BAWarp disk imager - Page 17

BlueAchenar · 22 January 2009, 21:49

Dinner is still hot.

The problem now is this. Something in the OS is turning the motor off when I use Enable() (not always). What besides trackdisk.device could do that?

BlueAchenar · 22 January 2009, 23:05

Quote:

Originally Posted by BlueAchenar

Dinner is still hot.

The problem now is this. Something in the OS is turning the motor off when I use Enable() (not always). What besides trackdisk.device could do that?

Finally fixed. Instead of fighting against the trackdisk.device I opted to cooperate with it.

BlueAchenar · 22 January 2009, 23:16

Ok. Looks good enough to me. Testing needed as always.

@Supamax: Remember, the polled transfer is too slow in your miggie for now and the rawread data correction only works well in polled mode. Try to generate raw extended adfs with -e.

There's an optimization I forgot. I could make the program do with 5 reads, what rawread needs 9 to do.

Later, I need some rest, but I'll explain first:

Rawread reads 3 times the same track. It it fails to correct the data reads more 3. And again more 3.

I read 5 times the same data. The beautiful thing is I can use the rawread functions like this. Try to correct data with reads 0,1,2 ; 1,2,3 ; 2,3,4 ; 3,4,5 ; 0,1,3 ; 0,1,4 ; etc... Beautiful isn't it? It will make it faster than rawread for difficult disks.

BlueAchenar · 22 January 2009, 23:23

---Quote (Originally by BlueAchenar)---
Finally fixed. Instead of fighting against the trackdisk.device I opted to cooperate with it.

---End Quote---

Prowler:

Quote:

Sounds like a good plan.

Does this mean v0.3 is just around the corner?

I wish!!!

Supamax · 22 January 2009, 23:25

Quote:

Originally Posted by BlueAchenar

@Supamax: Remember, the polled transfer is too slow in your miggie for now and the rawread data correction only works well in polled mode. Try to generate raw extended adfs with -e.

Should I use "-e" only, or "-d -e" ??

Quote:

I read 5 times the same data. The beautiful thing is I can use the rawread functions like this. Try to correct data with reads 0,1,2 ; 1,2,3 ; 2,3,4 ; 3,4,5 ; 0,1,3 ; 0,1,4 ; etc... Beautiful isn't it? It will make it faster than rawread for difficult disks.

I didn't understand, but... as you say, o Master

Supamax · 22 January 2009, 23:39

Ok, I'm trying the last version: AleWarp-0.2-6.

1) With "alewarp filename -d -e" it hangs with the usual "Reading cylinder 0, side 0...".
First of all, good news: the mouse doesn't freeze any more. The whole machine seems not freezed any more.
I kept the LMB pressed for a while, with no effect. But, after let's say 45 sec./1 min. the drive stopped, program exited on CLI and I see:
Data correction failed!!!0... <------ It would seem that it wrote above the previous msg "Reading cylinder 0, side 0..."
The program exited earlier than expected!!!

So, perhaps it detected the LMB pressed, after all

2) With "alewarp filename -d" it works continuosly (not hiccups, yeah!).
Sometimes showing "Writing 61600 data bytes to file...", most of the time not...
Dumped successfully!

3) With "alewarp filename" it works continuosly too, BUT it writes 41xxx..47xxx data bytes instead of 61600, so it's probably not working well yet.

BlueAchenar · 22 January 2009, 23:44

Quote:

Originally Posted by Supamax

Should I use "-e" only, or "-d -e" ??

I didn't understand, but... as you say, o Master

Simple really, imagine a function CorrectData(read1, read2, read3).

RawRead does this in the worst case scenario:

Read track 3 times
CorrectData(0, 1, 2)
Read track 3 times
CorrectData(0, 1, 2)
Read track 3 times
CorrectData(0, 1, 2)

I can do this:
Read track 5 times
CorrectData(0, 1, 2)
CorrectData(0, 1, 3)
CorrectData(0, 1, 4)
CorrectData(0, 2, 3)
CorrectData(0, 2, 4)
CorrectData(0, 3, 4)
CorrectData(1, 2, 3)
CorrectData(1, 2, 4)
CorrectData(1, 3, 4)
CorrectData(2, 3, 4)

So with only one read operation, I can try to use rawread's correction 10 times instead of 3.

About the options. Use all you want.

I'll optimise the polled reading to work in your A600.

BlueAchenar · 22 January 2009, 23:51

Quote:

Originally Posted by Supamax

Ok, I'm trying the last version: AleWarp-0.2-6.

With "alewarp filename -d -e" it hangs with the usual "Reading cylinder 0, side 0...".
First of all, good news: the mouse doesn't freeze any more. The whole machine seems not freezed any more.
I kept the LMB pressed for a while, with no effect. But, after let's say 45 sec./1 min. the drive stopped, program exited on CLI and I see:
Data correction failed!!!0... <------ It would seem that it wrote above the previous msg "Reading cylinder 0, side 0..."
The program exited earlier than expected!!!

So, perhaps it detected the LMB pressed, after all

WAIT DO NOT REPLY YET ... checking more...

Couldn't help it.

I told you not to use DMA with that method of data correction.
Here's an optimized polled one for you.

Edit: By the way, if the mouse moves, its because error correction is being applied.

Supamax · 22 January 2009, 23:54

Ok, now I'll try version 0.2-7.

1) "alewarp filename"
Works Ok, writes around 63xxx (changes every time)

NOTE: if I press the LMB to stop it, it doesn't exit in a "clean" mode. A Recoverable Alert pops up:

Code:

Recoverable Alert.         Press left mouse button to continue.
Error:  0100 0009  Task:  0006BF20

However I can press the LMB and all goes well

2) "alewarp filename -e"
Works Ok, and it's fast! Not slower than the .raw (non-adf) mode... are you sure it's doing error correction? Too fast

EDIT: mmmm, perhaps I spoke too quickly. The first tracks are read fast, but as soon as it proceeds (now at track 31) it seems to have slowed down...
No, some tracks are read faster, other tracks slower.

BlueAchenar · 23 January 2009, 00:23

Quote:

Originally Posted by Supamax

Ok, now I'll try version 0.2-7.

1) "alewarp filename"
Works Ok, writes around 63xxx (changes every time)

NOTE: if I press the LMB to stop it, it doesn't exit in a "clean" mode. A Recoverable Alert pops up:

Code:

Recoverable Alert.         Press left mouse button to continue.
Error:  0100 0009  Task:  0006BF20

However I can press the LMB and all goes well

2) "alewarp filename -e"
Works Ok, and it's fast! Not slower than the .raw (non-adf) mode... are you sure it's doing error correction? Too fast

EDIT: mmmm, perhaps I spoke too quickly. The first tracks are read fast, but as soon as it proceeds (now at track 31) it seems to have slowed down...
No, some tracks are read faster, other tracks slower. I suppose this is due to the MFM raw reading?

I'm preparing a few CDs to give to someone and therefore a little distracted.

The better the track is read the faster error correction will be. Nothing to be done about that. That's why some are slower than the others.

About the LMB, will look into that.

Supamax · 23 January 2009, 00:26

Do you need dumps?
- original adf
- extended adf created by AleWarp
- .raw created by AleWarp (without -e)

BlueAchenar · 23 January 2009, 00:33

Quote:

Originally Posted by Supamax

Do you need dumps?
- original adf
- extended adf created by AleWarp
- .raw created by AleWarp (without -e)

Could you send me the extended adf please? Have you tried it in UAE?

Supamax · 23 January 2009, 00:42

Quote:

Originally Posted by BlueAchenar

Could you send me the extended adf please? Have you tried it in UAE?

I was hoping and excited and drooling all over the floppies and... it doesn't work

.
You find it in The Zone!

I mean: it's created fine (2+ MB), but there must be something wrong in its structure.
It's the same Relokick floppy I used before.
It doesn't load in WinUAE, and if I check it with X-Copy (under WinUAE) it makes WinUAE *very* sluggish and show all red 2's instead of green 0's.

But you are near!! Very near!!

BlueAchenar · 23 January 2009, 00:47

Quote:

Originally Posted by Supamax

But you are near!! Very near!!

It's probably some structure quirk. I'll look into it tomorrow (or until I feel the urge to see what's wrong

). I need some rest for today. No source code today. It needs a cleanup.

Supamax · 23 January 2009, 00:48

Quote:

Originally Posted by BlueAchenar

It's probably some structure quirk. I'll look into it tomorrow (or until I feel the urge to see what's wrong

). I need some rest for today. No source code today. It needs a cleanup.

Ok, sleep well and may the Force be with you

BlueAchenar · 23 January 2009, 00:50

Quote:

Originally Posted by Supamax

Ok, sleep well and may the Force be with you

I'm not going to bed yet.

Maybe in 1 or 2 hours.

BlueAchenar · 23 January 2009, 02:06

Well, I cleaned up the code a little. No bugs corrected from the last version yet.

BlueAchenar · 23 January 2009, 12:21

3 bugs corrected in this one:

- Correct trackbits in extended ADF
- Polled transfers could lose 1 byte per read in slow amigas
- Correct track length in ADF header (this one also exists in rawread)

BlueAchenar · 23 January 2009, 12:34

Last known bug corrected:

- Unexpected quits are now graceful!

Well, now its back to error correction for me. I need a new DMA raw dump of a ADOS disk, but no hurry this time.

BlueAchenar · 23 January 2009, 14:49

New correction method "Improved RawRead". This is now active by default. It should make much better ADFs of non DOS disks. (Option -c selects the wanted method now).

Edit: If someone tests one of the generated ADFs on UAE please let me know.

22 January 2009, 23:39	#326
Supamax Da Digger :) Join Date: Nov 2008 Location: Monza, Italy Posts: 2,822	Ok, I'm trying the last version: AleWarp-0.2-6. 1) With "alewarp filename -d -e" it hangs with the usual "Reading cylinder 0, side 0...". First of all, good news: the mouse doesn't freeze any more. The whole machine seems not freezed any more. I kept the LMB pressed for a while, with no effect. But, after let's say 45 sec./1 min. the drive stopped, program exited on CLI and I see: Data correction failed!!!0... <------ It would seem that it wrote above the previous msg "Reading cylinder 0, side 0..." The program exited earlier than expected!!! So, perhaps it detected the LMB pressed, after all 2) With "alewarp filename -d" it works continuosly (not hiccups, yeah!). Sometimes showing "Writing 61600 data bytes to file...", most of the time not... Dumped successfully! 3) With "alewarp filename" it works continuosly too, BUT it writes 41xxx..47xxx data bytes instead of 61600, so it's probably not working well yet. Last edited by Supamax; 23 January 2009 at 00:00.

22 January 2009, 23:54	#329
Supamax Da Digger :) Join Date: Nov 2008 Location: Monza, Italy Posts: 2,822	Ok, now I'll try version 0.2-7. 1) "alewarp filename" Works Ok, writes around 63xxx (changes every time) NOTE: if I press the LMB to stop it, it doesn't exit in a "clean" mode. A Recoverable Alert pops up: Code: Recoverable Alert. Press left mouse button to continue. Error: 0100 0009 Task: 0006BF20 However I can press the LMB and all goes well 2) "alewarp filename -e" Works Ok, and it's fast! Not slower than the .raw (non-adf) mode... are you sure it's doing error correction? Too fast EDIT: mmmm, perhaps I spoke too quickly. The first tracks are read fast, but as soon as it proceeds (now at track 31) it seems to have slowed down... No, some tracks are read faster, other tracks slower. Last edited by Supamax; 23 January 2009 at 00:21.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Vectrex collection with 3D Imager, light-pen and complete set of games	Tolls	Retrogaming General Discussion	5	24 April 2011 13:45
Disk-2-Disk by Central Coast Software (full or demo version)	mark_k	request.Apps	3	09 July 2006 14:10
How can a floppy be a PC 720k disk and an Amiga 880k disk at the same time?	dreamkatcha	New to Emulation or Amiga scene	33	24 July 2004 14:25
Newbie question: how to access files on an Amiga disk using PC disk drive?	Nethead	New to Emulation or Amiga scene	5	04 July 2003 23:07
G Soft Audio Imager	Syko	support.Hardware	0	13 November 2002 18:48

22 January 2009, 21:49	#321
BlueAchenar Not dead yet!!! :D Join Date: Feb 2008 Location: Aveiro / Portugal Age: 42 Posts: 690	Dinner is still hot. The problem now is this. Something in the OS is turning the motor off when I use Enable() (not always). What besides trackdisk.device could do that?

23 January 2009, 00:26	#331
Supamax Da Digger :) Join Date: Nov 2008 Location: Monza, Italy Posts: 2,822	Do you need dumps? - original adf - extended adf created by AleWarp - .raw created by AleWarp (without -e)

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)