Turns out to be useful for complex cases.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Especially where we have just done a read and spin off a duplex to do a read
as well.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

eax/ebx/ecx/edx are 32-bit registers, and we need 8 bits from most
significand bits so there is no need to mask out 0xFF.
Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Otherwise, it's a PITA to close a duplexed connection.  If necessary we
can introduce a half-close to de-deplex later.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Also take out additional feature information from processor information
and feature bits since they can be tested using cpuid_test_feature() and
cpuid_has_feature()
Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

For now, we will just test for normal features, later on it'll be added
back.
Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Both run-set_alloc and run-15-timeout used the same port, so they sometimes
got stuck when running in parallel.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Apparently init --user adopts orphans.  To quote the author
Stewart Smith:

  As much as one can be happy in Ubuntu breaking something that has been
  true for what must be approaching 40 years, yep, I'm happy for you to
  make the changes.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Trivial, but they make coding easier and more predictable.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

MSVC only supports inline assembly and does not support the keyword
volatile for assembly.
Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

This is a much more cleaner way to do it and _should_ be easier for people
to use.
Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Signed-off-by: Ahmed Samy <f.fallen45@gmail.com>

Useful for getsockname().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

I really wanted an array of bytes in there, so make it more flexible.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Not a perfect solution (we'd ideally want to go to another plan
immediately, but that would involve a malloc).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Don't call through io_loop, but have it pass the connection back
and call manually.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Don't call from the plan-construction function, call after it returns.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Without this, closing an fd results in a spin...
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Overloading io_close() as a callback is ugly: create an explicit
io_close_cb().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Use a -1 for error codes.  This makes it easier to write your own io funcs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Rather than insisting on supplying them on every call to io_new_conn().
Also, this way it can be changed on a connection.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

And rename debug_io_plan() to io_plan_debug() so it can be exposed.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Debugging an async library is a pain: it's nice to force it into a
linear call chain to try to track problems.

Ugly code, though.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>