A hack, anyone, to turn on dma ?

[PCR]

Franc Zabkar wrote:
| On Wed, 25 Jun 2008 21:32:27 -0400, "PCR" <pcrrcp@netzero.net> put
| finger to keyboard and composed:
|
|>http://www.google.com/search?hl=en&q=TSSTcorp+CD/DVDW+SH-S182F++dma&st
art=10&sa=N
|>Google has about 132 for "TSSTcorp CD/DVDW SH-S182F dma". I haven't
|>read them all. Here is one...
|>
|>http://www.dvdplusvideo.com/dvdguide005.html
|>Enabling Direct Memory Access or "DMA" for faster DVD copying
|>
|>That article is speaking of WinXP. It says...
|>
|>"For repeated DMA errors. Windows XP will turn off DMA mode for a
|>device after encountering certain errors during data transfer
|>operations. If more that six DMA transfer timeouts occur, Windows
|>will turn off DMA and use only PIO mode on that device.
|>In this case, the user cannot turn on DMA for this device. The only
|>option for the user who wants to enable DMA mode is to uninstall and
|>reinstall the device."
|
| I found the following Win XP registry hack, but it doesn't appear to
| apply to Win9x.
|
| Getting back to DMA mode in Windows XP:
|
http://sniptools.com/tipstricks/getting-back-to-dma-mode-in-windows-xp#comment1891

I can't seem to make it fit Win98 either. In Win98, Device Manager
doesn't seem capable of setting PIO or DMA for a controller-- but only
for a device. It does seem that Shadow has DMA set in his Secondary IDE
Controller to which the DVD is connected, anyhow...

[HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Class\hdc\0002]
"DevLoader"="*IOS"
"DriverDesc"="Secondary IDE controller (dual fifo)"
"InfPath"="VIAVSD.INF"
"InfSection"="VIAIDE"
"PortDriver"="ESDI_506.pdr"
"ProviderName"="Microsoft"
"DriverDate"="2001\\10\\18"
"IDEDMADRIVE0"=hex:01
"IDEDMADRIVE1"=hex:01
"AutoInsertNotificationDrive0"=hex:01

BUT, going by "DMACurrentlyUsed=00", it ISN'T set in the DVD device.
Device Manager gives him the checkbox for it, but it won't stay set...

[HKEY_LOCAL_MACHINE\Enum\SCSI\TSSTCORPCD/DVDW_SH-S182FS\MF&CHILD0001&PCI
&VEN_1106&DEV_0571&SUBSYS_80ED1043&REV_06&BUS_00&DEV_0F&FUNC_0100]
"Capabilities"=hex:14,00,00,00
"AutoInsertNotification"=hex:00
"SCSITargetID"="0"
"SCSILUN"="0"
"RevisionLevel"="SB02"
"ProductId"="CD/DVDW SH-S182F"
"Manufacturer"="TSSTcorp"
"DeviceType"=hex:05
"Removable"=hex:01
"CurrentDriveLetterAssignment"="I"
"IDEMaster"=hex:01
"DMACurrentlyUsed"=hex:00
"HardwareID"="TSSTCORPCD/DVDW_SH-S182FS,GenCD,SCSI\\TSSTCORPCD/DVDW_SH-S
182FS"
"Class"="CDROM"
"ClassGUID"="{4d36e965-e325-11ce-bfc1-08002be10318}"
"Driver"="CDROM\\0000"
"Mfg"="(Dispositivo padrão de CD-ROM)"
"DeviceDesc"="TSSTcorp CD/DVDW SH-S182F"
"ConfigFlags"=hex:00,00,00,00
"Disconnect"=hex:01
"SyncDataXfer"=hex:00

I see you've made some suggestions. I'm still thinking.

| - Franc Zabkar
| --
| Please remove one 'i' from my address when replying by email.

--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net

 

[PCR]

Shadow wrote:
| On Wed, 25 Jun 2008 23:18:48 -0400, "PCR" <pcrrcp@netzero.net> wrote:
|
|>Shadow wrote:
|>| On Wed, 25 Jun 2008 21:32:27 -0400, "PCR" <pcrrcp@netzero.net>
|>| wrote:
|>|
|>I'm just not sure that constitutes a complete un/re-install--
|>especially as you point it to C:\Windows\Inf for drivers. I'm
|>speaking of JUST the DVD-- not all the other stuff. MSInfo, History,
|>for it showed...
| Its plug and pray. You plug it in, windows installs it.
|>
|>TSSTcorp CD/DVDW SH-S182F
|>Configuração original Tue Jun 24 22:06:37 2008 para Date
|>
|>I tend to doubt that actually is the date you first installed it. Do
|>you have a CD with its drivers? Are there instructions anywhere on
|>how to uninstall it? I might want to completely uninstall & then
|>reinstall it-- maybe unless that involves actually unplugging it.
|>That might be a little too much trouble.

| I've built computers for over 14 years (mine only). No
| trouble. But the DVDBurner has no drivers, it uses the windows native
| ones.

Alright. That's right. It's the same for my CD-R/W, actually-- "no
drivers are required or have been loaded for this device", Device
Manager says. Sorry. Well, what do you see at...

(1) "Device Manager, CDROM"
(2) D-Clk your DVD there for Properties, Driver tab.
Mine is not a DVD, but is... IDE-CD R/RW 4x4x24.
(3) "Update Driver" button, Next"
(4) "Bolt 'Display a list of all the drivers...', Next"
What shows in the Models box?
Mine says... (CD-ROM Drive) [4-23-1999]
(5) Bolt "Show all hardware"
What shows in the Manufacturers & Models boxes?
Mine shows... (Standard CDROM device) & (CDROM drive).

If yours is not the "standard", perhaps make it so & reboot. (Jot down
exactly what yours is first, so that you may switch back to it later if
desired.) Any difference?

|>
|>Why did you point it to C:\Windows\Inf for drivers? Where did you
|>point it the very first time when you originally installed it?

| I have no other place to point to ..... there are no
| non-windows drivers.

OK, OK-- sorry, that's right. Let's see what it looks like at those
Update Driver requestors.

|>
|>| Here is my key. DMA is currently "off"
|>| .........................................................
|>|
|>[HKEY_LOCAL_MACHINE\Enum\SCSI\TSSTCORPCD/DVDW_SH-S182FS\MF&CHILD0001&P
CI
|>&VEN_1106&DEV_0571&SUBSYS_80ED1043&REV_06&BUS_00&DEV_0F&FUNC_0100]
|>| "Capabilities"=hex:14,00,00,00
|>| "AutoInsertNotification"=hex:00
|>| "SCSITargetID"="0"
|>| "SCSILUN"="0"
|>| "RevisionLevel"="SB02"
|>| "ProductId"="CD/DVDW SH-S182F"
|>| "Manufacturer"="TSSTcorp"
|>| "DeviceType"=hex:05
|>| "Removable"=hex:01
|>| "CurrentDriveLetterAssignment"="I"
|>| "IDEMaster"=hex:01
|>| "DMACurrentlyUsed"=hex:00
|>
|>That looks like the big difference between yours & mine. Did you try
|>setting DMACurrentlyUsed to 01 & reboot? To do so...
| Yes, one of the first things I did.
| []'s

I should have known you would. Too bad it didn't work. I'm still
thinking. Try playing with that Update Driver requestor in the meantime.
Also, Zabcar seems to have lots more.


--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net

 

[Shadow]

On Thu, 26 Jun 2008 20:01:11 -0400, "PCR" <pcrrcp@netzero.net> wrote:

>Shadow wrote:
>| On Wed, 25 Jun 2008 23:18:48 -0400, "PCR" <pcrrcp@netzero.net> wrote:
>Alright. That's right. It's the same for my CD-R/W, actually-- "no
>drivers are required or have been loaded for this device", Device
>Manager says. Sorry. Well, what do you see at...
>
>(1) "Device Manager, CDROM"
>(2) D-Clk your DVD there for Properties, Driver tab.
> Mine is not a DVD, but is... IDE-CD R/RW 4x4x24.
>(3) "Update Driver" button, Next"
>(4) "Bolt 'Display a list of all the drivers...', Next"
> What shows in the Models box?
> Mine says... (CD-ROM Drive) [4-23-1999]
Gives the model of the dvd
>(5) Bolt "Show all hardware"
> What shows in the Manufacturers & Models boxes?
> Mine shows... (Standard CDROM device) & (CDROM drive).
again
>
>If yours is not the "standard", perhaps make it so & reboot. (Jot down
>exactly what yours is first, so that you may switch back to it later if
>desired.) Any difference?
it uses the mscdrom.inf from 5-5-1999

I just went crazy, dropped to DOS, renamed my windows folder
"badboy", reinstalled windows in "windows", and ---- the cdrom does
not accept dma, the box is there , but will not stay ticked. So it is
something quite native. The new windows did not even know the DVD
existed, it was a clean install.
So although samsung says the drive is compatible with windows
98, someone made a big booboo with the drivers. Either VIA or Samsung
when they designed the hardware.
I will have to boot linux for my backups. What I hate about
K3b is that it does not support checking the compilation, also it
knackers my long file names.
If you are interested, I can grope around in /proc and see if
linux is using any dma..... I'll do that

 

[Shadow]

What I got from linux::
(Sorry about formatting)

name value min max mode
---- ----- --- --- ----
current_speed 66 0 70 rw
dsc_overlap 0 0 1 rw
init_speed 12 0 70 rw
io_32bit 1 0 3 rw
keepsettings 0 0 1 rw
nice1 1 0 1 rw
number 2 0 3 rw
pio_mode write-only 0 255 w
unmaskirq 1 0 1 rw
using_dma 1 0 1 rw

.........................................
[root@localhost ~]# hdparm -v /dev/hdc

/dev/hdc:
IO_support = 1 (32-bit)
unmaskirq = 1 (on)
using_dma = 1 (on)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 256 (on)
HDIO_GETGEO failed: Inappropriate ioctl for device

................................................
[root@localhost ~]# hdparm -i /dev/hdc

/dev/hdc:

Model=TSSTcorpCD/DVDW SH-S182F, FwRev=SB02, SerialNo=
Config={ Fixed Removeable DTR<=5Mbs DTR>10Mbs nonMagnetic }
RawCHS=0/0/0, TrkSize=0, SectSize=0, ECCbytes=0
BuffType=unknown, BuffSize=0kB, MaxMultSect=0
(maybe): CurCHS=0/0/0, CurSects=0, LBA=yes, LBAsects=0
IORDY=on/off, tPIO={min:227,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2
AdvancedPM=no

* signifies the current active mode
...............................................

driver:
ide-cdrom version 4.61
.....................................................
dma:
2: floppy
4: cascade
(this is the same as windows. I cant figure out where the DVD dma
controler is , if the only one that appears is a floppy one...)
......................................................
From dmesg:
The controler:
VP_IDE: VIA vt8237 (rev 00) IDE UDMA133 controller on pci0000:00:0f.1
ide0: BM-DMA at 0xfc00-0xfc07, BIOS settings: hdaMA, hdbio
ide1: BM-DMA at 0xfc08-0xfc0f, BIOS settings: hdcMA, hddio
................................
The drive:
hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache, UDMA(33)
.....
I could probably dig for more.But this is not a linux group

I still can't understand why windows will not set dma. The
controler/drive has "dma" written all over it ....

 

[MEB]

Haven't scrolled back through this discussion [likely some of this was
preented]::
http://downloads.viaarena.com/drivers/4in1/VIA_Hyperion 4IN1_V448release.doc
SOME versions had bugs, some lacked support, some were specific, etc,, test
installing various drivers would require ensuring the previous had been
*completely* removed {particularly in the INF folder}.

98 supports DMA but may disable it if your 98 system fails certain tests.
Specifically when the drivers:
1. query the motherboard chip set,
2. query the drive itself, and,
3. test a short pattern of disk reads and writes to see if they are reliable
at DMA speeds.
If any of the three or combination fail, DMA will be disabled upon reboot.
After errors, driver or software, the system may downgrade [permanently].
Particularly true for newer OSs [XP, VISTA].
ONE WAY that WILL kill DMA is to attempt to read a badly burnt or scratched
disk multiple times... another, errors during the copy/burn process

There MAY be a difficulty between the default CDROM drivers supplied
[default IDE INFs] and the actual drive when accessed.
There MAY be BIOS issues, using ACPI in some BIOSs may allow DMA to be used
if not already set.

* It may depend upon other drives attached to the channel, try:
DVD-RW as Device 0 (1st) as Master and the other CD-RW/hard drive as Device
1 (2nd) as Slave.

You might try the potential fixes here:
http://www.mdgx.com/98-2.htm#W98DMA - it starts at the VERY top of the DMA
info, scroll down to see the help and potential fixes.
Make sure you look at:
http://download.microsoft.com/download/win98SE/Update/5638-6151/W98/EN-US/243450USA8.EXE -
contains an updated esdi_506
You can find more DMA fix info here:
MDGx Windows 98/98 SE + DOS 7.10 Tricks, Secrets, BUGs + FIXes
http://www.mdgx.com/newtip98.htm - use find DMA on the page in the browser's
search/find.

Cleaning:
And as I have previously posted, it might be advisable to open the drive
case and look at the laser lens. When they get blocked
[dust/dirt/hair/paper, whatever], smogged, or *smoked*, the system will
downgrade speeds and abilities, as will burning software. Sometimes you can
carefully blow them out, sometimes you have to CAREFULLY clean with a
de-natured alcohol moistened q-tip...

--
MEB
http://peoplescounsel.orgfree.com
--
_________


"Franc Zabkar" <fzabkar@iinternode.on.net> wrote in message
news:do9664lnmd6sapkft6ug08ncsj4en3feut@4ax.com...
| On Wed, 25 Jun 2008 21:30:50 -0300, Shadow <sh@dow> put finger to
| keyboard and composed:
|
| >On Wed, 25 Jun 2008 17:47:23 -0400, "glee" <glee29@spamindspring.com>
| >wrote:
| >
| >>"Shadow" <sh@dow> wrote in message
| >>news:kia56414frgspmk6rfqrga8b32k74an5kc@4ax.com...
| >>
| >>Yes. It shows DMA enabled for both Secondary IDE Channels. I am
starting to think
| >>the drive itself is going bad.
|
| > Well, I just verified that. I had a 3.904 Gb backup to do, did
| >it with Nero, in windows 98SE, and it took 93 minutes.
| > I then fired up a livecd PClinux and burnt the same project in
| >10 minutes and 4 seconds. using K3b.
| > Both times using LG 16x media and programs set to "max speed".
| >Both programs reported they were burning at 16x
| >
| > So it has to be a driver problem.
|
| I find it hard to accept that turning off the DMA setting in Device
| Manager is the root cause of a 9x performance difference.
|
| Does your drive/controller/driver have problems reading as well?
|
| - Franc Zabkar
| --
| Please remove one 'i' from my address when replying by email.

 

[MEB]

Dang forgot this, IF there are only two drives, both on the same channel,
move the DVD/CDROM drive to the second channel as master...

--
MEB
http://peoplescounsel.orgfree.com
--
_________

"MEB" <meb@not here@hotmail.com> wrote in message
news:umyKnlM2IHA.548@TK2MSFTNGP06.phx.gbl...
| Haven't scrolled back through this discussion [likely some of this was
| preented]::
|
http://downloads.viaarena.com/drivers/4in1/VIA_Hyperion 4IN1_V448release.doc
| SOME versions had bugs, some lacked support, some were specific, etc,,
test
| installing various drivers would require ensuring the previous had been
| *completely* removed {particularly in the INF folder}.
|
| 98 supports DMA but may disable it if your 98 system fails certain tests.
| Specifically when the drivers:
| 1. query the motherboard chip set,
| 2. query the drive itself, and,
| 3. test a short pattern of disk reads and writes to see if they are
reliable
| at DMA speeds.
| If any of the three or combination fail, DMA will be disabled upon
reboot.
| After errors, driver or software, the system may downgrade [permanently].
| Particularly true for newer OSs [XP, VISTA].
| ONE WAY that WILL kill DMA is to attempt to read a badly burnt or
scratched
| disk multiple times... another, errors during the copy/burn process
|
| There MAY be a difficulty between the default CDROM drivers supplied
| [default IDE INFs] and the actual drive when accessed.
| There MAY be BIOS issues, using ACPI in some BIOSs may allow DMA to be
used
| if not already set.
|
| * It may depend upon other drives attached to the channel, try:
| DVD-RW as Device 0 (1st) as Master and the other CD-RW/hard drive as
Device
| 1 (2nd) as Slave.
|
| You might try the potential fixes here:
| http://www.mdgx.com/98-2.htm#W98DMA - it starts at the VERY top of the DMA
| info, scroll down to see the help and potential fixes.
| Make sure you look at:
|
http://download.microsoft.com/download/win98SE/Update/5638-6151/W98/EN-US/243450USA8.EXE -
| contains an updated esdi_506
| You can find more DMA fix info here:
| MDGx Windows 98/98 SE + DOS 7.10 Tricks, Secrets, BUGs + FIXes
| http://www.mdgx.com/newtip98.htm - use find DMA on the page in the
browser's
| search/find.
|
| Cleaning:
| And as I have previously posted, it might be advisable to open the drive
| case and look at the laser lens. When they get blocked
| [dust/dirt/hair/paper, whatever], smogged, or *smoked*, the system will
| downgrade speeds and abilities, as will burning software. Sometimes you
can
| carefully blow them out, sometimes you have to CAREFULLY clean with a
| de-natured alcohol moistened q-tip...
|
| --
| MEB
| http://peoplescounsel.orgfree.com
| --
| _________
|
|
| "Franc Zabkar" <fzabkar@iinternode.on.net> wrote in message
| news:do9664lnmd6sapkft6ug08ncsj4en3feut@4ax.com...
| | On Wed, 25 Jun 2008 21:30:50 -0300, Shadow <sh@dow> put finger to
| | keyboard and composed:
| |
| | >On Wed, 25 Jun 2008 17:47:23 -0400, "glee" <glee29@spamindspring.com>
| | >wrote:
| | >
| | >>"Shadow" <sh@dow> wrote in message
| | >>news:kia56414frgspmk6rfqrga8b32k74an5kc@4ax.com...
| | >>
| | >>Yes. It shows DMA enabled for both Secondary IDE Channels. I am
| starting to think
| | >>the drive itself is going bad.
| |
| | > Well, I just verified that. I had a 3.904 Gb backup to do, did
| | >it with Nero, in windows 98SE, and it took 93 minutes.
| | > I then fired up a livecd PClinux and burnt the same project in
| | >10 minutes and 4 seconds. using K3b.
| | > Both times using LG 16x media and programs set to "max speed".
| | >Both programs reported they were burning at 16x
| | >
| | > So it has to be a driver problem.
| |
| | I find it hard to accept that turning off the DMA setting in Device
| | Manager is the root cause of a 9x performance difference.
| |
| | Does your drive/controller/driver have problems reading as well?
| |
| | - Franc Zabkar
| | --
| | Please remove one 'i' from my address when replying by email.
|
|
|

 

[Franc Zabkar]

On Fri, 27 Jun 2008 21:38:11 -0300, Shadow <sh@dow> put finger to
keyboard and composed:

> I just went crazy, dropped to DOS, renamed my windows folder
>"badboy", reinstalled windows in "windows", and ---- the cdrom does
>not accept dma, the box is there , but will not stay ticked. So it is
>something quite native. The new windows did not even know the DVD
>existed, it was a clean install.

> So although samsung says the drive is compatible with windows
>98, someone made a big booboo with the drivers. Either VIA or Samsung
>when they designed the hardware.

According to your other posts, it appears that your PCI chipset
registers are being correctly configured for UDMA mode by VIA's
driver. At least that's how it looks to me.

For comparison purposes, I have been looking at the PCI registers for
my own SiS 5513 IDE controller. I have three IDE devices, one of which
(a Ricoh MP6200A CD-writer) doesn't support DMA mode (it fails the Set
Features test at Glen's MS URL), and the PCI registers accurately
reflect this, ie UDMA mode is disabled for the writer, but enabled for
the others. The DMA checkbox also will not stay ticked.

In short, my chipset registers are consistent with the DMA checkbox
whereas yours are not.

> I will have to boot linux for my backups. What I hate about
>K3b is that it does not support checking the compilation, also it
>knackers my long file names.
> If you are interested, I can grope around in /proc and see if
>linux is using any dma..... I'll do that

It would be interesting to see what information your IDE devices
return in response to an ATAPI Identify Device command. Unfortunately
the only utilities I can find that give this kind of information are
suitable for ATA devices (ie hard drives) only. For example, Everest
Home Edition tells you which UDMA modes a drive can support and which
mode is currently active. I found something called BusTrace that
appears to be able to do the job, but it is not free:

http://www.bustrace.com/bustrace7/apps/ataidentify/index.htm

Just out of curiosity, I checked to see whether your writer has the
latest firmware. According to the following URL, it does (SB02, 09
March 2007):

http://www.cdrinfo.com/Sections/Firmware/SingleModel.aspx?DriveId=1413

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Franc Zabkar]

On Fri, 27 Jun 2008 23:07:35 -0300, Shadow <sh@dow> put finger to
keyboard and composed:

>dma:
> 2: floppy
> 4: cascade
>(this is the same as windows. I cant figure out where the DVD dma
>controler is , if the only one that appears is a floppy one...)

AIUI, the IDE controller does not do DMA via the motherboard's DMA
chips. It is a bus mastering controller which means it performs those
DMA functions itself. In contrast, the floppy controller can't do DMA
on its own, so it needs a DMA channel to access one of the
motherboard's 8237 DMA chips (or compatible equivalent).

If it helps, I'll scan the relevant section of IBM's PC AT Technical
Reference manual and post it on my web space.

>.....................................................
>From dmesg:
>The controler:
>VP_IDE: VIA vt8237 (rev 00) IDE UDMA133 controller on pci0000:00:0f.1
> ide0: BM-DMA at 0xfc00-0xfc07, BIOS settings: hdaMA, hdbio
> ide1: BM-DMA at 0xfc08-0xfc0f, BIOS settings: hdcMA, hddio
>...............................
>The drive:
>hdc: ATAPI 48X DVD-ROM DVD-R-RAM CD-R/RW drive, 2048kB Cache, UDMA(33)
>....
>I could probably dig for more.But this is not a linux group
>
>I still can't understand why windows will not set dma. The
>controler/drive has "dma" written all over it ....

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Shadow]

Re: A hack, anyone, to turn on dma ? SOLVED ???

Thank you all for your suggestions. I now have DMA, on by
default, I did not even have to check the box.
How ? I jumpered the drive to slave .(It was master, on
secondary ide, is now slave)
Why ?
I have no idea. I tried rejumpering it to master and it
insists it's not capable of DMA. Jumpered to slave ... my DMA back.

Windows must have some horrible obscure place where it keeps a
log of bad CDRoms, I must have used a bad CDrom, and it registered the
first channel of ide 1 as "bad" with the samsung writer.

The only file that seems to be updated is mscdrom.pnf. Shall I
delete it and see what happens ?

 

[Lee]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Jun 28, 11:22 am, Shadow <sh@dow> wrote:
>         Thank you all for your suggestions. I now have DMA, on by
> default, I did not even have to check the box.
>         How ? I jumpered the drive to slave .(It was master, on
> secondary ide, is now slave)
>         Why ?
>         I have no idea. I tried rejumpering it to master and it
> insists it's not capable of DMA. Jumpered to slave ... my DMA back.
>
>         Windows must have some horrible obscure place where it keeps a
> log of bad CDRoms, I must have used a bad CDrom, and it registered the
> first channel of ide 1 as "bad" with the samsung writer.
>
>         The only file that seems to be updated is mscdrom.pnf. Shall I
> delete it and see what happens ?
>        

The PNF files are of no consequence in Win98, they are simply machine
versions of the similar named INF files and they have already been
used and thus useless space takers to be deleted in the first place.
Win98 will mount an inf file and produce the corresponding PNF file
anytime it needs to use an INF file. NT systems have the pnf
extension officially registered as a "Preconfigured INF file" or
somesuch thing.

Your CDROM drive obviously likes to be the slave device, some want to
be masters. The usual fix for getting DMA boxes to stay checked for
CDROM drives is to just fix the same DMA issue for the systems IDE
drives, BUT...

Windows 98 has a built in broken DMA system so as to be able to use
Bus Mastering software to get it all going correctly (I assume, there
being no other excuse). I refer you to Mshdc.inf, and it's
[ESDI_AddReg] section:
HKR,,DriverDesc,,"ESDI Port Driver"
HKR,,DevLoader,,*IOS
HKR,,PortDriver,,ESDI_506.pdr

And Diskdrv.inf and it's [DiskReg] section:
HKR,,,,%DiskClassName%
HKR,,EnumPropPages,,"iosclass.dll,EnumPropPages"
HKR,,SilentInstall,,1
HKR,,NoInstallClass,,1
HKR,,Icon,,"3"

If you will add:
HKR,,IDEDMADrive0,3,01
HKR,,IDEDMADrive1,3,01

to both sections, and then save the inf files back to your inf
folder. Then remove your IDE drives from the Device Mangler and then
reboot, Windows will re-find them and install them with the DMA boxes
already checked.
http://www.microsoft.com/whdc/archive/idedma.mspx

This is a "fix" for DMA issues over at msfn forums and it works -
Windows 98 was built with a broken DMA to begin with. If you fix your
inf files and put them in the Windows\Options\Cabs folder then the
next time install Windows over the top of itself it will be set up
with a working DMA system, it's automatic.

 

[PCR]

Re: A hack, anyone, to turn on dma ? SOLVED ???

Shadow wrote:
| Thank you all for your suggestions. I now have DMA, on by
| default, I did not even have to check the box.

Amazing! I was about to give up hope & bow out off the thread! You did
well! Now, I am beginning to recall ancient threads in which glee &
others would suggest switching master/slave relationships on cables &
even moving from one cable to another-- but I doubt it was only the DMA
setting involved. I think that was for bigger failures, like a device
not working at all. Very good-- very, very good!

| How ? I jumpered the drive to slave .(It was master, on
| secondary ide, is now slave)

Is it alone on the cable & set to slave now? Or did you make something
else on that cable to be the master? It seems odd it could be alone &
set to slave & work. Did you have to move its position on the cable from
center connector to end or visa versa?

| Why ?
| I have no idea. I tried rejumpering it to master and it
| insists it's not capable of DMA. Jumpered to slave ... my DMA back.

Clearly what you have done is a cure to get that checkbox to work! Are
you getting the expected big speed improvement now?

| Windows must have some horrible obscure place where it keeps a
| log of bad CDRoms, I must have used a bad CDrom, and it registered the
| first channel of ide 1 as "bad" with the samsung writer.

One would think the IDE channel wouldn't/shouldn't be marked as bad for
DMA-- it should be the Samsung DVD ONLY to be so marked! Is there
anything else you could plug in there as master to see whether it
suffers the same fate of losing its DMA capability?

| The only file that seems to be updated is mscdrom.pnf. Shall I
| delete it and see what happens ?

I don't know. Some wise ones might say to stop fixing a thing as soon as
it is fixed. But here is mine -- which seems to be updated -- & I don't
even have a DVD at all...

mscdrom.PNF
Loc: C:\WINDOWS\INF
Size: 45,804 bytes
Mod: Thursday, October 10, 2002 08:26:02 PM

That file appears NOT to be in my Win98SE .cabs. It also does not appear
in my SFCLog.txt, making it difficult to divine where it can have come
from.

The .inf is as follows & does exist in the .cabs...

MSCDROM.INF
Loc: C:\WINDOWS\INF
Size: 37,408 bytes
Mod: Friday, April 23, 1999 10:22:00 PM

Cabinet PRECOPY2.CAB
04-23-1999 10:22:00p A--- 37,408 mscdrom.inf

|

--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net

 

[Franc Zabkar]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Sat, 28 Jun 2008 14:22:01 -0300, Shadow <sh@dow> put finger to
keyboard and composed:

> Thank you all for your suggestions. I now have DMA, on by
>default, I did not even have to check the box.
> How ? I jumpered the drive to slave .(It was master, on
>secondary ide, is now slave)
> Why ?
> I have no idea. I tried rejumpering it to master and it
>insists it's not capable of DMA. Jumpered to slave ... my DMA back.

Very strange. Does it burn discs any faster now?

> Windows must have some horrible obscure place where it keeps a
>log of bad CDRoms, I must have used a bad CDrom, and it registered the
>first channel of ide 1 as "bad" with the samsung writer.

Surely any such "log" would have been wiped out by your reinstallation
of Windows?

> The only file that seems to be updated is mscdrom.pnf. Shall I
>delete it and see what happens ?
>

Just to be safe, don't delete it, rename it to mscdrom.pn_

If you are curious, you might like to try this utility:
http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.EXE
http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.TXT

The .bin samples on my web page ...

http://www.users.on.net/~fzabkar/IDE-identify/

.... represent the ATA Identify Drive and ATAPI Identify Packet Device
information blocks for several of my devices.

In your case I would reset your PC, boot to a DOS prompt, run
ideinfo.exe and capture the above data to a file. This should tell you
how the BIOS has set up your drives. I would then type "win" to launch
Windows, allow it to do its thing, and then exit. You should be
returned to the DOS prompt where you can run ideinfo.exe again. Rename
your old file otherwise ideinfo will overwrite it without asking.
Compare the two files, paying attention to *words* 49 (Capabilities),
63 (Multiword DMA modes), and 88 (Ultra DMA modes).

fc /b old_file new_file

This should tell you what Windows has done, if anything.

For an explanation of the above data, see pages 108 - 111 and 89 - 92
of the following document.

Draft copy of ATA-ATAPI-5 standard, revision 3 dated 29 February 2000:
http://www.t10.org/t13/project/d1321r3-ATA-ATAPI-5.pdf

You can use Debug to view your data files:

C:\>debug idedrive.1
-d 100 2ff
-q

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Shadow]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Sun, 29 Jun 2008 14:41:36 +1000, Franc Zabkar
<fzabkar@iinternode.on.net> wrote:

>On Sat, 28 Jun 2008 14:22:01 -0300, Shadow <sh@dow> put finger to
>keyboard and composed:
>
>> Thank you all for your suggestions. I now have DMA, on by
>>default, I did not even have to check the box.
>> How ? I jumpered the drive to slave .(It was master, on
>>secondary ide, is now slave)
>> Why ?
>> I have no idea. I tried rejumpering it to master and it
>>insists it's not capable of DMA. Jumpered to slave ... my DMA back.
>
>Very strange. Does it burn discs any faster now?
I had not tested !!!!!
4mins 48 sec to burn 3.1GB data.
Remember it took 90 minutes to burn 3.9GB before.
Using imgburn freeware. Funny it took longer to verify than to
burn. 5 minutes and 15 seconds.
When I burnt under linux my drive was heavily fragmented, I
had just done a defrag before burning under windows, even so, it
appears windows burns faster. Remember , 10 minutes to burn 3.9GB
under linux.
>
>> Windows must have some horrible obscure place where it keeps a
>>log of bad CDRoms, I must have used a bad CDrom, and it registered the
>>first channel of ide 1 as "bad" with the samsung writer.
>
>Surely any such "log" would have been wiped out by your reinstallation
>of Windows?
>
>> The only file that seems to be updated is mscdrom.pnf. Shall I
>>delete it and see what happens ?
>>
>
>Just to be safe, don't delete it, rename it to mscdrom.pn_
I zipped all pnf and deleted the originals. It made no
difference. Lee was right.
>
>If you are curious, you might like to try this utility:
> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.EXE
> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.TXT

DRIVE 0 Adapter 0 at base address 1f0h Device is not removable.
Disk Reports BIOS Reports
Default Current
# of Cylinders______:16383 16383 1024
# of Heads__________: 16 16 255
# of Sectors/Track__: 63 63 63
Model Number________: ST3120023A
Serial Number_______: 3KA0X43M
Controller Rev. #___: 3.33
Double Word Transfer: No Buffer size (kB)____: 2048
# of ECC bytes______: 4 DMA support_________: Yes
IORDY support_______: Yes IORDY can be disabled: Yes
Standby timer support: Yes Secure Mode Supported: No
# of secs/interrupt_: 16 Current setting 16
LBA support Yes 114473.5MB of LBA addressable, 8063.5MB
in CHS mode
ATA versions supported: ATA-5 ATA-4 ATA-3 ATA-2 ATA-1
Unknown ATA version
Power management features supported
Security features supported
Security frozen
Security level high
SMART features supported
ATA compliant max. PIO transfer mode: 2
SW DMA transfer cycle timing modes:
MW DMA transfer cycle timing modes: 2, 1, 0
ATA-2 PIO transfer modes supported: 4, 3
Min MW DMA transfer cycle time/word: 120 ns 16.7MB/s
Mfg Recommended MW DMA transfer Cycle Time 120 ns 16.7MB/s
Min PIO transfer Cycle Time w/o Flow Control 240 ns 8.3MB/s
Min PIO transfer Cycle Time w IORDY Flow Control 120 ns 16.7MB/s

Adapter 0 at 1f0h Drive 1 not found Last status 0h
Adapter 1 at 170h Drive 0 not found Last status 7fh
DRIVE 1 Adapter 1 at base address 170h

ATAPI, type of device: CD-ROM device with 12 byte command packet
Model Number________: TSSTcorpCD/DVDW SH-S182F
Serial Number_______:
Controller Rev. #___: SB02
Double Word Transfer: No Buffer size (kB)____: 0
# of ECC bytes______: 0 DMA support_________: Yes
IORDY support_______: Yes IORDY can be disabled: Yes
Standby timer support: No Secure Mode Supported: No
# of secs/interrupt_: 0
LBA support Yes
ATA compliant max. PIO transfer mode: 2
SW DMA transfer cycle timing modes:
MW DMA transfer cycle timing modes: 2, 1, 0
ATA-2 PIO transfer modes supported: 4, 3
Min MW DMA transfer cycle time/word: 120 ns 16.7MB/s
Mfg Recommended MW DMA transfer Cycle Time 120 ns 16.7MB/s
Min PIO transfer Cycle Time w/o Flow Control 227 ns 8.8MB/s
Min PIO transfer Cycle Time w IORDY Flow Control 120 ns 16.7MB/s

Adapter 2 at 1e8h Drive 0 not found Last status ffh
Adapter 2 at 1e8h Drive 1 not found Last status ffh
Adapter 3 at 168h Drive 0 not found Last status ffh
Adapter 3 at 168h Drive 1 not found Last status ffh
Adapter 4 at 160h Drive 0 not found Last status ffh
Adapter 4 at 160h Drive 1 not found Last status ffh
Adapter 5 at 150h Drive 0 not found Last status ffh
Adapter 5 at 150h Drive 1 not found Last status ffh
Adapter 6 at 120h Drive 0 not found Last status ffh
Adapter 6 at 120h Drive 1 not found Last status ffh
>
>The .bin samples on my web page ...
ideinfo -F outputs a list of Z at DOS prompt after leaving windows
Z Z Z Z etc for drive 0 . The cdrom is correctly dumped. To get the
output above I just redirected.
ideinfo >> ideinfo.log
>
> http://www.users.on.net/~fzabkar/IDE-identify/
>
>... represent the ATA Identify Drive and ATAPI Identify Packet Device
>information blocks for several of my devices.
>
>In your case I would reset your PC, boot to a DOS prompt, run
>ideinfo.exe and capture the above data to a file. This should tell you
>how the BIOS has set up your drives. I would then type "win" to launch
>Windows, allow it to do its thing, and then exit. You should be
>returned to the DOS prompt where you can run ideinfo.exe again. Rename
>your old file otherwise ideinfo will overwrite it without asking.
>Compare the two files, paying attention to *words* 49 (Capabilities),
>63 (Multiword DMA modes), and 88 (Ultra DMA modes).
>
> fc /b old_file new_file
I use diff from my old cracker days ....
They are identical for the DVD
>
>This should tell you what Windows has done, if anything.
Nothing

 

[Shadow]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Sat, 28 Jun 2008 13:01:31 -0700 (PDT), Lee <melee5@my-deja.com>
wrote:

>The PNF files are of no consequence in Win98, they are simply machine
>versions of the similar named INF files and they have already been
>used and thus useless space takers to be deleted in the first place.
>Win98 will mount an inf file and produce the corresponding PNF file
>anytime it needs to use an INF file. NT systems have the pnf
>extension officially registered as a "Preconfigured INF file" or
>somesuch thing.
You are right
>
>Your CDROM drive obviously likes to be the slave device, some want to
>be masters.
Would you consider it to be curable or just a personality
trait ?

> The usual fix for getting DMA boxes to stay checked for
>CDROM drives is to just fix the same DMA issue for the systems IDE
>drives, BUT...
>
>Windows 98 has a built in broken DMA system so as to be able to use
>Bus Mastering software to get it all going correctly (I assume, there
>being no other excuse). I refer you to Mshdc.inf, and it's
>[ESDI_AddReg] section:
>HKR,,DriverDesc,,"ESDI Port Driver"
>HKR,,DevLoader,,*IOS
>HKR,,PortDriver,,ESDI_506.pdr
>
>And Diskdrv.inf and it's [DiskReg] section:
>HKR,,,,%DiskClassName%
>HKR,,EnumPropPages,,"iosclass.dll,EnumPropPages"
>HKR,,SilentInstall,,1
>HKR,,NoInstallClass,,1
>HKR,,Icon,,"3"
>
>If you will add:
>HKR,,IDEDMADrive0,3,01
>HKR,,IDEDMADrive1,3,01
>
>to both sections, and then save the inf files back to your inf
>folder. Then remove your IDE drives from the Device Mangler and then
>reboot, Windows will re-find them and install them with the DMA boxes
>already checked.
>http://www.microsoft.com/whdc/archive/idedma.mspx
>
>This is a "fix" for DMA issues over at msfn forums and it works -
>Windows 98 was built with a broken DMA to begin with. If you fix your
>inf files and put them in the Windows\Options\Cabs folder then the
>next time install Windows over the top of itself it will be set up
>with a working DMA system, it's automatic.
I will not try this right now, (too tired of hacking) but it
has been archived. I will not forget.
Thanks for the info.

 

[Shadow]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Sat, 28 Jun 2008 16:24:10 -0400, "PCR" <pcrrcp@netzero.net> wrote:

>Shadow wrote:
>| Thank you all for your suggestions. I now have DMA, on by
>| default, I did not even have to check the box.
>
>Amazing! I was about to give up hope & bow out off the thread! You did
>well! Now, I am beginning to recall ancient threads in which glee &
>others would suggest switching master/slave relationships on cables &
>even moving from one cable to another-- but I doubt it was only the DMA
>setting involved. I think that was for bigger failures, like a device
>not working at all. Very good-- very, very good!
>
>| How ? I jumpered the drive to slave .(It was master, on
>| secondary ide, is now slave)
>
>Is it alone on the cable & set to slave now? Or did you make something
>else on that cable to be the master? It seems odd it could be alone &
>set to slave & work. Did you have to move its position on the cable from
>center connector to end or visa versa?
It's all alone. I tried with 3 different cables, using the end
connector and the middle one, it makes no difference. It just has to
be jumpered as slave.
>
>| Why ?
>| I have no idea. I tried rejumpering it to master and it
>| insists it's not capable of DMA. Jumpered to slave ... my DMA back.
>
>Clearly what you have done is a cure to get that checkbox to work! Are
>you getting the expected big speed improvement now?
Yes. under 5 minutes for 3.1GB versus 93 minutes for 3.9Gb
(recording times)
>
>| The only file that seems to be updated is mscdrom.pnf. Shall I
>| delete it and see what happens ?
I deleted it. No difference
>
>I don't know. Some wise ones might say to stop fixing a thing as soon as
>it is fixed. But here is mine -- which seems to be updated -- & I don't
>even have a DVD at all...
>
>mscdrom.PNF
>Loc: C:\WINDOWS\INF
>Size: 45,804 bytes
>Mod: Thursday, October 10, 2002 08:26:02 PM
That was when you installed your cdrom. As Lee pointed out
these are generated files. You can delete them.
>
>That file appears NOT to be in my Win98SE .cabs. It also does not appear
>in my SFCLog.txt, making it difficult to divine where it can have come
>from.
>
>The .inf is as follows & does exist in the .cabs...
>
>MSCDROM.INF
>Loc: C:\WINDOWS\INF
>Size: 37,408 bytes
>Mod: Friday, April 23, 1999 10:22:00 PM
Same as mine. Mine is 05/05/1999 because its a Brazilian
Windows. Brazilians are always late.
>
>Cabinet PRECOPY2.CAB
>04-23-1999 10:22:00p A--- 37,408 mscdrom.inf
>
>|

 

[PCR]

Re: A hack, anyone, to turn on dma ? SOLVED ???

Shadow wrote:
| On Sat, 28 Jun 2008 16:24:10 -0400, "PCR" <pcrrcp@netzero.net> wrote:
|
|>Shadow wrote:
|>| Thank you all for your suggestions. I now have DMA, on by
|>| default, I did not even have to check the box.
|>
|>Amazing! I was about to give up hope & bow out off the thread! You did
|>well! Now, I am beginning to recall ancient threads in which glee &
|>others would suggest switching master/slave relationships on cables &
|>even moving from one cable to another-- but I doubt it was only the
|>DMA setting involved. I think that was for bigger failures, like a
|>device not working at all. Very good-- very, very good!
|>
|>| How ? I jumpered the drive to slave .(It was master, on
|>| secondary ide, is now slave)
|>
|>Is it alone on the cable & set to slave now? Or did you make something
|>else on that cable to be the master? It seems odd it could be alone &
|>set to slave & work. Did you have to move its position on the cable
|>from center connector to end or visa versa?

| It's all alone. I tried with 3 different cables, using the end
| connector and the middle one, it makes no difference. It just has to
| be jumpered as slave.

Interesting. Seems odd. Anyhow, you've got it working. So, no sense in
playing any more with that. My own CD-ROM is alone on its cable. Some
day I must go look at its jumper setting.

It might be interesting to know what happens if you remove the DVD &
plug something else in as a Master on that cable-- does it lose its DMA
too? However, as you've got things going so nice now, I hesitate to
suggest you try. You might end up with some sloppiness in Device Manager
as a minimum ill effect such as a "ghost" that shows up in Safe Mode.

|>
|>| Why ?
|>| I have no idea. I tried rejumpering it to master and it
|>| insists it's not capable of DMA. Jumpered to slave ... my DMA back.
|>
|>Clearly what you have done is a cure to get that checkbox to work! Are
|>you getting the expected big speed improvement now?

| Yes. under 5 minutes for 3.1GB versus 93 minutes for 3.9Gb
| (recording times)

That's quite an amazing improvement! It was well worth your effort &
persistence! You did really well!

|>
|>| The only file that seems to be updated is mscdrom.pnf. Shall I
|>| delete it and see what happens ?

| I deleted it. No difference

Very good. I see Lee knew all about that. I suppose it will harmlessly
return the next time Windows needs to run the .inf.

|>
|>I don't know. Some wise ones might say to stop fixing a thing as soon
|>as it is fixed. But here is mine -- which seems to be updated -- & I
|>don't even have a DVD at all...
|>
|>mscdrom.PNF
|>Loc: C:\WINDOWS\INF
|>Size: 45,804 bytes
|>Mod: Thursday, October 10, 2002 08:26:02 PM

| That was when you installed your cdrom. As Lee pointed out
| these are generated files. You can delete them.

Oops, that's right-- it's the same file for a CD-ROM or DVD-ROM. OK,
thanks.

|>
|>That file appears NOT to be in my Win98SE .cabs. It also does not
|>appear in my SFCLog.txt, making it difficult to divine where it can
|>have come from.
|>
|>The .inf is as follows & does exist in the .cabs...
|>
|>MSCDROM.INF
|>Loc: C:\WINDOWS\INF
|>Size: 37,408 bytes
|>Mod: Friday, April 23, 1999 10:22:00 PM

| Same as mine. Mine is 05/05/1999 because its a Brazilian
| Windows. Brazilians are always late.

Hmm. That explains the strange language in you MSInfo32 postings too!

|>
|>Cabinet PRECOPY2.CAB
|>04-23-1999 10:22:00p A--- 37,408 mscdrom.inf
|>
|>|

--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net

 

[Franc Zabkar]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Sun, 29 Jun 2008 08:51:28 -0300, Shadow <sh@dow> put finger to
keyboard and composed:

>On Sun, 29 Jun 2008 14:41:36 +1000, Franc Zabkar
><fzabkar@iinternode.on.net> wrote:

>>Does it burn discs any faster now?

> 4mins 48 sec to burn 3.1GB data.
> Remember it took 90 minutes to burn 3.9GB before.

>>If you are curious, you might like to try this utility:
>> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.EXE
>> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.TXT
>
>DRIVE 0 Adapter 0 at base address 1f0h Device is not removable.
>Model Number________: ST3120023A

>Double Word Transfer: No Buffer size (kB)____: 2048
># of ECC bytes______: 4 DMA support_________: Yes

>ATA versions supported: ATA-5 ATA-4 ATA-3 ATA-2 ATA-1

>ATA compliant max. PIO transfer mode: 2
>SW DMA transfer cycle timing modes:
>MW DMA transfer cycle timing modes: 2, 1, 0
>ATA-2 PIO transfer modes supported: 4, 3
>
>Adapter 0 at 1f0h Drive 1 not found Last status 0h
>Adapter 1 at 170h Drive 0 not found Last status 7fh


>DRIVE 1 Adapter 1 at base address 170h
>
>ATAPI, type of device: CD-ROM device with 12 byte command packet
>Model Number________: TSSTcorpCD/DVDW SH-S182F

>Double Word Transfer: No Buffer size (kB)____: 0
># of ECC bytes______: 0 DMA support_________: Yes

>ATA compliant max. PIO transfer mode: 2
>SW DMA transfer cycle timing modes:
>MW DMA transfer cycle timing modes: 2, 1, 0
>ATA-2 PIO transfer modes supported: 4, 3

The above reports are incomplete. This is because the author wrote the
program in 1996, well before UDMA made it into the ATAPI-4 spec in
1998.

http://en.wikipedia.org/wiki/ATAPI#ATA_standards_versions.2C_transfer_rates.2C_and_features

For info on the UDMA modes, you need to consult word #88 in your raw
512 bytes of Identify Packet Device data.

>>The .bin samples on my web page ...

>ideinfo -F outputs a list of Z at DOS prompt after leaving windows
>Z Z Z Z etc for drive 0 . The cdrom is correctly dumped. To get the
>output above I just redirected.
>ideinfo >> ideinfo.log
>>
>> http://www.users.on.net/~fzabkar/IDE-identify/
>>
>>... represent the ATA Identify Drive and ATAPI Identify Packet Device
>>information blocks for several of my devices.
>>
>>In your case I would reset your PC, boot to a DOS prompt, run
>>ideinfo.exe and capture the above data to a file. This should tell you
>>how the BIOS has set up your drives. I would then type "win" to launch
>>Windows, allow it to do its thing, and then exit. You should be
>>returned to the DOS prompt where you can run ideinfo.exe again. Rename
>>your old file otherwise ideinfo will overwrite it without asking.
>>Compare the two files, paying attention to *words* 49 (Capabilities),
>>63 (Multiword DMA modes), and 88 (Ultra DMA modes).
>>
>> fc /b old_file new_file
> I use diff from my old cracker days ....
> They are identical for the DVD
>>
>>This should tell you what Windows has done, if anything.
> Nothing

I tried the same test after un-ticking the checkboxes for both my
UDMA-capable CD-ROM and hard drive. Windows did not appear to make any
relevant changes to the IDE controller's chipset registers (they
remained set to UDMA mode), nor did Windows make any changes to the
drives' active UDMA modes. I don't understand this at all. BTW, my
HD's active UDMA mode (as reported by Everest Home Edition) is UDMA 2
(ATA-33), whereas its maximum possible UDMA mode is 4 (ATA-66). This
is consistent with the limitations of my SiS 5597 chipset (ATA-33 SiS
5513 IDE controller).

In short, despite un-ticking my DMA checkboxes, all my hardware
appeared to remain configured for UDMA mode 2.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[PCR]

Re: A hack, anyone, to turn on dma ? SOLVED ???

Franc Zabkar wrote:
| On Sun, 29 Jun 2008 08:51:28 -0300, Shadow <sh@dow> put finger to
| keyboard and composed:
|
|>On Sun, 29 Jun 2008 14:41:36 +1000, Franc Zabkar
|><fzabkar@iinternode.on.net> wrote:
|
|>>Does it burn discs any faster now?
|
|> 4mins 48 sec to burn 3.1GB data.
|> Remember it took 90 minutes to burn 3.9GB before.
|
|>>If you are curious, you might like to try this utility:
|>> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.EXE
|>> http://www.users.on.net/~fzabkar/IDE-identify/IDEINFO.TXT
|>
|>DRIVE 0 Adapter 0 at base address 1f0h Device is not removable.
|>Model Number________: ST3120023A
|
|>Double Word Transfer: No Buffer size (kB)____: 2048
|># of ECC bytes______: 4 DMA support_________: Yes
|
|>ATA versions supported: ATA-5 ATA-4 ATA-3 ATA-2 ATA-1
|
|>ATA compliant max. PIO transfer mode: 2
|>SW DMA transfer cycle timing modes:
|>MW DMA transfer cycle timing modes: 2, 1, 0
|>ATA-2 PIO transfer modes supported: 4, 3
|>
|>Adapter 0 at 1f0h Drive 1 not found Last status 0h
|>Adapter 1 at 170h Drive 0 not found Last status 7fh
|
|
|>DRIVE 1 Adapter 1 at base address 170h
|>
|>ATAPI, type of device: CD-ROM device with 12 byte command packet
|>Model Number________: TSSTcorpCD/DVDW SH-S182F
|
|>Double Word Transfer: No Buffer size (kB)____: 0
|># of ECC bytes______: 0 DMA support_________: Yes
|
|>ATA compliant max. PIO transfer mode: 2
|>SW DMA transfer cycle timing modes:
|>MW DMA transfer cycle timing modes: 2, 1, 0
|>ATA-2 PIO transfer modes supported: 4, 3
|
| The above reports are incomplete. This is because the author wrote the
| program in 1996, well before UDMA made it into the ATAPI-4 spec in
| 1998.
|
|
http://en.wikipedia.org/wiki/ATAPI#ATA_standards_versions.2C_transfer_rates.2C_and_features
|
| For info on the UDMA modes, you need to consult word #88 in your raw
| 512 bytes of Identify Packet Device data.
|
|>>The .bin samples on my web page ...
|
|>ideinfo -F outputs a list of Z at DOS prompt after leaving windows
|>Z Z Z Z etc for drive 0 . The cdrom is correctly dumped. To get the
|>output above I just redirected.
|>ideinfo >> ideinfo.log
|>>
|>> http://www.users.on.net/~fzabkar/IDE-identify/
|>>
|>>... represent the ATA Identify Drive and ATAPI Identify Packet Device
|>>information blocks for several of my devices.
|>>
|>>In your case I would reset your PC, boot to a DOS prompt, run
|>>ideinfo.exe and capture the above data to a file. This should tell
|>>you how the BIOS has set up your drives. I would then type "win" to
|>>launch Windows, allow it to do its thing, and then exit. You should
|>>be returned to the DOS prompt where you can run ideinfo.exe again.
|>>Rename your old file otherwise ideinfo will overwrite it without
|>>asking. Compare the two files, paying attention to *words* 49
|>>(Capabilities), 63 (Multiword DMA modes), and 88 (Ultra DMA modes).
|>>
|>> fc /b old_file new_file
|> I use diff from my old cracker days ....
|> They are identical for the DVD
|>>
|>>This should tell you what Windows has done, if anything.
|> Nothing
|
| I tried the same test after un-ticking the checkboxes for both my
| UDMA-capable CD-ROM and hard drive. Windows did not appear to make any
| relevant changes to the IDE controller's chipset registers (they
| remained set to UDMA mode), nor did Windows make any changes to the
| drives' active UDMA modes. I don't understand this at all. BTW, my
| HD's active UDMA mode (as reported by Everest Home Edition) is UDMA 2
| (ATA-33), whereas its maximum possible UDMA mode is 4 (ATA-66). This
| is consistent with the limitations of my SiS 5597 chipset (ATA-33 SiS
| 5513 IDE controller).
|
| In short, despite un-ticking my DMA checkboxes, all my hardware
| appeared to remain configured for UDMA mode 2.

Did you go through the same rigmarole I did with that? Unchecking DMA
for my IDE-CD R/RW 4x4x24 was uneventful enough. It was gone after the
requested reboot. Fine... BUT -- checking it back on -- I got a
requestor saying I should ask the manufacturer first to see whether DMA
is allowable! Sheesh! I tried a second time. It seemed to take, but
afterwards I got a message that the machine must shut down-- that's
different from a request to reboot! Fortunately, turning back on now,
the thing is set for DMA.

I tracked it with InCtrl5, & the only definitely relevant keys to change
were...

HKEY_LOCAL_MACHINE\Enum\SCSI\IDE-CD__R/RW_4X4X24_____C\MF&CHILD0001&PCI&
VEN_1106&DEV_0571&SUBSYS_00000000&REV_06&BUS_00&DEV_07&FUNC_0100
DMACurrentlyUsed 01 << Binary. It's 00 with DMA unchecked.

HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Class\hdc\0002
IDEDMADRIVE0 01 << Binary. It's 00 with DMA unchecked.

Unfortunately, Shadow had already played with the first to no happy
conclusion. And the second was always set right for him. Also, these
keys changed...

HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Dis
cardable\PostSetup\Component
Categories\{00021493-0000-0000-C000-000000000046}\Enum
Implementing <a bigish binary field>

HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Dis
cardable\PostSetup\Component
Categories\{00021494-0000-0000-C000-000000000046}\Enum
Implementing <another bigish binary field>

But I'm not sure what those do. Can they be the ones that supposedly
deal with the DC-ROM chip? (I tend to think the answer is no!)

No Wininit.ini file was generated to justify a request to reboot.
However, the registry key SetupProgramRan was modified &/or deleted both
times. It's possible another registry key/two went undetected by
InCtrl5, because I didn't complete the run until after the reboot. By
then, it may have come & gone!

| - Franc Zabkar
| --
| Please remove one 'i' from my address when replying by email.

--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net

 

[Franc Zabkar]

Re: A hack, anyone, to turn on dma ? SOLVED ???

On Tue, 1 Jul 2008 20:05:16 -0400, "PCR" <pcrrcp@netzero.net> put
finger to keyboard and composed:

>Franc Zabkar wrote:

>| In short, despite un-ticking my DMA checkboxes, all my hardware
>| appeared to remain configured for UDMA mode 2.
>
>Did you go through the same rigmarole I did with that? Unchecking DMA
>for my IDE-CD R/RW 4x4x24 was uneventful enough. It was gone after the
>requested reboot. Fine... BUT -- checking it back on -- I got a
>requestor saying I should ask the manufacturer first to see whether DMA
>is allowable!

Ditto.

>Sheesh! I tried a second time. It seemed to take, but
>afterwards I got a message that the machine must shut down-- that's
>different from a request to reboot! Fortunately, turning back on now,
>the thing is set for DMA.
>
>I tracked it with InCtrl5, & the only definitely relevant keys to change
>were...
>
>HKEY_LOCAL_MACHINE\Enum\SCSI\IDE-CD__R/RW_4X4X24_____C\MF&CHILD0001&PCI&
>VEN_1106&DEV_0571&SUBSYS_00000000&REV_06&BUS_00&DEV_07&FUNC_0100
>DMACurrentlyUsed 01 << Binary. It's 00 with DMA unchecked.
>
>HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Class\hdc\0002
>IDEDMADRIVE0 01 << Binary. It's 00 with DMA unchecked.
>
>Unfortunately, Shadow had already played with the first to no happy
>conclusion. And the second was always set right for him. Also, these
>keys changed...
>
>HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Dis
>cardable\PostSetup\Component
>Categories\{00021493-0000-0000-C000-000000000046}\Enum
>Implementing <a bigish binary field>
>
>HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Dis
>cardable\PostSetup\Component
>Categories\{00021494-0000-0000-C000-000000000046}\Enum
>Implementing <another bigish binary field>
>
>But I'm not sure what those do. Can they be the ones that supposedly
>deal with the DC-ROM chip? (I tend to think the answer is no!)

The same keys changed for me. I have no idea what the Explorer
settings do, or whether they are significant.

The only other things that occurred to me were ...

(1) that the OS tells the drive to do PIO transfers even though the
drive is configured for UDMA. Is that possible, or does the OS just
issue a read or write command and leave the rest up to the controller?

(2) that the bus master controller is commanded to do PIO transfers
via some other IO ports.

When I look at the IO ports assigned to my IDE controller in Device
Manager, I see ports 0x4000 - 0x4007 assigned to my primary IDE
controller and 0x4008 - 0x400F assigned to the secondary. These are in
addition to the standard ports (1F0, 170). In the OP's case, the
relevant range appears to be 0xFC00 - 0xFC0F. These registers are
described in the following document:

SFF Committee Information Specification for Bus Master Programming
Interface for IDE ATA Controllers Rev 1.0 (SFF-8038i Rev 1.0):
http://ata.freedoors.org/idework/specs/8038-r01.pdf

Note the "Bus Master IDE Status Register" at offsets 02 (0x4002,
primary channel) and 0A (0x400A, secondary channel).

Bits 6 and 5 have the following functions:

=====================================================================
6 Drive 1 DMA Capable: This read/write bit is set by device dependent
code (BIOS or device driver) to indicate that drive 1 for this channel
is capable of DMA transfers, and that the controller has been
initialized for optimum performance.

5 Drive 0 DMA Capable: This read/write bit is set by device dependent
code (BIOS or device driver) to indicate that drive 0 for this channel
is capable of DMA transfers, and that the controller has been
initialized for optimum performance.
=====================================================================

I thought the above looked promising, but when I looked at these
registers using Debug, both in real DOS mode and in a Windows DOS box,
the value of each was 0x04, ie bits 5 and 6 were zero. I verified that
both bits could be set to 1 in real DOS mode. This suggests that
neither drive on either IDE channel is DMA capable, but clearly this
is not the case. Another explanation could be that Windows turns both
bits on in protected (GUI) mode but turns them off again when it exits
to real mode ???

Of course none of the above answers the question why Windows thinks
that the OP's optical drive is not capable of DMA when it obviously
is.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[PCR]

Re: A hack, anyone, to turn on dma ? SOLVED ???

Franc Zabkar wrote:
| On Tue, 1 Jul 2008 20:05:16 -0400, "PCR" <pcrrcp@netzero.net> put
| finger to keyboard and composed:
|
|>Franc Zabkar wrote:
|
|>| In short, despite un-ticking my DMA checkboxes, all my hardware
|>| appeared to remain configured for UDMA mode 2.
|>
|>Did you go through the same rigmarole I did with that? Unchecking DMA
|>for my IDE-CD R/RW 4x4x24 was uneventful enough. It was gone after the
|>requested reboot. Fine... BUT -- checking it back on -- I got a
|>requestor saying I should ask the manufacturer first to see whether
|>DMA is allowable!
|
| Ditto.

Uhuh. It was more than a little scary! For two reboots I thought I had
caught Shadow's disease!

|>Sheesh! I tried a second time. It seemed to take, but
|>afterwards I got a message that the machine must shut down-- that's
|>different from a request to reboot! Fortunately, turning back on now,
|>the thing is set for DMA.
|>
|>I tracked it with InCtrl5, & the only definitely relevant keys to
|>change were...
|>
|>HKEY_LOCAL_MACHINE\Enum\SCSI\IDE-CD__R/RW_4X4X24_____C\MF&CHILD0001&PC
I&
|>VEN_1106&DEV_0571&SUBSYS_00000000&REV_06&BUS_00&DEV_07&FUNC_0100
|>DMACurrentlyUsed 01 << Binary. It's 00 with DMA unchecked.
|>
|>HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Class\hdc\0002
|>IDEDMADRIVE0 01 << Binary. It's 00 with DMA unchecked.
|>
|>Unfortunately, Shadow had already played with the first to no happy
|>conclusion. And the second was always set right for him. Also, these
|>keys changed...
|>
|>HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\D
is
|>cardable\PostSetup\Component
|>Categories\{00021493-0000-0000-C000-000000000046}\Enum
|>Implementing <a bigish binary field>
|>
|>HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\D
is
|>cardable\PostSetup\Component
|>Categories\{00021494-0000-0000-C000-000000000046}\Enum
|>Implementing <another bigish binary field>
|>
|>But I'm not sure what those do. Can they be the ones that supposedly
|>deal with the DC-ROM chip? (I tend to think the answer is no!)
|
| The same keys changed for me. I have no idea what the Explorer
| settings do, or whether they are significant.

Yea, me too. I looked at all the registry keys that seem connected to
those-- CLSID keys. Although I really can't figure it, they don't seem
to be wanting to set a bit in any CD-ROM chip. If that gets done, there
must be some other mechanism that does it.

| The only other things that occurred to me were ...
|
| (1) that the OS tells the drive to do PIO transfers even though the
| drive is configured for UDMA. Is that possible, or does the OS just
| issue a read or write command and leave the rest up to the controller?

I'd have to look that up. I'm thinking ... it could be ... the CD-ROM
device just needs a "DMA bit" to be set, whether it comes set or Windows
sets it. If set, could be software on a CD-ROM chip knows it should use
DMA & write to/from the shared memory area. And Windows will do the
same. That shared memory area probably is set up by Windows. If the bit
is not set, the CD-ROM chip software (if it exists & I think it must to
some extent) & Windows must have to do something else-- which is a LOT
slower going by Shadow's report.

| (2) that the bus master controller is commanded to do PIO transfers
| via some other IO ports.
|
| When I look at the IO ports assigned to my IDE controller in Device
| Manager, I see ports 0x4000 - 0x4007 assigned to my primary IDE
| controller and 0x4008 - 0x400F assigned to the secondary. These are in
| addition to the standard ports (1F0, 170). In the OP's case, the
| relevant range appears to be 0xFC00 - 0xFC0F. These registers are
| described in the following document:

Are you looking at Device Manager, Computer, Input/Output? For that, I
have...

0170-0177 Via Bus Master PCI IDE Controller
0170-0177 Secondary IDE Controller (dual fifo)

01F0-01F7 Via Bus Master PCI IDE Controller
01F0-01F7 Primary IDE Controller (dual fifo)

0376-0376 Via Bus Master PCI IDE Controller
0376-0376 Secondary IDE Controller (dual fifo)

03F6-03F6 Via Bus Master PCI IDE Controller
03F6-03F6 Primary IDE Controller (dual fifo)

And, after that, I have a veritable ton of "aliases" for those. But I
don't have 0x4000-0x4007 or 0x4008-0x400F set for them or for anything.
The final aliases I have for those are...

FF76-FF76 Alias of Via Bus Master PCI IDE Controller
FF76-FF76 Alias of Secondary IDE Controller (dual fifo)

FFF6-FFF6 Alias of Via Bus Master PCI IDE Controller
FFF6-FFF6 Alias of Primary IDE Controller (dual fifo)

| SFF Committee Information Specification for Bus Master Programming
| Interface for IDE ATA Controllers Rev 1.0 (SFF-8038i Rev 1.0):
| http://ata.freedoors.org/idework/specs/8038-r01.pdf
|
| Note the "Bus Master IDE Status Register" at offsets 02 (0x4002,
| primary channel) and 0A (0x400A, secondary channel).
|
| Bits 6 and 5 have the following functions:
|
| =====================================================================
| 6 Drive 1 DMA Capable: This read/write bit is set by device dependent
| code (BIOS or device driver) to indicate that drive 1 for this channel
| is capable of DMA transfers, and that the controller has been
| initialized for optimum performance.
|
| 5 Drive 0 DMA Capable: This read/write bit is set by device dependent
| code (BIOS or device driver) to indicate that drive 0 for this channel
| is capable of DMA transfers, and that the controller has been
| initialized for optimum performance.
| =====================================================================
|
| I thought the above looked promising, but when I looked at these
| registers using Debug, both in real DOS mode and in a Windows DOS box,
| the value of each was 0x04, ie bits 5 and 6 were zero. I verified that
| both bits could be set to 1 in real DOS mode. This suggests that
| neither drive on either IDE channel is DMA capable, but clearly this
| is not the case. Another explanation could be that Windows turns both
| bits on in protected (GUI) mode but turns them off again when it exits
| to real mode ???

I doubt it. GUI mode still is active when a Windows DOS box is open. I'm
not sure what to make of it.

| Of course none of the above answers the question why Windows thinks
| that the OP's optical drive is not capable of DMA when it obviously
| is.

As you know, when he set the DVD to be a slave -- even though it is
alone on its IDE cable -- suddenly DMA works for it. It's very puzzling!

| - Franc Zabkar
| --
| Please remove one 'i' from my address when replying by email.

--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
Should things get worse after this,
PCR
pcrrcp@netzero.net