[Dave-H]

dencorso, on 19 April 2012 - 06:38 PM, said:

OK. So let's investigate the pendrive some more.
1) How did you make it bootable originally (they usually are not)?

I used the HP "USB Disk Storage Format Tool" HPUSBFW.EXE.

dencorso, on 19 April 2012 - 06:38 PM, said:

2) Please download ChipGenius, run it with the pendrive inserted (under XP SP3) and post its report on the pendrive.

Do you remember this?!
Well, it's the same pen drive!
I remember ChipGenius from back then.
Anyway, here's what it says about the pen drive now -

Description: [I:][J:]USB Mass Storage Device(Integral/Integral ICE Flash Drive/ICE Flash Drive)
Device Type: Mass Storage Device

Protocal Version: USB 2.00
Current Speed: High Speed
Max Current: 200mA

USB Device ID: VID = 13FE PID = 1D21
Serial Number: 909A1F2A0006

Device Vendor: Integral
Device Name: ICE Flash Drive
Device Revision: 0110

Manufacturer: Integral/Integral
Product Model: ICE Flash Drive/ICE Flash Drive
Product Revision: PMAP/PMAP

Chip Vendor: Phison
Chip Part-Number: PS2251-56(PS2156) - F/W 01.01.10 [2006-06-26]
Flash ID Code: 98D384A5 - Toshiba [MLC-2K]

Tools on web: http://dl.mydigit.ne.../up/phison.html

dencorso, on 19 April 2012 - 06:38 PM, said:

3) When you insert it, under already running Win XP SP3, and then open Windows Explorer and right-click on the pendrive's letter, is there an "Eject" entry on the pop-up menu (as indicated by the red arrow in the attached image)?

Yes, there is an "eject" entry in the list.

[dencorso]

OMG, it's the Integral ICE USB 2.0 Flash Drive!!! How could I ever have forgotten about it?

Well, I confess I'm baffled. What I can suggest, right now, is: no matter what has happened, the partition containing Win 98 and your data is OK. So, this is a great opportunity to acquire a good old-fashioned and trustworthy partition image, as it is now. Then we can move on to more hardcore experimentation.

I just thought of the "MBR Unbalance" jaclaz has discussed extensively in the past. That may be it. I think I've just corrected for it that MBR I'd sent you before. So the new version is attached. Just copy it to LBA 0 (the first sector of the pendrive), the same way you did it before, and attempt to boot from it. Let's see what happens.

Attached File(s)

•  DAVEMBR2.zip (390bytes)
  Number of downloads: 4

[jaclaz]

Hmmm, you guys had way too much fun while I was away.

@Dave-H
Start form scratch.
Just partition/format the thing with RMPREPUSB.

For the record RMPREPUSB was developed - among other reasons - to correct a number of misbehaviours of the HP Format utility, and applies a couple of "tricks" sometimes useful.
Using the HP Format utility in 2012 is "looking (UNNEEDEDLY) for troubles.

jaclaz

[Dave-H]

Hi again jaclaz, I didn't realise you had been away.

I have downloaded and installed RMPrepUSB 2.1.630 (not the beta version) ready to be deployed if necessary.

I tried den's latest MBR file, and sorry to report it's made no difference.
It still gives "OS not found" if I try and boot directly, and via Plop it will only go to the command prompt without an immediate write error.

I noticed this time that if I do boot to the command prompt with it, I cannot access my system drive (normally drive C: of course, but in this case drive D:).
I can access all the other drives fine, but with drive D: although I can get to a D:> prompt, if I try and actually access the drive I get the error message "Invalid Media Type Reading Drive D". Very strange.

I get the write error on the C: drive again if I try and copy anything to it from one of the dives I can access.

I'm beginning to think that jaclaz is right, and we're running out of options to actually fix this, educational though it has been, and maybe it is time to just reformat the pen drive, copy my backup to it, and hopefully then be able to boot to Windows 98 from it as before.
What do you reckon den?

[dencorso]

Dave-H, on 21 April 2012 - 09:37 AM, said:

I'm beginning to think that jaclaz is right, and we're running out of options to actually fix this, educational though it has been, and maybe it is time to just reformat the pen drive, copy my backup to it, and hopefully then be able to boot to Windows 98 from it as before.
What do you reckon den?

Yeah, enough is enough! Just make sure your backup is really up-to-date, or make a further one, just in case. Then go ahead, as jaclaz suggested. Let's get it working once more.

@jaclaz: Welcome back, my friend!

[Dave-H]

OK, I've re-prepared the drive using the RM tool, copied my backup to it, and I now have a working bootable pen drive again, which will boot to Windows 98!

I guess that the HP and RM utilities must do something to the drive which Plop undid, and we just weren't able to redo it, whatever it was!
Incidentally it's no longer triggering the XP AutoPlay routine when I insert it, which must mean something.

I've tried again booting with the Plop floppy disk, and accessing the pen drive using its driver.
Unfortunately, this is just producing the write failure error message again.
It boots fine without Plop, but won't boot with it.
I does look to me as if the Plop USB driver does indeed make the drive read only, in which case we were onto a loser with it right from the start!

Any other options for USB 2.0 in this scenario?

This post has been edited by Dave-H: 22 April 2012 - 11:07 AM

[jaclaz]

Dave-H, on 21 April 2012 - 09:37 AM, said:

I'm beginning to think that jaclaz is right,

beginning?

The PLoP DOS USB driver is by design Read Only.

Once it has been used for the first part of booting it is normally replaced by ANOTHER driver (capable of R/W operations).
This happens "by default" when you boot a NT based system (which "protected mode" part loads it's own drivers) :
http://www.rmprepusb.../tutorials/plop
(or any Linux distro)
up to you to find a way that may work in DOS (and/or Windows 9x/Me), here is some loose info:
http://forum.plop.at....php?topic=92.0

BTW there is NO actual need for you to use PLoP, it was dencorso's idea to try using it..... (but - to be fair - only in order to test the USB booting speed of your mothrboard ).


jaclaz

This post has been edited by jaclaz: 22 April 2012 - 10:30 AM

[dencorso]

jaclaz, on 22 April 2012 - 10:29 AM, said:

BTW there is NO actual need for you to use PLoP, it was dencorso's idea to try using it..... (but - to be fair - only in order to test the USB booting speed of your mothrboard ).

Well, my idea was simply to chronometer how long does it take to boot to plain DOS (the command prompt) on a direct boot vs. booted by Plop (if one selects fast the USB option, that would do... but configuring Plop to boot hiddenusb would probably be best). USB 2.0 is so much faster than USB 1.1 that any imprecisions in the measuring the boot incept point would be irrelevant. So, either it would boot faster with Plop (meaning that the motherboard boots at USB 1.1 speeds) or it would boot slower with Plop (meaning the motherboard boots at USB 2.0 speeds), because of the overhead of using the floppy. But I ought to have suggested to Dave to use another pendrive, made bootable to DOS only, just for the purpose of this test. Lesson learned. Sorry, Dave, for the trouble I caused.
Yet, we still don't know whether Dave's board boots at USB 1.1 or 2.0 speeds, and we never will, unless we measure the boot times.

[jaclaz]

dencorso, on 22 April 2012 - 08:20 PM, said:

Well, my idea was simply to chronometer how long does it take to boot to plain DOS

Sure, you meant well : but the result is currently :

• (possibly) *fast* NOT working

vs:

• (possibly) *slow* working

I would rather have the Win9x actually working (no matter whether *slow* or *fast*) and only later see if it can be sped up.....

jaclaz

This post has been edited by jaclaz: 23 April 2012 - 03:30 AM

[Dave-H]

Thanks guys!
Well I do have it working again, just rather slowly.
I'm pretty sure my motherboard would only be natively USB 1 as it dates from 2003.

If it's possible to get USB 2 speeds out of it in any way that would be good, if only that it might improve the stability of Windows 98 when running from the pen drive.
The guy who designed Plop does say in response to a forum query about this that he was considering releasing a driver that gave read and write access for this specific configuration only (because of the need to be able to fit it all on a floppy disk) but that was written back in 2009, so I assume it's never happened.

I also looked here as it was referenced in the same forum, and I did download the package, but I'm afraid my eyes glazed over when I started looking at the documentation, and I never determined whether there was anything there that would help with this.

[jaclaz]

Dave-H, on 23 April 2012 - 06:47 AM, said:

Thanks guys!
Well I do have it working again, just rather slowly.
I'm pretty sure my motherboard would only be natively USB 1 as it dates from 2003.

And the easiest way to check if something can be done is simply to check whether you are using:

• OHCI = USB 1.1
  
• UHCI = USB 1.0
  
• EHCI = USB 2.0

drivers (i.e which kind of chipset you have).
If you have EHCI chipset BUT your BIOS does not use them with USB 2.0 speed, THEN PLoP (or other solutions) my be useful, if the chipset is OHCI or UHCI, there is no way you can speed it up.

jaclaz

[Dave-H]

This is the relevant section from my motherboard manual -

The Intel E7505 chipset is a high-performance chipset with a performance
and feature-set designed dual processor servers. The E7505 chipset consists
of three major components: the Memory Controller Hub (MCH), the I/O
Controller Hub 4 (ICH4) and the PCI-X 64-bit Hub 2.0 (P64H2).

The MCH has four hub interfaces, one to communicate with the ICH4 and
three for high-speed I/O communications. The MCH employs a 144-bit wide
memory bus for a DDR-266 memory interface, which provides a total bandwidth
of 4.27 GB/s. The ICH4 interface is a 266 MB/sec point-to-point
connection using an 8-bit wide, 66 MHz base clock at a 4x data transfer
rate. The P64H2 interface is a 1 GB/s point-to-point connection using a 16-
bit wide, 66 MHz base clock at an 8x data transfer rate.

The ICH4 I/O Controller Hub provides various integrated functions, including
a two-channel Ultra ATA/100 bus master IDE controller, USB 2.0 host controllers,
an integrated LAN controller, a System Management Bus controller
and an AC'97 compliant interface.

The P64H2 PCI-X Hub provides a 16-bit connection to the MCH for highperformance
IO capability and two independent 64-bit PCI-X interfaces.

I was a bit surprised to find that it does actually mention USB 2.0 controllers, but presumably that doesn't necessarily mean that it would use USB 2.0 if booting from a USB device. It could just mean that it has support for USB 2.0 when used with a suitable operating system.
If it is already using USB 2.0 then presumably it's already working as well as it's ever going to do!

This post has been edited by Dave-H: 23 April 2012 - 12:17 PM

[dencorso]

Dave-H, on 23 April 2012 - 12:16 PM, said:

It could just mean that it has support for USB 2.0 when used with a suitable operating system.

I remembered your motherboard has a ICH4 southbridge, so I knew it has USB 2.0 capable hardware.
But many boards having USB 2.0 capable hardware do boot at USB 1.1 speeds (for compatibility [with what?], I guess).
And some netbooks, like the Eee PC 900 have an even stranger behaviour.
There's a setting in the Advanced BIOS setup called, rather cryptically:

"OS Installation", which options are "Finished" and "Start"...

Now, then, if one sets it to "Start", it boots at USB 1.1 speeds, and setting it to "Finished" causes the machine to boot at USB 2.0 speeds. And this is rather counter-intuitive, because, when installing Win XP on it, one tends to set that to "Start", which causes the installation process to last for ages... The right way to do it, of course, is to install XP with the "Finished" setting from the start!!!
Yet, the Eee PC 900 Manual insists one *must* use the "Start" setting to install XP. Go figure!

[jumper]

dencorso, on 23 April 2012 - 09:36 PM, said:

Now, then, if one sets it to "Start", it boots at USB 1.1 speeds, and setting it to "Finished" causes the machine to boot at USB 2.0 speeds. And this is rather counter-intuitive, because, when installing Win XP on it, one tends to set that to "Start", which causes the installation process to last for ages... The right way to do it, of course, is to install XP with the "Finished" setting from the start!!!
Yet, the Eee PC 900 Manual insists one *must* use the "Start"setting to install XP. Go figure!

Although we understand what you mean, "USB 1.1 speeds" and "USB 2.0 speeds" is incorrect usage. There is low-speed and full-speed (USB 1.x and 2.0 controllers), and hi-speed (most, but not all, USB 2.0 controllers). I have one PCI USB 2.0 controller that only supports low and full speeds.

Perhaps XP doesn't install the driver correctly if the EHCI controller is already configured for hi-speed. Or maybe full-speed is considered more reliable than hi-speed by the Eee PC engineers.

This post has been edited by jumper: 26 April 2012 - 12:20 AM

[dencorso]

jumper, on 26 April 2012 - 12:13 AM, said:

Although we understand what you mean, "USB 1.1 speeds" and "USB 2.0 speeds" is incorrect usage. There is low-speed and full-speed (USB 1.x and 2.0 controllers), and hi-speed (most, but not all, USB 2.0 controllers). I have one PCI USB 2.0 controller that only supports low and full speeds.

Perhaps XP doesn't install the driver correctly if the EHCI controller is already configured for hi-speed. Or maybe full-speed is considered more reliable than hi-speed by the Eee PC engineers.

Low-speed (= 1.5 MBit/s or 187.5 KB/s) and Full-speed (= 12 Mbit/s or 1.5 MB/s) *are* the USB 1.x speeds. Hi-speed (480 Mbit/s or 60 MB/s) is the USB 2.0 speed. Low-speed, Full-speed, High-speed and now Super-speed (600 MB/s, for USB 3.0) are official but confusing terms. Moreover, USB 3.0 devices in compatibility mode report they are "USB 2.1", so perhaps the less confusing terms would be "USB1" or "USB2" or "USB3" speeds.

Now, any USB 2.0 device which can do *only* Low-speed and Full-speed is either broken or a fake. In my experience (with USB hubs, mainly), they always were relabelled fakes. But, with Chinese hubs costing US$ 1 or less apiece, the risk of buying fake hubs does not worry me much. Controllers are another story... if yours has a true USB 2.0 chip, like a VIA 6202, then it's broken, not a fake.

I'm positive that win XP SP2 and SP3 install flawlessly on Asus Eee PCs 701 or 900, with the BIOS set to "Finished". I've done it more times than I can count. Probably the problem existed with the very earliest BIOS versions, and ASUS never bothered the modify their manuals.

[PeteB]

I was surprised to find years ago that there was such a range in performance for flash drives even within say USB 2.0 class.
Read Access time benchmark:
http://www.tomshardw...-3.13,2289.html

Read throughput:
http://www.tomshardw...erage,2292.html

Many more:
http://www.tomshardw...hmarks,109.html

I would agree that it is more likely that the USB does not default to full speed on boot, but it is still surprising that there is such a wide range in speed.
Are you able to disable legacy USB support in the BIOS or are there any options as far as default speed?

[dencorso]

A simple question you may already have answered elsewhere, Dave, but bear with me:
Does your pendrive get recognized right either by ASPIEHCI.SYS/GUEST.EXE or by USBASPI.EXE/DI1000DD.SYS and work OK?

[Dave-H]

PeteB, on 28 April 2012 - 02:46 PM, said:

Are you able to disable legacy USB support in the BIOS or are there any options as far as default speed?

Yes I can disable legacy USB support in the BIOS, but the pen drive isn't then detected of course.
There are no USB speed adjustment options.

dencorso, on 30 April 2012 - 03:34 AM, said:

A simple question you may already have answered elsewhere, Dave, but bear with me:
Does your pendrive get recognized right either by ASPIEHCI.SYS/GUEST.EXE or by USBASPI.EXE/DI1000DD.SYS and work OK?

Sorry, you've lost me!

None of those file exist on my system.

This post has been edited by Dave-H: 30 April 2012 - 04:45 AM