I've tried getting my DEC Multia to boot off any modern Distro (CD via SCSI), all of which but one BSD had the same problem: aboot hangs as soon as it starts the kernel. Sometimes the drive gets accessed for a few minutes in random intervals, but nothing happens. Debian 2.2r3 works though, so maybe the solution can be found in it... The issue seems known to other Distro maitainers, but a) It is technically a Gentoo problem, too and b) If I got any answers at all (besides "Me, too"), it was "Use MILO instead", which is nice (and probably works, I've not had the time to try yet), but aboot should work (So maybe the Gentoo Alpha Maintainers are more inclined or capable to look into this :-P)...
I don't think that the alpha folks have ever even seen this ancient bug.
is this bug reproducible with new releases and versions? If so, please reopen including info about the system.
O.k. I'm trying a 2009 CD install now and the last messages are.... jumping to bootstrap code and it stops. Is the TGA fbdev driver compiled into the kernel on the install CD ? I'll try passing some flags to see if that helps.
Yep, passing "console=ttyS0" gets me going.
You can also use "boot -flags 1" in SRM since we have two preset kernel command lines (the second enabling serial console). Also, please don't hijack ancient bugs that probably have nothing to do with your current situation.
But it still doesn't explain why I HAVE! to use serial console. I shouldn't need to as theres'a perfectly good tgafb driver that could be used here. Anyway, given this bug was exactly the symptoms I had I used it. If you feel this is fixed by saying "use serial console" that's fine.
Well, I don't have a tga card to reproduce this bug and neither does any of the other alpha devs. And without it, debugging is pretty much impossible. Unless you want to do it yourself, which I doubt.
Hi Alan. :) (In reply to comment #6) > But it still doesn't explain why I HAVE! to use serial console. I shouldn't > need to as theres'a perfectly good tgafb driver that could be used here. This is probably true. > Anyway, given this bug was exactly the symptoms I had I used it. I don't see the similarities. The original reporter says 'aboot hangs as soon as it starts the kernel.' Your report sounds completely different. Maybe file a new bug? > If you feel this is fixed by saying "use serial console" that's fine. I don't think it's fixed, but I can't imagine it's worth too much time to debug Multia-specific problems either. I guess it could be a cast of tgafb not being in the install CD kernel. Tobias?
Mmm. Quick to judge. I actually wrote the original XFree86/X.Org TGA driver, so I'm more than willing to debug.
(In reply to comment #8) > Hi Alan. :) > > (In reply to comment #6) > > But it still doesn't explain why I HAVE! to use serial console. I shouldn't > > need to as theres'a perfectly good tgafb driver that could be used here. > > This is probably true. > > > Anyway, given this bug was exactly the symptoms I had I used it. > > I don't see the similarities. The original reporter says 'aboot hangs as soon > as it starts the kernel.' Your report sounds completely different. Maybe file a > new bug? No, because aboot did appear to hang because there's nothing displayed on the TGA console. > > If you feel this is fixed by saying "use serial console" that's fine. > > I don't think it's fixed, but I can't imagine it's worth too much time to debug > Multia-specific problems either. Why ? > I guess it could be a cast of tgafb not being in the install CD kernel. Tobias? Which is what I suspected in the original post.
>> I don't think it's fixed, but I can't imagine it's worth too much time to debug >> Multia-specific problems either. > > Why ? Well, mainly because very few people still have Multias and use them for anything (fewer than have EV6s). None of us have one, and even if I did have one, I'd, for instance, be more interested in fixing a bug that affects EV6+ or UP1500 users than a Multia bug.
Well, the fact that Gentoo doesn't boot out of the box on Multia's maybe putting people off using it. I have one and am willing to fix, but as mentioned I suspect it's a simple "not compiled tgafb into install CD". If that's it, why can't we just update the CD with this. I also have UP1100, UP1500 and a PWS600a that I'll be investing time in to make sure it all works again.
Oh, and I have TGA & TGA2 cards to test in these.
I just checked and we don't have CONFIG_FB_TGA=y in the live CDs kernel. I never suspected that not having that would make the console dysfunctional (I had figured if no support for the installed card was there, the VGA emulation I know of all the alphas I ever had physical access to would kick in). You learn something every day :) And sorry for being a bit gruff - mostly, old bugs are reopened by people who wouldn't know the difference between an ES40 and an Miata if it fell on their foot. Anyway, I'll make sure we'll have CONFIG_FB_TGA=y in this week's autobuild and you can test with that.
Cool, I'll most certainly test it when it's up. Meanwhile my Multia is currently installing over the serial line.
