scsi v43.46 beta need testing

[alexh]

Originally Posted by 8bitbubsy

Is this only for A600 and A1200?

Yes, despite the name it is only the driver for onboard IDE.

Originally Posted by 8bitbubsy

Will this one work with the SCSI-II DMA port on a Blizzard 2060?

No.

Originally Posted by 8bitbubsy

Or do I need the Phase5 SCSI drivers only?

Shouldn't need any SCSI drivers. They should be in the ROM on the 2060.

[_mandark_]

I can confirm that beta 4 works very well on my A1200 with a 4 GiB Compact Flash card!
A1200, 68030, FFS, OS 3.9 BB4

[phipscube]

I've now tried all different kinds of combinations to get my partitions working with beta4. Put simply I can't

Firstly, I got a new HDD, a 320GB Western Digital. I partitoned the drive through OS4.0 Boot Disk (floppy/CD combo).

DH0: - 1GB - FFS (will be OS 3.9 system)
DH1: - 2GB - SFS (will be OS 4.x system)
DH2: - 1GB - SWAP
DH3: - 100GB - left uninitialised (will be used for OS 3.9 progs)
DH4: - 194GB - left uninitialised (will be used for OS 4.x progs)

I then booted into OS3.9 with floppy/CD combo allowing the original OS 3.9 ROM to softboot from disk.

DH0 appeared ok, DH1 and 2 not shown (but expected this as they are setup for OS4.x) DH3 and 4 uninitialised. So all ok.

I then installed OS 3.9 to DH0.

soft Re-booted with default 3.9 ROMS still in memory. OS 3.9 in all its 4 colour glory is presented. Drives are still as above.

loaded on IDEFIX Express drivers from original disk supplied with adapter.
Soft reboot, now my CD drive is working.

Downloaded PFS3 from Aminet (V5.3) and burnt to a CD along with some other useful utils.

Started install of PFS3, needs a library to continue (multiuser.library). I obtained this off aminet too and copied it into my system libs folder. PFS3 installed fine. I copied over PFS3 and PFS3DS from L to my emergency floppy so I can install easier in future.

Installed BB1 and BB2. Before last reset I changed Startup-sequence Setpach line to "Setpatch NOROMUPDATE QUIET" to deactivate the default 3.9 Soft ROMS built into OS 3.9. Now the system will boot directly from my custom ROMS with beta4 SCSI.

Cold Reboot, system boots fine (and much faster).

I can now prepare the OS 3.9 partitions with HDToolbox 45.6 (the default HDToolbox 45.2 is updated to 45.6 when BB2 is installed). I copied HDToolbox 45.6 to emergency floppy for future ease of use.

In HDToolbox 45.6 my HDD was reported as 127GB, I expected this. My partitions are still recognised correctly though.

I deleted partition DH3 and then created it again within 45.6. I loaded in the PFS3DS filesystem (identifier set as 0x50465303, default mask 0x7FFFFFFE, Buffers 200, Blocksize 512).

Saved to disk, rebooted machine, formatted DH3 (quick format). The partition was now seen as "PROGRAMS". So this is a 100GB partition that i've set up as PFS3DS filesystem in HDtoolbox 45.6 and then formatted. Its showing as 94GB available (filesystem overhead??).

Soft reboot of machine, system loads up great, DH3 = Uninitialised! BANG!

Cold Reboot of machine, still Uninitialised.

Changed setpatch line to "Setpatch QUIET" to enable the standard soft ROMS.

Cold Reboot, DH3 is seen as "PROGRAMS" and is fine.

Tried to change the partition to FFS, then formatted. DH3 is now seen fine both with Beta4 of my custom ROMS and standard soft ROMS.

Tried to put back to PFS3DS on DH3 - Uninitialised (even after a format)
Tried to put on SFS on DH3 - Whole system freaked out and I couldn't boot after that at all. Had to take HDD out and wipe it on a PC (through a USB - IDE adapter).

Did all the above again but this time avoided the SFS step Suffice to say the exact same happened regarding Uninitialised DH3.

So now I had PFS3DS and HDToolbox 45.6 on a floppy I decided to boot from Emergency Disk/OS3.9 CD and wipe DH0.

Setup DH0 with PFS3DS in HDtoolbox 45.6 and formatted. (system is running from soft ROMS)

Installed OS3.9 on DH0 with PFS3DS filesystem. OK

Rebooted with soft ROMS, ok.

Adjusted Setpatch command to "Setpatch NOROMUPDATE QUIET".

Cold reboot.... System Request "No disk present in device DH0"

Booted from Emergency Disk, put setpatch back to boot from Soft ROMS.

Machine boots fine from DH0, set DH3 up as PFS3DS, ok. seen fine after a soft reboot.

Took my old 80GB drive, set up a partiton on it as PFS3DS. Under soft ROMS its seen, under my Custom ROMS with beta4 SCSI its uninitialised. I didn't want to wipe my sys partition on this drive as its set up nicely for testing etc (and burning new roms in remus).

I then tried to put my 320GB back in and set all partitons up completely from OS4.0 Mediatoolbox.

DH0 - 1GB - PFS3DS
DH1 - 2GB - SFS/00
DH2 - 1GB - SWAP
DH3 - 100GB - PFS3DS
DH4 - 194GB - left uninitialised (AFAIK this size partition will only be supported with JXFS in OS4.1 Classic so am waiting to set that up when I get OS 4.1 next week).

Quick formatted, DH0, DH1, DH3.

Did all above steps, loaded OS 3.9 on etc. with soft ROMS, System boots, DH3 recognised. With my custom ROMS, System Request "No disk present in device DH0".

So thats where I am now, and i'm pretty sure beta4 (and I assume all others from beta2 up) has a problem with any filesystem other than FFS and SFS (as boot partition ONLY). All instances of PFS3DS in my case failed with Beta4.

I'm going to make another set of custom roms exactly as mine are now but instead of beta4 I will burn in Beta1 and see if anything changes.

[Minuous]

@Cosmos:

Have you looked at the Boing Bag #3 thread recently? There is an A3000 user who is having problems with scsi.device, is this now fixed in the latest beta?

[Cosmos]

@Minuous

I'm going to read this thread ! But remember that my scsi beta support only A600 and A1200 for the moment...


@phipscube

Yes, try with the beta 1

I'll certainly remove Doobrey's code in the beta 5...

[_mandark_]

Hmm, I just run into following issue: I wanted to change some parameters of my partitions in HDToolbox (A1200, OS 3.9, 4 GiB CF-Card, beta 4), clicked onto "Save" and got the message from HDToolbox:
"Error 7 in Drive description"

Originally I partitioned and formatted the drive with WinUAE 2.3.2, uaehdf.device. In HDToolbox it reads out the following values of the drive:

Manufacturers Dame: Generic
Drive Name: Compact Flash
Revision: 0.00

My A1200 with your scsi.device beta 4 reads out:
(this is also shown as the name of the drive when starting HDToolbox)

Manufacturers Dame: CF
Drive Name: Card 4 GB
Revision: Ver6

scsi.device Ver 43.43 reads out the same values as beta4, but editing partitions and saving the choices works as expected.

Just verified this issue on WinUAE. If I add the CF card as IDE0 and use HDToolbox through scsi.device beta4 I got the same "Error 7 in Drive description" as on my real A1200 when I want to install/modify my CF Card. I can also confirm this issue exists since beta1.

[_mandark_]

Just played around some more with my new CF card in WinUAE and on my real A1200 (ACA1230-25)... and yet I found another issue

I created a 200 MiB system (OS 3.9 BB4) and a 3,7 GiB Data partition (with WinUAE 2.3.2, uaehf.device). Formatted both partition with FastFilesystem 15.6 then. I got to see interesting results with different versions of scsi.device:

scsi.device_40.12:
fdh0, fdh1 are working as expected
scsi.device_43.43:
A1200 hangs in a reboot-loop. A1200 is only able to boot up if fdh1 is deselected in early startup-menu
scsi.device_43.45:
fdh0, fdh1 are working as expected
scsi.device_43.46b4:
fdh0 is working as expected, but fdh1 is shown uninitialised. formatting fdh1 is not working (starts but never finishes...)
scsi.device_44.2:
fdh0 is working as expected, but fdh1 is shown uninitialised. formatting fdh1 is not working (starts but never finishes...)



Which leaves scsi.device_43.45 as the only one fully working on my A1200 with this CF-Card. Creating and formatting partitions work here. But as soon I boot up my A1200 with a different scsi.device I get the issues described. Booting with scsi.device_43.45 again, makes the 'lost' fdh1 re-appear and everything is fine...

Strange, because I didn't have this issues on my other A1200 (Blizzard IV 1230, 16 MiB RAM, different 4 GiB CF Card).

[Cosmos]

And with the beta 1 ??

[mfilos]

I had the same issues yesterday as well trying to format an 8GB CF.
Tested with Doobrey's scsi.device 44.2 and Cosmos 43.46b4 and both act the same:
- Partition 1 --> 950MB (showed as uninitialized upon boot and formatted just fine)
- Partition 2 --> ~7GB (showed as uninitialized upon boot and started formatting then hang).
After that system couldn't even boot with the CF on... unless I cleaned it with Diskpart under Windows to be able to reuse it.

(I haven't tried with b1 yet)

[Cosmos]

The scsi v40.12 (Kickstart 3.1) and the v43.45 (Boing Bag 2) use big loops for copying files. It's IDE, not SCSI, all is done by the CPU !

On the A4000D, these loops give a faster transfer rate when using .w (16 bit) than .l (32 bit)

On the A600/A1200, .l is faster...


Maybe I found why big loops .w are faster than .l in the read/write function on the A4000D scsi driver : PM me for explanations if you know how are working U901 & U902.
These two Gals need a simple fix (I hope) for working faster using .l !


I use this funny trick to get all ax+0 / ax+1 / ax+2 / ax+3 using big loops .l all the time in my scsi beta :

- ax+0 => loop .l
- ax+1 => .b + .w => loop .l => .b
- ax+2 => .w => loop .l => .w
- ax+3 => .b => loop .l => .w + .b

... and I wanna use it into the A4000D scsi driver, of course !!

[Minuous]

Cosmos, does your scsi.device 43.46b4 incorporate all the improvements/fixes from Doobrey's scsi.device 44.2? If so, maybe you should bump the version number (to eg. 44.3)?

[alexh]

Doobrey's 44.2 is out of sync. 43.45 came after.

[_mandark_]

Originally Posted by Cosmos

And with the beta 1 ??

Ok, did some tests with the different betas:

beta1: works flawlessly
beta2: 2nd partition is shown uninitialised. Changing drive parameters with HDToolbox works.
beta3: 2nd partition is shown uninitialised. "Error 7 in Drive description" when changing drive parameters with HDToolbox.
beta4: Same as beta3

[Cosmos]

scsi.device v43.46 (A600 & A1200) beta 5

- fix 'error 7'
- Doobrey's code removed

[mfilos]

Awesome news!!!
Thx a lot Cosmos my friend

[Cosmos]

If you use Pfs3 with a CF or an HD larger than 4 Gb, try this new version of Pfs3ds (v18.4 = last version from Piru with some bugs fix)

[mfilos]

Wow. Thanks for sharing that Cosmos!
Will there be 030/040/060 versions as well from Piru? :O

[_mandark_]

beta5 works very well!
I got no uninitialised partitions any longer, using FFS or PFS. the "Error 7 in Drive description" in HDToolbox is also gone.

[altcomputing]

Works finally! All the bugs i reported months ago have disappeared!

But the performance also dropped from almost 3MB/s to merely 2.4MB/s!

[Cosmos]

Please, give me your benchs for the five betas !