[REQ][8.09.rc1] DP_Install_Tool - "Unsigned driver" error in DPInst

I had to edit my reply because my english is sorta always turning out like Flemish.
I blame it on my Flemishness.

What kills signing?
There are many factors, and the code aware/driver aware geeks (regulars) at DriverPacks know a good many few of those and have helped work around the pitiful situation hardware vendor driver writers have put all of YOU in.

Driver writer teams apparently take an existing "model" (based on an still older guideline.) and most often just ADD their HWID to the INF, while the core system level driver is also changed.

They rewrite core system of driver because the chip combo they want to support has another mask..
Also when it was proprietary re-manufactored to their company needs.
I know there are system file differences written for the purpose of adding cross-checks in system file, so that a new HWID added SOMETIMES means nothing more than that the HWID was added -embedded-checked- IN the system file.

ALSO NOTE; in CPU fabrication one knows stepping. A stepping is done by re-etching part of the mask, but a new "stepping" of a CPU does not require rewrite of system level driver (an internal optimisation, as it were).
A proprietary build is not a stepping, it is truly a different chip when the code has to allow other interoperability function for the new HWID this mask gets. 
(Hence, the often VERY diffuse line between driver upgrade/new driver.)

The new driver for the new chip was often written with a lot of the old info.. and they left old INFo in the INF.

The default HWID is then not "truly" supported.
Often times, the old default "hwid" no longer works with THEIR new SYS file.

MSFT claims Vista has over 2.5 million drivers.
XP, well, I don't know.
OUR problem is that a HWID does not make a driver, but a driver can break a HWID supported by an older driver.

take an example.
a driver supplied by gigabyte motherboard.
157 HWID on a type of controller. (that's mass, chipset)
The controller chip vendor, default HWID for that type, have 23.
DriverPack MassStorage, if only lists 23, gets numerous requests to add non supported, and if (IF..) when we are lucky we find that the HWIDS were all supported by same core system file.
Great, we take the driver with 157, and do not have to break signing.

Dream on.
Same motherboard vendor has at least 18 versions of drivers on the CD.
Some have overlap, some have poison overlap.
Poison overlap happens when a newer driver portends (pretends?) to support a HWID left in from the 'boilerplate" txt the driver writer worked on while writing the new driver.

The sad fact was/is that because the manufactors have a driver tested against THAT motherboard, it works.
OK, maybe it was good enough for that mobo..
BUT, the damn thing bluescreens on its contemporary siblings, and on its 'fore-fathers'.
And, sometimes that driver Bluescreens on its new mask, or in a new combo (combo of reliable chip with another chip.)

hah, 2.5 million drivers.
a gazzillion HWIDs
NO true database on masks, on core incompatibilities, on cross-chip conflicts a new mask had to fix.
You know, If I were not with my heart for DriverPacks, I'd given up a long time.

Last edited by Jaak (2008-12-07 16:24:51)

Hi Jaak,

Thanks for the warm welcome! I guess you could say I have the balls, but... whether I have the means, that is another question I am quite new to all this! This is currently something I have some time to spend on (for the next month or two), so I don't mind a bit of fishing for answers. I generally try to post only when I have run out of patience.

I wasn't aware of the new RC, thanks for letting me know. I guess I'll try to remember to check more often! I will post again when I've tried the new packs.

With the new base RC and latest test packs, the soundcard, CPU, and RAID all install fine, only the ethernet does not (still with the same error about signing). Perhaps I will test with this PC again soon with new test packs to see if the problem goes away.

Maybe there is a way to achieve what I want to do with the official releases and I am oblivious to it? My goal is to be able to have many up to date virtual machines with various operating systems (Xp home, xp pro, vista 32, vista 64, etc). When I build a new PC, I PXE boot linux, mount the relevant VM disk as a network block device, and dd the partition to the new hard disk. I then reboot and PXE boot LiveXP, run OfflineSysPrep to inject mass storage drivers. Finally, I reboot into windows and run DP_Install_Tool. So far I have everything working except the DP_Install_Tool.

Re your 2nd post: I think I can get past the quirkiness (I just had to read some things twice haha), thanks for the info - I had no idea they made it so difficult for you guys.

Last edited by bombastic (2008-12-07 15:54:30)

Hey YOU.
you do have the balls.

One thing jumping out (what I read, and I do not claim to fully understand.) is off line sys-prep, a non-disc based iso off line, and then a path/signing issue.

I don't have the hardware to run that sort of setup.
I think you will get a few helpful hints by other people.
We do have a sub forum called universal imaging.

---
and I still think you have the balls to test testpacks, so why fish?
I left an angelworm.

Fredric Brown must love me.
Placet is a lonely planet is the best.. no, angelworm was.

Last edited by Jaak (2008-12-07 16:49:28)