Add iphlpapi support

[LuoZijun]

Issue: #561

Files:

C Header	Rust	Review
shared/ifdef.h	shared/ifdef.rs	TODO
shared/ifmib.h	shared/ifmib.rs	TODO
shared/ipifcons.h	shared/ipifcons.rs	TODO
shared/ipmib.h	shared/ipmib.rs	TODO
shared/Iprtrmib.h	shared/iprtrmib.rs	TODO
shared/mprapidef.h	shared/mprapidef.rs	TODO
shared/nldef.h	shared/nldef.rs	TODO
shared/tcpestats.h	shared/tcpestats.rs	TODO
shared/tcpmib.h	shared/tcpmib.rs	TODO
shared/udpmib.h	shared/udpmib.rs	TODO
um/IPExport.h	um/ipexport.rs	TODO
um/IPTypes.h	um/iptypes.rs	TODO
um/iphlpapi.h	um/iphlpapi.rs	TODO

Stage:

• Dependency Tree ( build.rs )
• Header Translation
• Code Style

[LuoZijun]

@retep998 I know this PR have many code style problem, i hope we can put it in the final stage to solve it.

[brooss]

time_t is 32bits on x86

[retep998]

@brooss time_t is 64bit on x86.

[brooss]

Ok, possibly I'm wrong, but from my testing of GetAdapterAddresses() IP_ADAPTER_INFO.LeaseObtained and IP_ADAPTER_INFO.LeaseExpires are 32bit on x86 and 64bit on x86_64.

[retep998]

How exactly were you testing that? VC++ by default uses a 64bit time_t on 32bit. It is possible to get a 32-bit time_t by defining a certain macro, but that's a non-default configuration which I don't support. I don't know what MinGW does.

[brooss]

I'm testing from Rust. Call GetAdaptersInfo() with a zero sized buffer and:

GetAdaptersInfo fills in this variable with the required size, and returns an error code of ERROR_BUFFER_OVERFLOW.

let mut buflen: ULONG = unsafe{mem::zeroed()};
let _ = unsafe{GetAdaptersInfo(null_mut(), &mut buflen)};

buflen should then be a multiple of mem::sizeof::<IP_ADAPTER_INFO>().

I get multiples of 640bytes returned by GetAdaptersInfo on x86.

If time_t is 32bit then sizeof::<IP_ADAPTER_INFO>() is 640bytes and the Lease times look correct to me. If time_t is 64bit then the sizeof::<IP_ADAPTER_INFO>() is 8bytes too long, LeaseObtained has a nonsense time and LeaseExpires is zero.

Just checked on mingw32 and sizeof (IP_ADAPTER_INFO); also shows 640 bytes.

What does sizeof (IP_ADAPTER_INFO); give on vc++?

[retep998]

sizeof(IP_ADAPTER_INFO) in VC++ is 648, but GetAdaptersInfo specifies a value that is a multiple of 640 for me as well. If I call the function with a large enough buffer, then the Next field points to an address that is 640 bytes after the base address.

Also MSDN has this to say:

When using Visual Studio 2005 and later, the time_t datatype defaults to an 8-byte datatype, not the 4-byte datatype used for the LeaseObtained and LeaseExpires members on a 32-bit platform. To properly use the IP_ADAPTER_INFO structure on a 32-bit platform, define _USE_32BIT_TIME_T (use -D _USE_32BIT_TIME_T as an option, for example) when compiling the application to force the time_t datatype to a 4-byte datatype.

So while time_t itself should remain 64bit, IP_ADAPTER_INFO should probably use __time32_t.

[brooss]

So while time_t itself should remain 64bit, IP_ADAPTER_INFO should probably use __time32_t.

👍

[crmaxx]

@LuoZijun on current nightly I have next issue

$ cargo build
   Compiling winapi v0.3.4 (https://github.com/LuoZijun/winapi-rs?branch=iphlpapi#9dddf9d6)
error: no rules expected the token `Value`
  --> C:\Users\mzhukov\.cargo\git\checkouts\winapi-rs-3464f7d97049bddf\9dddf9d\src\shared\ifdef.rs:79:5
   |
79 |     Value: ULONG64,
   |     ^^^^^

error: no rules expected the token `ullAlign`
   --> C:\Users\mzhukov\.cargo\git\checkouts\winapi-rs-3464f7d97049bddf\9dddf9d\src\shared\iprtrmib.rs:152:5
    |
152 |     ullAlign: ULONGLONG,
    |     ^^^^^^^^

error: no rules expected the token `s`
  --> C:\Users\mzhukov\.cargo\git\checkouts\winapi-rs-3464f7d97049bddf\9dddf9d\src\shared\udpmib.rs:52:5
   |
52 |     s: c_int,
   |     ^

error: no rules expected the token `s`
   --> C:\Users\mzhukov\.cargo\git\checkouts\winapi-rs-3464f7d97049bddf\9dddf9d\src\shared\udpmib.rs:103:5
    |
103 |     s: c_int,
    |     ^

error: aborting due to 4 previous errors

error: Could not compile `winapi`.

[LuoZijun]

@crmaxx

Yeah, I guess this is an anonymous macro problem.
It seems that anonymous structure members should use BITFIELD!( not BITFIELD!{.
@retep998 Am i right ?

I developed it on macOS, so fixing the problem took me some time.
I will try to fix that on this weekend.

[retep998]

@LuoZijun According to CONTRIBUTING.md

Macro invocations at the item level should use {}. Macro invocations at the expression level should use ().

Since BITFIELD! is invoked at the item level...

[LuoZijun]

@retep998 Thank you, got it.

[awakecoding]

Any update on this? We would need the same API

[devnought]

Like @awakecoding, also looking to use this. Any updates? @LuoZijun ?

[glebpom]

Hi! Is it going to be merged? Is any help needed?

[retep998]

There is still some work done that needs to be done with this PR before it can be merged, but it appears the original author of this PR has abandoned it, so anyone is free to resume work on this PR.

[mattlknight]

@retep998 I would like to take ownership of this, or at least, help complete this PR. That being said, I am very noob when it comes to FFI stuff. I'm working on gettting this completed in my local fork now, contuining where @LuoZijun left off. Trying to figure out the output of "cargo test" now.

[mattlknight]

@awakecoding @devnought @glebpom I have opened a new PR #675 for this and continued where @LuoZijun left off. Any comments/critiques welcome. I'm now checking to see what is missing, if anything, otherwise the PR is done.

[retep998]

This PR has been superseded by #802