Sfor

I'm working with a 1.5TB Seagate drive and a 1TB WD internal drives, currently. The biggest partition size is 512GB. No problems detected, so far. The Seagate proclaims, their sector size translation firmware with SmartAllign technology is significantly faster, then solutions used in other drives. So, the difference between working with 512B and 4096B sectors is not a big one. They claim, their translation firmware is almost as fast as the advanced format alligning software running on a PC.

Right. Then it is safe to use 2TB drives, as long as they are working with 512B sectors. I know the Seagate has a technology of firmware related sector size translation. So it is not necesary to use any additional software with their drives. I'm curious if it is safe to buy just any 2TB drive. Or, perhaps one should carefully examine drives to be 512B sector size compatible.

Thanks to the ESDI_506.PDR patch and updated HDD controller drivers it is possible to work with hard drives bigger than 137GB. But it appears there is another problem with drives above 2TB. So I have a few questions related to this problem. - Does the problem start at 2TB or just above it? - What is the cause (system, driver, BIOS)? - Can it be avoided by reducing the drive capacity through a proper drive configuration? - Is it related to the sector size greater than 512 bytes? ------------------------------------------------- Conclusions: - Problem starts with drives larger than 2TB. - The cause is the old partition structure standard. - Windows 98 has an issue with partitions larger than 1TB.

I was able to install JRE 6u24 with a help of the /s switch. But, I see no difference. On every tested Firefox version, the result is a failed attempt to install Java Runtime Enviroment plugin. In case of Firefox 3.5.19: firefoxjre_exe.exe performed an illegal operation (or something like that). So, in the end, the problem is the Firefox trying to install the Java Runtime Enviroment plugin.

I was unable to install JRE 6u24. At first it complained about gdiplus.dll file missing. Then, after adding a gdiplus.dll from Open Office 2.4.1 installer crashed.

I have problems with getting JRE working with Firefox. There are many versions of Firefox 3 and JRE above 20. I'm looking for a usable combination. There are posts suggesting the Firefox 3.5 works better with Java, than 3.6 does. But, I had no luck to get it working in either one. My test case is the Ultra VNC Java client. It is working correctly with Firefox 2.0.0.17 and JRE 6u7. KernelEX 4.5 final. - JRE 6u7 - Firefox 2.0.0.17 works correctly, Firefox 3.5.19 works correctly, Firefox 3.6.17 fails on installing JRE plugin - JRE 6u20 - does not install - JRE 6u21 - Firefox 2.0.0.17 fails on installing JRE plugin, Firefox 3.5.19 fails on installing JRE plugin, Firefox 3.6.17 fails on installing JRE plugin - JRE 6u22 - Firefox 2.0.0.17 fails on installing JRE plugin, Firefox 3.5.19 fails on installing JRE plugin, Firefox 3.6.17 fails on installing JRE plugin - JRE 6u24 - silent instalation required (/s switch), Firefox 2.0.0.17 fails on installing JRE plugin, Firefox 3.5.19 fails on installing JRE plugin

AC97 devices are connected through the chipset, so the first part of the PCI ID is the chipset ID. The true device ID and manufacturer ID are contained in the SUBSYS_ string. The first four characters are the device ID, and the last four are the manufacturer ID, as far as I remember. So, it is possible to have a correct driver, but with wrong motherboard chipset reference. That's why AC97 driver .INF files tend to have a lot of PCI ID string entries related to multiple chipsets.

Well. The problem is more complicated, apparently. I found three different causes of the shutdown/reboot problems. - the integrated Broadcom Ethernet Controller. (not installing drivers or disabling in BIOS solves the problem) - USB controllers. (disabling BIOS legacy and IRQ support solves the problem) - integrated Intel 865G controller (After installing Intel driver the computer fails to reboot - replaced with other AGP card solved the problem) In the end, I just lost one PCI slot and BIOS legacy USB support. I think I can live with it.

I've tried the APM mode, already. For some reason the PNP Bios does not work, in such a case. The ethernet adapter along with many other devices are not detected at all. So, the APM mode seems to be useless.

I'm trying to install a Windows 98 on a HP/Compaq d530 with Asus P4SD motherboard. The biggest nuisance is the integrated Broadcom NetExtreme ethernet controller, so far. ACPI works well, without the Broadcom NetExtreme driver runnig. The system is shuting down and rebooting smoothly. After installing the Broadcom driver, the Windows 98 system freezes during shutdown or reeboot. I'm out of ideas, for the moment.

I have yet another problem with accessing files from Linux. This time I can not get the polish national characters in file names working correctly in Backup PC. This software uses smbclient, as far as I know. I have little experience with this scripting tool, I'm afraid. Anyways, the SMB access from Nautilus works perfectly. The only problem is Backup PC. I was able to get the national characters working correctly in the administative pannel in Backup PC by upgrading Ubuntu to 11.04. I was able to install a newer version of Backup PC, because of that. Older versions had a problem with processing national characters. Still I do not know, if the current problem is related to the Backup PC processing, or the SmbClient parameters, or something else.

Currently I do have just Seagate ST31000524AS and Western Digital WD10EARS. I had another Seagate 500GB SATA drive a while ago. But, the effect was exactly the same. I do not remember trying to connect both Seagate drives to the controller, at the same time. So, it is possible to say it was always a Seagate and WD pair. So, I can not use two other than Seagate drives, I'm afraid.

I did replace the driver with the one obtained from Dawicontrol page. It does have newer files and istalls without a problem on the Windows 98. Everything seems to be working fine until a second SATA drive gets connected. Then Windows 98 freezes during boot. So, no change for me. I do not know if it is safe to replace the Silicon Image BIOS with the Dawicontrol one. The chip is not exactly the same, I'm afraid.

Have you looked at Cygwin's ssh server? I have serious doubts if the Cygwin with ssh will be able to provide remote access to DOS full screen applications. In theory, since Cygwin is a linux API it should provide a remote acces just for linux applications. On the other hand it could be an option for the remote files transfer problem. I'm well aware of the Team Viewer. But, since I'm planning to use the solution in wider range I can not stick to "for noncommercial use" clausule. This solution is much too expensive to use it in small scale business. Besides, I have a working VPN solution, already. So, the strongest point of the Team Viewer does bring me nothing new. A full version of Real VNC could do the trick (if it works as advertised). It's price is just 30 USD for a licence, so it's way cheaper then the Team Viewer. The freeware version seems to be the most reliable VPN software for Windows 98, so far.

Well, I'm using the PUTTY to access linux console form Windows remotely. But, I see no possibility to access Windows, the same way. Some additional software would have to be running on the Windows remote end, I think. I was looking for a way to access remotely Windows text console for quite a time, already. The VNC does not work for the full screen text modes.

I'm looking for alternative ways to transmit and receive files from remote computers trough VPN connections. The problem is to get a reliable free software with Windows 98 support. 1) Microsoft Networking works, but it is relatively slow and exploited by quite a number of trojans and other unwanted stuff. I found no good way to block the Microsoft Networking services to work in just one direction, as I can not block incoming traffic with a firewall. The Linksys/Cisco RV042 router's firewall does not process the VPN inbound traffic, apparently. 2) Some VNC software do have the file transfer ability, but there are numerous issues with these. Also the file transfer speed seems to be quite a slow one. a ) Tight VNC - the version with Windows 98 support does have a problem with slow connections. The client keeps hanging if there is a need of the full screen refresh. The file transfers seems to be reliable, but slow. The Linux client can not transfer files, apparently. b ) Ultra VNC - very feature rich, yet very problematic on integrated Intel graphic adapters with Windows 98. The client hangs on Windows 98 while receiving files from remote locations (it is necesary to use the Java client instead). I found no good linux client supporting file transfers properly, so far. 3) The FTP seems to be the best choice when the transfer speed is the key factor. Well, the speed is importand, but the simplicity of usage is more important. I do not have much experience to be able to evaluate this idea. I had no opportunity to use Windows 98 based FTP servers or easy to use clients. So, I'm looking for additional ideas and softawe related to these topics. Since all the communication is going through the hardware secured VPN tunnels, the software does not need to use encryption.

It does not have a sense. According to the ATA specification the DMA transfer commands do not contain memory addresses. So, the DMA tranfers have to be done through motherboard DMA controller, as a memory adress has to be passed for a DMA transfer to be performed. I would rather say, the Windows displays just the DMA channels staticaly assigned to particular devices. All other channels can be used by any device, when necesary. So, the free motherboard DMA channels are used by any device in need of such a resource. Unlike IRQ, the available DMA channels are not staticaly assigned to particular devices.

The truth is, other devices are using DMA (the best example is the IDE controller). Windows 98 does not report the real DMA usage, simply.

Well. The Windows 98 has a problem with installation a Windows NT only drivers. If there are two INF files in the same folder, Windows 98 sometimes installs the Windows NT INF file instead of the 9X one. So, Windows 98 does not believe windows NT is a newer system. I had exacly the same problem with SiI3114 controller. So, I've switched to SiI3512. As, the 3112/3512 seems to have a better Windows 98 compatible driver support.

I do have a motherboard with Intel 845 chipset, as well. In the device manager there are just 4 entries. One for the FDD controller, one for DMA controller, ant two for Sound Blaster card. In case of IDE drives, all DMA related settings are available in Intel Application Accellerator pannel. However, it does not cover SATA PCI controllers. So, the effects you mentioned are quite normal, if you have a Intel Application Accellerator installed.

My SiI3512 is working fine, as long as there is just one SATA drive connected to it. But, there is no DMA checkbox available in the Device Manager. Still the performance is just a bit worse than the same drive connected through a SATA to ATA adapter with UDMA-5 working. The difference in speed is just a few percent caused by the bus mastering, probably. I do not have the SiI3114, any more. In any case, I was not able to make it work with windows 98. I do not know the bios version, but only the Si3114r5.inf had a proper identyfication strings for it. So, ir was the Soft Raid 5 version. I'm not using the java utility. The one provided with 3114 was not working for me.

There are driver sets available for both RAID and non-RAID controler versions. The problem is the compatibility with Windows 98 got dropped in both driver sets, quite a few years ago. So, the newer controller versions are no longer covered by the drivers for Windows 98. I do not know if reflashing the controller BIOS to older version can affect the driver compatibility problem. The description of the 3512 chip has got ilustrations from the 3112. The driver sets are the same, as well. So, basicaly, it is the same design. Also, file names in driver sets for 3512 are clearly 3112. There are no files for just the 3512, then. A wile ago, I was playing with a 3114 controller. While the driver was reported as working correctly, the drives were still in DOS compatibility mode. So, no exclamation mark in device manager does not mean, the driver is working correcly. Since there are driver sets available, it can not be the problem with the AHCI interface. This problem is only valid, when using generic drivers for SATA controllers.

Well, I've tried installing the non-RAID drivers forcefuly, already. The driver was not working and all the partition from the drive were in DOS compatibility mode. So, changing the device ID in the .inf file will not do, I think. I have no idea how can I reflash the BIOS in the controller. Unitek does not seem to provide any BIOS upgrades for their PCI-SATA-2R controller.

I was able to get the same version of drivers installed. But nothing have changed. I can not force the 1.5GBit transfer rate on the ST31000524AS. The other drive is WD10EARS, so it does have the jumpers to force the 1.5GBit transfer rate. The drives are both working correctly, when connected without the other one to the same controller. So, the 1.5GBit transfer negotiation should be working fine. The problem starts only when both are connected, in the moment the windows driver is initializing. In some point the HDD LED stops blinking and gets permanently lited. The system gets frozen, before GUI kicks in. As a workaround I'm using a IDE to SATA adapter in order to connect the ST31000524AS through the mainboard IDE port.

Apparently, there some differences in two files in my driver set: si3112r.mpd (version 1,0,57,0) silsupp.cpl (version 3,0,22,0) I'll try to make some experiments related to these files, later.