Petr

Then suggest another microcode to remove. It's no problem to change it. Or to use UPDATE.SYS from ME (I know, not for SESP). I has to be noticed that the microcode update is done by the BIOS too, and UPDATE.SYS will replace the microcode only if it has newer than is in the BIOS. I have no expericence with notebooks so I don't know if they have microcode support up to date or not. I do have experience with Gigabyte desktop boards with 440BX chipset, and only latest Coppermine P III processors does not have the latest version in one model. Petr

Maybe I have not expressed it clearly, I have compared original Windows 98 SE file (23-Apr-1999, 22:22, 2.31.0.0) with your (and SESP) file (02-Aug-1999, 2.31.0.1). Petr

I can be seen in properties of the file. Original UPDATE.SYS: 1632-0,2,611-0,b27,612-0,c6,616-0,c6,617-0,c6,619-0,d2,633-0,36,634-0,37,650-4,19,650-20,2e,650-80,2f,650-1,32,651-1,30,652-1,14,652-8,15,652-4,29,652-2,8,653-1,1,653-4,4,660-1,4,665-10,1,66a-2,3,66a-8,7,66a-20,6,671-4,14,672-4,21,672-1,4 Modified UPDATE.SYS: 650-1,40,651-1,40,652-1,2a,653-1,10,660-1,a,665-10,3,672-1,10,673-1,e,681-1,d,681-10,11,683-1,13,683-10,14,686-1,7,686-10,8,68a-10,1,6b1-10,1c,6b4-10,1,f07-1,12,f0a-1,13,f0a-4,14,f12-1,2d,f12-4,2e,f13-4,5,f24-4,1e,f27-4,37,f29-4,2e,f33-d,b (CPUID-Platform,revision) Or, in more detail form: Original UPDATE.SYS: 1632 0 2 10.6.1998 Pentium II OverDrive Processor PGA 611 0 b27 18.12.1996 Pentium Pro Processor B0 CPGA 612 0 c6 10.12.1996 Pentium Pro Processor C0 CPGA 616 0 c6 10.12.1996 Pentium Pro Processor sA0 CPGA 617 0 c6 10.12.1996 Pentium Pro Processor sA1 CPGA 619 0 d2 18.2.1998 Pentium Pro Processor sB1 CPGA 633 0 36 23.9.1998 Pentium II Processor (?) Klamath C0 PGA (?) 634 0 37 23.9.1998 Pentium II Processor (?) Klamath C1 PGA (?) 650 1 32 10.6.1998 Pentium II / Celeron Processor Deschutes / Covington dA0 SECC / SEPP 650 4 19 12.12.1997 Pentium II Xeon Processor (?) Deschutes dA0 SECC (?) 650 20 2e 11.2.1998 Mobile Pentium II Processor (?) Deschutes dA0 Micro-PGA1 (?) 650 80 2f 11.2.1998 ?????????????????????????????? Deschutes dA0 ?????????????? 651 1 30 10.6.1998 Pentium II / Celeron Processor Deschutes / Covington dA0 SECC/SECC2 / SEPP 652 1 14 10.6.1998 Pentium II Processor Deschutes dB0 SECC/SECC2 652 2 08 8.4.1998 Mobile Pentium II Processor (?) Deschutes dB0 Mini-Cartridge 652 4 29 16.9.1998 Pentium II Xeon Processor Drake B0 SECC 652 8 15 10.6.1998 Mobile Pentium II Processor (?) Deschutes dB0 MMC1/MMC2 (?) 653 1 01 14.10.1998 Pentium II Processor Deschutes dB1 SECC/SECC2 653 4 04 29.7.1998 Pentium II Xeon Processor Drake B1 SECC 660 1 04 5.6.1998 Intel Celeron Processor Mendocino mA0 SEPP 665 10 01 27.8.1998 Intel Celeron Processor Mendocino mB0 PPGA 66a 2 03 5.8.1998 Mobile Pentium II Processor mdxA0 Mini-Cartridge 66a 8 07 26.2.1999 Mobile Pentium II / Celeron Processor mdbA0 MMC1/MMC2 66a 20 06 19.8.1998 Mobile Pentium II / Celeron Processor mdpA0 Micro-PGA1 671 4 14 11.8.1998 ?????????????????????????????? SECC (?) 672 1 04 29.10.1998 Pentium III Processor Katmai kB0 SECC2 672 4 21 30.10.1998 Pentium III Xeon Processor Tanner B0 SECC, SECC330 Modified UPDATE.SYS: 650 1 40 25.5.1999 Pentium II / Celeron Processor Deschutes / Covington dA0 SECC / SEPP 651 1 40 25.5.1999 Pentium II / Celeron Processor Deschutes / Covington dA0 SECC/SECC2 / SEPP 652 1 2a 12.5.1999 Pentium II Processor Deschutes dB0 SECC/SECC2 653 1 10 28.6.1999 Pentium II Processor Deschutes dB1 SECC/SECC2 660 1 a 5.5.1999 Intel Celeron Processor Mendocino mA0 SEPP 665 10 3 5.5.1999 Intel Celeron Processor Mendocino mB0 PPGA 672 1 10 22.9.1999 Pentium III Processor Katmai kB0 SECC2 673 1 e 10.9.1999 Pentium III Processor Katmai kC0 SECC2 681 1 d 21.9.1999 Pentium III Processor Coppermine cA2 SECC/SECC2 681 10 11 21.9.1999 Pentium III Processor Coppermine cA2 FC-PGA 683 1 13 6.2.2001 Pentium III Processor Coppermine cB0 SECC2 683 10 14 6.2.2001 Pentium III / Celeron Processor Coppermine cB0 FC-PGA / PPGA 686 1 7 5.5.2000 Pentium III Processor Coppermine cC0 SECC2 686 10 8 5.5.2000 Pentium III / Celeron Processor Coppermine cC0 FC-PGA / PPGA 68a 10 1 2.11.2000 Pentium III / Celeron Processor Coppermine cD0 FC-PGA / PPGA 6b1 10 1c 15.2.2001 Pentium III / Celeron Processor Tualatin tA1 PPGA-370 6b4 10 1 10.1.2002 Pentium III / Celeron Processor Tualatin tB1 PPGA-370 f07 1 12 16.7.2002 Pentium 4 Processor Willamette B2 PPGA-423 INT2 f0a 1 13 16.7.2002 Pentium 4 Processor Willamette C1 PPGA-423 INT2 f0a 4 14 16.7.2002 Pentium 4 Processor Willamette C1 PPGA-478 FC-PGA2 f12 1 2d 2.5.2003 Pentium 4 Processor Willamette D0 PPGA-423 INT2 f12 4 2e 2.5.2003 Pentium 4 Processor Willamette D0 PPGA-478 FC-PGA2 f13 4 5 8.5.2003 Pentium 4 / Celeron Processor Willamette E0 PPGA-478 FC-PGA2 f24 4 1e 5.6.2003 Pentium 4 Processor Northwood B0 PPGA-478 f27 4 37 4.6.2003 Pentium 4 / Celeron Processor Northwood C1 PPGA-478 f29 4 2e 11.8.2004 Pentium 4 / Celeron Processor Northwood D1 PPGA-478 f33 d b 12.5.2004 Pentium 4 / Celeron Processor Prescott C0 All Or you can display microcodes yourself by using ctmc utility ftp://ftp.heise.de/pub/ct/ctsi/ctmc10.zip - just type "ctmc update.sys". Warning - ctmc utility does not support long filenames and does not work under WinXP. It also supports 2048 byte microcode only. Petr

I have just noticed that original version of ctl3dv2.dll is binary exactly the same as version in your zip file, just the version number is different, nothing else. Is it possible that somebody just changed the version number by a hex editor? Therefore there is no need to add this file to the SESP. Petr

Yes, I have not noticed that there is a new page. I'd say just be careful, it's easy to lose your data. Some time ago I discovered how to patch Award BIOSes to break 34GB and 64GB barriers - http://www.ryston.cz/petr/bios/ - but 137GB barrier is worse in its nature. I have no personal experience yet so I don't want to make you sure what is the right way. Petr

You have ASUS A7V400-MX, it means that you should use VIA 4-in-1 drivers probably. But I'm not sure if everything will work, it has to be tried. There is a patch of ESDI_506.PDR but it is not free of charge: http://members.aol.com/rloew1/Programs/Patch137.htm I have no personal experience so I cannot help you further. Petr

What do you mean ? I have crossed it without problems on very standard hardware it seems. <{POST_SNAPBACK}> The 137 GB barrier is caused not by the filesystem (FAT32 in your case), but by the device driver. Standard ESDI506.PDR driver supports standard 28 bit LBA (not 24, it was typo) addressing only, it means 268435456 sectors per 512 bytes = 137438953472 bytes. I already wrote you that you can be sure that you really crossed the barrier only if you filled the disk beyond it. Did you? And the problem is in ESDI506.PDR. So you have to check what driver do you use. Go to the device manager and look what files are asociated with your hard disk controller. Is there ESDI506.PDR? If you still don't believe, please read http://support.microsoft.com/kb/305098/EN-US/ e.g.: OK? Petr

AFAIK these are several different things. Size detection is easy, it is read from the harddrive by "IDENTIFY DEVICE" ATA command, words 60:61, and since the number of addressable sectors is 32-bit number, maximum supported size is 2048 GB. Fdisk does no formating, it just writes the Partition Table contained in the Master Boot Record (MBR). So far no problem with 128 GiB (137 GB) limit. This limit and therefore the problem of old OSes is caused by the fact that the old scheme of Logical Block Addressing (LBA) has 24 bits only. New LBA has 48-bits. In practice, it may look like everyting works smoothly even on system that does not support LBA-48. But only before you cross the 137 GB border. After this border, you will start to write again from the beginning of the disk - overwriting the MBR and making your disk totally unusable with all data lost. More information you can find here: http://www.48bitlba.com/ Petr

It may - (or not?) be good idea to add also updated driver with micocode updates for Intel processors. It is very simple patch replacing not so often used microcodes with new ones. UPDATE.SYS is now updated in each service pack for Windows 2000/XP/2003, but was never updated for Windows 98/SE/ME. For processors before P4 Prescott, each microcode update had the size 2000 bytes without header and 2048 with header, and it was specific for every stepping and for every platform (socket). Starting P4 Prescott, some updates are bigger and most updates are intended for all platforms with the same core stepping. It means different handling in the driver. The number of microcodes included in UPDATE.SYS for different windows versions is: Windows 98: 7 Windows 98 SE: 27 Windows ME: 49 Windows 2000 SP4: 81 Windows XP SP2: 94 Windows 2003 SP1: 101 There is only limited space in the current driver so some limitation has to be applied. Updated UPDATE.SYS for Windows 98 SE, version changed to 4.10.2223 http://old.winpack.org/nsp_petr/mc/update_sys-w98se.zip Contains 26 microcodes for Pentium II, Pentium III and Celeron with cores: Deschutes / Covington / Mendocino / Katmai / Coppermine / Tualatin in Slot 1 and FCPGA-370 packages and for Pentium 4 Willamette and Northwood. No Xeon, EE and mobile processors. Updated UPDATE.SYS for Windows Me, version changed to 4.90.3001 http://old.winpack.org/nsp_petr/mc/update_sys-wme.zip 48 microcodes, more platforms, also EE, mobile and Pentium M. It is functional in Windows 98 SE too. I've added also one 2048 byte microcode for Pentium 4 Prescott stepping C0, CPUID 0F33, I have not tested whether the code can use the right mask compare of the platform ID or used simlple comparison. It should not be so big problem to prepare patch of the original UPDATE.SYS not to distribute Microsoft code. Microcode updates are also contained in BIOSes (Award BIOSes shows PATCH ID: during boot), but very often they are not up to date. For example Gigabyte has no Northwood D1 microcodes in BIOS for my motherboards (GA-8PE667 Ultra2). The success of the microcode update can be seen either by Intel® Processor Frequency ID Utility: http://www.intel.com/support/processors/tools/frequencyid/ but please be sure to use the right version. Version 7.2 (fidenu32.msi), the newest, supports Win98SE to WXP and does NOT support Pentium 4. Version 7.1 (fidenu31.msi), the previous, supports Win98SE to WXP and does support Pentium 4. Version 5.6 was the last version that supported Windows 95 and 98 Gold. New Intel® Processor Identification Utility http://support.intel.com/support/processors/tools/piu/ supports the whole Pentium 4 processor family but only on Windows 2000 and XP. Always it is possible to use bootable (DOS) version of these utilities. In the registry, this is how looks successful update from microcode 21 (in BIOS) to 2E by UPDATE.SYS: [HKEY_LOCAL_MACHINE\Hardware\Description\System\CentralProcessor\0] "VendorIdentifier"="GenuineIntel" "Identifier"="x86 Family 15 Model 2 Stepping 9" "Update Status"=dword:00000000 "Update Signature"=hex:00,00,00,00,2e,00,00,00 "Previous Update Signature"=hex:00,00,00,00,21,00,00,00 I have not found any information about meaning of "Update Status", my observation is the following: [HKEY_LOCAL_MACHINE\Hardware\Description\System\CentralProcessor\0] "VendorIdentifier"="AuthenticAMD" "Identifier"="AMD-K6 3D processor" "Update Status"=dword:00000001 Status 1 seems to mean that the CPU does not support microcode update. [HKEY_LOCAL_MACHINE\Hardware\Description\System\CentralProcessor\0] "VendorIdentifier"="GenuineIntel" "Identifier"="x86 Family 15 Model 2 Stepping 7" "Update Status"=dword:00000002 Status 2 seems to mean that there is no suitable microcode in UPDATE.SYS present. Status 6 can be seen for the second logical CPU for Pentium 4 Hyper-Threading processors. (In WXP, not tested with W98)

I've found very nice plugin for Total Commander, showing both binary and string file version infomation, as well as many other useful things - e.g. DLL dependency. http://physio-a.univ-tours.fr/tcplugins/FILEINFO.htm

Thank you for the clarification, regsvr32.exe 5.0.2134.1 is really from the first Windows 2000 without any service pack. Petr

It may be a problem, but I have found that Windows 98SE can survive even with only one free IRQ for the PCI bus: IRQ 3: COM2 (exclusive) IRQ 4: COM1 (exclusive) IRQ 5: Winbond Memory Stick Storage (MS) Device Driver IRQ 6: Floppy Disk Controller IRQ 7: LPT1 IRQ 8: RTC IRQ 9: 3Com 3C996B Gigabit Server NIC IRQ 9: Intel® PRO/100 VE Network Connection IRQ 9: Win9x-ME Promise MBUltra133 (PDC20276) IDE Controller IRQ 9: Realtek AC'97 Audio IRQ 9: Gigabyte USB Enhanced Host Controller IRQ 9: Intel USB 2.0 Enhanced Host Controller IRQ 9: Intel® 82801DB/DBM USB Universal Host Controller - 24C2 IRQ 9: Intel® 82801DB/DBM USB Universal Host Controller - 24C4 IRQ 9: Intel® 82801DB/DBM USB Universal Host Controller - 24C7 IRQ 9: NEC USB Open Host IRQ 9: NEC USB Open Host IRQ 9: Intel® 82801DB/DBM SMBus Controller - 24C3 IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: IRQ ACPI pro řízení PCI IRQ IRQ 9: Přerušení SCI užívané sběrnicí ACPI IRQ 9: NVIDIA GeForce4 MX 440 with AGP8X IRQ 10: Standard MPU-401 IRQ 11: Winbond Smartcard Reader IRQ 12: PS/2 Mouse IRQ 14: Primary Ultra ATA Controller IRQ 14: Intel® 82801DB Ultra ATA Storage Controller - 24CB IRQ 15: Secondary Ultra ATA Controller IRQ 15: Intel® 82801DB Ultra ATA Storage Controller - 24CB and everything worked perfectly - with exception of one DOS diagnostic utility. More important may be the assignment of the four PCI interupt lines (INTA to INTD) but I'm not sure if it possible to display it in Windows. Petr

It adds this fix's info into IE's About Dialog Box. I see now. Thanks for the explanation, I have no idea what verinst does. Maybe it's from Win2000 SP4, I try to find the source. Win2000 SP4 contains version 5.0.2195.6662. I extracted this file from MOBILE.CAB, but its version is something like 4.10.1500.1 (Both Windows 98 and XP shows it as 4.10.1500.1). I've found the following article in the knowledgebase http://support.microsoft.com/kb/q234573/ explaining that the 5.0.1952.1 version from IE 5.0 is buggy and should be replaced by older version 5.0.1717.2. So no new file :-)

This patch contains only a few REGISTRY settings, they're in the SPUPDATE.INF. OK, and what does the command %winbootdir%\VERINST.EXE q240308 in the RUNPST.BAT file? I have no idea how verinst works. My Q313829 (ENGLISH) contains 3072 bytes VERINST.EXE. You are right, I was confused by the fact that in the Czech version of this patch is regular older 21504 bytes verinst.exe. I was sure that I have downloaded English version, but the name is the same and I lookd into wrong one. About Regsvr32.exe and Ctl3dv2.dll, I'm not sure now. (Ctl3dv2.dll sent by MDGx, Regsvr32.exe is probably from IE60SP1). Regsvr32.exe in IE 6.0SP1 English is 5.0.1586.1, i.e. even older (in Gsetup95.cab). This version is latest and compatible with IE 5.0, 5.5 and 6.0. Yes, but I supposed that you don't place updates from newer IE versions to SESP2. OK then. I have also noticed one file that was included with IE 5.0 but in Win98SE is older: iphlpapi.dll - Windows 98 SE (Net7.cab): 5.0.1717.2 iphlpapi.dll - IE 5.0 (mobile.cab): 5.0.1952.1 Yes, I read often statement like this: "I would definitely recommend you to install Adaptec's 4.60 ASPI version. 4.7x are known to have a lot of problems including installation." But it seems that problems were with versions 4.70 and 4.71, the lastest 4.71.2 shoud be OK - at least I read " ASPI 4.71a2 (currently the latest official release, seems to be another fairly good version)" about it. I remember that two years ago I had a confilct of Adaptec ASPI with some burner software, I don't remember which, and since it has no uninstaller, I had to write my own: rundll32 advpack.dll,LaunchINFSection %windir%\inf\adaptec.inf,Uninstall :-) I was just curious if there are really verified problems of 4.71.2 ASPI layer in comaprison with 4.60. Regards, Petr

Thank you for your detail explanation, and yes, it was typo, I meant INF, not INI file, annd I haven't noticed the BAT file. Petr

Hello Gape, what versioning do you recommend to use when referencing to file version numbers? There are two, sometimes different. 1. Binary version, part of VS_FIXEDFILEINFO structure. It is used by getver and by Windows XP for example. 2. Text version, part of StrinFileInfo structure. For comparison, these are "Binary" and "Text" versions for the same fils: all Win98 files are referenced 4.10.0.2222 (binary) and 4.10.2222 (text) Sometimes the version is just different: CRYPTDLG.DLL 5.0.1558.6072 (binary) and 5.00.1555.1 (text) Sometimes it gives additional information: SCHANNEL.DLL (40-bit) 5.00.1880.14 (binary) and 4.86.1964.1880 (text) SCHANNEL.DLL (128-bit) 5.00.1880.14 (binary) and 4.87.1964.1880 (text) Sometimes the text information is very long: hhctrlui.dll 5.2.3790.1830 5.2.3790.1830 (srv03_sp1_rtm.050324-1447) It should be also noted that getver used on Windows XP system gives incorrect version numbers for the following files: COMCTL32.DLL 5.82.2900.2180 instead of 5.81.4916.400 GDI32.DLL 5.1.2600.2180 instead of 4.10.0.2225 RPCRT4.DLL 5.1.2600.2180 instead of 4.71.3336.0 USER32.DLL 5.1.2600.2180 instead of 4.10.0.2231 And LE files (mostly VXD) does not show the version information at all. It is good to know for those who prepare the W98 updates on Windows XP system. Regards, Petr

Hello Gape, I have tried to understand to everything in SESP2RC3 and the following things are unclear to me: --------------------------------------------------- Maybe instead of 3 lines if exist %winbootdir%\system\iosubsys\SMARTVSD.VXD goto end if exist %winbootdir%\system\SMARTVSD.VXD copy %winbootdir%\system\SMARTVSD.VXD %winbootdir%\system\iosubsys if exist %winbootdir%\system\SMARTVSD.VXD del %winbootdir%\system\SMARTVSD.VXD would be better just one line and without "goto": if exist %winbootdir%\system\SMARTVSD.VXD move %winbootdir%\system\SMARTVSD.VXD %winbootdir%\system\iosubsys --------------------------------------------------- I see no Q240308 patch in the SP but it is referenced in the bat file: if errorlevel 0 %winbootdir%\VERINST.EXE q240308 --------------------------------------------------- 2 to 4 Digit Date Conversion Tool http://www.microsoft.com/windows98/downloa...ion/Default.asp seems to be just for Windows 98 Gold, Windows 98 SE seems to have the 4 digit year display in Control Panel/Regional Settings already set. At least in Czech version of Windows 98 SE. --------------------------------------------------- I was not able to locate the original source of the following files: regsvr32.exe - according to spupdate.inf it is from vs6sp6, but: SESP2RC3 contains version 5.0.2134.1 vs6sp6 contains version 5.00.1641.1 Verinst.exe - according to spupdate.inf it is from 313829, but: SESP2RC3 contains file 3072 bytes, 23.11.1998 9:33 Q313829 contains file 21504 bytes, 21.11.1997 8:10 Ctl3dv2.dll - I don't know what is the source of version 2.31.001. Do you know where they came from? --------------------------------------------------- Then some unclarities with IE: Sens.dll - according to spupdate.inf it is from 258191, but SESP2RC3 contains version 5.50.4807.2300 (from Internet Explorer 5.5 Service Pack 2) Q258191 contains version 5.0.2729.1800 Q319303 - http://support.microsoft.com/kb/q319303 seems to be upgrade just for Microsoft Internet Explorer version 6 but you don't include other updates for MSIE other versions than the 5.0 supplied with Win98SE? --------------------------------------------------- RICHED20.DLL SESP2RC3 contains version 5.30.23.1215 w2ksp4 contains version 5.30.23.1215 wxpsp2 contains version 5.30.23.1221 Maybe this newer version could be used? Or there is any incompatibity? --------------------------------------------------- New hotfixes from Microsoft for Windows 98 Second Edition 888113 - http://support.microsoft.com/kb/q888113 891781 - http://support.microsoft.com/kb/q891781 --------------------------------------------------- Adaptec's ASPI - I'm not sure why version 4.60 is included in SESP2RC3, according to the Adaptec website: http://www.adaptec.com/worldwide/support/d...=aspi_471a2.exe the right version for Windows 98 is 4.71.2, version 4.60 http://www.adaptec.com/worldwide/support/d...ekey=aspi32.exe is older and "requires that either an Adaptec host adapter or an Adaptec version of the ASPI layer be installed on your system" I remember that there were some rumors about a bug in 4.71.2, but this version was released on 23 Nov 2002 and is still recommended by Adaptec. --------------------------------------------------- Why not to add registry entries for QFECheck? They look like: ;qfecheck HKLM,%UpdateKey%\%SPKey%\%LocaleID%%UpdID%\,,,"%UpdName%" HKLM,%UpdateKey%\%SPKey%\%LocaleID%%UpdID%,%10%\SYSTEM32\DRIVERS\PORTCLS.SYS,,"4.10.0.2224" etc. --------------------------------------------------- And thank you for the SESP. Regards, Petr

Hello MDGx, hhctrlui.dll (English version) is version 5.2.3644.0 in your HHUPD.EXE, while in Windows 2003 SP1 Final it is also 5.2.3790.1830 as other files - is it intentional? And why are FIND.COM and WinME start.com included? I've found no reference in the INI file. Petr