Re: AmigaONE && Debian (unstable?)
On Tue, Jun 08, 2004 at 10:32:06AM +0200, Ole-Egil Hvitmyren wrote:
> Sven Luther wrote:
> >On Mon, Jun 07, 2004 at 11:14:28PM +0100, Ken Moffat wrote:
> >
> >>On Mon, 7 Jun 2004, Ole-Egil Hvitmyren wrote:
> >>
> >>
> >>>Since I'm no kernel hacker I'm not the guy to bug about this. Like I
> >>>said, I can help with patches to kernel-package and amiga-fdisk. Kernel
> >>>is not my field. I know MAI themselves were working on porting the 2.6
> >>>kernel, but I haven't heard a single word from them for almost a year.
> >>>
> >>
> >>Supposedly, it was two or three months from being sent to private
> >>testers in January, but we both know about the personnel changes since
> >>then. The 2.4 port still has issues with DMA which make it unready for
> >>mainstream.
> >
> >
> ><troll mode>
> >huh, and i thought it was the hardware who had a DMA problem.
> ></troll mode>
> >
> >Sorry, couldn't resist :), no offense intented, i think this whole issue
> >is more of a painfull mess, but let's not talk about this here.
>
> I don't see ANY damn reason why you need to drag that debate in here,
> but let's just say certain other OSes doesn't have a problem marking DMA
> buffers as non-cacheable.
Well, that is hardly any proof that there is no problem, just that the
other OSes (with an S) somehow work around the problem. My pegasos 1
has uptime of months, and is serving both to download my mail and as my
wife's primary workstation, running X and and gnome and other such
stuff. But then, it uses the hardware workaround and didn't need any
above tricks that i know of.
> It's stated pretty explicitly in the northbridge documentation that this
> is how it needs to work, saying the hardware is buggy because it follows
> its own documentation seems a TAD silly to me.
Well, it doesn't say so in the doc i have, so ...
> But you can't help it, can you? Every time someone mentions the AmigaOne
> it has to be "not very stable the time (two years ago, was it?) I saw
Well, it was in October, but as said, i believe it may have been because
a bad board, or something such, since i know other people (like you) are
running linux just fine on it.
> it" and so on. Yes, we've had some Linux problems. Quite a few,
> actually. But most of that seems to come from MAI and Eyetech wanting to
> get everything for free, and not doing anything to actively support the
> development of the Linux kernel.
Well, that may be one reason, but i guess it is not the only one.
> Yes, I know at least some things about the personell changes since then.
> But I didn't know they lost their entire kernel development division ;-)
I didn' even know that, nor do i care. , so sorry for the disgression,
and back to serious things ...
> And about the amiga-fdisk patch. I'll send it soon enough. But since it
> doesn't affect classic Amiga how is it to be verified? And considering
> your own bug report is now 255 days old (and counting), what good do you
> suppose it will do? I did have a discussion with cts some time in
> October, but he couldn't test it then, so...
Well, that build is an example of what you should not do. I fixed this
in oldenbourg in october, but somehow forgot to send in the patch, and
later lost my disk and was not worried, thinking it was ok, that
amiga-fdisk has it already patched. But then, parted is the important
one, so ...
> And technically I shouldn't be the one to file the bug-report, since I
> don't even have OS4 running myself, yet. Hopefully I'll have a CD within
> a few days. It would be nice to get it properly tested.
Therein lies your error. If you don't do it, who will, probably nobody.
> About parted:
> I can take a quick look. The patch should be fairly trivial (just one
> more type of block to avoid deleting, and one pointer in the RDSK that
> needs to be updated if the RDB is to be restructured). I attached the
If it is just the boot list thing, i added those, after discussion with
you or/and some other guy, if i rememebr well. I never had feedback on
if it workeed though.
> patch I made to this email so you can also compare to what you did in
Ok, i will have a look.
> parted. While I was at it I did put in some more work on amiga-fdisk,
> but maybe parted is a better place to spend time. At the very least
Yep, since debian-installer use libparted, and miga-fdisk is no more on
the initrd, i believe.
> amiga-fdisk should be compiled for more than m68k and ppc (I see an old
> outstanding on that one. I myself use it with the emulator "Amithlon" on
> my triple-boot PC here in the office from time to time), and it should
> be compilable with GCC 3.3.
It is, i sent at least that fix, and i also remember using it without
problem on my athlon box. In fact my initial pegasos linux iinstall used
amiga-fdisk on my x86 to partition an old scsi disk i had around, copy
the kernel on it, move it to the pegasos, and try running it. no tftp
booting at that time, like we have now. That made that i was the first
one (and maybe the only one) bit by the bug in amiga-fdisk whichused a
16bit half-word to store the sector size when calculating the
cyl/head/block size. And this almost 5 year after the last upstream
release. I guess nobody used amiga-fdisk to create a fully blank
partition table, so it never happened. And the authors, Roman or Geert i
think, mostly tested on older small disk where this was no problem.
Friendly,
Sven Luther
Reply to: