Mask and MaxTransfer

Mask and MaxTransfer

Post by Emilio Desalv » Sat, 20 Jan 2001 04:15:11



Hello!
This is probably a FAQ, or should be...

How do I define the proper values for Mask and MaxTransfer for a given
controller, file system and HD combination?
--
Emilio Desalvo

   /<>\   +-----------------------------------------------------+
  /____\  ! "Ms. Wolversham, you are authorized to return fire!"!
          ! LtCdr. Avshari, aboard HMS Bellerophon, RMN         !
  FNORD   +----http://www.geocities.com/CapeCanaveral/1566/-----+
               You do not need MAPS to reach me...

 
 
 

Mask and MaxTransfer

Post by Sebastien Vianna » Sat, 20 Jan 2001 17:42:09


Quote:> How do I define the proper values for Mask and MaxTransfer for a given
> controller, file system and HD combination?

Have a look at this guide page: http://www.xs4all.nl/~hjohn/SFS/troubles.htm
Lbas

 
 
 

Mask and MaxTransfer

Post by Patrick For » Sat, 20 Jan 2001 19:48:03



> How do I define the proper values for Mask and MaxTransfer for a given
> controller, file system and HD combination?

There are two essentail tools you can use. They are guesswork and
trial-and-error.

Here is what the PFS3 docs say. This is not FS specific, but should also
apply to FFS:

Mask
  The DMAMask indicates which memory areas the device can access with DMA. High
  speed devices like disk controllers can read and write data from memory
  without any CPU load. This is called Dynamic Memory Access or DMA.
  Unfortunately some devices can only access part of the memory. Any ZorroII
  board, for instance, can only access the lower 16MB of memory. The
  corresponding value for Mask is 0xfffffc. ZorroIII boards can access all
  memory, so for those 0xfffffffc should be used. Double check what kind of
  board is used, since if the Mask is set wrong it can result in serious
  problems with the disk.

  The mask is also used to control the alignment of any buffers used to access
  the device. Alignment to longword is recommended. Some controllers will work
  without alignment, but even those will perform a lot better with aligned
  memory. Longword alignment is selected by ending the mask with a 'c'.

  For most boards the best Mask settings are:
  Zorro II board:  use 0xfffffc.
  Zorro III board: use 0xfffffffc.

  If you don't have memory problems you should use the same mask as FFS/PFS2
  did. The mask is filesystem independent.

  For the A1200/A4000 internal IDE controller  use 0xfffffffc
  For ZorroIII boards (Fastlane, Phase5, 4091) use 0xffffffff
  For ZorroII boards and Oktagon 4008          use 0x7ffffffc

  If you experience problems with the Mask, you can try other values.

MaxTransfer
  This setting indicates how much data can be sent to the device in one go.
  Some drives cannot handle very large blocks of data. By specifying the maximum
  the device can handle here, problems are prevented. Most scsi devices can
  handle a MaxTransfer of 0xfffe00, which is 16MB. For some IDE devices
  0x1fe00 (130K) is necessary. If files greater than 64K get corrupted then
  reduce it to 0xfe00.

  The MaxTransfer must be divisible by the sector size (usually 512) due to a
  bug in the Workbench format code. This is done by ending the value with 'e00'.
  For example: instead of 0xffffff, use 0xfffe00.

  Use the same maxtransfer as FFS/PFS2 did. Maxtransfer is filesystem independent.

--
--

+ =================================+
  Patrick Ford                  
 Auckland, New Zealand                      

#The Unix Guru's View of Sex
unzip ; strip ; touch ; grep ; finger ; mount ; fsck ; more ; yes ; umount ; sleep

Harvesters, send spam directly to:






 
 
 

Mask and MaxTransfer

Post by Craig Blackma » Sun, 21 Jan 2001 05:35:04


On 19-Jan-01 03:48:03,  Re: Mask and MaxTransfer,   Patrick Ford typed:

Quote:>MaxTransfer
>  This setting indicates how much data can be sent to the device in one go.
>  Some drives cannot handle very large blocks of data. By specifying the
>  maximum the device can handle here, problems are prevented. Most scsi
>  devices can handle a MaxTransfer of 0xfffe00, which is 16MB. For some IDE
>  devices
>  0x1fe00 (130K) is necessary. If files greater than 64K get corrupted then
>  reduce it to 0xfe00.
>  The MaxTransfer must be divisible by the sector size (usually 512) due to a
>  bug in the Workbench format code. This is done by ending the value with
>  'e00'. For example: instead of 0xffffff, use 0xfffe00.

I have been using 0xffffff back when I bought a Z3 Fastlane and that's
what its docs called for. But I don't know for sure if that is correct for
today's equipment.

If I understand this correctly. The 'e00' is only critical with IDE drives
and not SCSI drives. Do you know if this is correct? I tend to believe
this since the Z3-CS/PPC docs don't don't do the 'e00' recommendation. However,
the Z3-CS/PPC docs do state that using '0xFFFFFE', which was set for non-ZorroIII
controllers, "causes considerable data loss".

Could they have really meant that 0xFFFFFE would slow data transfer since
that is optimum for an 16bit vice 32bit (?) controller?

Quote:>  Use the same maxtransfer as FFS/PFS2 did. Maxtransfer is filesystem
>  independent.

------

Since the Mask is used to set DMA access, I am curious as to how the
memory on the CS/PPC has affected this. The docs that come with the CS/PPC
have not been changed since the Z3 Fastlane docs. They actually are word
for word the same, as far as I can tell.

Does anyone know how DMA is affected between the CS/PPC memory and the
native 4000T bus? Also, since the Mask is set for the lower amount of
memory found on the Amiga motherboard, how does this affect accessing the
much larger amount of memory found on the CS/PPC.
_____________________________________________________________
 Cheers from Craig
of Tucson, Arizona, USA
--------------------------------------------------------------

A lot of good luck is undeserved, but then so is a lot of bad luck.

 
 
 

Mask and MaxTransfer

Post by Patrick For » Sun, 21 Jan 2001 11:56:00



> On 19-Jan-01 03:48:03,  Re: Mask and MaxTransfer,   Patrick Ford typed:

> >MaxTransfer
> >  This setting indicates how much data can be sent to the device in one go.
> >  Some drives cannot handle very large blocks of data. By specifying the
> >  maximum the device can handle here, problems are prevented. Most scsi
> >  devices can handle a MaxTransfer of 0xfffe00, which is 16MB. For some IDE
> >  devices
> >  0x1fe00 (130K) is necessary. If files greater than 64K get corrupted then
> >  reduce it to 0xfe00.

> >  The MaxTransfer must be divisible by the sector size (usually 512) due to a
> >  bug in the Workbench format code. This is done by ending the value with
> >  'e00'. For example: instead of 0xffffff, use 0xfffe00.

> I have been using 0xffffff back when I bought a Z3 Fastlane and that's
> what its docs called for. But I don't know for sure if that is correct for
> today's equipment.

I'm certain it should be 0x7fffffff ((seven Fs)

Quote:> If I understand this correctly. The 'e00' is only critical with IDE drives
> and not SCSI drives. Do you know if this is correct? I tend to believe
> this since the Z3-CS/PPC docs don't don't do the 'e00' recommendation. However,
> the Z3-CS/PPC docs do state that using '0xFFFFFE', which was set for non-ZorroIII
> controllers, "causes considerable data loss".

> Could they have really meant that 0xFFFFFE would slow data transfer since
> that is optimum for an 16bit vice 32bit (?) controller?

I don't think maxtransfer has much effect at all on speed. Whatever you set
either works, or corrupts large files.

Quote:> >  Use the same maxtransfer as FFS/PFS2 did. Maxtransfer is filesystem
> >  independent.

> ------

> Since the Mask is used to set DMA access, I am curious as to how the
> memory on the CS/PPC has affected this. The docs that come with the CS/PPC
> have not been changed since the Z3 Fastlane docs. They actually are word
> for word the same, as far as I can tell.

> Does anyone know how DMA is affected between the CS/PPC memory and the
> native 4000T bus? Also, since the Mask is set for the lower amount of
> memory found on the Amiga motherboard, how does this affect accessing the
> much larger amount of memory found on the CS/PPC.

I believe 7 and seven effs give access to all memory, and is correct for
any reputable Z3 or accelerator-mounted controller.
--
--

+ =================================+
  Patrick Ford                  
 Auckland, New Zealand                      

#The Unix Guru's View of Sex
unzip ; strip ; touch ; grep ; finger ; mount ; fsck ; more ; yes ; umount ; sleep

Harvesters, send spam directly to:






 
 
 

Mask and MaxTransfer

Post by Emilio Desalv » Sun, 21 Jan 2001 21:23:18


Quote:> > How do I define the proper values for Mask and MaxTransfer for a given
> > controller, file system and HD combination?
> There are two essentail tools you can use. They are guesswork and
> trial-and-error.

What? No WAGs?

Quote:> Here is what the PFS3 docs say. This is not FS specific, but should also
> apply to FFS:

Which is what I started reading after I posted the message... B-)))

At the moment I have a mask of 0x7FFFFFFE and an MT of 0x00FFFFFF,
which is probably a leftover from last century, I'll toy with it in
the weekend...

Given that we have a similar setup (excluding the RAM... B-))), what
are you using, Patrick?
--
Emilio Desalvo

   /<>\   +-----------------------------------------------------+
  /____\  ! "Ms. Wolversham, you are authorized to return fire!"!
          ! LtCdr. Avshari, aboard HMS Bellerophon, RMN         !
  FNORD   +----http://www.geocities.com/CapeCanaveral/1566/-----+
               You do not need MAPS to reach me...

 
 
 

Mask and MaxTransfer

Post by Patrick For » Mon, 22 Jan 2001 00:46:21



> > > How do I define the proper values for Mask and MaxTransfer for a given
> > > controller, file system and HD combination?
> > There are two essentail tools you can use. They are guesswork and
> > trial-and-error.

> What? No WAGs?

> > Here is what the PFS3 docs say. This is not FS specific, but should also
> > apply to FFS:

> Which is what I started reading after I posted the message... B-)))

Here is the complete explanation that leaves absolutely no room for
confusion. Why has it tken me so many years to find it explained so well?

http://www.xs4all.nl/~hjohn/SFS/mask.htm

Download SFS while you are there. If I knew then what I know now, I may
not have bought PFS3. It looks as if SFS has all the same benefits.

Quote:> Given that we have a similar setup (excluding the RAM... B-))), what
> are you using, Patrick?

I have A4000 with Warpengine '040 +SCSI and no IDE drives.
ATM  I have a couple--500MB and 1.5GB that I pulled out of my wife's machine that I will
probably put one of in mine.
--
--

+ =================================+
  Patrick Ford                  
 Auckland, New Zealand                      

#The Unix Guru's View of Sex
unzip ; strip ; touch ; grep ; finger ; mount ; fsck ; more ; yes ; umount ; sleep

Harvesters, send spam directly to:






 
 
 

Mask and MaxTransfer

Post by Craig Blackma » Mon, 22 Jan 2001 21:25:31


On 20-Jan-01 08:46:21,  Re: Mask and MaxTransfer,   Patrick Ford typed:

Quote:>Here is the complete explanation that leaves absolutely no room for
>confusion. Why has it tken me so many years to find it explained so well?
>http://www.xs4all.nl/~hjohn/SFS/mask.htm
>Download SFS while you are there. If I knew then what I know now, I may
>not have bought PFS3. It looks as if SFS has all the same benefits.

Two ticks against SFS. As far as I and some other folks can tell SFS is no
longer supported. I thought someone here confirmed that.

Also, PFS3 has a very good recovery tool called PFS Doctor which has
helped me fix dying drives enough to recovery everything on them.

PFS3 has been in development for quite a long time and PFS3 in particular
has a really great track record. BTW anyone heard anything from Michael
lately. I haven't been on the forum for a long time.

But I agree with Patrick about downloading it. Experiment with it on one
partition. If you like a more modern filesystem then go an buy PFS3 you
won't regret it.
_____________________________________________________________
        "Everyone is entitled to my opinion."

      Cheers from Craig of Tucson, Arizona, USA  
--------------------------------------------------------------

MS-DOS is to a real OS as "creation science" is to paleontology.

 
 
 

Mask and MaxTransfer

Post by Mike Wood » Tue, 23 Jan 2001 02:35:32


Quote:>On 20-Jan-01 08:46:21,  Re: Mask and MaxTransfer,   Patrick Ford typed:
>>Download SFS while you are there. If I knew then what I know now, I may
>>not have bought PFS3. It looks as if SFS has all the same benefits.

>Two ticks against SFS. As far as I and some other folks can tell SFS is no
>longer supported. I thought someone here confirmed that.

Neither is PFS as ram-jam have left the amiga market

Quote:>Also, PFS3 has a very good recovery tool called PFS Doctor which has
>helped me fix dying drives enough to recovery everything on them.

SFS writes disks in such a way that makes it very hard to damage them and it has
basic repairs features built in.

Quote:>But I agree with Patrick about downloading it. Experiment with it on one
>partition. If you like a more modern filesystem then go an buy PFS3 you
>won't regret it.

You cant but PFS anymore and SFS is a quite modern file system.

Mike Woods

 
 
 

Mask and MaxTransfer

Post by Craig Blackma » Tue, 23 Jan 2001 22:09:58


On 21-Jan-01 10:35:32,  Re: Mask and MaxTransfer,   Mike Woods typed:

Quote:>>On 20-Jan-01 08:46:21,  Re: Mask and MaxTransfer,   Patrick Ford typed:
>>>Download SFS while you are there. If I knew then what I know now, I may
>>>not have bought PFS3. It looks as if SFS has all the same benefits.

>>Two ticks against SFS. As far as I and some other folks can tell SFS is no
>>longer supported. I thought someone here confirmed that.
>Neither is PFS as ram-jam have left the amiga market

RAM-Jam was just a vendor not the developer.

The GREED website (developer) is still up and working. We haven't seen
much activity, I think, because we don't need to. Once a filesystem is
working without problems there isn't much of anything to do with it. But
Hugo and Michael have said they will update the website soon.

Quote:>>Also, PFS3 has a very good recovery tool called PFS Doctor which has
>>helped me fix dying drives enough to recovery everything on them.
>SFS writes disks in such a way that makes it very hard to damage them and it
>has basic repairs features built in.

LOL PFS3 had all of this for a long time. Basically, SFS was trying to
clone what PFS was already doing. PFS3 has a longer development
time. It also has a great tool that can help you with a physical glitch.
That was the point. PFS3 has more and does more than SFS.

Quote:>>But I agree with Patrick about downloading it. Experiment with it on one
>>partition. If you like a more modern filesystem then go an buy PFS3 you
>>won't regret it.
>You cant but PFS anymore and SFS is a quite modern file system.

As I said SFS is NOT supported anymore. It was NOT as fully developed as
PFS. No additional tools were ever developed for it. Its only advantage
is that it is free.

If you have the money PFS3 is a much better choice for the long run.
_____________________________________________________________
        "Everyone is entitled to my opinion."

      Cheers from Craig of Tucson, Arizona, USA  
--------------------------------------------------------------

What do we learn from history?  That we learn nothing from history.

 
 
 

Mask and MaxTransfer

Post by Patrick For » Tue, 23 Jan 2001 23:23:36


 >

Quote:> As I said SFS is NOT supported anymore. It was NOT as fully developed as
> PFS. No additional tools were ever developed for it. Its only advantage
> is that it is free.

I think SFS could be considered as a /sort/ of demo for PFS3. People should
download it, try it, and if they like it, buy PFS3.

The main resistance against PFS3 seems to be from some curious idea that it
is risky because it is not standard.
--
--

+=+=+=+=+=+=+=+=+=+=+=+=+=+=+==+=+=+=+=+=+=+=+=+=+=+=+=+=+
  Patrick Ford   Auckland, New Zealand                      

Harvesters, send spam directly to:




 
 
 

Mask and MaxTransfer

Post by Mike Wood » Wed, 24 Jan 2001 03:26:11


Quote:>On 21-Jan-01 10:35:32,  Re: Mask and MaxTransfer,   Mike Woods typed:
>RAM-Jam was just a vendor not the developer.
>The GREED website (developer) is still up and working. We haven't seen
>much activity, I think, because we don't need to. Once a filesystem is
>working without problems there isn't much of anything to do with it. But
>Hugo and Michael have said they will update the website soon.

Ah ok

Quote:>>SFS writes disks in such a way that makes it very hard to damage them and it
>>has basic repairs features built in.
>LOL PFS3 had all of this for a long time. Basically, SFS was trying to
>clone what PFS was already doing. PFS3 has a longer development
>time. It also has a great tool that can help you with a physical glitch.
>That was the point. PFS3 has more and does more than SFS.

SFS has had this for just as long and it has never attempted to be a clone of
PFS

Quote:>As I said SFS is NOT supported anymore. It was NOT as fully developed as
>PFS. No additional tools were ever developed for it. Its only advantage
>is that it is free.

SFS *does* have aditional tools in fact they are on the aminet.

Mike Woods

 
 
 

Mask and MaxTransfer

Post by Craig Blackma » Wed, 24 Jan 2001 22:15:39


On 22-Jan-01 11:26:11,  Re: Mask and MaxTransfer,   Mike Woods typed:

Quote:>SFS has had this for just as long and it has never attempted to be a clone of
>PFS

Mike, I don't think you are aware of how long PFS has been around. It has
been in some phase of development for about 7 or 8 years. SFS has been
around for what, 2 years. Everything they tried to do with SFS was what
PFS had been doing for years.

Like I said the good part is they were doing it for free.

BTW I can't find any additional recovery or repair tools for SFS on Aminet
that even begin to have a glimmer of what PFS Doctor has and does.

I agree with Patrick that SFS is a nice look at a more modern filesystem
and should be considered a primer for PFS3.
_____________________________________________________________
        "Everyone is entitled to my opinion."

      Cheers from Craig of Tucson, Arizona, USA  
--------------------------------------------------------------

I am going to be assertive, if that is okay with you.

 
 
 

Mask and MaxTransfer

Post by Mike Wood » Thu, 25 Jan 2001 05:16:43


Quote:>On 22-Jan-01 11:26:11,  Re: Mask and MaxTransfer,   Mike Woods typed:
>>SFS has had this for just as long and it has never attempted to be a clone
>>of PFS
>Mike, I don't think you are aware of how long PFS has been around. It has
>been in some phase of development for about 7 or 8 years. SFS has been
>around for what, 2 years. Everything they tried to do with SFS was what
>PFS had been doing for years.
>Like I said the good part is they were doing it for free.
>BTW I can't find any additional recovery or repair tools for SFS on Aminet
>that even begin to have a glimmer of what PFS Doctor has and does.
>I agree with Patrick that SFS is a nice look at a more modern filesystem
>and should be considered a primer for PFS3.

Well it's a case of each to his own.

Mike Woods

 
 
 

Mask and MaxTransfer

Post by Craig Blackma » Thu, 25 Jan 2001 08:18:10


On 23-Jan-01 13:16:43,  Re: Mask and MaxTransfer,   Mike Woods typed:

Quote:>>On 22-Jan-01 11:26:11,  Re: Mask and MaxTransfer,   Mike Woods typed:
>>>SFS has had this for just as long and it has never attempted to be a clone
>>>of PFS
>>Mike, I don't think you are aware of how long PFS has been around. It has
>>been in some phase of development for about 7 or 8 years. SFS has been
>>around for what, 2 years. Everything they tried to do with SFS was what
>>PFS had been doing for years.
>>Like I said the good part is they were doing it for free.
>>BTW I can't find any additional recovery or repair tools for SFS on Aminet
>>that even begin to have a glimmer of what PFS Doctor has and does.
>>I agree with Patrick that SFS is a nice look at a more modern filesystem
>>and should be considered a primer for PFS3.
>Well it's a case of each to his own.

I don't disagree with that statement but you should have the facts
straight so people can make an intelligent decision.
_____________________________________________________________
        "Everyone is entitled to my opinion."

      Cheers from Craig of Tucson, Arizona, USA  
--------------------------------------------------------------

Annoy the opposition -- cite fact.