Re: AmigaONE Debian (unstable?)
Hi, Geert Uytterhoeven wrote: Hmm, last time I used a floppy on my LongTrail, it did work (to my surprise, since a few years earlier it was broken ;-) Yes, CHRP had PC-style floppy controllers. And decent South Bridges used on PPC (e.g. W53C883) usually support 32-bit ISA DMA. The South Bridge on the AmigaOne is a VIA686B. It's a common PC component, and it implements an i8259 ISA DMA controller, which can't do DMA above 16MB. I think it _might_ have an option for 32 Bit ISA DMA, but due to the docs being so poor, and no other information being available, it's probably easier to do what I believe x86 Linux does on similar hardware - use ISA DMA below 16MB only. Regards, Ross..
Re: AmigaONE Debian (unstable?)
On Tue, Aug 03, 2004 at 03:16:21PM +1000, Ross Vumbaca wrote: Hi, Geert Uytterhoeven wrote: Hmm, last time I used a floppy on my LongTrail, it did work (to my surprise, since a few years earlier it was broken ;-) Yes, CHRP had PC-style floppy controllers. And decent South Bridges used on PPC (e.g. W53C883) usually support 32-bit ISA DMA. The South Bridge on the AmigaOne is a VIA686B. It's a common PC component, and it implements an i8259 ISA DMA controller, which can't do DMA above 16MB. Hmm, the DMA controller is called the 8237. The 8259 is the PIC, aka Painful Interrupt Controller. Once upon a while, Intel introduced a couple of PCI/ISA bridge with enhanced DMA capabilities: - the 82378ZB with scatter/gather DMA with 32 bit addressing capability - the 82379AB with an I/O APIC (meaning Awfully Painful...) and another kind of enhanced DMA without scatter/gather but with guess what ... 27 bit (yes!) addressing capability. If you want it, I still have the pdf from March 1996 on my disk. Note that 27 bit was not bad since very few machines at the time could even have such a tremendous (128MB) amount of RAM. The PreP specification said that the ISA bridge should be the Intel 82378ZB, so Intel stopped producing it and no other Intel bridges since then have to my knowledge implemented the same ISA DMA engine. Only Winbond (again to my knowledge) copied it in the 83C553 and 83C554 bridges. I think it _might_ have an option for 32 Bit ISA DMA, but due to the docs being so poor, and no other information being available, it's probably easier to do what I believe x86 Linux does on similar hardware - use ISA DMA below 16MB only. I have Winbond 83C553 or 83C554 on my boards, but I have no devices that use ISA DMA (well I could connect a floppy but what for?). Regards, Gabriel
Re: AmigaONE Debian (unstable?)
Hi, Gabriel Paubert wrote: Hmm, the DMA controller is called the 8237. The 8259 is the PIC, aka Painful Interrupt Controller. Sorry! You are right, I always get those to numbers confused ;). Once upon a while, Intel introduced a couple of PCI/ISA bridge with enhanced DMA capabilities: - the 82378ZB with scatter/gather DMA with 32 bit addressing capability - the 82379AB with an I/O APIC (meaning Awfully Painful...) and another kind of enhanced DMA without scatter/gather but with guess what ... 27 bit (yes!) addressing capability. If you want it, I still have the pdf from March 1996 on my disk. Note that 27 bit was not bad since very few machines at the time could even have such a tremendous (128MB) amount of RAM. The Articia data sheet specifies only an i8237, and the registers it lists are what the original i8237 had, but I'd be interested in the pdf file, to see where abouts the extra registers are placed (maybe it will help me make sense of the brief mention of the 32bit DMA registers). The PreP specification said that the ISA bridge should be the Intel 82378ZB, so Intel stopped producing it and no other Intel bridges since then have to my knowledge implemented the same ISA DMA engine. Only Winbond (again to my knowledge) copied it in the 83C553 and 83C554 bridges. Our board isn't a PReP. I think it's supposed to be a POP. I have Winbond 83C553 or 83C554 on my boards, but I have no devices that use ISA DMA (well I could connect a floppy but what for?). A floppy is still a good fall back for certain situations.. Regards, Ross..
Re: AmigaONE Debian (unstable?)
Quoting Sven Luther [EMAIL PROTECTED]: On Mon, Aug 02, 2004 at 09:31:42PM +0200, Turbo Fredriksson wrote: Quoting Sven Luther [EMAIL PROTECTED]: What Debian packages? ;) The debian kernel packages. There was/is no kernel packages! There's only a kernel and a root image: Well, there should be. Doh! Where have you BEEN the last few days? Irritating people and 'not-reading' mails? What have I been stating over and over again in this tread!?!? Well, you could have forwarded me a working patch, so ... I'm just going to kill-file you now. -- Marxist domestic disruption nuclear Waco, Texas Noriega president Legion of Doom Semtex Qaddafi Albanian BATF ammunition 747 subway plutonium [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
On Mon, Aug 02, 2004 at 06:37:27PM +1000, Ross Vumbaca wrote: Hi, Sven Luther wrote: Bah, that means that the A1 people will again do their own hack for debian support probably. Ah well, i don't really care. There is no A1 people. There is just myself, Ken, and Turbo who own AmigaOnes, AFAIK. Everyone else who ever did anything to do with Linux on the A1, no longer does it. So, please avoid the errors of your forgoers, and contribute those changes back to debian, even if it really is too late for sarge. MAI does seem to do Linux stuff in-house, but since they are not particularly communicative and do not appear to be interested in Debian or in sharing their stuff, I wouldn't put much stock in it.. He. Btw, my quick and dirty port of the Debian Woody installer to the AmigaOne was a hack, but I did do it AGES ago, when there was nothing on the AmigaOne (except an old version of SuSe), and the only kernel we had was the quick and dirty 2.4.19. I didn't get paid to do that work, and I didn't get any free hardware to do it either (unlike some people), so give me a break. Ah, i understand and am not blaming you. But anyway, i told here numerous times what would be the proper way to have your stuff integrated in debian, and nothing ever happened that way. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
Hi, Sven Luther wrote: So, please avoid the errors of your forgoers, and contribute those changes back to debian, even if it really is too late for sarge. I did, if you take a look at the debian-bf list, a short while after I produced the patches for the Woody installer (last year). But no one was really interested in it. Regarding the kernel, it was never in a state to move upstream.. Regards, Ross..
Re: AmigaONE Debian (unstable?)
Hi, Turbo Fredriksson wrote: There's a number of stuff that I'm not sure about... Remember that the A1 version of 2.4.22 is based on 2.4.22-ben1 or -ben2. That was 2.4.21-benh not 2.4.22. 2.4.22 they supposedly based off a stock kernel, but I kinda doubt that too (no idea what they based it off). Regards, Ross..
Re: AmigaONE Debian (unstable?)
Hi, Turbo Fredriksson wrote: Finally, i will be releasing a 2.4.26 powerpc kernel package next week, so please provide a diff against this one (kernel-source + powerpc patch). Fair enough. I'll do 2.4.25 packages localy to test the kernel with, and when you release 2.4.26 I'll 'port' everything to that... I took a quick look at the patches.. The true AmigaOne patches are actually very tiny, and anything to do with Macs can be ignored (especially most if not all of debian-powerpc.patch.txt). Note that my hack/patch to floppy.h that you had in the 2.4.22 diff, needs to be used if you want a working floppy drive. The PPC version of floppy.h seemed to be rather Apple specific, and dumped all the necessary stuff to use PC floppy controllers without DMA (there is a problem using the ISA DMA with the floppy on A1). Regards, Ross..
Re: AmigaONE Debian (unstable?)
On Mon, Aug 02, 2004 at 07:01:43PM +1000, Ross Vumbaca wrote: Hi, Sven Luther wrote: So, please avoid the errors of your forgoers, and contribute those changes back to debian, even if it really is too late for sarge. I did, if you take a look at the debian-bf list, a short while after I produced the patches for the Woody installer (last year). But no one was bf was dead once woody released. It should even have died before it, but d-i was not ready in time for woody. It is a real mess, and it is no wonder nobody likes to even look at it. really interested in it. Regarding the kernel, it was never in a state to move upstream.. Yeah, but what about the debian packages ? And you could have asked for help to clean it up, couldn't you ? Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Sun, Aug 01, 2004 at 11:03:32AM +0200, Turbo Fredriksson wrote:
Quoting Sven Luther [EMAIL PROTECTED]:
The 'linux-kernel-di-powerpc' package from the 'debian installer' will
depend on this (and the 'kernel-{image,modules}-2.4.25-amigaone' packages
which will in turn be built using the source kernel package above) to make
the installer aware of AmigaONE.
Ok, so what is your plan ? Build your own forked debian-installer, or
contribute your changes back to the main debian kernel ?
Naturaly this is a semi-official Debian GNU/Linux effort! EVERYTHING will end
up in 'official distribution'. That's the whole point.
Yeah, but you suffer from the let's make it perfect before we merge it in
syndrom.
Also notice that powerpc is phasing out the 2.4.x kernels in favor of the
2.6.x ones, which will be installable by default, so if you would be able to
forward patch those, it would be really great, and we may even consider
adding
them to the debian kernel package, altough i fear what Christoph will have
to
say from a 1.8MB patch. Is all this really necessary ?
The only part that's relevant is a patch that's ~28Kb. The 1.8Mb you (might
have)
seen at the URL was faulty. It contained a lot of .orig and .rej that I've
forgot
to remove from the 'old' directory, but DID remove from the 'new'...
Porting this to 2.6 shouldn't be IMPOSSIBLE. Have no 'intention' to do this at
this moment. If everything works OK with 2.4, then I might take on the task
on porting to 2.6, but I rather have a 'real kernel hacker' do this for me :)
Please, try the 2.6 port, and abandon the 2.4 stuff. This will most assuredly
never go into the upstream 2.4 kernel, and both debian and upstream powerpc
kernel developers want to abandon 2.4 in favor of 2.6 kernels.
I had a quick look at it, and there is two remarks i can do :
1) please call those amigaone-2.4.25.diff, as they should.
I'm already doing that. Just not on the URL shown, but in the package (and/or
the patch you'll get 'eventually' - i.e. when I've had a chanse to test it for
real :).
Ok.
2) it would be nice to remove all those .orig files cluttering the patch,
and you may bring the size down a bit. Also, please resolve :
linux-2.4.25.PowerPC/arch/ppc/boot/Makefile.rej
Sorry, forgot that.
Also, are those patches against the kernel-source-2.4.25 package, or against
the kernel-source + kernel-patch-2.4.25-powerpc package ?
The latter... However, there's the 'amigaone-powerpc.patch.txt' which I'm not
sure about...
My main problem with those stuff was the fear that they might include the
whole of the linuxppc_2_4 patches also, like the early peg1 patches did.
The raw/semiraw AmigaONE patch 'created' a powerpc patch, where (a large)
parts
of this is availible in the kernel-patch-2.4.25-powerpc ppc patch. At the URL,
Yeah, exactly. Lot of work to extract from it a few handfull of the actual
patches.
I've put the missing parts in the file amigaone-powerpc.patch.txt. I'm not
sure
exactly what that is (yet), but it doesn't look AmigaONE specific. More like a
'backport' of some kind... It also contain two .rej files, which I was not
able
to resolve (quickly). That's how I ended up with the conlusion that it
was/must
be a backport..
If someone regognise them, good. Othervise I'll have to dig through kernel
source(s) to see if I can find where they originate...
Maybe just use those parts you identified, and do the porting again with them,
using the bigger patch as reference. Ususally works better this way.
Still, please try out 2.6.7+ directly, it should be much easier and
worthwhile.
Finally, i will be releasing a 2.4.26 powerpc kernel package next week, so
please provide a diff against this one (kernel-source + powerpc patch).
Fair enough. I'll do 2.4.25 packages localy to test the kernel with, and when
you release 2.4.26 I'll 'port' everything to that...
Ok.
Everything here is already done localy. I just have get my AmigaONE
installed
so I can start building packages for the PPC :)
Bah, please don't go the same way this happened last year, and have a
thought
at providing the packages back upstream, altough i believe it may well be
too
late for the sarge release by now, at least for debian-installer which is
very
near the release.
I was 'donated' this machine in the sole purpose in getting OFFICIAL support
for
the AmigaONE in Debian GNU/Linux. Holding back on ANYTHING will void our
agreement.
So, why didn't you do so ? In all this past year i never saw a patch and
bugreport from you ? And it took me from december to around now to get proper
d-i support for the pegasos.
I alredy expected that it would be to late for sarge, no worries. My main goal
is to get it SUPPORTED, not in getting it supported NOW...
Ok.
Btw, what happened last year that you refer to?
there was already a thread about amigaone, and nothing
Re: AmigaONE Debian (unstable?)
Hi, Sven Luther wrote: Yeah, but what about the debian packages ? And you could have asked for help to clean it up, couldn't you ? What Debian packages? ;) I worked on this with others in late 2002, early 2003. At that time, there were other people working on this, so it was assumed it would get cleaned up by us. I then went on a long holiday of a few months since I just finished uni. I came back in late 2003, and not much had progressed, with time people left, and I wasn't particularly interested anymore, since I was working on AmigaOS4. Remember - I only ever did this in my spare time, I never got paid, I never received hardware inducements, I was under no agreement to produce anything. Anyway enough of that talk, hopefully something fruitful can be done now... Regards, Ross..
Re: AmigaONE Debian (unstable?)
On Mon, Aug 02, 2004 at 07:07:14PM +1000, Ross Vumbaca wrote: if you want a working floppy drive. The PPC version of floppy.h seemed to be rather Apple specific, and dumped all the necessary stuff to use PC floppy controllers without DMA (there is a problem using the ISA DMA with the floppy on A1). Macs don't use the normal Linux floppy driver at all, but rather the swim3 driver.
Re: AmigaONE Debian (unstable?)
Quoting Ross Vumbaca [EMAIL PROTECTED]: Sven Luther wrote: So, please avoid the errors of your forgoers, and contribute those changes back to debian, even if it really is too late for sarge. I did, if you take a look at the debian-bf list, a short while after I produced the patches for the Woody installer (last year). But no one was really interested in it. Regarding the kernel, it was never in a state to move upstream.. This is not the first time 'they' have ignored 'important patches'. I wrote RAID installation on the root partition two-three years ago. Debian STILL (!!) don't got working RAID root installation working... And yes, I'm aware that D-I have support for this, but it isn't working. -- Albanian spy cryptographic security 747 attack Peking colonel NORAD Nazi domestic disruption Delta Force ammonium arrangements FBI [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
Quoting Sven Luther [EMAIL PROTECTED]: What Debian packages? ;) The debian kernel packages. There was/is no kernel packages! There's only a kernel and a root image: majorskan:/mnt/chroots/ppc-a1/image# find -type f ./boot/boot.cat ./boot/boot.img ./dists/woody/main/disks-powerpc/current/amigaone/drivers.tgz ./dists/woody/main/disks-powerpc/current/amigaone/images-1.44/rescue.bin ./dists/woody/main/disks-powerpc/current/amigaone/rescue.bin ./kernel/kernel.img ./kernel/modules.tgz This is it! NOTHING else have ever been produced! This is the full content of Ross' boot CD. That's why I'm here. To make this 'propperly' and officially. Create EVERYTHING that's ever going to be inside Debian GNU/Linux. This includes (but not limited to) debian installer support and kernel package(s). You said you wanted to have it ready to move upstream, meaning the upstream kernel trees, i guess. This will never happen for 2.4 kernels now. And then there was amiga-fdisk. I'm only interessted in 2.4 (at this moment). That's what's bootable, I'll leave kernel hacking to those that feel urged to do this. I'll get working on 2.6 when/if I/we have everything else... -- Waco, Texas nitrate plutonium Noriega security counter-intelligence fissionable critical Ft. Meade Treasury ammonium explosion FSF Iran Qaddafi [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
Quoting Sven Luther [EMAIL PROTECTED]: Naturaly this is a semi-official Debian GNU/Linux effort! EVERYTHING will end up in 'official distribution'. That's the whole point. Yeah, but you suffer from the let's make it perfect before we merge it in syndrom. True enough. Maybe that's a flaw, but I reccon the PPC people @ kernel.org won't accept anything in the state it's in now. I was 'donated' this machine in the sole purpose in getting OFFICIAL support for the AmigaONE in Debian GNU/Linux. Holding back on ANYTHING will void our agreement. So, why didn't you do so ? In all this past year i never saw a patch and bugreport from you ? And it took me from december to around now to get proper d-i support for the pegasos. Did what? Produce patches? Didn't you read the thread? I got it three days ago! -- smuggle FSF terrorist Rule Psix Panama $400 million in gold bullion president congress bomb AK-47 spy security Khaddafi Nazi 767 [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
On Mon, Aug 02, 2004 at 09:31:42PM +0200, Turbo Fredriksson wrote: Quoting Sven Luther [EMAIL PROTECTED]: What Debian packages? ;) The debian kernel packages. There was/is no kernel packages! There's only a kernel and a root image: Well, there should be. majorskan:/mnt/chroots/ppc-a1/image# find -type f ./boot/boot.cat ./boot/boot.img ./dists/woody/main/disks-powerpc/current/amigaone/drivers.tgz ./dists/woody/main/disks-powerpc/current/amigaone/images-1.44/rescue.bin ./dists/woody/main/disks-powerpc/current/amigaone/rescue.bin ./kernel/kernel.img ./kernel/modules.tgz This is it! NOTHING else have ever been produced! This is the full content of Ross' boot CD. Well, you could have forwarded me a working patch, so ... That's why I'm here. To make this 'propperly' and officially. Create EVERYTHING that's ever going to be inside Debian GNU/Linux. This includes (but not limited to) debian installer support and kernel package(s). Then please do the 2.6 port of the kernel. You said you wanted to have it ready to move upstream, meaning the upstream kernel trees, i guess. This will never happen for 2.4 kernels now. And then there was amiga-fdisk. I'm only interessted in 2.4 (at this moment). That's what's bootable, I'll leave kernel hacking to those that feel urged to do this. I'll get working on 2.6 when/if I/we have everything else... :(. As said, 2.6 will be the default powerpc kernel for sarge and beyond, so ... Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Mon, Aug 02, 2004 at 09:24:34PM +0200, Turbo Fredriksson wrote: Quoting Ross Vumbaca [EMAIL PROTECTED]: Sven Luther wrote: Bah, that means that the A1 people will again do their own hack for debian support probably. Ah well, i don't really care. There will be NO 'debian only hack'! Not from me. And everything I do or receive will (!) one way or the other end up where it (should) belong. Well, nowhere then, since your patch will never be accepted in the 2.4 official trees. The only reason why the's been talk about Debian in this regard, is because I'm a Debian GNU/Linux developer, I use Debian GNU/Linux on EVERYTHING (servers and workstations and what not - if I dared, I'd use it on my Palm :). It's quite natural that I'll do this as a Debian GNU/Linux project. So, what is the problem ? You can have those patches in debian, and thus support from debian-installer, and then either also submit them upstream, or in the case of 2.6 kernels, work with the debian-kernel team to help you get them in shape and submitted upstream. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
Quoting Ken Moffat [EMAIL PROTECTED]: There's a number of stuff that I'm not sure about... Remember that the A1 version of 2.4.22 is based on 2.4.22-ben1 or -ben2. Ah. That explain quite a lot I guess. I already suspected a 'backport' of some kind, but a 'fork' is more or less the same thing - a nuisance :) Umm, looks like you've got Configure.help.orig in your non-A1 tree, probably from a failed patch, and that forms 99% of this file - I think you've put up the wrong stuff here. Taken care of - see thread... -- FBI Legion of Doom Ft. Bragg Nazi $400 million in gold bullion subway AK-47 767 Noriega domestic disruption CIA class struggle smuggle Albanian Semtex [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Ken Moffat wrote:
On Sat, 31 Jul 2004, Sven Luther wrote:
On Sat, Jul 31, 2004 at 04:24:09PM +0100, Ken Moffat wrote:
On Sat, 31 Jul 2004, Sven Luther wrote:
What graphic card and thus fbdev are you using ?
The notorious radeon 9200se, but under 2.6.7 I'm not using a console fb
yet, just a vga console. X11R6.7.0 just recognises the card and works
without any of the old magic incantations.
Interesting, i was not able to use the vgaconsole on my pegasos. Do you have
any patches for that ?
From the A1 2.4 stuff -
diff -Nuar linux-2.4.25-vanilla/drivers/video/vgacon.c
linux-2.4.25-amigaone/drivers/video/vgacon.c
--- linux-2.4.25-vanilla/drivers/video/vgacon.c 2002-08-24
13:27:39.0 +0100
+++ linux-2.4.25-amigaone/drivers/video/vgacon.c 2004-06-19
18:00:53.0 +0100
@@ -212,7 +212,11 @@
else/* If not, it is color. */
{
vga_can_do_color = 1;
+#ifdef CONFIG_AMIGAONE
+ vga_vram_base = 0xFD0B8000; /* Mapped by the startup */
+#else
vga_vram_base = 0xb8000;
+#endif
vga_video_port_reg = 0x3d4;
vga_video_port_val = 0x3d5;
if ((ORIG_VIDEO_EGA_BX 0xff) != 0x10)
That's not the correct fix. VGA_MAP_MEM() should take into account the ISA
memory space offset.
And in fact it does, cfr. include/asm-ppc/vga.h, just make sure to initialize
vgacon_remap_base with the correct ISA memory space offset.
There's a similar patch for vga16fb.c -
Vga16fb doesn't use VGA_MAP_MEM(), it just uses ioremap() on the VGA memory.
But arch/ppc/mm/pgtable.c has a special case for mapping memory below 16 MiB,
and adds _ISA_MEM_BASE to such addresses. So you have to make sure to
initialize _ISA_MEM_BASE with the correct ISA memory space offset as well.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [EMAIL PROTECTED]
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say programmer or something like that.
-- Linus Torvalds
Re: AmigaONE Debian (unstable?)
On Sun, 1 Aug 2004, Geert Uytterhoeven wrote:
That's not the correct fix. VGA_MAP_MEM() should take into account the ISA
memory space offset.
And in fact it does, cfr. include/asm-ppc/vga.h, just make sure to initialize
vgacon_remap_base with the correct ISA memory space offset.
There's a similar patch for vga16fb.c -
Vga16fb doesn't use VGA_MAP_MEM(), it just uses ioremap() on the VGA memory.
But arch/ppc/mm/pgtable.c has a special case for mapping memory below 16 MiB,
and adds _ISA_MEM_BASE to such addresses. So you have to make sure to
initialize _ISA_MEM_BASE with the correct ISA memory space offset as well.
Gr{oetje,eeting}s,
Geert
Thanks.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
Quoting Sven Luther [EMAIL PROTECTED]:
The 'linux-kernel-di-powerpc' package from the 'debian installer' will
depend on this (and the 'kernel-{image,modules}-2.4.25-amigaone' packages
which will in turn be built using the source kernel package above) to make
the installer aware of AmigaONE.
Ok, so what is your plan ? Build your own forked debian-installer, or
contribute your changes back to the main debian kernel ?
Naturaly this is a semi-official Debian GNU/Linux effort! EVERYTHING will end
up in 'official distribution'. That's the whole point.
Also notice that powerpc is phasing out the 2.4.x kernels in favor of the
2.6.x ones, which will be installable by default, so if you would be able to
forward patch those, it would be really great, and we may even consider adding
them to the debian kernel package, altough i fear what Christoph will have to
say from a 1.8MB patch. Is all this really necessary ?
The only part that's relevant is a patch that's ~28Kb. The 1.8Mb you (might
have)
seen at the URL was faulty. It contained a lot of .orig and .rej that I've
forgot
to remove from the 'old' directory, but DID remove from the 'new'...
Porting this to 2.6 shouldn't be IMPOSSIBLE. Have no 'intention' to do this at
this moment. If everything works OK with 2.4, then I might take on the task
on porting to 2.6, but I rather have a 'real kernel hacker' do this for me :)
I had a quick look at it, and there is two remarks i can do :
1) please call those amigaone-2.4.25.diff, as they should.
I'm already doing that. Just not on the URL shown, but in the package (and/or
the patch you'll get 'eventually' - i.e. when I've had a chanse to test it for
real :).
2) it would be nice to remove all those .orig files cluttering the patch,
and you may bring the size down a bit. Also, please resolve :
linux-2.4.25.PowerPC/arch/ppc/boot/Makefile.rej
Sorry, forgot that.
Also, are those patches against the kernel-source-2.4.25 package, or against
the kernel-source + kernel-patch-2.4.25-powerpc package ?
The latter... However, there's the 'amigaone-powerpc.patch.txt' which I'm not
sure about...
The raw/semiraw AmigaONE patch 'created' a powerpc patch, where (a large) parts
of this is availible in the kernel-patch-2.4.25-powerpc ppc patch. At the URL,
I've put the missing parts in the file amigaone-powerpc.patch.txt. I'm not sure
exactly what that is (yet), but it doesn't look AmigaONE specific. More like a
'backport' of some kind... It also contain two .rej files, which I was not able
to resolve (quickly). That's how I ended up with the conlusion that it was/must
be a backport..
If someone regognise them, good. Othervise I'll have to dig through kernel
source(s) to see if I can find where they originate...
Finally, i will be releasing a 2.4.26 powerpc kernel package next week, so
please provide a diff against this one (kernel-source + powerpc patch).
Fair enough. I'll do 2.4.25 packages localy to test the kernel with, and when
you release 2.4.26 I'll 'port' everything to that...
Everything here is already done localy. I just have get my AmigaONE
installed
so I can start building packages for the PPC :)
Bah, please don't go the same way this happened last year, and have a thought
at providing the packages back upstream, altough i believe it may well be too
late for the sarge release by now, at least for debian-installer which is very
near the release.
I was 'donated' this machine in the sole purpose in getting OFFICIAL support for
the AmigaONE in Debian GNU/Linux. Holding back on ANYTHING will void our
agreement.
I alredy expected that it would be to late for sarge, no worries. My main goal
is to get it SUPPORTED, not in getting it supported NOW...
Btw, what happened last year that you refer to?
--
PLO supercomputer iodine Delta Force jihad BATF colonel Albanian 747
AK-47 Khaddafi fissionable Iran congress smuggle
[See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
Quoting Ken Moffat [EMAIL PROTECTED]:
might be because you don't have ide=nodma in the bootargs. That way
lies pain and an eventual reinstall.
We suspect a faulty memory. I'll try your solution, and we'll see...
Update: Nope, 'ida=nodma' didn't do any difference. I'll have to wait for
the new memory until I can proceed.
No kidding, was it only 1.8MB ? it felt more.
Sorry, it WAS more. Should have been 3.0Mb! That's including the CVS
directories and all the CVS key words which (it seemed) was quite easy
to filter out. Applying a bit of filtering, I ended up with 762Kb. Much
better :)
Yes, I did look through it and I've extracted the real changes and put
them at http://homepage.ntlworld.com/zarniwhoop.
I've put mine in 'digestable' chunks at http://www.bayour.com/AmigaONE/.
There's a number of stuff that I'm not sure about...
* linux-2.4.22_full.patch.txt
This is a raw diff between upstream and the CVS version.
* linux-2.4.22_semi.patch.txt
This is the un-digastable patch (special params to diff to filter out
MOST of the CVS stuff though).
* syskonnect.patch.txt
* ieee1394.patch.txt
These seems to be totaly unrelated to the AmigaONE. Don't know where
it comes from, but I guess from the 'bleeding edge upstream 2.4'...
* powerpc.patch.txt
Most of this is also availible in the 'kernel-patch-2.4.25-powerpc'
Debian GNU/Linux source package (which I thought I'd add the AmigaONE
patch(es) to).
Not all of it is rejected (after adding the Debian GNU/Linux PPC patch)
though! Maybe that's the 'bleeding edge PowerPC patch'?
* strange.patch.txt
This file is stuff that I found ... 'strange'. They look like spelling
errors etc, but I don't know for sure (yet - have to look closer)...
* amigaone-2.4.22.patch.txt
This seems to be ALL that's AmigaONE related. It applies _almost_ clean
to the linux-2.4.25 tree (which is the kernel that's used in the source
package mentioned above).
Trying to apply this on 2.4.25 shows that I've done something wrong when
I 'digested' the patch. I've corrected this below...
* amigaone-2.4.25.patch.txt
This is the AmigaONE patch modified slightly for 2.4.25... It's ONLY the
AmigaONE relevant part.
The 'linux-2.4.25.PowerPC/' directory I'm comparing against is the
original/upstream 2.4.25 kernel tree with the PowerPC patch found in the
2.4.25 ppc kernel package, NOTHING else...
The only one I'm really interested in and which will end up in the source
package mentionend above is the last one. The others could 'be nice', but
they are not relevant for this task...
Seem to work mostly ok with
2.4.25 and later (read the docs, most people who have commented get
adequate results with the work-around for disk dma. Not perfect by any
means, but it's a start). You'll need mkimage from u-boot to build the
uImage. But, you'll need to use an older kernel to install the system
before you can do this.
I've packaged part of uboot (1.0.0) - the mkimage binary. The rest seemed
irrelevant at this time.
The 'linux-kernel-di-powerpc' package from the 'debian installer' will
depend on this (and the 'kernel-{image,modules}-2.4.25-amigaone' packages
which will in turn be built using the source kernel package above) to make
the installer aware of AmigaONE.
Everything here is already done localy. I just have get my AmigaONE installed
so I can start building packages for the PPC :)
--
Qaddafi Cocaine ammonium Legion of Doom 767 nitrate security Albanian
FSF president congress kibo Honduras counter-intelligence Rule Psix
[See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
On Sat, Jul 31, 2004 at 11:45:38AM +0200, Turbo Fredriksson wrote:
Quoting Ken Moffat [EMAIL PROTECTED]:
might be because you don't have ide=nodma in the bootargs. That way
lies pain and an eventual reinstall.
We suspect a faulty memory. I'll try your solution, and we'll see...
Update: Nope, 'ida=nodma' didn't do any difference. I'll have to wait for
the new memory until I can proceed.
No kidding, was it only 1.8MB ? it felt more.
Sorry, it WAS more. Should have been 3.0Mb! That's including the CVS
directories and all the CVS key words which (it seemed) was quite easy
to filter out. Applying a bit of filtering, I ended up with 762Kb. Much
better :)
Yes, I did look through it and I've extracted the real changes and put
them at http://homepage.ntlworld.com/zarniwhoop.
I've put mine in 'digestable' chunks at http://www.bayour.com/AmigaONE/.
There's a number of stuff that I'm not sure about...
* linux-2.4.22_full.patch.txt
This is a raw diff between upstream and the CVS version.
* linux-2.4.22_semi.patch.txt
This is the un-digastable patch (special params to diff to filter out
MOST of the CVS stuff though).
* syskonnect.patch.txt
* ieee1394.patch.txt
These seems to be totaly unrelated to the AmigaONE. Don't know where
it comes from, but I guess from the 'bleeding edge upstream 2.4'...
* powerpc.patch.txt
Most of this is also availible in the 'kernel-patch-2.4.25-powerpc'
Debian GNU/Linux source package (which I thought I'd add the AmigaONE
patch(es) to).
Not all of it is rejected (after adding the Debian GNU/Linux PPC patch)
though! Maybe that's the 'bleeding edge PowerPC patch'?
* strange.patch.txt
This file is stuff that I found ... 'strange'. They look like spelling
errors etc, but I don't know for sure (yet - have to look closer)...
* amigaone-2.4.22.patch.txt
This seems to be ALL that's AmigaONE related. It applies _almost_ clean
to the linux-2.4.25 tree (which is the kernel that's used in the source
package mentioned above).
Trying to apply this on 2.4.25 shows that I've done something wrong when
I 'digested' the patch. I've corrected this below...
* amigaone-2.4.25.patch.txt
This is the AmigaONE patch modified slightly for 2.4.25... It's ONLY the
AmigaONE relevant part.
The 'linux-2.4.25.PowerPC/' directory I'm comparing against is the
original/upstream 2.4.25 kernel tree with the PowerPC patch found in the
2.4.25 ppc kernel package, NOTHING else...
The only one I'm really interested in and which will end up in the source
package mentionend above is the last one. The others could 'be nice', but
they are not relevant for this task...
Seem to work mostly ok with
2.4.25 and later (read the docs, most people who have commented get
adequate results with the work-around for disk dma. Not perfect by any
means, but it's a start). You'll need mkimage from u-boot to build the
uImage. But, you'll need to use an older kernel to install the system
before you can do this.
I've packaged part of uboot (1.0.0) - the mkimage binary. The rest seemed
irrelevant at this time.
The 'linux-kernel-di-powerpc' package from the 'debian installer' will
depend on this (and the 'kernel-{image,modules}-2.4.25-amigaone' packages
which will in turn be built using the source kernel package above) to make
the installer aware of AmigaONE.
Ok, so what is your plan ? Build your own forked debian-installer, or
contribute your changes back to the main debian kernel ?
Also notice that powerpc is phasing out the 2.4.x kernels in favor of the
2.6.x ones, which will be installable by default, so if you would be able to
forward patch those, it would be really great, and we may even consider adding
them to the debian kernel package, altough i fear what Christoph will have to
say from a 1.8MB patch. Is all this really necessary ?
I had a quick look at it, and there is two remarks i can do :
1) please call those amigaone-2.4.25.diff, as they should.
2) it would be nice to remove all those .orig files cluttering the patch,
and you may bring the size down a bit. Also, please resolve :
linux-2.4.25.PowerPC/arch/ppc/boot/Makefile.rej
Also, are those patches against the kernel-source-2.4.25 package, or against
the kernel-source + kernel-patch-2.4.25-powerpc package ?
Finally, i will be releasing a 2.4.26 powerpc kernel package next week, so
please provide a diff against this one (kernel-source + powerpc patch).
Everything here is already done localy. I just have get my AmigaONE installed
so I can start building packages for the PPC :)
Bah, please don't go the same way this happened last year, and have a thought
at providing the packages back upstream, altough i believe it may well be too
late for the sarge release by now, at least for debian-installer which is very
near the release.
Friendly,
Sven Luther
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Turbo Fredriksson wrote:
Quoting Ken Moffat [EMAIL PROTECTED]:
might be because you don't have ide=nodma in the bootargs. That way
lies pain and an eventual reinstall.
We suspect a faulty memory. I'll try your solution, and we'll see...
Update: Nope, 'ida=nodma' didn't do any difference. I'll have to wait for
the new memory until I can proceed.
Make sure you're using registered memory if you've got more than one
stick of it. If you need more advice on which memory works/doesn't work
(the old single-sided/double-sided type of issue, I think), bug me
privately - I've got some old private A1 archives I can dig through.
No kidding, was it only 1.8MB ? it felt more.
Sorry, it WAS more. Should have been 3.0Mb! That's including the CVS
directories and all the CVS key words which (it seemed) was quite easy
to filter out. Applying a bit of filtering, I ended up with 762Kb. Much
better :)
Yes, I did look through it and I've extracted the real changes and put
them at http://homepage.ntlworld.com/zarniwhoop.
I've put mine in 'digestable' chunks at http://www.bayour.com/AmigaONE/.
For 2.4, there doesn't seem a lot of point. If you look at the code, a
lot of it is somewhat rough (not a comment on the people who wrote it,
in their situation I'd probably have done something similar if I was
able to).
There's a number of stuff that I'm not sure about...
Remember that the A1 version of 2.4.22 is based on 2.4.22-ben1 or
-ben2.
* linux-2.4.22_full.patch.txt
This is a raw diff between upstream and the CVS version.
* linux-2.4.22_semi.patch.txt
This is the un-digastable patch (special params to diff to filter out
MOST of the CVS stuff though).
* syskonnect.patch.txt
* ieee1394.patch.txt
These seems to be totaly unrelated to the AmigaONE. Don't know where
it comes from, but I guess from the 'bleeding edge upstream 2.4'...
see above
* powerpc.patch.txt
Most of this is also availible in the 'kernel-patch-2.4.25-powerpc'
Debian GNU/Linux source package (which I thought I'd add the AmigaONE
patch(es) to).
Not all of it is rejected (after adding the Debian GNU/Linux PPC patch)
though! Maybe that's the 'bleeding edge PowerPC patch'?
* strange.patch.txt
This file is stuff that I found ... 'strange'. They look like spelling
errors etc, but I don't know for sure (yet - have to look closer)...
again, I put these down to using the 2.4-ppc tree instead of Marcelo's.
* amigaone-2.4.22.patch.txt
This seems to be ALL that's AmigaONE related. It applies _almost_ clean
to the linux-2.4.25 tree (which is the kernel that's used in the source
package mentioned above).
Trying to apply this on 2.4.25 shows that I've done something wrong when
I 'digested' the patch. I've corrected this below...
* amigaone-2.4.25.patch.txt
This is the AmigaONE patch modified slightly for 2.4.25... It's ONLY the
AmigaONE relevant part.
Umm, looks like you've got Configure.help.orig in your non-A1 tree,
probably from a failed patch, and that forms 99% of this file - I think
you've put up the wrong stuff here.
The 'linux-2.4.25.PowerPC/' directory I'm comparing against is the
original/upstream 2.4.25 kernel tree with the PowerPC patch found in the
2.4.25 ppc kernel package, NOTHING else...
The only one I'm really interested in and which will end up in the source
package mentionend above is the last one. The others could 'be nice', but
they are not relevant for this task...
I'm not overly interested in the debian kernel packages for the A1, at
least not until the A1 is in the main kernel. However, if I can sort
out the rtc stuff and track down another error which might be kernel
related, I'll be putting up some new patches for 2.4.27 (hint - use
dumpe2fs to have a look at the dates for your rootfs: I've gone beyond
1903 now, but I'm managing to trash the year back to 1984). At that
point I'll probably revisit my rescue CD (also on my zarniwhoop page).
Seem to work mostly ok with
2.4.25 and later (read the docs, most people who have commented get
adequate results with the work-around for disk dma. Not perfect by any
means, but it's a start). You'll need mkimage from u-boot to build the
uImage. But, you'll need to use an older kernel to install the system
before you can do this.
I've packaged part of uboot (1.0.0) - the mkimage binary. The rest seemed
irrelevant at this time.
That's all you need. 1.0.0 is old, but I imagine that part hasn't
changed.
The 'linux-kernel-di-powerpc' package from the 'debian installer' will
depend on this (and the 'kernel-{image,modules}-2.4.25-amigaone' packages
which will in turn be built using the source kernel package above) to make
the installer aware of AmigaONE.
Everything here is already done localy. I just have get my AmigaONE installed
so I can start building packages for the PPC :)
Ken
--
das eine Mal als Tragödie, das andere Mal als
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Sven Luther wrote: Also notice that powerpc is phasing out the 2.4.x kernels in favor of the 2.6.x ones, which will be installable by default, so if you would be able to forward patch those, it would be really great, and we may even consider adding them to the debian kernel package, altough i fear what Christoph will have to say from a 1.8MB patch. Is all this really necessary ? For 2.6 I'm down to about 200KB, which I think includes some of the original A1 files that I'm not using (e.g. _time merged into _setup as a preliminary to using todc_time). This is without e.g. the floppy stuff and the so-called dma fixes. If it can be made to work, I guess 200KB or less (which is still rather a lot). Bah, please don't go the same way this happened last year, and have a thought at providing the packages back upstream, altough i believe it may well be too late for the sarge release by now, at least for debian-installer which is very near the release. Friendly, Sven Luther Agreed, if it doesn't eventually get merged then it isn't worth doing. I've heard the people at mai are issuing a binary 2.6.6 to their testers, whoever they might be, but they aren't exactly communicative. My own attempts on 2.6.7 are at a very early stage and have certain difficulties (page-up in 'less' locks it, coming out of X gives a non-legible screen, and the rtc stuff isn't working properly yet). But then, I was never cut out to be a kernel hacker. Looks like I'm unlikely to get back to 2.6 before September, unfortunately. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Sat, Jul 31, 2004 at 03:14:01PM +0100, Ken Moffat wrote: On Sat, 31 Jul 2004, Sven Luther wrote: Also notice that powerpc is phasing out the 2.4.x kernels in favor of the 2.6.x ones, which will be installable by default, so if you would be able to forward patch those, it would be really great, and we may even consider adding them to the debian kernel package, altough i fear what Christoph will have to say from a 1.8MB patch. Is all this really necessary ? For 2.6 I'm down to about 200KB, which I think includes some of the original A1 files that I'm not using (e.g. _time merged into _setup as a preliminary to using todc_time). This is without e.g. the floppy stuff and the so-called dma fixes. If it can be made to work, I guess 200KB or less (which is still rather a lot). Ok. Still too much though, but i guess there is no chance of it changing. Bah, please don't go the same way this happened last year, and have a thought at providing the packages back upstream, altough i believe it may well be too late for the sarge release by now, at least for debian-installer which is very near the release. Friendly, Sven Luther Agreed, if it doesn't eventually get merged then it isn't worth doing. I've heard the people at mai are issuing a binary 2.6.6 to their testers, whoever they might be, but they aren't exactly communicative. He. My own attempts on 2.6.7 are at a very early stage and have certain difficulties (page-up in 'less' locks it, coming out of X gives a non-legible screen, and the rtc stuff isn't working properly yet). What graphic card and thus fbdev are you using ? But then, I was never cut out to be a kernel hacker. Looks like I'm Bah, it is like that that one learns. I started the same way, cleaning up the old POP patches. unlikely to get back to 2.6 before September, unfortunately. Too late for sarge anyway, but tell us when you have progress. That said, consider using the debian-kernel mailing list for this kind of stuff. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Sven Luther wrote: On Sat, Jul 31, 2004 at 03:14:01PM +0100, Ken Moffat wrote: On Sat, 31 Jul 2004, Sven Luther wrote: Also notice that powerpc is phasing out the 2.4.x kernels in favor of the 2.6.x ones, which will be installable by default, so if you would be able to forward patch those, it would be really great, and we may even consider adding them to the debian kernel package, altough i fear what Christoph will have to say from a 1.8MB patch. Is all this really necessary ? For 2.6 I'm down to about 200KB, which I think includes some of the original A1 files that I'm not using (e.g. _time merged into _setup as a preliminary to using todc_time). This is without e.g. the floppy stuff and the so-called dma fixes. If it can be made to work, I guess 200KB or less (which is still rather a lot). Ok. Still too much though, but i guess there is no chance of it changing. Bah, please don't go the same way this happened last year, and have a thought at providing the packages back upstream, altough i believe it may well be too late for the sarge release by now, at least for debian-installer which is very near the release. Friendly, Sven Luther Agreed, if it doesn't eventually get merged then it isn't worth doing. I've heard the people at mai are issuing a binary 2.6.6 to their testers, whoever they might be, but they aren't exactly communicative. He. My own attempts on 2.6.7 are at a very early stage and have certain difficulties (page-up in 'less' locks it, coming out of X gives a non-legible screen, and the rtc stuff isn't working properly yet). What graphic card and thus fbdev are you using ? The notorious radeon 9200se, but under 2.6.7 I'm not using a console fb yet, just a vga console. X11R6.7.0 just recognises the card and works without any of the old magic incantations. But then, I was never cut out to be a kernel hacker. Looks like I'm Bah, it is like that that one learns. I started the same way, cleaning up the old POP patches. unlikely to get back to 2.6 before September, unfortunately. Too late for sarge anyway, but tell us when you have progress. That said, consider using the debian-kernel mailing list for this kind of stuff. Actually, I'm mostly not a debian user, I'm only really here for help with my iBook. In the meantime, anything I do with the A1 kernel will initially be announced on the a1linux group at yahoo, and at amigaworld.net, I don't see any point cluttering this list with it until there is something more usable for people to look at. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Sat, Jul 31, 2004 at 04:24:09PM +0100, Ken Moffat wrote: On Sat, 31 Jul 2004, Sven Luther wrote: On Sat, Jul 31, 2004 at 03:14:01PM +0100, Ken Moffat wrote: On Sat, 31 Jul 2004, Sven Luther wrote: Also notice that powerpc is phasing out the 2.4.x kernels in favor of the 2.6.x ones, which will be installable by default, so if you would be able to forward patch those, it would be really great, and we may even consider adding them to the debian kernel package, altough i fear what Christoph will have to say from a 1.8MB patch. Is all this really necessary ? For 2.6 I'm down to about 200KB, which I think includes some of the original A1 files that I'm not using (e.g. _time merged into _setup as a preliminary to using todc_time). This is without e.g. the floppy stuff and the so-called dma fixes. If it can be made to work, I guess 200KB or less (which is still rather a lot). Ok. Still too much though, but i guess there is no chance of it changing. Bah, please don't go the same way this happened last year, and have a thought at providing the packages back upstream, altough i believe it may well be too late for the sarge release by now, at least for debian-installer which is very near the release. Friendly, Sven Luther Agreed, if it doesn't eventually get merged then it isn't worth doing. I've heard the people at mai are issuing a binary 2.6.6 to their testers, whoever they might be, but they aren't exactly communicative. He. My own attempts on 2.6.7 are at a very early stage and have certain difficulties (page-up in 'less' locks it, coming out of X gives a non-legible screen, and the rtc stuff isn't working properly yet). What graphic card and thus fbdev are you using ? The notorious radeon 9200se, but under 2.6.7 I'm not using a console fb yet, just a vga console. X11R6.7.0 just recognises the card and works without any of the old magic incantations. Interesting, i was not able to use the vgaconsole on my pegasos. Do you have any patches for that ? That said, apparently X11R6.7.0 don't really do the right thing. But then, I was never cut out to be a kernel hacker. Looks like I'm Bah, it is like that that one learns. I started the same way, cleaning up the old POP patches. unlikely to get back to 2.6 before September, unfortunately. Too late for sarge anyway, but tell us when you have progress. That said, consider using the debian-kernel mailing list for this kind of stuff. Actually, I'm mostly not a debian user, I'm only really here for help with my iBook. In the meantime, anything I do with the A1 kernel will initially be announced on the a1linux group at yahoo, and at amigaworld.net, I don't see any point cluttering this list with it until there is something more usable for people to look at. Bah, that means that the A1 people will again do their own hack for debian support probably. Ah well, i don't really care. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Sven Luther wrote:
On Sat, Jul 31, 2004 at 04:24:09PM +0100, Ken Moffat wrote:
On Sat, 31 Jul 2004, Sven Luther wrote:
What graphic card and thus fbdev are you using ?
The notorious radeon 9200se, but under 2.6.7 I'm not using a console fb
yet, just a vga console. X11R6.7.0 just recognises the card and works
without any of the old magic incantations.
Interesting, i was not able to use the vgaconsole on my pegasos. Do you have
any patches for that ?
From the A1 2.4 stuff -
diff -Nuar linux-2.4.25-vanilla/drivers/video/vgacon.c
linux-2.4.25-amigaone/drivers/video/vgacon.c
--- linux-2.4.25-vanilla/drivers/video/vgacon.c 2002-08-24 13:27:39.0
+0100
+++ linux-2.4.25-amigaone/drivers/video/vgacon.c2004-06-19
18:00:53.0 +0100
@@ -212,7 +212,11 @@
else/* If not, it is color. */
{
vga_can_do_color = 1;
+#ifdef CONFIG_AMIGAONE
+ vga_vram_base = 0xFD0B8000; /* Mapped by the startup */
+#else
vga_vram_base = 0xb8000;
+#endif
vga_video_port_reg = 0x3d4;
vga_video_port_val = 0x3d5;
if ((ORIG_VIDEO_EGA_BX 0xff) != 0x10)
Dunno if the mapping is the same on the pegasus.
There's a similar patch for vga16fb.c -
diff -Nuar linux-2.4.25-vanilla/drivers/video/vga16fb.c
linux-2.4.25-amigaone/drivers/video/vga16fb.c
--- linux-2.4.25-vanilla/drivers/video/vga16fb.c2001-12-05
22:01:07.0 +
+++ linux-2.4.25-amigaone/drivers/video/vga16fb.c 2004-06-19
18:00:45.0 +0100
@@ -32,7 +32,11 @@
#define dac_reg(0x3c8)
#define dac_val(0x3c9)
+#ifdef CONFIG_AMIGAONE
+#define VGA_FB_PHYS 0xFD0A
+#else
#define VGA_FB_PHYS 0xA
+#endif
#define VGA_FB_PHYS_LEN 65536
/* - */
Actually, I'm mostly not a debian user, I'm only really here for help
with my iBook. In the meantime, anything I do with the A1 kernel will
initially be announced on the a1linux group at yahoo, and at
amigaworld.net, I don't see any point cluttering this list with it until
there is something more usable for people to look at.
Bah, that means that the A1 people will again do their own hack for debian
support probably. Ah well, i don't really care.
Friendly,
Sven Luther
Point taken, I'll try to follow debian-kernel when I get back to the
A1.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Sat, 31 Jul 2004, Ken Moffat wrote: * amigaone-2.4.25.patch.txt This is the AmigaONE patch modified slightly for 2.4.25... It's ONLY the AmigaONE relevant part. Umm, looks like you've got Configure.help.orig in your non-A1 tree, probably from a failed patch, and that forms 99% of this file - I think you've put up the wrong stuff here. Sorry, drop my unwarranted assertion about a failed patch (just found I got .orig files here after patching up 2.4.27-rc4), but this file _is_ spurious. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Thu, 29 Jul 2004, Turbo Fredriksson wrote:
I got my AmigaOne yesterday, and I'm currently trying to install it with
the ISO @ SourceForge. However, I'm having some troubles with the IDE disk,
cable or something completley different - 'device not ready for command' or
something like that.
might be because you don't have ide=nodma in the bootargs. That way
lies pain and an eventual reinstall.
Quoting Sven Luther [EMAIL PROTECTED]:
I think the kernel patch for those is more big than the pegasos one, and
i am not entirely sure in which subarch category the AmigaOne falls.
There's a kernel CVS module ('linux-2.4.22') at the SourceForge site and
diff'ing that with the original 2.4.22 kernel source tree gives me a 1.8Mb
patch. Most of this is just '$Id: $' changes...
I sent a mail to Ross about this, I'll see what that can give us. If nothing
else, I'll extract the relevant part myself. Don't want to do that - 1.8Mb
patch is A LOT to look through :)
No kidding, was it only 1.8MB ? it felt more. Yes, I did look through
it and I've extracted the real changes and put them at
http://homepage.ntlworld.com/zarniwhoop. Seem to work mostly ok with
2.4.25 and later (read the docs, most people who have commented get
adequate results with the work-around for disk dma. Not perfect by any
means, but it's a start). You'll need mkimage from u-boot to build the
uImage. But, you'll need to use an older kernel to install the system
before you can do this.
I guess a complete new one. |Turbo, how do you boot linux on amigaone, and
where do you get the kernel image for it. What is the status about
initrd booting ?
To simply everything for me, I boot the boot image through TFTP. Booting
from CD, Floppy and HD is also supported (and VERY simmilar to TFTP booting).
For an initrd, you can put both the kernel and the initrd into the
uImage. However, since the maximum size for an eltorito image is 2.88MB
you can't get very much in there.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
I got my AmigaOne yesterday, and I'm currently trying to install it with
the ISO @ SourceForge. However, I'm having some troubles with the IDE disk,
cable or something completley different - 'device not ready for command' or
something like that.
Quoting Sven Luther [EMAIL PROTECTED]:
I think the kernel patch for those is more big than the pegasos one, and
i am not entirely sure in which subarch category the AmigaOne falls.
There's a kernel CVS module ('linux-2.4.22') at the SourceForge site and
diff'ing that with the original 2.4.22 kernel source tree gives me a 1.8Mb
patch. Most of this is just '$Id: $' changes...
I sent a mail to Ross about this, I'll see what that can give us. If nothing
else, I'll extract the relevant part myself. Don't want to do that - 1.8Mb
patch is A LOT to look through :)
I guess a complete new one. |Turbo, how do you boot linux on amigaone, and
where do you get the kernel image for it. What is the status about
initrd booting ?
To simply everything for me, I boot the boot image through TFTP. Booting
from CD, Floppy and HD is also supported (and VERY simmilar to TFTP booting).
--
Panama radar strategic BATF Cuba Clinton critical Honduras Nazi
Qaddafi spy FBI PLO tritium KGB
[See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
On Thu, Jul 29, 2004 at 06:21:17PM +0200, Turbo Fredriksson wrote:
I got my AmigaOne yesterday, and I'm currently trying to install it with
BAh, you should have gotten a pegasos 2 with full 2.6.7 kernel and
debian-installer support :)
the ISO @ SourceForge. However, I'm having some troubles with the IDE disk,
cable or something completley different - 'device not ready for command' or
something like that.
Mmm. I cannot help you much, the only amigaone i ever saw was oopsing about
disk stuff, but this was one year ago and probably due to deffective hardware
(broken board, bad memory, bad ide cable, what else do i know ?).
Quoting Sven Luther [EMAIL PROTECTED]:
I think the kernel patch for those is more big than the pegasos one, and
i am not entirely sure in which subarch category the AmigaOne falls.
There's a kernel CVS module ('linux-2.4.22') at the SourceForge site and
diff'ing that with the original 2.4.22 kernel source tree gives me a 1.8Mb
patch. Most of this is just '$Id: $' changes...
Yeah, i had to fight with this kind of nonsense when i first started the
pegasos 1 patch. Finally of the mulit-megabit patch one double a4 page of
usefull patch resulted. But then, the pegasos being a chrp, it is much nearer
to already supported hardware.
I sent a mail to Ross about this, I'll see what that can give us. If nothing
else, I'll extract the relevant part myself. Don't want to do that - 1.8Mb
patch is A LOT to look through :)
Meld is your friend here. and probably there is some diff option to ginore the
Id lines.
I guess a complete new one. |Turbo, how do you boot linux on amigaone, and
where do you get the kernel image for it. What is the status about
initrd booting ?
To simply everything for me, I boot the boot image through TFTP. Booting
from CD, Floppy and HD is also supported (and VERY simmilar to TFTP booting).
Err, from what kernel source you get it, and where do you get the compiled
version, not to speak about the debian installation initrd, but i guess you
still use those adapted boot-floppies.
Friendly,
Sven Luther
Re: AmigaONE Debian (unstable?)
On Tue, 2004-06-08 at 03:32, Ole-Egil Hvitmyren wrote: 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. Hehehehehe, that's getting funny :) (Note that I don't know anything about the specific problem you are dealing with here ...) Ok, so Linux do support non-coherent DMA quite well, it's atcually widely used by various sorts of embedded PPC CPUs like 4xx, 8xx, ... _HOWEVER_, doing that in a northbridge for anything but an embedded CPU, especially a CPU of the 6xx/7xxx family is just insane. It's basically incompetent northbridge design. Linux uses BATs to map the entire linar memory aperture, on those CPUs, which is a significant performance gain, and allows to simplify some low level memory management issues. However, for various reasons I can explain separately, that means that on those CPU, we cannot easily map arbitrary pieces of memory as non-cacheable in a reliable way (and yes, that is a problem with AGP on some machines). 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. No, that means the HW is a Piece Of Shit ! 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 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. I'm not from any of these but I've been one of the maintainers of the PPC kernel for long enough to have a chance to play with a wide range of northbridges. Cache coherency is a basic feature of anything claiming to be used as a desktop machine. Ben.
Re: AmigaONE Debian (unstable?)
Hi, Christoph Hellwig writes: Ole-Egil Hvitmyren wrote: A significant amount of people have already written to tell me that if they had been given boards for free it would have happened. Now, would it, or would it NOT, help to have a BIT of money in that regard? Although IBM has a lot of money they haven't sent me a single piece of hardware either ;-) You can all have a free RS/6000-250 from me if you want. Regards, Jens. -- J'qbpbe, le m'en fquz pe j'qbpbe! Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
Re: AmigaONE Debian (unstable?)
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. 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. 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 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. 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 ;-) 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... 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. 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 patch I made to this email so you can also compare to what you did in 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 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. -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/ amiga-fdisk-0.04-bootstrapcode-patch.gz Description: application/gzip
Re: AmigaONE Debian (unstable?)
Ole-Egil Hvitmyren wrote: and it should be compilable with GCC 3.3. Oooh, I didn't notice that. It's been fixed :-) -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
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
Re: AmigaONE Debian (unstable?)
Hi, Jens Schmalzing wrote: Turbo Fredriksson writes: Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? Implement it. If the kernel source needs patching, try to get the patch into the main kernel.org tree, or the Debian kernel-source package, or the Debian kernel-patch-powerpc package. If there is a kernel-image package already that should work (-powerpc, I presume?) but doesn't, find out how the .config itself, or the support scripts mkinitrd and mkvmlinuz need to be changed. Massive thread ensues Hey guys, I've worked quite a lot with the A1 and Linux. I see that there is a bit of confusion in this thread, so I'll mention/clear some points (hopefully). First, the A1 is a PPC board, and there is a whole web site with info, pictures etc about the board here http://amigaone-linux.sf.net. The web site was set up for AmigaOne specific kernel development (the plan was that the patches would then move upstream). Note that the information is slightly out of date. The source tree for the AmigaOne is present at that web site. (Yeah we need to make a patch against a stock tree). The firmware of the AmigaOne is known as UBoot. This used to be known as PPCBoot. It is open source firmware. The web site is http://uboot.sf.net The majority of AmigaOne owners that run Linux on their board, run Debian. This is because I made an unofficial port of the Debian Woody bf installer for the AmigaOne early last year, and apart from that (and an old SuSe version), there wasn't anything else until recently (there is also a YDL installer now, but it is un-maintained and will gradually grow out of date). I sent my patches to the bf dev list, but no one seemed particuarly interested, no matter. At the time that the installer was built, there was no boot loader, so the AmigaOne would boot Linux in a similar way to the PReP PPC machines. This is the best that UBoot can offer. You merge the kernel (and optionally a ramdisk) into a file, create a special image (using a tool mkimage), and you dd it to a partition. This has vastly changed now. An AmigaOS4 developer added Linux booting support to the OS4 boot loader. So there is a boot loader for the AmigaOne that can boot Linux by reading a kernel image off an ext2/ext3 partition. The boot loader uses some functions from GRUB, and will be open source too. Stuff like yaboot does not work of course, since we don't have an OF. One condition of the boot loader is that it only operates with RDB partitions - no big deal, that's the Amiga partition scheme. For those that don't want Amiga partition schemes, UBoot can be used with the old method (i.e like a PReP machine). It would be nice to get 'official' support for the AmigaOne into Debian though, and I'm happy to take part in co-operation with Turbo et al. Regards, Ross..
Re: AmigaONE Debian (unstable?)
Hi, Sven Luther wrote: BTW, i am curious, what is the latest status of the AmigaOne thingy, i almost don't hear about those anymore, and since there were problems with the buggy northbridge, i had a feeling that production stopped or something such, but then my info comes mostly from anti-amigaone guys, like you can imagine, so i may be misinformed. Production of the boards didn't stop, and they even developed a new type of board known as the uA1 (micro A1). It is a Mini-ITX board, with PowerPC CPU, etc etc. Quite a cool piece of hardware, there are no other PowerPC Mini-ITX products AFAIK, also Mini-ITX boards are usually a low end VIA CPU. These boards are using relatively powerful CPUs such as the IBM 750FX @ 800MHZ. This is all well known known information. Are your MOS colleagues not in touch with reality? ;) Regards, Ross..
Re: AmigaONE Debian (unstable?)
On Wed, Jun 09, 2004 at 03:04:14AM +1000, Ross Vumbaca wrote: Hi, Sven Luther wrote: BTW, i am curious, what is the latest status of the AmigaOne thingy, i almost don't hear about those anymore, and since there were problems with the buggy northbridge, i had a feeling that production stopped or something such, but then my info comes mostly from anti-amigaone guys, like you can imagine, so i may be misinformed. Production of the boards didn't stop, and they even developed a new type of board known as the uA1 (micro A1). It is a Mini-ITX board, with PowerPC CPU, etc etc. Quite a cool piece of hardware, there are no other PowerPC Mini-ITX products AFAIK, also Mini-ITX boards are usually a low end VIA CPU. These boards are using relatively powerful CPUs such as the IBM 750FX @ 800MHZ. This is all well known known information. Are your MOS colleagues not in touch with reality? ;) Nope, but i usually don't concern myself with this stuff, already keeping up with the actuality of the debian community is quite time consuming, that i usually refrain from interesting myself with this whole sad brotherly war. Oh well, let's close this parenthesis, we are disgressing anyway. Friendly, Sven Luther
AmigaONE Debian (unstable?)
Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? There's a woody install/boot CD (basically a special kernel - base kernel with AmigaONE patch?) on http://amigaone-linux.sourceforge.net/. From what I remember, all one have to do is packaging the kernel flavor, and make sure that the boot/install CD is including this when built... -- Iran bomb ammonium 747 Cocaine fissionable BATF jihad explosion CIA Kennedy Cuba Rule Psix class struggle security [See http://www.aclu.org/echelonwatch/index.html for more about this]
Re: AmigaONE Debian (unstable?)
Hi, Turbo Fredriksson writes: Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? Implement it. If the kernel source needs patching, try to get the patch into the main kernel.org tree, or the Debian kernel-source package, or the Debian kernel-patch-powerpc package. If there is a kernel-image package already that should work (-powerpc, I presume?) but doesn't, find out how the .config itself, or the support scripts mkinitrd and mkvmlinuz need to be changed. Regards, Jens. -- J'qbpbe, le m'en fquz pe j'qbpbe! Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
Re: AmigaONE Debian (unstable?)
Jens Schmalzing wrote: Hi, Turbo Fredriksson writes: Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? Implement it. If the kernel source needs patching, try to get the patch into the main kernel.org tree, or the Debian kernel-source package, or the Debian kernel-patch-powerpc package. If there is a kernel-image package already that should work (-powerpc, I presume?) but doesn't, find out how the .config itself, or the support scripts mkinitrd and mkvmlinuz need to be changed. The firmware is U-Boot, so the kernel needs a bit of patching. Most of the guys working on the Linux kernel got tied up in writing stuff for AmigaOS4. It seems to me that unless one happens to be IBM or one of the other big players getting more than a few lines into any of the places you mention is quite an undertaking :-( U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) Things I know that needs a patch: kernel-package (subarch in script plus new config-file. I've been using kernel-package on this subarch for almost a year, so this _shouldn't_ be difficult to implement ;-) ) kernel-patch-powerpc (unfortunately, we don't have anything later than 2.4.22 at the moment. Trying to follow the kernel development with a small team and no inside contacts feels like ice skating uphill) debian-installer (at least we had to make our own version of bf, I presume d-i won't accept an unknown machine from /proc/cpuinfo either?) amiga-fdisk (new extensions to RDB standard means running amiga-fdisk on the harddrive will disable dual-booting. I can provide the necessary bits. Quite small patch, actually) mkuboot (or something along those lines, mkimage from U-Boot packaged so people can run it on their own machines instead of mkvmlinuz). So the question basically boils down to this: Where is the right place to START? Btw, does anyone here know what excactly is responsible for setting correct timestamp on /proc? It comes up with a date some time in 1903 even though the clock is otherwise correct here, which causes havoc in certain places (for instance if machine was not shut down properly then /var/lock is never cleaned properly on bootup, and therefore xdm refuses to start). Debugging kernels is hard enough as it is without crap like that ;-) -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 11:51:43AM +0200, Ole-Egil Hvitmyren wrote: The firmware is U-Boot, so the kernel needs a bit of patching. Most of the guys working on the Linux kernel got tied up in writing stuff for AmigaOS4. It seems to me that unless one happens to be IBM or one of the other big players getting more than a few lines into any of the places you mention is quite an undertaking :-( Not really. I don't remember having seen your patch on linux-kernel or linuxppc-devel either. Send me a pointer to the patches and I'll review then and try to get them in. If you actually want the stuff maintained in newer releases I'd need some hardware, though ;-)
Re: AmigaONE Debian (unstable?)
Christoph Hellwig wrote: On Mon, Jun 07, 2004 at 11:51:43AM +0200, Ole-Egil Hvitmyren wrote: The firmware is U-Boot, so the kernel needs a bit of patching. Most of the guys working on the Linux kernel got tied up in writing stuff for AmigaOS4. It seems to me that unless one happens to be IBM or one of the other big players getting more than a few lines into any of the places you mention is quite an undertaking :-( Not really. I don't remember having seen your patch on linux-kernel or linuxppc-devel either. Send me a pointer to the patches and I'll review then and try to get them in. 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. If you actually want the stuff maintained in newer releases I'd need some hardware, though ;-) Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) As I am on the list, please do NOT CC me on replies :-) -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
Hi, Ole-Egil Hvitmyren writes: kernel-patch-powerpc (unfortunately, we don't have anything later than 2.4.22 at the moment. Trying to follow the kernel development with a small team and no inside contacts feels like ice skating uphill) Well, a number of people that can be considered inside contacts follow this list. mkuboot (or something along those lines, mkimage from U-Boot packaged so people can run it on their own machines instead of mkvmlinuz). The idea of mkvmlinuz is to have something that takes an uncompressed kernel and a ramdisk, and flanges the two together to produce a single bootable binary for a given PowerPC sub-architecture. It should be possible to integrate mkimage into this. So the question basically boils down to this: Where is the right place to START? Forward port your 2.4 patch to 2.6, I would say. Then split it into digestable chunks and try to get them in mainline. Regards, Jens. -- J'qbpbe, le m'en fquz pe j'qbpbe! Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
Re: AmigaONE Debian (unstable?)
The firmware is U-Boot, so the kernel needs a bit of patching. Most of the guys working on the Linux kernel got tied up in writing stuff for AmigaOS4. It seems to me that unless one happens to be IBM or one of the other big players getting more than a few lines into any of the places you mention is quite an undertaking :-( Nah, it shouldbn't be that bad. Also, I've proposed them in the past to deal with it if they sent me a board, but they have been trying to bullshit me into buying the board from them and them paying me back based on time spent or some kind of crappy agreement where I would have been fucked in the end, so I'm not eager to help on this one. U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) That's crap. Fine for embedded maybe, but it should really be able to load a plain vmlinux. Things I know that needs a patch: kernel-package (subarch in script plus new config-file. I've been using kernel-package on this subarch for almost a year, so this _shouldn't_ be difficult to implement ;-) ) kernel-patch-powerpc (unfortunately, we don't have anything later than 2.4.22 at the moment. Trying to follow the kernel development with a small team and no inside contacts feels like ice skating uphill) I maintain the port to the whole PowerMac line alone. Bullshit. So the question basically boils down to this: Where is the right place to START? The right place to start is to dig out the kernel patches, port them to 2.6 and submit them to me or paulus (if possible via linuxppc-dev mailing list) for review and possible merge upstream. Ben.
Re: AmigaONE Debian (unstable?)
Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) That has nothing to do with big players at all. Ben.
Re: AmigaONE Debian (unstable?)
Hi, Benjamin Herrenschmidt writes: U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) That's crap. Fine for embedded maybe, but it should really be able to load a plain vmlinux. Does that include a ramdisk? My understanding was that you need either a bootloader or a bootable image like the one describer above in order to boot an initrd kernel. On sub-architectures that don't have a bootloader, such as PReP or CHRP, this leaves only the bootable image. Regards, Jens. -- J'qbpbe, le m'en fquz pe j'qbpbe! Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
Re: AmigaONE Debian (unstable?)
On Mon, 2004-06-07 at 11:11, Jens Schmalzing wrote: Hi, Benjamin Herrenschmidt writes: U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) That's crap. Fine for embedded maybe, but it should really be able to load a plain vmlinux. Does that include a ramdisk? My understanding was that you need either a bootloader or a bootable image like the one describer above in order to boot an initrd kernel. On sub-architectures that don't have a bootloader, such as PReP or CHRP, this leaves only the bootable image. Can't uboot load the ramdisk separately like yaboot can ? It's also possible to extend on the existing miboot.image which is parseable and contains both kernel ramdisk, rather than defining yet another image format ... Ben;
Re: AmigaONE Debian (unstable?)
Hi, Benjamin Herrenschmidt writes: Can't uboot load the ramdisk separately like yaboot can ? Duh, I must have misunderstood this then. So uboot is the bootloader for AmigaOne? Then it should be taught to load an uncompressed kernel and a separate ramdisk from a real root filesystem. Also, it should be packaged for Debian, probably in a separate package from the other PowerPC bootloaders. Regards, Jens. -- J'qbpbe, le m'en fquz pe j'qbpbe! Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
Re: AmigaONE Debian (unstable?)
Jens Schmalzing wrote: Hi, Benjamin Herrenschmidt writes: Can't uboot load the ramdisk separately like yaboot can ? Duh, I must have misunderstood this then. So uboot is the bootloader for AmigaOne? Then it should be taught to load an uncompressed kernel and a separate ramdisk from a real root filesystem. Also, it should be packaged for Debian, probably in a separate package from the other PowerPC bootloaders. We One item of note. In the case of the AmigaONE it's burned into EEPROM. (Least it was when I was playing around with it which was some time ago). These boxes originally had an implementation of OF which was .. well it wasn't quite perfect, but it could load yaboot at least... You are quite right tho, it would be desireable for it to know how to load a kernel and ramdisk from a file system. Regards, Tom Regards, Jens.
Re: AmigaONE Debian (unstable?)
Benjamin Herrenschmidt wrote: Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) That has nothing to do with big players at all. Are you saying I have more money than IBM? I would certainly hope not ;-) Seriously, I think Eyetech and MAI are being complete assholes on this one, but I would like to add that it would have been really cool if someone from the outside had half a chance in heck of figuring all of this out and actually get there in time to have a patch for something that is semi-current. Every time I manage to make something that looks like a patch for some version of Linux that version is obsoleted by something that breaks whatever I did. I may not be the smartest guy around, but at least I'm man enough to face the fact. I am not the right person to ask this of :-) I guess more manpower is the only way to go. Anyway, I have done a bit of work on two of the 4 or 5 (can't remember, it's been HOURS already :-) ) points I wrote down earlier, possibly enough to have it ready in less than 5 minutes (amiga-fdisk and kernel-package). The rest of the points are somewhat more work, but I'm going to contribute if there's more than me out there. At least it is a START, right? :-) But I guess the number of people with AmigaOnes who actually bothered to subscribe here is pretty low, even if I did hint a few _thousand_ times elsewhere that this was a good list for debian and powerpc related questions. -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
Jens Schmalzing wrote: Hi, Benjamin Herrenschmidt writes: Can't uboot load the ramdisk separately like yaboot can ? Duh, I must have misunderstood this then. So uboot is the bootloader for AmigaOne? Then it should be taught to load an uncompressed kernel and a separate ramdisk from a real root filesystem. Also, it should be packaged for Debian, probably in a separate package from the other PowerPC bootloaders. No, Ben is the one who misunderstood. U-Boot is FIRMWARE, not BOOTLOADER. Putting OS specific filesystem code into FIRMWARE on a board that is currently running MORE than one operating system is more or less the definition of optimism. Please, please, please do not waste any more time discussing that option. It's already crammed up with x86 emulators and whatnots ;-) -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 10:35:54PM +0200, Ole-Egil Hvitmyren wrote: Benjamin Herrenschmidt wrote: Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) That has nothing to do with big players at all. Are you saying I have more money than IBM? I would certainly hope not ;-) Are you saying you can get a patch included by paying money?
Re: AmigaONE Debian (unstable?)
Jens Schmalzing wrote: Hi, Benjamin Herrenschmidt writes: U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) That's crap. Fine for embedded maybe, but it should really be able to load a plain vmlinux. Does that include a ramdisk? My understanding was that you need either a bootloader or a bootable image like the one describer above in order to boot an initrd kernel. On sub-architectures that don't have a bootloader, such as PReP or CHRP, this leaves only the bootable image. It isn't actually a bootable image, it's just the vmlinux (and optional ramdisk) with a header that tells the firmware (U-Boot) how to interpret the data. For instance, what sort of compression has been used, whether or not it includes a ramdisk (or you could have a ramdisk in a separate file, no problem). The resulting file can then be loaded into memory from partition (RDB or PC partition map), network, etc. From memory the file can be booted. So it looks and feels a lot like PREP to the user, but it is in fact a lot simpler. We're already running mkvmlinuz on our kernels already, I fail to see how another subarch that looks almost like PREP can be considered so bad. Yes, we are currently lacking a bootloader that understands filesystems AND can load Linux. Which means you can load AmigaOS4 from a filesystem, or you can load Linux from a partition much like on my PReP. The bootloader that is used when booting AmigaOS4 behaves much like the typical x86 bootloaders, in that it has a second stage in the partition map (RDB extension called BRB, boot record block) which is loaded by a minimal loader in the actual firmware. This second stage then understands the AmigaOS4 filesystem (called FFS2, as in second fast-filesystem). so far there has been much talk and little action around the load file from inside ext2 school of design. Unfortunately I am not the right man for that kind of action. There is a multitude of boards out there that use U-Boot (it supports ARM, MIPS, PPC and x86 already), so I think support for this will have to come sooner or later anyway. Or is there some rule that says no embedded-like boards are to be supported that I missed? :-) Here's the source to mkimage.c: http://cvs.sourceforge.net/viewcvs.py/u-boot/u-boot/tools/mkimage.c?rev=1.12view=auto To my quick and shifty eyes it seemed quite straightforward, but I'm just in from a long bikeride so will make a SLIGHT reservation on that one. It has a reference to (#include image.h near the top) http://cvs.sourceforge.net/viewcvs.py/u-boot/u-boot/include/image.h?rev=1.9view=auto and it uses crc32 functions from: http://cvs.sourceforge.net/viewcvs.py/u-boot/u-boot/lib_generic/crc32.c?rev=1.3view=auto But since that one includes common.h which again includes almost half of U-Boot I think we'll stop there :-) -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
Christoph Hellwig wrote: On Mon, Jun 07, 2004 at 10:35:54PM +0200, Ole-Egil Hvitmyren wrote: Benjamin Herrenschmidt wrote: Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) That has nothing to do with big players at all. Are you saying I have more money than IBM? I would certainly hope not ;-) Are you saying you can get a patch included by paying money? I think you're reading a bit much into a small joke which even had a smiled on it. A significant amount of people have already written to tell me that if they had been given boards for free it would have happened. Now, would it, or would it NOT, help to have a BIT of money in that regard? I cannot BELIEVE how silly this discussion got, and so quickly as well. It's likely to be me, so I'll just shut the fuck up now. Good bye. -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 11:16:40PM +0200, Ole-Egil Hvitmyren wrote: I think you're reading a bit much into a small joke which even had a smiled on it. A significant amount of people have already written to tell me that if they had been given boards for free it would have happened. Now, would it, or would it NOT, help to have a BIT of money in that regard? I cannot BELIEVE how silly this discussion got, and so quickly as well. It's likely to be me, so I'll just shut the fuck up now. Good bye. Although IBM has a lot of money they haven't sent me a single piece of hardware either ;-) Okay, let's stop it now..
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 10:04:52PM +0200, Christoph Hellwig wrote: I cannot BELIEVE how silly this discussion got, and so quickly as well. It's likely to be me, so I'll just shut the fuck up now. Good bye. Although IBM has a lot of money they haven't sent me a single piece of hardware either ;-) Okay, let's stop it now.. Well, not yet. My offer to review the patches when ported to 2.6 still stands. Without hardware I simply can't gurantee it stays working so someone with hardware will have to regularly test it and if needed fix it.
Re: AmigaONE Debian (unstable?)
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. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Mon, 7 Jun 2004, Ole-Egil Hvitmyren wrote: U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) The mainline 2.4 kernel is already patched to use mkimage to create a uImage, it's just getting the right code in for the kernel that matters. I've put a patch to compile u-boot with gcc-3.3.1+ on my A1 page at http://homepage.ntlworld.com/zarniwhoop. Btw, does anyone here know what excactly is responsible for setting correct timestamp on /proc? It comes up with a date some time in 1903 even though the clock is otherwise correct here, which causes havoc in certain places (for instance if machine was not shut down properly then /var/lock is never cleaned properly on bootup, and therefore xdm refuses to start). Debugging kernels is hard enough as it is without crap like that ;-) Heh, I'd forgotten about that one. Supposedly, it comes down to what the bootscripts are doing and when (i.e. I'm told it doesn't happen on the patched YDL). What's really amusing is that the invalid time (negative, at a guess) is moving forward, but in all conscience this is a minor detail of getting a good kernel. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Mon, 7 Jun 2004, Christoph Hellwig wrote: Not really. I don't remember having seen your patch on linux-kernel or linuxppc-devel either. Send me a pointer to the patches and I'll review then and try to get them in. If you actually want the stuff maintained in newer releases I'd need some hardware, though ;-) Can I ask the idiot's question ( where can I subscribe to linuxppc-devel ) please ? Google seems to think an embedded '-' is some sort of wildcard these days. Ken -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Mon, 7 Jun 2004, Ken Moffat wrote: Can I ask the idiot's question ( where can I subscribe to linuxppc-devel ) please ? Google seems to think an embedded '-' is some sort of wildcard these days. Ken Sorry, I see it's -dev, so lists.linuxppc.org -- das eine Mal als Tragödie, das andere Mal als Farce
Re: AmigaONE Debian (unstable?)
On Mon, 2004-06-07 at 15:35, Ole-Egil Hvitmyren wrote: Benjamin Herrenschmidt wrote: Which is why it would help to be IBM or one of the other big players. Does it look like I'm shitting hardware? ;-) That has nothing to do with big players at all. Are you saying I have more money than IBM? I would certainly hope not ;-) big players don't bother with board support for X Y or Z, most of that kind of stuff is volunteer work in linux. AmigaOne could have fairly decent and maintained support by just providing a board and eventually some techs infos to an interested kernel hacker. Paying him would be even better but I'm sure some would do the work just for a board or 2 ;) Seriously, I think Eyetech and MAI are being complete assholes on this one, but I would like to add that it would have been really cool if someone from the outside had half a chance in heck of figuring all of this out and actually get there in time to have a patch for something that is semi-current. Every time I manage to make something that looks like a patch for some version of Linux that version is obsoleted by something that breaks whatever I did. I may not be the smartest guy around, but at least I'm man enough to face the fact. I am not the right person to ask this of :-) I guess more manpower is the only way to go. Anyway, I have done a bit of work on two of the 4 or 5 (can't remember, it's been HOURS already :-) ) points I wrote down earlier, possibly enough to have it ready in less than 5 minutes (amiga-fdisk and kernel-package). The rest of the points are somewhat more work, but I'm going to contribute if there's more than me out there. At least it is a START, right? :-) But I guess the number of people with AmigaOnes who actually bothered to subscribe here is pretty low, even if I did hint a few _thousand_ times elsewhere that this was a good list for debian and powerpc related questions. -- AmigaOne dev list FAQ (when I say F, I mean F): http://www.samfundet.no/~olegil/amiga/ -- Benjamin Herrenschmidt [EMAIL PROTECTED]
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 09:20:14AM +0200, Jens Schmalzing wrote: Hi, Turbo Fredriksson writes: Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? Implement it. I think the kernel patch for those is more big than the pegasos one, and i am not entirely sure in which subarch category the AmigaOne falls. I guess a complete new one. |Turbo, how do you boot linux on amigaone, and where do you get the kernel image for it. What is the status about initrd booting ? Then you would need debian-installer support, and that's it, but it is maybe a bit late for going into that right now. BTW, i am curious, what is the latest status of the AmigaOne thingy, i almost don't hear about those anymore, and since there were problems with the buggy northbridge, i had a feeling that production stopped or something such, but then my info comes mostly from anti-amigaone guys, like you can imagine, so i may be misinformed. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 11:51:43AM +0200, Ole-Egil Hvitmyren wrote: Jens Schmalzing wrote: Hi, Turbo Fredriksson writes: Anyone have an idea how to get 'official' support for the AmigaONE in Debian GNU/Linux? Implement it. If the kernel source needs patching, try to get the patch into the main kernel.org tree, or the Debian kernel-source package, or the Debian kernel-patch-powerpc package. If there is a kernel-image package already that should work (-powerpc, I presume?) but doesn't, find out how the .config itself, or the support scripts mkinitrd and mkvmlinuz need to be changed. The firmware is U-Boot, so the kernel needs a bit of patching. Most of the guys working on the Linux kernel got tied up in writing stuff for AmigaOS4. It seems to me that unless one happens to be IBM or one of the other big players getting more than a few lines into any of the places you mention is quite an undertaking :-( Well, if you never submit patches, there is serious little chance that they will get accepted. I still remember those third party amiga-fdisk packages that never got contributed back in any way for example. U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) See, another thing that could have been usefull for everyone involved and was hidden. Things I know that needs a patch: kernel-package (subarch in script plus new config-file. I've been using kernel-package on this subarch for almost a year, so this _shouldn't_ be difficult to implement ;-) ) kernel-patch-powerpc (unfortunately, we don't have anything later than 2.4.22 at the moment. Trying to follow the kernel development with a small team and no inside contacts feels like ice skating uphill) At least 2.4.25/26 should be there, there will be no older kernel version in sarge. debian-installer (at least we had to make our own version of bf, I presume d-i won't accept an unknown machine from /proc/cpuinfo either?) Yep, but it is in rc status now, january was the right time to work on this. amiga-fdisk (new extensions to RDB standard means running amiga-fdisk on the harddrive will disable dual-booting. I can provide the necessary bits. Quite small patch, actually) amiga-fdisk was last modified in 2001 or so, and your changes never contributed back. But if you want debian-installer support, you need parted. I believe that parted amiga partition support i wrote is already amigaone friendly, never tested, obviously, as there is no amiaos4 yet, and the only amigaone i saw was not even able to boot debian without oopsing, but then it was probably a early devel board or something. mkuboot (or something along those lines, mkimage from U-Boot packaged so people can run it on their own machines instead of mkvmlinuz). Why not contribute the code necessary to mkvmlinuz ? So the question basically boils down to this: Where is the right place to START? A bit late i believe. All the above need to be done, but it is hardly shortly before the sarge release that it is the right thing for that. Now, just start submitting bug reports with patches for the above problems would be a good start. About d-i, sincerely i wish you luck, i believe that you probably need many weeks if not month of work to have it done, especially at this stage and due to the modular development style of d-i. But please go ahead. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
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. Friendly, Sven Luther
Re: AmigaONE Debian (unstable?)
On Mon, Jun 07, 2004 at 06:11:11PM +0200, Jens Schmalzing wrote: Hi, Benjamin Herrenschmidt writes: U-Boot comes with a tool that does more or less the same as mkvmlinuz, called mkimage (takes in an elf and an optional ramdisk image and outputs a bootable image) That's crap. Fine for embedded maybe, but it should really be able to load a plain vmlinux. Does that include a ramdisk? My understanding was that you need either a bootloader or a bootable image like the one describer above in order to boot an initrd kernel. On sub-architectures that don't have a bootloader, such as PReP or CHRP, this leaves only the bootable image. Supposedly, a yaboot2 should be able to work without problems on all of those, once it is ready though. Friendly, Sven Luther
