Uninsatll/reinstall HD to recover DMA?

[Terry Pinnell]

I'm trying to restore DMA to one of my HDs. At this page
http://www.microsoft.com/whdc/device/storage/IDE-DMA.mspx
I read:

"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."

This is what I suspect has happened to my 200GB drive J: (which is on
the IDE Primary Slave channel).
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

But I have a couple of queries before I attempt that:

Does "...uninstall and reinstall the device," refer to the following
Device Manager entry:

IDE ATA/ATAPI controllers
- Primary IDE Channel <---- THIS ONE

Or is it
Disk drives
- Maxtor 6L200P0 <---- THIS ONE

And - my key question! - can I use Uninstall and then let XP
automatically reinstall WITHOUT RISK TO MY DATA?

 

[Terry Pinnell]

<snip>

After reassurances elsewhere I went ahead and uninstalled the
Controller. It did indeed reinstall fine. But the bad news is that
it's still showing Current Transfer Mode = PIO
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

I seem to have done what was advised here:
http://www.microsoft.com/whdc/devic...ge/IDE-DMA.mspx

I've also had these two links from Wesley Vogel, to articles that I'll
study:

IDE ATA and ATAPI disks use PIO mode after multiple time-out or CRC
errors occur
http://support.microsoft.com/?kbid=817472

DMA Reverts to PIO
http://www.michna.com/kb/WxDMA.htm

Any other suggestions on how I might get DMA back would be greatly
welcomed. Anyway, there goes Sunday!

 

[Rod Speed]

<snip>

After reassurances elsewhere I went ahead and uninstalled the
Controller. It did indeed reinstall fine. But the bad news is that
it's still showing Current Transfer Mode = PIO
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

I seem to have done what was advised here:
http://www.microsoft.com/whdc/devic...ge/IDE-DMA.mspx

I've also had these two links from Wesley Vogel, to articles that I'll
study:

IDE ATA and ATAPI disks use PIO mode after multiple time-out or CRC
errors occur
http://support.microsoft.com/?kbid=817472

DMA Reverts to PIO
http://www.michna.com/kb/WxDMA.htm

Any other suggestions on how I might get DMA back would be greatly
welcomed. Anyway, there goes Sunday!

Presumably its still getting the high error rate and so is turning DMA off
again.

So you need to fix the high error rate and uninstall again.

Try a new cable.

 

[Peter]

After reassurances elsewhere I went ahead and uninstalled the

Controller. It did indeed reinstall fine. But the bad news is that
it's still showing Current Transfer Mode = PIO
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

I seem to have done what was advised here:
http://www.microsoft.com/whdc/devic...ge/IDE-DMA.mspx

I've also had these two links from Wesley Vogel, to articles that I'll
study:

IDE ATA and ATAPI disks use PIO mode after multiple time-out or CRC
errors occur
http://support.microsoft.com/?kbid=817472

DMA Reverts to PIO
http://www.michna.com/kb/WxDMA.htm

Any other suggestions on how I might get DMA back would be greatly
welcomed. Anyway, there goes Sunday!

What settings for IDE devices do you have set in your motherboard BIOS?

 

[Arno Wagner]

After reassurances elsewhere I went ahead and uninstalled the
Controller. It did indeed reinstall fine. But the bad news is that
it's still showing Current Transfer Mode = PIO
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

I seem to have done what was advised here:
http://www.microsoft.com/whdc/devic...ge/IDE-DMA.mspx

I've also had these two links from Wesley Vogel, to articles that I'll
study:

IDE ATA and ATAPI disks use PIO mode after multiple time-out or CRC
errors occur
http://support.microsoft.com/?kbid=817472

DMA Reverts to PIO
http://www.michna.com/kb/WxDMA.htm

Any other suggestions on how I might get DMA back would be greatly
welcomed. Anyway, there goes Sunday!

Check your cabeling and get current drivers. From my experience
bad cables and bad drivers are the two things that are likely
to cause the errors you are experiencing.

Arno

 

[Terry Pinnell]

Check your cabeling and get current drivers. From my experience
bad cables and bad drivers are the two things that are likely
to cause the errors you are experiencing.

Arno

Thanks all. Had enough for today. Back on the case tomorrow, and will
follow up those suggestions - although cables are pretty new.

 

[Rod Speed]

Thanks all. Had enough for today. Back on the case tomorrow, and
will follow up those suggestions - although cables are pretty new.

Might have been bad out of the box. Thats the main problem with
the technology used, it isnt hard to bend the things that bite the cable
or loosen a connector by pulling on the ribbon to get it off a drive.

 

[Arno Wagner]

Thanks all. Had enough for today. Back on the case tomorrow, and will
follow up those suggestions - although cables are pretty new.

Form my experience cables come defective or are not that easy
to break. And out-of-spec cables (i.e. not flat and/or longer
than 45 cm) may cause problems as well.

Arno

 

[Terry Pinnell]

Check your cabeling and get current drivers. From my experience
bad cables and bad drivers are the two things that are likely
to cause the errors you are experiencing.

I'm about to shut down and start fiddling around with cables etc. But
in parallel, could I get some help on what disk controller driver
update I need and where to find it please?

Key parts of my PC spec: 4-year old AMD Athlon XP1800+ 512MB PC2100
DDR memory; m/b = ASUS A7A266-E with ACPI BIOS v 1011; System Chipset
= M1647 ALiMAGiK 1 AGP System Controller, Buses = AGP/PCI/USB, L2
On-board Cache = 256kB ECC synchronous write-back, HDD = 2 x 200 GB
MAXTOR 6L200P0 and 1 x 60 GB MAXTOR 6L060J3, UDMA-133 EIDE with 2MB
buffer.

Repeating that earlier screenshot of Device Manager:
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg
If I look at Properties|Driver for the 'ALi M5229 PCI Bus Master IDE
Controller', I see the following:
ALi M5229 PCI Bus Master IDE Controller
Driver files:
C:\WINDOWS\System32\DRIVERS\aliide.sys
C:\WINDOWS\System32\DRIVERS\atapi.sys
C:\WINDOWS\System32\DRIVERS\pciidex.sys
Provider: Acer Laboratories Inc.
File version: 1.20
Copyright: Copyright (C) Acer Laboratories Inc. 2000
Digital Signer: Microsoft Windows XP Publisher

There are 2 alternatives offered if I look at the list using Update
Driver:
ALi M5229 PCI Bus Master IDE Controller (Microsoft)
Standard Dual Channel PCI IDE Controller
ALi M5229 PCI Bus Master IDE Controller (Acer Labs Inc)

I'm out of my depth here. For example, I don't really see what
difference there is between the top one (currently in use) and the
bottom, apart from one seems to be from MS and the other form Acer.

So, I'd much appreciate advice on what PCI Controller driver I need
and where I can download it please.

 

[Rod Speed]

I'm about to shut down and start fiddling around with cables etc.
But in parallel, could I get some help on what disk controller
driver update I need and where to find it please?

Basically off the motherboard manufacturer's web site.

Key parts of my PC spec: 4-year old AMD Athlon XP1800+ 512MB PC2100
DDR memory; m/b = ASUS A7A266-E with ACPI BIOS v 1011; System Chipset
= M1647 ALiMAGiK 1 AGP System Controller, Buses = AGP/PCI/USB, L2
On-board Cache = 256kB ECC synchronous write-back, HDD = 2 x 200 GB
MAXTOR 6L200P0 and 1 x 60 GB MAXTOR 6L060J3, UDMA-133 EIDE with 2MB
buffer.

Repeating that earlier screenshot of Device Manager:
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg
If I look at Properties|Driver for the 'ALi M5229 PCI Bus Master IDE
Controller', I see the following:
ALi M5229 PCI Bus Master IDE Controller
Driver files:
C:\WINDOWS\System32\DRIVERS\aliide.sys
C:\WINDOWS\System32\DRIVERS\atapi.sys
C:\WINDOWS\System32\DRIVERS\pciidex.sys
Provider: Acer Laboratories Inc.
File version: 1.20
Copyright: Copyright (C) Acer Laboratories Inc. 2000
Digital Signer: Microsoft Windows XP Publisher

There are 2 alternatives offered if I look at the list using Update
Driver:
ALi M5229 PCI Bus Master IDE Controller (Microsoft)
Standard Dual Channel PCI IDE Controller
ALi M5229 PCI Bus Master IDE Controller (Acer Labs Inc)

I'm out of my depth here. For example, I don't really see what
difference there is between the top one (currently in use) and the
bottom, apart from one seems to be from MS and the other form Acer.

So, I'd much appreciate advice on what PCI Controller driver I need
and where I can download it please.

I've never seen that effect with an XP system.

Arno is a linux fanatic.

 

[Odie Ferrous]

I'm trying to restore DMA to one of my HDs. At this page
http://www.microsoft.com/whdc/device/storage/IDE-DMA.mspx
I read:

"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."

This is what I suspect has happened to my 200GB drive J: (which is on
the IDE Primary Slave channel).
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg

But I have a couple of queries before I attempt that:

Does "...uninstall and reinstall the device," refer to the following
Device Manager entry:

IDE ATA/ATAPI controllers
- Primary IDE Channel <---- THIS ONE

Or is it
Disk drives
- Maxtor 6L200P0 <---- THIS ONE

And - my key question! - can I use Uninstall and then let XP
automatically reinstall WITHOUT RISK TO MY DATA?

Terry,

You're using Maxtor drives.

Take them down to the dump, and replace them with Seagates. 120GB or
300GB - nothing else.


Odie

 

[Terry Pinnell]

Basically off the motherboard manufacturer's web site.




I've never seen that effect with an XP system.

Arno is a linux fanatic.

Hair tearing/teeth gnashing time ;-(

I replaced cable with brand new one, but no change.

Uninstalled/reinstalled Controller umpteen times.

Followed article mentioned earlier and deleted 'checksum' registry
entries.

Still no joy.

FWIW, here are the XP Home registry entries directly after
uninstalling IDE Controller and letting XP reinstall it (which takes 2
reboots). Reminder: The 200 GB drive under discussion is the 'Primary
Slave', on which CHKDSK was recently run from XP.

SCREENSHOT:
-----------
http://www.terrypin.dial.pipex.com/Images/DMA-Registry1.jpg

DETAIL:
------
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E96A-E325-11CE-BFC1-08002BE10318}]
"Class"="hdc"
@="IDE ATA/ATAPI controllers"
"Icon"="-9"
"Installer32"="SysSetup.Dll,HdcClassInstaller"
"TroubleShooter-0"="hcp://help/tshoot/tsdrive.htm"

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E96A-E325-11CE-BFC1-08002BE10318}\0000]
"InfPath"="mshdc.inf"
"InfSection"="aliide_Inst"
"ProviderName"="Microsoft"
"DriverDateData"=hex:00,80,62,c5,c0,01,c1,01
"DriverDate"="7-1-2001"
"DriverVersion"="5.1.2600.1106"
"MatchingDeviceId"="pci\\ven_10b9&dev_5229"
"DriverDesc"="ALi M5229 PCI Bus Master IDE Controller"

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E96A-E325-11CE-BFC1-08002BE10318}\0001]
"InfPath"="mshdc.inf"
"InfSection"="atapi_Inst_primary"
"ProviderName"="Microsoft"
"DriverDateData"=hex:00,80,62,c5,c0,01,c1,01
"DriverDate"="7-1-2001"
"DriverVersion"="5.1.2600.1106"
"MatchingDeviceId"="primary_ide_channel"
"DriverDesc"="Primary IDE Channel"
"MasterDeviceType"=dword:00000001
"SlaveDeviceType"=dword:00000001

"SlaveDeviceDetectionTimeout"=dword:00000001 <-- I wonder if
this is the entry that causes the drop back to PIO mode? And if so,
whether I could safely change it to dword:00000000, or delete it?

"MasterDeviceTimingMode"=dword:00010010

"SlaveDeviceTimingMode"=dword:00000010 <-- I wonder if
this specifies 'PIO' (and the line above 'DMA')?

"MasterIdDataCheckSum"=dword:00014d8d
"SlaveIdDataCheckSum"=dword:00019948
"EnumPropPages32"="storprop.dll,IdePropPageProvider"
"MasterDeviceTimingModeAllowed"=dword:ffffffff

"SlaveDeviceTimingModeAllowed"=dword:0000001f <-- I read
somewhere that another thing to try is looking for low values like
this and restoring them to default (presumably like the entry above).
But it refused to accept this change when I tried it!

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E96A-E325-11CE-BFC1-08002BE10318}\0002]
"InfPath"="mshdc.inf"
"InfSection"="atapi_Inst_secondary"
"ProviderName"="Microsoft"
"DriverDateData"=hex:00,80,62,c5,c0,01,c1,01
"DriverDate"="7-1-2001"
"DriverVersion"="5.1.2600.1106"
"MatchingDeviceId"="secondary_ide_channel"
"DriverDesc"="Secondary IDE Channel"
"MasterDeviceType"=dword:00000002
"SlaveDeviceType"=dword:00000002
"MasterDeviceTimingMode"=dword:00002010
"MasterIdDataCheckSum"=dword:000207ff
"SlaveDeviceTimingMode"=dword:00002010
"SlaveIdDataCheckSum"=dword:00019485
"EnumPropPages32"="storprop.dll,IdePropPageProvider"
"MasterDeviceTimingModeAllowed"=dword:ffffffff
"SlaveDeviceTimingModeAllowed"=dword:ffffffff

Expletives deleted <g>

 

[Arno Wagner]

Hair tearing/teeth gnashing time ;-(

I replaced cable with brand new one, but no change.

45cm 80 pin ''standard'' cable?

Uninstalled/reinstalled Controller umpteen times.

Followed article mentioned earlier and deleted 'checksum' registry
entries.

This is _very_ unlikely to be a software problem. Unless the controller
BIOS does some bad hardware initialisation...

Still no joy.

There is one more component that often has problems that surface
in various surprising ways: The PSU. Can you test the disk with
a different PSU?

Arno

 

[Arno Wagner]

You're using Maxtor drives.

Take them down to the dump, and replace them with Seagates. 120GB or
300GB - nothing else.

If cooled well, they are not that bad. I have about 40 running in an
air-conditioned server room for > 2 years now. Drive temperature is
< 30C, even under high load. The only failures I had are likely drives
damaged in transport and one I am unsure about.

Maxtor drives get really hot not cooled very well. I had Maxtor disks
fail after an hour or so because of overheating when I used them in
'free air' mode (lying besides the PC on the ground), with light load
and 25C room temperature. They heated up to >70C in that time (my
observation on when they fail is 70C...80C). I had the same with a
Maxtor disk in an Agrosy aluminium enclosure. Samsungs and Seagates
stay at someting like room-temperature +15C in this enclosure, even
under load.

Still, I have current backups of all the important stuff (on tape).
No matter what the drive is, backup is non-optional. You can save on
the recovery effort with good drives, but not on the backup effort.

Arno

 

[Terry Pinnell]

45cm 80 pin ''standard'' cable?
Yep.



This is _very_ unlikely to be a software problem. Unless the controller
BIOS does some bad hardware initialisation...


There is one more component that often has problems that surface
in various surprising ways: The PSU. Can you test the disk with
a different PSU?

Thanks, appreciate your patience.

I was about to try replacing it with a MAXTOR 60 GB I have at hand to
see if I still get the 'controller errors'. But instead I first did
something I should have tried sooner: boot up to a copy of XP Home I
made months ago in another partition. In that environment, I am still
in PIO mode on that drive. Does that conclusively rule out software?
And imply it must be either
- the drive itself is fundamentally flawed, despite its newness?
- the power supply?

- and, I suppose, the m/b?

 

[Eric Gisin]

Repeating that earlier screenshot of Device Manager:
http://www.terrypin.dial.pipex.com/Images/Drives1.jpg
If I look at Properties|Driver for the 'ALi M5229 PCI Bus Master IDE
Controller', I see the following:
ALi M5229 PCI Bus Master IDE Controller
Driver files:
C:\WINDOWS\System32\DRIVERS\aliide.sys
C:\WINDOWS\System32\DRIVERS\atapi.sys
C:\WINDOWS\System32\DRIVERS\pciidex.sys
Provider: Acer Laboratories Inc.
File version: 1.20
Copyright: Copyright (C) Acer Laboratories Inc. 2000
Digital Signer: Microsoft Windows XP Publisher

There are 2 alternatives offered if I look at the list using Update
Driver:
ALi M5229 PCI Bus Master IDE Controller (Microsoft)
Standard Dual Channel PCI IDE Controller
ALi M5229 PCI Bus Master IDE Controller (Acer Labs Inc)

I'm out of my depth here. For example, I don't really see what
difference there is between the top one (currently in use) and the
bottom, apart from one seems to be from MS and the other form Acer.

Windows ships with ALi IDE drivers (Microsoft),
you installed (Acer Labs Inc). Is there a newer version?
The "Standard ... IDE" is the generic IDE driver. Give it a try.

 

[Arno Wagner]

Yep.

Then the cable is very likely not the problem.

Thanks, appreciate your patience.

I was about to try replacing it with a MAXTOR 60 GB I have at hand to
see if I still get the 'controller errors'. But instead I first did
something I should have tried sooner: boot up to a copy of XP Home I
made months ago in another partition. In that environment, I am still
in PIO mode on that drive. Does that conclusively rule out software?

Cannot say for sure, but at least it makes it less likely.

And imply it must be either
- the drive itself is fundamentally flawed, despite its newness?

Possible. The only think I can think of is semiconductor failure, very
unlikely today, but it still happens. Maybe static electricity damaged
a bus-driver. That could also have happened on the controller
side. The bus-drivers are hardened against this, but it is still a
possibility.

- the power supply?

Yes. I think you should try that next. It could generate enough
noise to interfere with data transfer on the cables.

- and, I suppose, the m/b?

Well, the controller, yes. Might also be an unlucky combination
of a not too serious problem with controller and disk that only
surfaces when you combine both.

Some more things to try:

- Move the disk to the other controller channel (other connector).
- Try with your other disk.

Arno

 

[Rod Speed]

Yep.

You sure you have it the right way around, with the
correct connector used at the motherboard end ?

Thanks, appreciate your patience.

I was about to try replacing it with a MAXTOR 60 GB I have at
hand to see if I still get the 'controller errors'. But instead I first
did
something I should have tried sooner: boot up to a copy of XP Home
I made months ago in another partition. In that environment, I am still
in PIO mode on that drive. Does that conclusively rule out software?

Not really if you cant be sure that it hasnt always been used in PIO mode,
or that the driver is warty and you are now in the situation where it
gets a significant error rate and so disables DMA for that reason.

And imply it must be either
- the drive itself is fundamentally flawed, despite its newness?

Possible, but not very likely. The obvious test for that possibility
is to put the drive alone in a completely different system and
see if you get DMA disabled in that system with that drive.

- the power supply?

Its possible. Certainly worth trying another if you can do that.

- and, I suppose, the m/b?

Yes, its certainly possible that bad caps on the motherboard
are what are producing the high error rate that is causing the
driver to disable DMA mode. Unfortunately this possibility is
the hardest to test easily. I'd certainly visually inspect it for
obvious bad caps. Presumably you realise that that is electrolytic
caps with bulged tops. Just some of the caps, not all of them.


I'd certainly check the asus web site for an updated driver for
that motherboard, mainly because that is one of the easiest
checks to do. Check for a later bios flash too. You can
certainly get some odd effects with a wart in the bios.

 

[Terry Pinnell]

You sure you have it the right way around, with the
correct connector used at the motherboard end ?
Yes!




Not really if you cant be sure that it hasnt always been used in PIO mode,

I can be sure of that. It was certainly not running like this until
very recently. Apart from the slowness of the drive itself, a side
effect seems to be that the CPU slows down; everything else runs
slowly while the drive is in use.

or that the driver is warty and you are now in the situation where it
gets a significant error rate and so disables DMA for that reason.


Possible, but not very likely. The obvious test for that possibility
is to put the drive alone in a completely different system and
see if you get DMA disabled in that system with that drive.

Or do as I've just done, and replace it with another drive. That's
back to UDMA5!

 

[Terry Pinnell]

[...]

Hair tearing/teeth gnashing time ;-(

I replaced cable with brand new one, but no change.

45cm 80 pin ''standard'' cable?

Yep.

Then the cable is very likely not the problem.

Thanks, appreciate your patience.

I was about to try replacing it with a MAXTOR 60 GB I have at hand to
see if I still get the 'controller errors'. But instead I first did
something I should have tried sooner: boot up to a copy of XP Home I
made months ago in another partition. In that environment, I am still
in PIO mode on that drive. Does that conclusively rule out software?

Cannot say for sure, but at least it makes it less likely.

And imply it must be either
- the drive itself is fundamentally flawed, despite its newness?

Possible. The only think I can think of is semiconductor failure, very
unlikely today, but it still happens. Maybe static electricity damaged
a bus-driver. That could also have happened on the controller
side. The bus-drivers are hardened against this, but it is still a
possibility.

- the power supply?

Yes. I think you should try that next. It could generate enough
noise to interfere with data transfer on the cables.

- and, I suppose, the m/b?

Well, the controller, yes. Might also be an unlucky combination
of a not too serious problem with controller and disk that only
surfaces when you combine both.

Some more things to try:

- Move the disk to the other controller channel (other connector).
- Try with your other disk.

Replaced the newish 200 GB with its 4-year old 60 GB predecessor - and
UDMA 5 is back!

So, although I suppose it *could* still be some obscure software issue
connected with the very large size of the drive, the most likely
explanation appears to be a faulty drive, yes?

Only bought it in October, from Misco, UK. Wonder what my chances of
getting a replacement are...?