Event queue doorbells must use UAR pages 0-127 depending on event queue
number; other doorbells must use pages 128+ (and we choose to use page
128).

Infiniband devices no longer block waiting for link-up in
register_ibdev().

Hermon driver needs to create an event queue and poll for link-up events.

Infiniband core needs to reread MAD parameters when link state changes.

IPoIB needs to cope with Infiniband link parameters being only partially
available at probe and open time.

PXE is a catch-all image format with no signature checks.  If an
unsupported image file is loaded, it will be treated as a PXE image.  In
most cases, the image will be too large to be loaded as a PXE image (which
has to fit in base memory), so the error returned to the user will be that
the segment could not fit within the memory region.

Add an explicit check to pxe_image.c to reject images larger than base
memory with ENOEXEC.

Add ENOEXEC to the error string table.

03-19-2008, Hilko Bengen, Cleanups and fixes for newer cards
(successfully tested with 8110SC-d onboard NIC)

gPXE is not compliant with the HTTP/1.1 specification (RFC 2616),
since it lacks support for "Transfer-Encoding: chunked".  gPXE is,
however, compliant with the HTTP/1.0 specification (RFC 1945), which
does not require "Transfer-Encoding: chunked" to be supported.

The only HTTP/1.1 feature that gPXE uses is the "Host:" header, but
servers universally accept that one from HTTP/1.0 clients as an
optional extension (it is obligatory for HTTP/1.1).  gPXE does not,
for example, appear to support connection caching.  Advertising as a
HTTP/1.0 client will typically make the server close the connection
immediately upon sending the last data, which is actually beneficial
if we aren't going to keep the connection alive anyway.

In particular, expose the system UUID as a setting ("smbios/uuid").

The PXE spec is (as usual) unclear on precisely when ProxyDHCPREQUESTs
should be issued.  We adapt the following, slightly paranoid approach:

  If an offer contains an IP address, then it is a normal DHCPOFFER.

  If an offer contains an option #60 "PXEClient", then it is a
  ProxyDHCPOFFER.  Note that the same packet can be both a normal
  DHCPOFFER and a ProxyDHCPOFFER.

  After receiving the normal DHCPACK, if we have received a
  ProxyDHCPOFFER, we unicast a ProxyDHCPREQUEST back to the ProxyDHCP
  server on port 4011.  If we time out waiting for a ProxyDHCPACK, we
  treat this as a non-fatal error.

This avoids confusing other code by insisting that we have e.g. a
filename consisting of 128 zero bytes.

Add documentation for the FILE_EXEC and FILE_CHECK_API extension
calls, and update the documentation for the FILE_READ call.

Change the PXE return code for EWOULDBLOCK from PXENV_STATUS_FAILURE
to PXENV_STATUS_TFTP_OPEN.  This code is only used by the FILE_READ
PXEXT call, and is necessary to distinguish "error" from "no data" in
that call.

(The only other nonblocking call is UDP_READ, where the caller doesn't
care about the distinction, however, gPXE doesn't use EWOULDBLOCK
internally to represent this condition in that code.)

Add FILE_API_CHECK to the PXEXT API so the NBP can query the
availability and status of the API.

Copy DHCP responses to a standard malloc()ed buffer, rather than
retaining the I/O buffer that they arrived in.

read_user() assumes that zero-length buffers don't exist, and optimises
around this.

Allow for settings to be described by something other than a DHCP option
tag if desirable.  Currently used only for the MAC address setting.

Separate out fake DHCP packet creation code from dhcp.c to fakedhcp.c.

Remove notion of settings from dhcppkt.c.

Rationalise dhcp.c to use settings API only for final registration of the
DHCP options, rather than using {store,fetch}_setting throughout.

DHCP code was using an incorrect check for whether to construct a
DHCPDISCOVER or DHCPREQUEST packet.

It's probably easier to have a named setting for next-server, rather
than to explain to people why they need to pretend it is option 175.3.

Using -MM strips out any system header files from the dependency list,
which means that we could safely use an external stdarg.h if we wanted
to.

The home-grown versions don't work properly for static variadic
functions, when gcc can choose to use a non-standard calling convention.

Add dedicated functions create_dhcpdiscover(), create_dhcpack() and
create_proxydhcpack() for use by external code such as the PXE preboot
code.

Register ProxyDHCP options under the global scope "proxydhcp".

Unregister previously-acquired DHCP and ProxyDHCP settings when DHCP
succeeds.

IPv4 may not always be dragged into the link, so add a specific
configuration option for it.

Convert DHCP option applicators in dns.c and iscsi.c to settings
applicators.

Kill off DHCP option applicators.

Add a configuration settings block for each net device.  This will
provide the parent scope for settings applicable only to that network
device (e.g. non-volatile options stored on the NIC, options obtained via
DHCP, etc.).

Expose the MAC address as a setting.

Add the notion of the settings hierarchy, complete with
register/unregister routines.

Rename set->store and get->fetch to avoid naming conflicts with get/put
as used in reference counting.

Add the concept of an abstract configuration setting, comprising a (DHCP)
tag value and an associated byte sequence.

Add the concept of a settings namespace.

Add functions for extracting string, IPv4 address, and signed and
unsigned integer values from configuration settings (analogous to
dhcp_snprintf(), dhcp_ipv4_option(), etc.).

Update functions for parsing and formatting named/typed options to work
with new settings API.

Update NVO commands and config UI to use new settings API.

Russ Nelson sent an email requesting the license change to
GPL Version 2.

Allow setting names such as "ip:hex".

Timers are sometimes required before the call to initialise(), so we
cannot rely on initialise() to set up the timers before use.

Also fix a potential integer overflow issue in generic_currticks_udelay()