NetBSD/Amiga

[strim]

Originally Posted by ahhyes

I'm using the compact flash drive from amigakit, does this still apply?

MaxTransfer might apply, depending on a CF card used. Some cards need it, some don't.

NetBSD makes no use of this setting (it doesn't need it, as the ata/wd driver is sane enough to pick a proper value automatically, in contrast to AmigaOS scsi.device). It is important to have it set correctly during transfer of miniroot (on AmigaOS).

BTW. As far as I know Mask value is unused on IDE controllers even on AmigaOS...

[Rotzloeffel]

Originally Posted by strim

MaxTransfer might apply, depending on a CF card used. Some cards need it, some don't.

NetBSD makes no use of this setting (it doesn't need it, as the ata/wd driver is sane enough to pick a proper value automatically, in contrast to AmigaOS scsi.device). It is important to have it set correctly during transfer of miniroot (on AmigaOS).

Yes! It is VERY important so ensure a correct transfer of the extracted miniroot.fs and the other problems he told us

Originally Posted by strim

BTW. As far as I know Mask value is unused on IDE controllers even on AmigaOS...

IHMO you should set it! I have no idea because it only tells the controller to use chip-Ram 24-bit DMA or other Ram and as far as I read, there is no known device wich needs chip or 24-bit instead of any Ram.

[strim]

Originally Posted by Rotzloeffel

IHMO you should set it! I have no idea because it only tells the controller to use chip-Ram 24-bit DMA or other Ram

Amiga IDE controllers don't use DMA at all, that's why this value doesn't make any sense for them .

and as far as I read, there is no known device wich needs chip or 24-bit instead of any Ram.

I can imagine you could have Zorro II DMA SCSI controller with badly written driver that would try to do DMA to Fast outside 24-bit area. Then you would use the mask value to prevent it from accessing this Fast RAM...

[ahhyes]

Originally Posted by strim

Amiga IDE controllers don't use DMA at all, that's why this value doesn't make any sense for them ..

So then it makes no difference whether I set this value or not?

[Rotzloeffel]

Originally Posted by strim

Amiga IDE controllers don't use DMA at all, that's why this value doesn't make any sense for them .

I can imagine you could have Zorro II DMA SCSI controller with badly written driver that would try to do DMA to Fast outside 24-bit area. Then you would use the mask value to prevent it from accessing this Fast RAM...

Aha! Thanks for the explanation! But never heard about such Controllers.

BTW! Good morning strim

[Rotzloeffel]

Originally Posted by ahhyes

So then it makes no difference whether I set this value or not?

set it to the value I wrote or leave it as it actually is. The maxTransfer value is the value you must change!

And remember! a allready extracted miniroot.fs on your harddrive and also the copied adf-files and all others are eventually corrupt! You need to extract it again from a CD-Rom to ensure that it is not corrupt.

[ahhyes]

ahh i had gotten mixed up with mask and xfer (tired after work. lol). my bad. I had set both of these to zero when i created the partitions.maybe that is why i am having problems.

[Rotzloeffel]

Originally Posted by ahhyes

ahh i had gotten mixed up with mask and xfer (tired after work. lol). my bad. I had set both of these to zero when i created the partitions.maybe that is why i am having problems.

I am sure that this is your actual problem...... With computers it is the same as in life! Don´t drink and drive

[ahhyes]

Originally Posted by Rotzloeffel

I am sure that this is your actual problem...... With computers it is the same as in life! Don´t drink and drive

What about drinking and operating a computer *raises bier stein*

[Toni Wilen]

Max transfer requirement is not drive/CF bug, it is ROM built-in IDE scsi.device "feature" (requires ATA-1 behavior that was removed in ATA-2, note that some drives still do support ATA-1 behavior)

http://eab.abime.net/showpost.php?p=759918&postcount=25

[Rotzloeffel]

Originally Posted by Toni Wilen

Max transfer requirement is not drive/CF bug, it is ROM built-in IDE scsi.device "feature" (requires ATA-1 behavior that was removed in ATA-2, note that some drives still do support ATA-1 behavior)

http://eab.abime.net/showpost.php?p=759918&postcount=25

we become a little bit offTopic atually but thank you for this link! So this is the reason to set maxtransfer to 255 blocks instead of 256 with 512 bytes as one piece and therefore 0x1FE00 ?

[ahhyes]

Actually I provided a screenshot in an earlier post:

http://ahhyes.net/IMAG0030.jpg

I thought I had set those to zero.. I think I just left the defaults as per in the screenshot

[Rotzloeffel]

Originally Posted by ahhyes

Actually I provided a screenshot in an earlier post:

http://ahhyes.net/IMAG0030.jpg

I thought I had set those to zero.. I think I just left the defaults as per in the screenshot

jepp and these are wrong in your case! Why are you using custom-filesystem instead of FastFilesystem?

uhhh Maxtransfer is only needed for your Amiga-OS Paritition!!!!!!!!

[Rotzloeffel]

your screenshot is showing your SWAP Paritition! There is no need for changing maxtransfer ! it is not needed and ignored by netBSD!

you have to change it only on your Amiga-OS Parititions!

[ahhyes]

yes, but what about when writing the miniroot fs data to that partition whilst still in amiga OS? there is still the potential for this to be a problem?

[ahhyes]

Originally Posted by Rotzloeffel

jepp and these are wrong in your case! Why are you using custom-filesystem instead of FastFilesystem?

So I can install NetBSD. The install document gives this information. I am wondering whether the maxtransfer could be affecting things when running the xstreamtodev utility under amigaos to dump the miniroot fs data.

I'll go back and check whether those values are messed up for any Amiga partitions, I was curious about the specific issue as mentioned above.

[Rotzloeffel]

Originally Posted by ahhyes

So I can install NetBSD. The install document gives this information. I am wondering whether the maxtransfer could be affecting things when running the xstreamtodev utility under amigaos to dump the miniroot fs data.

I'll go back and check whether those values are messed up for any Amiga partitions, I was curious about the specific issue as mentioned above.

the reason for this is IMHO that you got a corrupt miniroot.fs after unpacking it from CD to your Amiga-OS Paritition with a wrong maxtransfer value which causes IMHO your boot problems! Like a corrupt startup-sequence for example.

[ahhyes]

Thanks for the suggestions! It's certainly given me something to go back and check. I ended up putting the amiga away after a lot of frustration and completely random behaviour that ended up driving me insane.

I think I will just wipe all the partitions and start from scratch with a clean slate.

[Twiggy]

Unfortunately I have a 3COM 3C589 and NetBSD still doesn't support it. Is there a list of networks cards that work with NetBSD on the Amiga?

[phx]

The supported Zorro network cards are listed here:
http://wiki.netbsd.org/ports/amiga/

But your 3COM may be a PCMCIA card? Hmm... 3C589 belongs to the Eltherlink III family. Maybe the ep(4) driver works?

Try to build a kernel with

ep* at pcmcia? function?

But that driver should already be in the GENERIC kernel...