which are hightly kernel specific. But it brings in a whole new set of chicken/egg problems. (like, if the kernel build supervisor is maintained by the same tool, then there will be problems.) Greg Cheers, -Peter To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
Greg Lehey wrote: I think the real problem is that the klds get built with make world and not with a kernel build. How about changing that? I've got the opposite problem on another machine: I did a make world, but not a reboot, and now my Linux emulation is broken. Same problem with loader, with the added irony that loader is located under sys/... -- Daniel C. Sobral(8-DCS) d...@newsguy.com d...@freebsd.org Proof of Trotsky's farsightedness is that _none_ of his predictions have come true yet. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
I think the real problem is that the klds get built with make world and not with a kernel build. How about changing that? I've shot myself in the foot this way too. I'd really like to see modules built as part of the kernel and not userland. As time goes on, I think we will push more and more of the kernel into modules (rather like Solaris). This distinction of where the live kernel's bits live isn't useful. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: HEADS UP: bdevsw has changed its character
In message 19224.926238...@critter.freebsd.dk Poul-Henning Kamp writes: : In general any commit to sys/kern/* or sys/sys/* should make you pay : a LOT of attention. Also, when the number of commits in these areas has been as high as it has been lately, one should think twice before updating... Warner To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message