WebSite Information > Article Discussions
eComStation 2.2 Beta First Impressions
Doug Bissett:
--- Quote ---I'm seeing the occasional problem of the system boots to a working Desktop but the USB mouse is totally unresponsive,
--- End quote ---
This is one of the symptoms that the USB drivers can cause. I have reverted to the 186 version, and have had no more problems with that sort of thing.
I have noticed that boot times seem to be shorter with eCS 2.2 (once I get it going properly). I haven't actually timed it. If you are seeing a problem at the time the logo is on screen, it could be a number of things. The most likely, is that BootDly.SYS is inserting a ridiculously long delay after a SCSI driver is loaded. Inspect your CONFIG.SYS for such a thing. AFAICT, the delay may be needed if you boot from a SCSI device, otherwise, it is not really needed. Some experimenting may be needed to find a more reasonable delay, if it is needed at all. My SCSI card runs my scanner. No delay is necessary for that.
Doug Bissett:
I decided to tackle my old test machine. The machine has a CD drive, not a DVD drive. It also seems that it is difficult to find an IDE DVD drive. So, I got a USB DVD drive.
The first problem is that the machine will not boot from a USB CD/DVD drive. Okay, I made the diskettes. After remaking one bad diskette, I got it to boot, and loaded all four diskettes. It asked for the driver update diskette, which I didn't try, but also gave me a message that the code had not loaded properly, for some reason. Okay, I pressed a key, and the preboot menu showed up. So, I proceeded. The install started, and came to an abrupt halt, when the disk drive could not be "seen". The Dani driver problem (I thought somebody said that only affected a few laptops). So, start over, and select the IBM1S506 driver. Now, the USB DVD drive could not be read. After I powered everything off, and started again, the USB drive was okay, so I proceeded. All went well until it did the first reboot. Now, the USB drive could not be read, again. Okay, I know that is one of the symptoms of the USB drivers, so I rebooted to an alternate boot system, and put the 186 USB drivers in. While I was at it, I put the Dani driver, and DaniFLT.FLT, from eCS 2.1 in and changed the CONFIG.SYS. Rebooted again, and off it went, for a couple of minutes. Then it stopped. No indication of any problems, but it was going nowhere.I realized that I had forgotten to REM the NTFS.IFS and HPFS.IFS lines in CONFIG.SYS, so I rebooted, did that and rebooted again. It continued, and now I am waiting as it is apparently installing CUPS. I think i will kill it because it should not take that long, even if it is downloading it. Hmmm. It took the rest button to break it loose.
More about this saga, if I ever get the time to do it.
Gregg Young:
I finally got the time to do some in depth testing. I avoided the genmac problems. Who knew that having an "UNSUPPORTED" nic would be a good thing. It would be nice to have the option to install the base genmac driver. NAP doesn't install my unsupported driver correctly but MPTS does. The USB & DANI drivers work for me.
The install went fine except for the error between boots "Operating system missing, system halted". Rebooting fixed this. This is caused by a space in the drive label name. It is probably a quoting issue in how the new setboot (eCS 2.1 and higher) passes the name to boot manager.
Drag text doesn't install correctly. The DTProgram template gets the Ftp class instead of the DTProgram class. Not a huge issue and uninstalling DTProgram and reinstalling it fixes it. Split view is also problematic independent of rather you like it or not. Because split view is installed by running a REXX script to set the "Drives" default view to split view any drive that is added after the install (all network and virtual drives) don't default to it. I couldn't get the script to set the additional drives either. Not sure why. The other problem as already discussed is the lack of an easy way to change to the standard behavior.
PNG Desktop was much more of a problem. If I installed it the phase 3 boot always failed to populate the desktop forcing an additional reboot. The Amouse object failed to install in the settings folder and since mouse settings aren't really optional this represents a reason to remove this completely if it can't easily be fixed. The issue of drive labels verses letter which has previously been discussed needs to be addressed also. Lastly, the Drag Text file template icon is lost if you open its properties.
I think this is a very good showing for a beta. I have submitted bug reports for the issues listed above.
Pete:
Hi Doug
I tried to resolve the usb issues I'm seeing by deleting USBOHCD.SYS and USBEHCD.SYS then installing the files from the 10.186 build by Lars Erdmann
Rebooted to a working Desktop with Mouse "active"
I hoped the above had cured the problem but find it has not - 2 of 5 boots have resulted in no mouse "active" when the Desktop has loaded.
I then tried replacing USBD.SYS and USBHID.SYS with the IBM 10.162 builds
Rebooted with no problems, mouse works.
A few reboots later - and no changes to config.sys file - I find boot stopping after loading the genmac marvbell 88e8056 driver.
This is about the time that the usbmouse.sys driver should load...
I tried turning off my Logitech M570 Trackball - just purchased to replace my ancient wired Microsoft Trackball - and then turning it on. No difference...
I pulled the wireless receiver from the USB socket - Boot continued; I reinserted the wireless receiver and the mouse worked fine when the Desktop had loaded. Interesting... I suspect the USB drivers are not even "seeing" the attached mouse until the wireless receiver is reinserted. But, Why not?
This combination of Lars Erdmanns controller drivers build 10.186 and IBMs 10.162 USBD.SYS and USBHID.SYS works fine on this hardware when booting from my eCS2.1 volume. This seems to indicate the problem is not just related to USB drivers.
Regards
Pete
Doug Bissett:
I think that Lars has been doing great work with the USB stuff. However, it seems that the poorly written, and even more poorly followed USB specs are causing problems for him. Throwing in "unusual" devices, like a trackball, doesn't help. It may become a case where users will need to mix, and match, the different parts to find a combination that works. At least Lars is willing to continue trying.
GENMAC also has some very strange problems (that likely won't get fixed), but I haven't seen any evidence that either driver will interfere with the other (doesn't mean that it doesn't happen).
ECS 2.2 beta seems to have picked a couple of bad combinations of both of those drivers, and there are some other problems that may affect what you are seeing. Right now, I am waiting to see what the next beta does. The current one seems to have a number of problems, that have been identified, but working around one, may be causing others to show up. It becomes a chicken and the egg problem, which can waste a whole lot of time.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version