Hard 'C' Question

[AtW]

Assembly is a little more past it than C. What are you doing that can't be achieved using intrinsics BTW? The only time ASM seems needed these days is in SIMD, where you can get massive improvements.

Intrinsics suck arse - I tried them first but found bug in VS 2008 that was causing wrong calculations, confirmed by Microsoft, and yes I was trying to use new SSE4.2 SIMD...

[d000hg]

Oh, I thought you meant standard 64-bit ASM which seemed weird. If you are doing parallel data processing then ASM makes sense, it's really interesting.
Have you tried the Intel compiler?

[Sysman]

Assembly is a little more past it than C. What are you doing that can't be achieved using intrinsics BTW? The only time ASM seems needed these days is in SIMD, where you can get massive improvements.

From my stash of quotes:

Watching discussions of string and macro handling in C. And complaints. C is supposed to be assembler. If that bugs you, use something else.

[AtW]

Oh, I thought you meant standard 64-bit ASM which seemed weird. If you are doing parallel data processing then ASM makes sense, it's really interesting.
Have you tried the Intel compiler?

Yes Intel compiler is the best. Visual Studio C++ dropped support for inline asm in 64-bit compiles

[scotspine]

Err .... you create a U/m include file for your struct.

There it might not include your LIST_ENTRY field, you just have to
cope with this in the K/M data copy routines.

Or, you can cut and paste the LIST_ENTRY definition in to the U/m include file.

Or create a dummy version of your own devising using PVOID.

Basic C programming.

[scotspine]

"so .Netters need not apply" - meaning we have no knowledge of machine code, assembly, instruction sets, direct mode addressing, indirection, bitshift etc? kids stuff!!

[suityou01]

Err .... you create a U/m include file for your struct.

There it might not include your LIST_ENTRY field, you just have to
cope with this in the K/M data copy routines.

Or, you can cut and paste the LIST_ENTRY definition in to the U/m include file.

Or create a dummy version of your own devising using PVOID.

Basic C programming.

Sharing memory between user mode and kernel mode is basic C programming?

[suityou01]

"so .Netters need not apply" - meaning we have no knowledge of machine code, assembly, instruction sets, direct mode addressing, indirection, bitshift etc? kids stuff!!

pretty much, yes.

HTH

[suityou01]

Still not working out for me.

I have a structure in Kernel mode defined as

NET_BUFFER_LIST*

now my user mode app does not know what a NET_BUFFER_LIST is so I declared it as PVOID.

All other members in the struct after the NET_BUFFER_LIST has garbled data.

So the question is

What's the difference in size between a PVOID and a NET_BUFFER_LIST*

Nothing shirley as they're both pointers?

[OrangeHopper]

Pointers can be different lengths from one program to another. It depends on how you have compiled them.

Have you checked what byte alignment is in place? I seem to recall that intel chips allow different policies on demand. Been a while so the up-to-date experts such as scots could correct me on this one.