Re: wi(4) or wicontrol broken
> > At that point, I had to boot bsd.old because it simply refused to boot. > > Based on the incorrect assumption that the snapshot was pre -release I > > decided to download the kernel from 3.8-release and install that. The kernel > > now booted but wi didn't work. I started looking at the differences between > > 3.7-stable and 3.8-release and noticed a few changes in wicontrol and if_wi* > > and thought the problem was somewhere there to be found. > > You are talking nonsense. 3.8 is not on any ftp site. So how can you > download and install it? It was 3:00 (morning) when I wrote that, cut me some slack... Of course I couldn't download the kernel, what I meant was download (via CVS) the sources for OPENBSD_3_8_BASE, compile and install the kernel. > You have everything but your versions straight. Clearly. Theo cleared that up already. By the way, I did a full make build from OPENBSD_3_8_BASE and wicontrol doesn't work. Certainly it could be due to user error, but I won't know for certain until I get my CD in the mail.
Re: wi(4) or wicontrol broken
On Sun, 16 Oct 2005, Ludwig Mises wrote: > > You did not report any problem in a decent way. I'm almost 100% sure you > > created the problem yourself by not following the proper upgrade path. > > > > This is very probable. Actually, it gets more complex than this... I > installed the snapshot from 10/14 and after rebooting the kernel wouldn't > boot. It dropped into the debugger only after a few lines of booting, > somewhere around: > > bios0: ROM list: 0xc/0xc000 0xe8000/0x4000! > cpu0 at mainbus0 > > At that point, I had to boot bsd.old because it simply refused to boot. > Based on the incorrect assumption that the snapshot was pre -release I > decided to download the kernel from 3.8-release and install that. The kernel > now booted but wi didn't work. I started looking at the differences between > 3.7-stable and 3.8-release and noticed a few changes in wicontrol and if_wi* > and thought the problem was somewhere there to be found. You are talking nonsense. 3.8 is not on any ftp site. So how can you download and install it? > Now, after having the versions straight, I am running 3.8-release and wi is > working, after changing the #define for SIOCSWAVELAN to that which is in > 3.8-current. You have everything but your versions straight. -Otto
Re: wi(4) or wicontrol broken
Ludwig Mises wrote: This is very probable. Actually, it gets more complex than this... No, it doesn't. Once you 'upgraded' from -current to -release, you guaranteed that you'd see problems that the developers don't care about. You may have any number of such problems, all stemming from kernel, userland, and build tools not matching each other. At that point, I had to boot bsd.old because it simply refused to boot. Based on the incorrect assumption that the snapshot was pre -release I decided to download the kernel from 3.8-release and install that. The kernel now booted but wi didn't work. See above. Now, after having the versions straight, I am running 3.8-release and wi is working, after changing the #define for SIOCSWAVELAN to that which is in 3.8-current. That's because you're not actually running 3.8-release. I'll bet quite a bit that wi(4) works just fine in 3.8-release, *and* in the 10/14 snapshot, without anyone recompiling anything to change the #define for SIOCSWAVELAN. You've completely voided the reason a person might choose to run -release or -stable instead of -current, your best bet is to use the snapshot and not 3.8-release as built on top of the snapshot. Sorry to have wasted your time and for not providing a proper dmesg. Nobody wants that dmesg. -- Matthew Weigel
Re: wi(4) or wicontrol broken
> You did not report any problem in a decent way. I'm almost 100% sure you > created the problem yourself by not following the proper upgrade path. This is very probable. Actually, it gets more complex than this... I installed the snapshot from 10/14 and after rebooting the kernel wouldn't boot. It dropped into the debugger only after a few lines of booting, somewhere around: bios0: ROM list: 0xc/0xc000 0xe8000/0x4000! cpu0 at mainbus0 At that point, I had to boot bsd.old because it simply refused to boot. Based on the incorrect assumption that the snapshot was pre -release I decided to download the kernel from 3.8-release and install that. The kernel now booted but wi didn't work. I started looking at the differences between 3.7-stable and 3.8-release and noticed a few changes in wicontrol and if_wi* and thought the problem was somewhere there to be found. Now, after having the versions straight, I am running 3.8-release and wi is working, after changing the #define for SIOCSWAVELAN to that which is in 3.8-current. Sorry to have wasted your time and for not providing a proper dmesg. I wasn't able to do so with the upgrade to the 10/14 snapshot because it simply wouldn't boot (and obviously wi in the snapshot does work, but I never got to try it because the kernel wouldn't boot).
Re: wi(4) or wicontrol broken
On Sun, 16 Oct 2005, Ludwig Mises wrote: > > Of course it works perfectly in -current. > > > By the way, thank you for answering this question. Will this be backported > to 3.8-stable? This not, that neither. You did not report any problem in a decent way. I'm almost 100% sure you created the problem yourself by not following the proper upgrade path. -Otto
Re: wi(4) or wicontrol broken
> Of course it works perfectly in -current. By the way, thank you for answering this question. Will this be backported to 3.8-stable?
Re: wi(4) or wicontrol broken
> The problem is you made an assumption which is false. Yes, I made an assumption. Is the snapshot of 10/14 3.8-current? If so, I'll just upgrade from my now 3.8-release system with the snapshot. I made the mistake of thinking that -release would be more recent than the snapshot, of course now I realise my folly.
Re: wi(4) or wicontrol broken
> > You tried to go back... (3.8-current -> OPENBSD_3_8_BASE which > > corresponds to 3.8-release). If you want 3.8-release, the easiest thing to > > do will be to wait for the actual release, and then install from scratch. > > > Are you saying that the snapshot of 10/14 is 3.8-current? If so, my bad. Let > me rephrase the question. Why does wicontrol no longer work in the snapshot > of 10/14? Of course it works perfectly in -current. The problem is you made an assumption which is false.
Re: wi(4) or wicontrol broken
> You tried to go back... (3.8-current -> OPENBSD_3_8_BASE which > corresponds to 3.8-release). If you want 3.8-release, the easiest thing to > do will be to wait for the actual release, and then install from scratch. Are you saying that the snapshot of 10/14 is 3.8-current? If so, my bad. Let me rephrase the question. Why does wicontrol no longer work in the snapshot of 10/14?
Re: wi(4) or wicontrol broken
On Sat, Oct 15, 2005 at 11:53:42PM -0600, Ludwig Mises wrote: > I upgraded to OPENBSD_3_8_BASE (after first upgrading to the snapshot from > 10/14) and wicontrol no longer seems to work: quoting http://www.openbsd.org/faq/faq5.html#Flavors : "One should also understand that the update process is supported in only one direction: from older to newer, and from -stable to -current. You can not run 3.7-current (or a snapshot), then decide you are living too dangerously, and step back to 3.7-stable. You are on your own if you choose any path other than the supported option of reloading your system from scratch, do not expect assistance from the OpenBSD development team." You tried to go back... (3.8-current -> OPENBSD_3_8_BASE which corresponds to 3.8-release). If you want 3.8-release, the easiest thing to do will be to wait for the actual release, and then install from scratch. -- steven Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
wi(4) or wicontrol broken
I upgraded to OPENBSD_3_8_BASE (after first upgrading to the snapshot from 10/14) and wicontrol no longer seems to work: # wicontrol wi0 wicontrol: SIOCGWAVELAN (0xfd0b): Invalid argument I notice that there was one change to sbin/wicontrol/wicontrol.c and a few changes to sys/dev/ic/if_wi.c but I'm not sure which of these is the cause of the problem. Any ideas?