dencorso

Now you know why I use Corsair, OCZ or Patriot. It varies from year to year... Corsair was the best for many years, but Patriot is now the best, IMHO. OCZ had a really great one in the OCZ ATV Turbo (the yellow and black ones, which came in 4 GB or 8 GB sizes) but they don't make it anymore. The common, blue and orange OCZ ATVs still exist, but are not so good. But really great pendrives are quite expensive. All three above named manufacturers, plus SanDisk have some intermediate price pendrives for everyday use which are quite fine, though. I don't recommend any other manufacturer, but those four.

Did you add a manufacturer driver to 98se already? If so, you've got to remove it (and that means finding which oemxx.inf it became) before redetectig it. Also it might be needed to hunt down its entry in the registry.

That's not so. Even if it were USB 3.0 it would look into USBSTOR.INF, which defines storage devices not usb protocol. Does the pendrive work on XP or 2k?

You see: machine translators are no good. I don't know Chinese language. But you dont know English language. We are even! Impossible display / Impossible use / full capacity. What part of "impossible" did you not understand? Impossible. Cannot be done. Even Bill Gates cannot do it.

Use CONFIG.SYS attached to this post. CONFIG.7z

zip, attach and post your config.sys

Try this. Let me know what happens. PS: by comparing this usbstor.inf with the one you have, you'll find out how it's done. usbstor.7z

No (不是 / 无). Never (从不). Impossible (不可能).

You using Win98SE or WinME?

The Durutti Column - Sketch For The Dawn Their sound remains every bit as eerie and outlandish today as it did way back when. http://www.youtube.com/watch?v=MzhWWpEA4_A

Moreover, without the RAM Limitation Patch, as you go beyond 1 GiB RAM, as is your case, submix8c, Win 98SE tends to oscillate somewhat in the amount of ram it detects, from boot to boot. It's not uncommon to see 1023 MiB, then 1024 MiB, then back to 1023 MiB. If you kick MaxPhysPage a little beyond 40000, then you'll always see at least 1024 MiB, but it will still oscillate around it. As you approach the maximum detectable memory (which also varies a little from board to board) it may oscilate in 2 MiB steps, instead of 1 MiB steps. I reported it here, from the time before I moved on to RLoew's great patch, but at the time I was just puzzled by it. Nowadays I know that's the normal behaviour of the unpatched MS vmm.vxd (even when you use the highest version available of it, which is v. 4.10.0.2226, for Win 98SE). On rereading that post now, 4 years later, I can add that there is a subtle balance among the sizes of the Page File, the ramdisk (unless it's a non-XMS one) and the MaxFileCache sizes, that must be reached for the machine to work, let alone for it to work stably, which was elucidated by RLoew some time afterwards (see this and this). All three posts I pointed to in this post are the result of original research, the first mine, the other two by RLoew, and contain information that, AFAIK, is not findable anywhere else.

Rush - Closer to the Heart

No additional drivers are needed. Determine VID and PID (using USBVIEW or ChipGenius) and add them to USBSTOR.INF. But, with all due respect, that pendrive is the first one I hear NUSB did not detect... are you sure it's working? I have 12 pendrives and 3 external HDDs, plus four SDHC cards, two USB floppies and a SATA/IDE --> USB adapter, and NUSB detects them all correctly.

And MDGx's site offers a lot more of related tweaks.

Yes. Dr. Hoiby's patch is 0x01 --> 0x11 at offset 0x38DD. But the header checksum must be corrected, and that adds 3 more changed bytes. This would be the minimum required for a clean patch. Now, since I've changed the version number, there are some more bytes changed: 0 --> 5 in hexa, 0 --> 5 in ASCII, NT --> 9x in ASCII (yes, I hadn't mentioned this one...) makes four, the VS_VERSION_INFO and StringFileInfo checksums at the Version resource (both catered for by eXeScope, taking care to disallow filesize change) add two more bytes. So this makes it a 10 bytes change, in total. While the Version resource changes are not required, I think it nice to have an easy way to ascertain which file it is, just by looking at the Properties Tab.

@Prozactive: Well, then you have all the relevant updates present and the bug remains. Please try to document in details how to reproduce it, so that we may attempt to swat it. @LoneCrusader: Please find below the required patch pattern. I know you know how to use it, but for the benefit of those who don't, here's how: In order to use the enclosed patch pattern you need to grab the original EXPLORER.EXE v. 4.72.3612.1700 from IE401SP2, size 171.280 bytes, PE Timestamp Mon Feb 08 1999 21:04:25, MD5=2E52FD906097ACB2ADAE3E7E9380F921 CRC32=69912E1E, from inside Ie4shl95.cab, and rename it, say, EXPLORER.IE4. Then download the freeware command-line utils.zip, from KanastaCorp, grab inside it just PATCH.EXE and drop it into the %windir%\command\ folder. Then download the attached file containing explorer401sp2.pat. Now create a temporary folder, say, C:\TEMP\PATCH and put EXPLORER.IE4 and explorer401sp2.pat in it. Then open a DOS box, go to C:\TEMP\PATCH and, from there run the following command: PATCH -p EXPLORER.IE4 explorer401sp2.pat EXPLORER.EXE A new EXPLORER.EXE will be created, and it will be v. 4.72.3612.1705. This file is the EXPLORER.EXE from IE401SP2 patched for the 256 colors icons in system tray. You may use Dr. Hoiby's test to check whether it operates correctly. No icon or bitmap has been modified in this patched version, as per your requirement. explorer401sp2.7z

Because I set MaxPhysPage=40000 in system.cb You can edit it to be MaxPhysPage=48000 and it'll see more RAM

Can't you read English? Stop bumping threads you don't understand. Extract and drop the attached file into C:\WINDOWS: system.zip

Did you create SYSTEM.CB or not?

That's a pity. Never had any problem with it. Then your only easy option is to get Matt Pietrek's PEDUMP.EXE I explained where to get the latest version here. You'll download PE.EXE, and this is a SFX installer, so you can simply open it with WinRAR, 7-zip or your favorite extractor program and get PEDUMP.EXE from inside it whithou any need to run the installer. After you get it try runing, from a DOS box <pedump <path/nameoffile.ext> | find /i "TimeDateStamp" | find /v "00000000">, where <nameoffile.ext> is the name of the PE executable which date you are interested in, and .ext can be .exe, .dll, .ils, .sys, .mpd or a lot of other file extensions. It only works on PE executables, but if you provide it with a file which isn't a PE executable it'll duly complain and exit. I have myself puzzled over this question for a long time too, and here I give you the result of my musings. If you go to MDGx site, the last two lines of this page state the following: Then, by using PEDUMP as described above, getver and dir I compiled this table for the patched versions 1700 and 1710: explorer.exe v. 4.72.3612.1710 size 171.280 PE Timestamp Mon Feb 08 1999 21:04:25 explorer.exe v. 4.72.3612.1700 size 171.280 PE Timestamp Sat Jan 30 1999 00:00:13 Now, you can get IE4SHL95.CAB from three different sources, AFAIK: IE55SP2, IE55SP1 and IE401SP2. From each you can extract a version of explorer.exe, but you'll find that those from IE55SP2 and IE55SP1 are identical, according to fc /b. So this leaves us with just two different versions, which analysis is the following: explorer.exe v. 4.72.3612.1700 size 171.280 PE Timestamp Mon Feb 08 1999 21:04:25 from IE401SP2 explorer.exe v. 4.72.3612.1700 size 171.280 PE Timestamp Sat Jan 30 1999 00:00:13 from IE55SP1/2 So far, these are the hard facts. Below is the explanation I concocted that, IMHO, satisfies all known facts. I believe that explorer.exe from IE401SP2, originally versioned as 4.72.3612.1700 *IS* the unmodded original from which modded explorer.exe v. 4.72.3612.1710 was created, by adding the 256 colors patch and updating some of the icons. Its version was changed to reflect the fact that its compilation date is *newer* than that of the explorer.exe found in IE55SP1/2. A quick and dirty comparison of the relevant files using first eXeScope and then WinHex seems to support my conclusions. So, AFAIK, you already have the file you are looking for. But this is just my opinion... My forensics, at the time, were more thorough than what I implied in the above quote. By now I'm positive the above is the right interpretation of the facts. v 1710 and v. 1700 from IE401SP2 differ in 39% of their bytes, which is consistent with a file patched and resource edited. Visual inspection with Beyond Compare shows all diferences are in the resource section, but for the 256 color patch.

There are no maybes. Either SYSTEM.CB contains MaxPhysPage=48000 and MaxFileCache=131072, or it contains just MaxPhysPage=48000 and Xenó86's VCache is used. Patching IO.SYS by hand may be required, in case xRayeR's patcher isn't liking the IO.SYS given to it. Of course, loading HIMEM.EXE and HIMEM.SYS together will do no good.

Well, 1710 is a later build of 1700. They differ in the PE Timestamp, among other things. And erpdude8, who first found it, also added the 256 colors Tray Icons fix to it. OK, so you have it, and have NUSB installed. Do you also have this update installed? It's the only one I think might make a difference.

The unreferenced antivirus must be F-Prot, BitDefender or GData, judging by the detection string (one of the only pieces of information the OP offered)... Moreover, as submix8c most aptly pointed out: So, I respectfully suggest you all move on to other matters and forget about this, unless (or until) the OP deigns to show up again and demonstrates he/she is at least reading this thread.

Sometimes communication does go awry no matter what we do, doesn't it? That said, let me re-pose you two a question: the final Win Explorer for 98SE is v. 4.72.3612.1710... is that the one you're both using? I do remember having seen optical drive phantoms but I never investigated it further, since I use optical media mainly for backups, so that I don't insert and remove CDs/DVDs much too often. But I think it was more common with previous versions of explorer than with the final one...

This way.