This version is from 2017 and should be stale enough to not cause
an outrage. All the relevant distros have it or a newer version.
We also already depend on some symbols defined in 3.0.5 anyway,
so let's take the opportunity to reduce our missing_efi.h
baggage.
The uefi_call_wrapper exists to convert to the right calling convention
and presumably predates compilers that can do so natively. The only
architecture where this is even needed is x86_64.
But because we are building with GNU_EFI_USE_MS_ABI defined, the
EFIAPI macro tells the compiler to use the right calling convention
for EFI functions. Our shim callback (which is called by EFI itself)
already relies on this.
This also adds a safety check to make se we are compiling with
GNU_EFI_USE_MS_ABI defined and also adds it to the compiler args
unconditionally. It is only used with x86_64 anyways, so it should
be fine to do so. EFI_FUNCTION_WRAPPER is unused in gnu-efi, so
it is dropped.
We are already using void in several places and having a screaming
typedef for void feels pointless. There are also CONST, IN, OUT
and OPTIONAL which we aren't using either.
This leaves missing_efi.h to keep it in line with how they are
defined in gnu-efi and/or the specs.
A little helper function and some unusual formatting makes this
whole thing a lot easier on the eyes. Also, right-aligning the
properties for better readability at runtime.
Let's define two helpers strdupa_safe() + strndupa_safe() which do the
same as their non-safe counterparts, except that they abort if called
with allocations larger than ALLOCA_MAX.
This should ensure that all our alloca() based allocations are subject
to this limit.
afaics glibc offers three alloca() based APIs: alloca() itself,
strndupa() + strdupa(). With this we have now replacements for all of
them, that take the limit into account.
Register a LINUX_EFI_INITRD_MEDIA_GUID DevicePath with a LoadFile2Protocol interface and serve the initrd to a supported Linux kernel (Version 5.8+)
Leave the x86 code for older kernels in place until supported kernels become more mainstream
We already did that for sd-boot, hence do it for sd-stub the same way.
Also, move the __attribute__ stuff to the beginning of the statement,
rather than the middle. Mostly just because we usually put it first for
implementations for identifiers (for prototypes we put it last).
These functions are used pretty much independently of locale, i.e. the
only info relevant is whether th locale is UTF-8 or not. Hence let's
give this its own pair of .c/.h files.
We inquire the EFI var for this at two places, let's add a helper that
queries it and gracefully handles it if we can't get it, by returning a
zero mask, i.e. no features supported.
Let's make some stuff const. Most importanly call AllocatePages() with
a pointer to an EFI_PHYSICAL_ADDRESS instead of a pointer to a
pointer. On 64bit this makes no difference, but on i386 this is simply
not correct, since EFI_PHYSICAL_ADDRESS is 64bit there, even though
pointers are 32bit.
let's move showing of the splash screen to the earliest place we know
the splash bmp address. After all a splash screen is all about showing
as early as we can. This matters as doing TPM stuff or packing up a
large cpio might take time.
While we are at it, move the conditionalization of the splash screen
into the function instead of doing it ahead of calling it. This should
encapsulate things more nicely.
Zbigniew Jędrzejewski-Szmek
Dimitri Papadopoulos
Jan Janssen
Max Resch
Lennart Poettering
Yu Watanabe
Luca Boccassi