Realtime clock - IRQs

[Guest]

Is there a difference in how Win98 and Win2K programically use IRQ's and/or the realtime clock

The context of my questions

I have a developer telling me that an application wrote for win98 will be quite an effort to migrate to Win2K because Win2K doesn't have the same interupts available as was in Win98.

This is somewhat confusing to me because I always thought this was handled at a processor/BIOS level

Can anyone offer any clarification

Thanks

Todd

 

[Lanwench [MVP - Exchange]]

Well, in 9x the application has full access to the hardware - in NT-based
systems the HAL handles that - not sure if this is what the developer was
getting at.

 

[Bob I]

I suppose that the issue of Windows 2000 sharing "software" interupts
throws this developers software into a tizzy as it doesn't find, and
can't have complete control over them.

 

[Dan Seur]

Lanwench & BobI are correct. So is the developer.

To clarify, a DOS-based system like W98 allows applications
to use instructions that directly manipulate real hardware.
No hardware instruction in the set of all machine instructions
is "off limits."

An NT-class system like W2k gives each application a "virtual
machine", which is a software model of the real hardware, and
prevents the app from actually "touching" hardware with certain
types of instructions. W2k will refuse to execute such an app
when it encounters such an instruction. This is in order to
protect apps from each other and to ensure the several
apps that may be executing more or less simultaneously (in
subsecond "time slices") from disrupting each other or the
system itself. This is a very important reliability feature.

It's a big job, often, for a developer to modify a large app
to strip out the "illegal" code and substitute, for each instance,
higher-level code written to comply with the so-called Application
Program Interface" of an NT-class system. This conversion is
often just about as arduous as the initial product was to write.