Date: Fri, 8 Aug 2014 16:50:45 +0200
Subject: [PATCH 1/4] Restrict GNU_EFI_USE_MS_ABI GCC version test to x86_64
The version test only applies to x86_64 builds, so no need to do it
for other archs.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Thu, 31 Jul 2014 13:42:23 -0700
Subject: [PATCH 4/4] Use Shell protocols to retrieve argc/argv, when
available.
New header files efishellintf.h efishellparm.h are coming from EDK
II, initial location and license at top of files. Only modifications:
- efishellintf.h: s/EFI_FILE_PROTOCOL/EFI_FILE/ + expand BITx macros (1<<x)
- efishellparm.h: typedef VOID *SHELL_FILE_HANDLE to avoid including
ShellBase.h
- both: removed extern EFI_GUID variable decls
This also adds apps/t8.c, a simple demo.
Signed-off-by: David Decotigny <decot@googlers.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Mon, 28 Jul 2014 21:28:50 -0700
Subject: [PATCH 3/3] make cmdline parsing a 1st class citizen
Refactor ParseCmdline and apps/Alloc+FreePages to factorize
boilerplate and move the new parser to the main API.
Signed-off-by: David Decotigny <decot@googlers.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Mon, 28 Jul 2014 21:01:35 -0700
Subject: [PATCH 1/3] Fix cmdline parser
The cmdline parser would not return the correct number of args, would
allocate one too many. Also make it clear from the declaration that we
expect a suitably lare argv.
Signed-off-by: David Decotigny <decot@googlers.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Mon, 21 Jul 2014 14:26:23 -0400
Subject: [PATCH] inc/efistdarg.h: Use gcc builtins instead of stdarg.h or broken stubs
We cannot use stdarg.h, as this breaks applications compiling
with -nostdinc because those will not find the header.
We also cannot use the stubs, as they just produce broken code,
as seen in the gummiboot 45-1 Debian release.
Signed-off-by: Julian Klode <julian.klode@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Mon, 2 Jun 2014 23:26:48 +0200
Subject: [PATCH] always observe EFIAPI calling convention when calling
STO.SetAttribute
We have to consider the following cases wrt. the PRINT_STATE.Output and
PRINT_STATE.SetAttr EFIAPI function pointers, especially when building for
x86_64 with gcc:
(1) The compiler is new enough, and EFIAPI actually ensures the Microsoft
calling convention. In this case everything happens to work fine even
if we forget uefi_call_wrapper(), because the wrapper would expand to
a normal C function call anyway.
(2) Otherwise (ie. gcc is old), EFIAPI expands to nothing, and we must
take into account the called function's origin:
(2a) If the callee that is declared EFIAPI is *defined* inside gnu-efi,
then EFIAPI means nothing for the callee too, so caller and callee
only understand each other if the caller intentionally omits
uefi_call_wrapper().
(2b) If the callee that is declared EFIAPI is defined by the platform
UEFI implementation, then the caller *must* use
uefi_call_wrapper().
The PRINT_STATE.Output EFIAPI function pointer is dereferenced correctly:
the PFLUSH() distinguishes cases (2a) from (2b) by using IsLocalPrint().
However use of the PRINT_STATE.SetAttr EFIAPI function pointer is not
always correct:
- The PSETATTR() helper function always relies on the wrapper (case (2b)).
This is correct, because PRINT_STATE.SetAttr always points to a
platform-provided function.
- The DbgPrint() function contains two incorrect calls: they mistakenly
assume case (2a) (or case (1)), even though the pointer always points to
a platform function, implying (2b). (The error is masked in case (1).)
Fix them.
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Equivalent to PoolPrint but using a va_list parameter
Signed-off-by: Sylvain Chouleur <sylvain.chouleur@intel.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Consequently, when your program calls DbgPrint() via the DEBUG() macro,
it fails to set up the stack correctly (it does not pass the arguments
through the ellipsis (...) according to the EFIAPI calling convention).
However, va_start() inside DbgPrint() *assumes* that stack.
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
and uses a pointer to argv[] like it's argv[]. The compiler only issues
warnings though. Here is a patch to remove compiler warnings and make the
code behave.
Signed-off-by: Bernard Burette <bub75@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
"LibLocateHandleByDiskSignature()" because the "Start" variable is
give a value which is not used.
Signed-off-by: Bernard Burette <bub75@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
The path is in the two files inc/efidevp.h and lib/dpath.c.
It also add the Sata device path and removes the "/?" path for unknown device paths.
Signed-off-by: Bernard Burette <bub75@users.sf.net>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Fri, 21 Feb 2014 06:03:00 +0100
Subject: [PATCH] Makefile: Don't clean what doesn't already exists
Trying to recurse into subdirectories of object files may lead to an
error if the directory doesn't exist. Even when cleaning.
Signed-off-by: Sylvain Gault <sylvain.gault@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Date: Wed, 19 Feb 2014 05:18:14 +0100
Subject: [PATCH] make install: Don't overwrite up-to-date files
make install used to copy files unconditionnally to their destination.
However, if the destination is used by another Makefile, it will always
see modified files. "install" target now only update the files when they
need to.
Signed-off-by: Sylvain Gault <sylvain.gault@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Status must be initialized before calling GrowBuffer() as it may
otherwise be uninitialized or set to EFI_BUFFER_TOO_SMALL by other
functions.
Signed-off-by: Gene Cumm <gene.cumm@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
directory) and TOPDIR (top of source tree) to separate the
build directory from the source tree.
Signed-off-by: Gene Cumm <gene.cumm@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
It contains verbatim copy of the comment at beginning of file from
elilo.
(there was no COPYING file in the elilo source that the comment refers
to.)
Signed-off-by: Jerry Hoemann <jerry.hoemann@hp.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
the kernel arch. The information actually needed is the default target
of gcc.
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Signed-off-by: Sylvain Gault <sylvain.gault@gmail.com>
and SetNVVariable to store non volatile variable.
Signed-off-by: Sylvain Chouleur <sylvain.chouleur@gmail.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
e.g. tries to copy StrCmp, gcc will give you warnings about those
calls, and the warnings are right. These clutter up other things you
might miss that you should be more concerned about.
You could work around it through vigorous typecasting to non-const
types, but why should you have to? All of these functions are
rigorously defined as not changing their input - it is const, and should
be marked as such.
Signed-off-by: Peter Jones <pjones@redhat.com>
Please enter the commit message for your changes. Lines starting
Used to test/find memory fragmentation issues linux.
Signed-off-by: Jerry Hoemann <jerry.hoemann@hp.com>
Signed-off-by: Nigel Croxon <nigel.croxon@hp.com>
Don't allow duplicate symbols, abort upon first error.
also make sure the linker script comes last for apps.
Signed-off-by: David Decotigny <decot@googlers.com>
make -C apps would fail on tcc.c because uefi_call_wrapper() doesn't
deal correctly with efi_call0-type invocation. Fixed by this patch.
Signed-off-by: David Decotigny <decot@googlers.com>
GCC 4.8.0 adds some optimizations that will use movups/movaps (and use
%xmm* registers) when they're faster, and of course that won't work at
all since UEFI firmwares aren't guaranteed to initialize the mmx/sse
instructions.
This will be even more annoying, since most UEFI firmwares don't
initialize the #DE or #UD trap handlers, and your backtrace will be a
random path through uninitialized memory, occasionally including
whatever address the IDT has for #UD, but also addresses like "0x4" and
"0x507" that you don't normally expect to see in your call path.
Signed-off-by: Peter Jones <pjones@redhat.com>
Some Makefiles tickle a bug in make 3.82 that cause libefi.a
and libgnuefi.a dependencies to expand to the odd values:
libefi.a: boxdraw.o) smbios.o) ...
libgnuefi.a(reloc_x86_64.o:
The patch replaces libgnuefi.a($(OBJS)) & libefi.a($(OBJS))
with an equivalent expansion that should work with any make
that supports $(patsubst).
Reorder variables in Make.defaults so that they are grouped by
functions. Also fixed ifeq (x,y) to have required syntax and make it
work for ARCH amd64->x86_64 renaming on BSD. Also provides top-level
Makefile with a "mkvars" target that displays effective variables.
Signed-off-by: David Decotigny <decot@googlers.com>
Instead of asking developers to explicitly pass the number of
parameters to the functions that get called, we determine them
automatically at preprocessing time. This should result in more
robust code.
Argument va_num is now ignored in x86_64 code, both with and
without HAVE_USE_MS_ABI.
Credits to the macro magic given in the comments.
Signed-off-by: David Decotigny <decot@googlers.com>
On x86_64 without HAVE_USE_MS_ABI support, uefi_call_wrapper() is a
variadic function. Parameters >=5 are copied to the stack and, when
passed small immediate values (and possibly other parameters), gcc
would emit a movl instruction before calling uefi_call_wrapper(). As a
result, only the lower 32b of these stack values are significant, the
upper 32b potentially contain garbage. Considering that
uefi_call_wrapper() assumes these arguments are clean 64b values
before calling the efi_callX() trampolines, the latter may be passed
garbage. This makes calling functions like
EFI_PCI_IO_PROTOCOL.Mem.Read()/Write() or BS->OpenProtocol() quite
unreliable.
This patch fixes this by turning uefi_call_wrapper() into a macro that
allows to expose the efi_callX() trampoline signatures to the callers,
so that gcc can know upfront that it has to pass all arguments to
efi_callX() as clean 64b values (eg. movq for immediates). The
_cast64_efi_callX macros are just here to avoid a gcc warning, they do
nothing otherwise.
Signed-off-by: David Decotigny <decot@googlers.com>
macros only exist in link.h on Linux. On FreeBSD, the equivalent macro is
__ElfN(). But the macro usage is redundant. You're only going to compile the
ia32 file for IA32 binaries and the x86_64 file for X64 binaries. If you had
just one file built for both cases, then using the macro might make more
sense.
- Removes the "#define foo_t efi_foo_t" macros from reloc_ia32.c and
reloc_x86_64.c.
- Modifies inc/x86_64/efibind.h and inc/ia32/efibind.h to use the new
definitions for uint64_t, int64_t and int8_t. The 64-bit types are now defined
as:
typedef int __attribute__((__mode__(__DI__))) int64_t;
typedef unsigned int __attribute__((__mode__(__DI__))) uint64_t;
This removes the conflict between the host types dragged in by elf.h and the
type definitions in efibind.h that made the #define foo_t efi_foo_t" hack
necessary. Also, int8_t is now defined as signed char instead of just char
(assuming char == signed char is apparently not good enough).
- Also modifies these files to use stdint.h instead of stdint-gcc.h. It's
unclear if this is completely correct, but stdint-gcc.h is not present with
all GCC installs, and if you use -std=c99 or later you will force this case to
be hit. This also can break clang, which doesn't have a stdint-gcc.h at all.
- Removes the #include of <link.h> from reloc_ia32.c and reloc_x86_64.c (since
with the previous changes it's not needed anymore).
- Places the #include of <elf.h> after #include <efi>/#include <efilib.h> so
that we know the types will always be defined properly, in case you build on a
system where <elf.h> doesn't automatically pull in the right header files to
define all the needed types. (This actually happens on VxWorks. It's harmless
elsewhere. If you don't care about VxWorks, you can leave this out.)
- Modifies setjmp_ia32.S and setjmp_x86_64.S so to change "function" to
@function. The clang compiler doesn't like the former. Clang and GCC both like
the latter.
- Modifles Make.defaults so that if ARCH is detected as "amd64," it's changed
to "x86_64." It happens that uname -m on 64-bit FreeBSD reports the former
rather than the latter, which breaks the build. This may also be the case on
some other OSes. There's a way to force uname(1) to return x86_64 as the
machine type, but this way is a little friendlier.
- Creates gnuefi/elf_ia32_fbsd_efi.lds which specifies the object file type as
elf-ia32-freebsd. This is required for building on FreeBSD/i386, not just
FreeBSD/amd64.
- Modifies apps/Makefile to always use
$(TOPDIR)/gnuefi/elf_$(ARCH)_fbsd_efi.lds when building on either 32-bit or
64-bit FreeBSD instead of just for the x86_64 case.
- Changed LDFLAGS in Make.defaults to include --no-undefined. This will cause
linking to fail if there are any unsatisfied symbols when creating foo.so
during any of the app builds, as opposed to just silently succeeding and
producing an unusable binary.
- Changed CFLAGS to include -ffreestanding -fno-stack-protector -fno-stack-
check. This prevents clang from inserting a call to memset() when compiling
the RtZeroMem() and RtSetMem() routines in lib/runtime/efirtlib.c and guards
against the native compiler in some Linux distros from adding in stack
checking code which relies on libc help that isn't present in the EFI runtime
environment.
This does the following:
- Cleans up the ia32 and x86-64 relocation code a bit (tries to break the
dependency between the host ELF headers and the EFI runtime environment)
- Avoids the dependency on stdint-gcc.h which may not always be available
- Allows GNU EFI to build out of the box on both FreeBSD/i386 and
FreeBSD/amd64
- Allows GNU EFI to build out of the box with either GCC or clang on
FreeBSD/i386 and FreeBSD/amd64 9.0 and later.
- Makes things a little easier to port to VxWorks
- Avoids creating un-runable binaries with unresolved symbol definitions
(which can be very confusing to debug)
