xref: /src/sys/external/bsd/gnu-efi/dist/ChangeLog

2013-06-12 Nigel Croxon <nigel.croxon@hp.com>
    Disable MMX and SSE
    
    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>

    Author: Nigel Croxon <nigel.croxon@hp.com>
    Date:   Wed Jun 12 10:29:40 2013 -0400

    bug in make 3.82 expand to odd values
    
    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).

    Author: Nigel Croxon <nigel.croxon@hp.com>
    Date:   Wed Jun 12 09:53:01 2013 -0400

    support .text.* sections on x86_64
    
    Group them in .text. Also add vague linkage sections in .text.
    
    Signed-off-by: David Decotigny <decot@googlers.com>

    Author: Nigel Croxon <nigel.croxon@hp.com>
    Date:   Wed Jun 12 09:51:36 2013 -0400

    cleanup and fix Make.defaults
    
    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>

    Author: Nigel Croxon <nigel.croxon@hp.com>
    Date:   Wed Jun 12 09:47:16 2013 -0400

    automatically determine number of uefi_call_wrapper() args on x86_64
    
    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>

    Author: Nigel Croxon <nigel.croxon@hp.com>
    Date:   Wed Jun 12 09:38:10 2013 -0400

    fix parameter-passing corruption on x86_64 for >= 5 args
    
    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>

    Author: noxorc <nigel.croxon@hp.com>
    Date:   Wed May 15 15:26:16 2013 -0400

    - Removes the ElfW() macro usage from reloc_ia32.c and reloc_x86_64.c. These
    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)

    Author: noxorc <nigel.croxon@hp.com>
    Date:   Wed May 8 16:29:45 2013 -0400

    Add the definitions for TCP, UDP and IP, for both IPv4 and IPv6.


2013-05-02 Nigel Croxon <nigel.croxon@hp.com>
	* Chnage from Matt Fleming <matt.fleming@intel.com>
	  - Preparation for adding the networking protocol definitions.
	    Add the service binding protocol.

2013-02-21 Nigel Croxon <nigel.croxon@hp.com>
	* Change from Peter Jones <pjones@redhat.com>
	  - Previously we were incorrectly passing 3 functions with
            the System V ABI to UEFI functions as EFI ABI functions.
            Mark them as EFIAPI so the compiler will (in our new
            GNU_EFI_USE_MS_ABI world) use the correct ABI.
          - These need to be EFIAPI functions because in some cases
            they call ST->ConOut->OutputString(), which is an EFIAPI
            function. (Which means that previously in cases that 
            needed "cdecl", these didn't work right.)
          - If the compiler version is new enough, and GNU_EFI_USE_MS_ABI
            is defined, use the function attribute ms_abi on everything
            defined with "EFIAPI".  Such calls will no longer go through
            efi_call*, and as such will be properly type-checked.
          - Honor PREFIX and LIBDIR correctly when passed in during the build.
          - Add machine type defines for i386, arm/thumb, ia64, ebc, x86_64.
          - __STDC_VERSION__ never actually gets defined unless there's a
            --std=... line.  So we were accidentally defining lots of c99
            types ourself. Since it's 2012, use --std=c11 where appropriate,
            and if it's defined and we're using gcc, actually include gcc's
            stdint definitions.
          - New test application added: route80h. This is a test program
            for PciIo.  It routes ioport 80h on ICH10 to PCI. This is also
            useful on a very limited set of hardware to enable use of
            a port 80h debug card.
          - New test applcation added: modelist. This lists video modes
            the GOP driver is showing us.
	* Change from Finnbarr Murphy
          - https://sourceforge.net/p/gnu-efi/feature-requests/2/
            Please add the following status codes to <efierr.h>
            EFI_INCOMPATIBLE_VERSION 25
            EFI_SECURITY_VIOLATION 26 
            EFI_CRC_ERROR 27 
            EFI_END_OF_MEDIA 28 
            EFI_END_OF_FILE 31 
            EFI_INVALID_LANGUAGE 32 
            EFI_COMPROMISED_DATA 33
	* Change from SourceForge.net Bug report
          - https://sourceforge.net/p/gnu-efi/bugs/5/
            BufferSize is a UINT64 *. The file shipped with GNU EFI is from
            1998 whereas the latest one is from 2004. I suspect Intel changed
            the API in order handle 64-bit systems.
        * Change from Felipe Contreras <felipe.contreras@gmail.com>
          - The current code seems to screw the stack at certain points.
            Multiple people have complained that gummiboot hangs right away,
            which is in part the fault of gummiboot, but happens only
            because the stack gets screwed. x86_64 EFI already aligns the
            stack, so there's no need for so much code to find a proper
            alignment, we always need to shift by 8 anyway.
        * Change from A. Steinmetz
          - https://sourceforge.net/p/gnu-efi/patches/1/
            The patch prepares for elilo to support uefi pxe over ipv6
            See uefi spec 2.3.1 errata c page 963 as reference.
            Verfied on an ASUS Sabertooth X79 BIOS Rev. 2104 system which
            is able to do an IPv6 UEFI PXE boot.
	* Release 3.0t

2012-09-21 Nigel Croxon <nigel.croxon@hp.com>
	* Change from Peter Jones <pjones@redhat.com>
	  - EFI Block I/O protocol versions 2 and 3 provide more information
	    regarding physical disk layout, including alingment offset at the
	    beginning of the disk ("LowestAlignedLba"), logical block size
	    ("LogicalBlocksPerPhysicalBlock"), and optimal block transfer size
	    ("OptimalTransferLengthGranularity").
	* Release 3.0r

2012-04-30 Nigel Croxon <nigel.croxon@hp.com>
	* Change from Matt Fleming <matt.fleming@intel.com>
	  -  The .reloc section is now 4096-byte boundary for x86_64.
	     Without this patch the .reloc section will not adhere to
	     the alignment value in the FileAlignment field (512 bytes by
	     default) of the PE/COFF header. This results in a signed
	     executable failing to boot in a secure boot environment.
	* Release 3.0q

2011-12-12 Nigel Croxon <nigel.croxon@hp.com>
        * Changes from Fenghua Yu <fenghua.yu@intel.com>
          - This fixes redefined types compilation failure for tcc.c on x86_64 machines.
        * Release 3.0p

2011-11-15 Nigel Croxon <nigel.croxon@hp.com>
        * Changes from Darren Hart <dvhart@linux.intel.com>
          - Conditionally assign toolchain binaries to allow overriding them.
          - Force a dependency on lib for gnuefi.
        * Release 3.0n

2011-08-23 Nigel Croxon <nigel.croxon@hp.com>
        * Changes from Peter Jones <pjones@redhat.com>
          - Add guarantee 16-byte stack alignment on x86_64.
          - Add routine to make callbacks work.
          - Add apps/tcc.efi to test calling convention.
        * Release 3.0m

2011-07-22 Nigel Croxon <nigel.croxon@hp.com>
	* Changed Makefiles from GPL to BSD.
	* Changes from Peter Jones <pjones@redhat.com>
	  - Add ifdefs for ia64 to mirror ia32 and x86-64 so that
	    one can build with GCC.
	  - Add headers for PciIo.
	  - Add the UEFI 2.x bits for EFI_BOOT_SERVICES
	  - Add an ignore for .note.GNU-stack section in X86-64 linker maps.
        * Release 3.0l

2011-04-07  Nigel Croxon <nigel.croxon@hp.com>
        * Change license from GPL to BSD.
        * Release 3.0j

2009-09-12  Julien BLACHE <jb@jblache.org>
	* Add support for FreeBSD.
	* Release 3.0i

2009-09-11  Julien BLACHE <jb@jblache.org>
	* Fix elf_ia32_efi.lds linker script to be compatible with the new
          linker behaviour. Patch from the RedHat bugzilla 492183.

2009-06-18  Nigel Croxon <nigel.croxon@hp.com>
        * Release 3.0h

2008-11-06  Nigel Croxon <nigel.croxon@hp.com>
        * Fix to not having any relocations at all. 

2008-09-18  Nigel Croxon <nigel.croxon@hp.com>
        * Use LIBDIR in makefiles
        * Add setjmp/longjmp
        * Fixes incorrect section attribute in crt0-efi-ia32.S
        * Adds value EfiResetShutdown to enum EFI_RESET_TYPE
        * Fixes a RAW warning in reloc_ia64.S
        * Adds the USB HCI device path structure in the headers
          patches were supplied by Peter Jones @ RedHat

2008-02-22  Nigel Croxon <nigel.croxon@hp.com>
	* Added '-mno-red-zone' to x68_64 compiles.
	  Patch provided by Mats Andersson.

2008-01-23  Nigel Croxon <nigel.croxon@hp.com>
        * release 3.0e to support x86_64
          EFI calling convention, the stack should be aligned in 16 bytes
          to make it possible to use SSE2 in EFI boot services.
          This patch fixes this issue. Patch provided by Huang Ying from Intel.

2007-05-11 Nigel Croxon <nigel.croxon@hp.com>
        * release 3.0d to support x86_64 from Chandramouli Narayanan
          from Intel and based on 3.0c-1

2006-03-21  Stephane Eranian <eranian@hpl.hp.com>
	* merged patch to support gcc-4.1 submitted by
	  Raymund Will from Novell/SuSE

2006-03-20  Stephane Eranian <eranian@hpl.hp.com>
	* updated ia-64 and ia-32 linker scripts to
	  match latest gcc. The new gcc may put functions in
	  .text* sections. patch submitted by H.J. Lu from Intel.

2004-11-19  Stephane Eranian <eranian@hpl.hp.com>
	* added patch to ignore .eh_frame section for IA-32. Patch
	  submitted by Jim Wilson

2004-09-23  Stephane Eranian <eranian@hpl.hp.com>
	* added patch to discard unwind sections, newer toolchains
	  complained about them. Patch submitted by Jesse Barnes from SGI.

2003-09-29  Stephane Eranian <eranian@hpl.hp.com>
	* updated elf_ia64_efi.lds to reflect new data sections 
	  created by gcc-3.3. Patch provided by Andreas Schwab from Suse.

2003-06-20  Stephane Eranian <eranian@hpl.hp.com>
	* updated elf_ia64_efi.lds and elf_ia32_efi.lds to include
	  new types data sections produced by recent version of gcc-3.x

2002-02-22  Stephane Eranian <eranian@hpl.hp.com>
	* release 3.0a
	* modified both IA-64 and IA-32 loader scripts to add support for the
	  new .rodata sections names (such as rodata.str2.8). Required
	  for new versions of gcc3.x.

2001-06-20  Stephane Eranian <eranian@hpl.hp.com>
	* release 3.0
	* split gnu-efi package in two different packages: the libary+include+crt and the bootloader.
	* removed W2U() hack and related files to get from wide-char to unicode. 
	* Use -fshort-wchar option for unicode.
	* restructured Makefiles now install under INSTALLROOT.

2001-04-06  Stephane Eranian <eranian@hpl.hp.com>

	* incorporated patches from David and Michael Johnston at Intel
	  to get the package to compile for IA-32 linux target.

	* Fixed ELILO to compile for Ia-32 (does not execute yet, though):
	  Makefile and start_kernel() function.

2001-04-06  Andreas Schwab <schwab@suse.de>

	* Fixed config.c  to
	  get the timeout directive to do something. implemented the global
	  root= directive.

	* Fix the efi_main() to deal with the -C option properly

2001-04-05  Stephane Eranian <eranian@hpl.hp.com>

	* update efi library to latest EFI toolkit 1.02 as distributed
	  by Intel. Fixed header + library files to compile with GCC

	* merged ELI and LILO (as of gnu-efi-1.1) together, mostly
	  taking the config file feature of ELI.

	* renamed LILO to ELILO to make the distinction

	* restructured code to make it easier to understand and maintain

	* fixed FPSWA driver checking and loading: we try all possible
	  files and let the driver itself figure out if it is the most
	  recent.
	* added support for compression (gzip) but keep support for plain
	  ELF image. ELILO autodetects the format

	* change the way the kernel is invoked. Now we call it in 
	  physical memory mode. This breaks the dependency between the
	  kernel code and the loader. No more lilo_start.c madness.

	* changed the way the boot_params are passed. We don't use the 
	  ZERO_PAGE_ADDR trick anymore. Instead we use EFI runtime memory.
	  The address of the structure is passed to the kernel in r28
	  by our convention.

	* released as gnu-efi-2.0

2001-04-03  David Mosberger  <davidm@hpl.hp.com>

	* gnuefi/reloc_ia32.c (_relocate): Change return type from "void"
	to "int".  Return error status if relocation fails for some
	reason.

	* gnuefi/elf_ia32_efi.lds: Drop unneeded ".rel.reloc" section.

	* gnuefi/crt0-efi-ia32.S (_start): Exit if _relocate() returns with
	non-zero exit status.

	* inc/ia32/efibind.h [__GNUC__]: Force 8-byte alignment for 64-bit
	types as that is what EFI appears to be expecting, despite the
	"#pragma pack()" at the beginning of the file!

2001-03-29  David Mosberger  <davidm@hpl.hp.com>

	* gnuefi/reloc_ia32.c: Add a couple of defines to work around
	libc/efilib collision on uint64_t et al.
	(_relocate): Use ELF32_R_TYPE() instead of ELFW(R_TYPE)().

	* gnuefi/crt0-efi-ia32.S (dummy): Add a dummy relocation entry.

2001-03-29  David Mosberger  <davidm@hpl.hp.com>

        * gnuefi/reloc_ia32.c: Add a couple of defines to work around
        libc/efilib collision on uint64_t et al.
        (_relocate): Use ELF32_R_TYPE() instead of ELFW(R_TYPE)().

        * gnuefi/crt0-efi-ia32.S (dummy): Add a dummy relocation entry.

2000-10-26  David Mosberger  <davidm@hpl.hp.com>
 
 	* gnuefi/elf_ia64_efi.lds: Mention .rela.sdata.
 
 	* Make.defaults (CFLAGS): Remove -nostdinc flags so we can pick
 	up the C compiler's stdarg.h.
 
 	* inc/stdarg.h: Remove this file.  It's not correct for gcc (nor
 	most other optimizing compilers).

2000-10-10  Stephane Eranian <eranian@hpl.hp.com>

	* cleaned up the error message and printing of those.
	* added support to load the FPSWA from a file in case support is not
	  present in the firmware already
	* fixed split_args() to do the right thing when you have leading spaces
	  before kernel name
	* changed the argify() function to rely on \0 instead of LoadOptionSize
	  as the field seems to be broken with current firmware
	* bumped version to 1.0

2000-10-04  David Mosberger  <davidm@hpl.hp.com>
 
	* gnuefi/reloc_ia64.S: Reserve space for up to 750 function descriptors.

	* gnuefi/elf_ia64_efi.lds: Add .sdata section for small data and
	put __gp in the "middle" of it.

	* gnuefi/crt0-efi-ia64.S (_start): Use movl/add to load
	gp-relative addresses that could be out of the range of the addl
	offset.
	* gnuefi/reloc_ia64.S (_relocate): Ditto.

	* apps/Makefile: Remove standard rules and include Make.rules instead.
	* lilo/Makefile: Ditto.

	* Make.rules: New file.

2000-08-04  Stephane Eranian <eranian@hpl.hp.com>
	* released version 0.9
	* incorporated ACPI changes for Asuza by NEC < kouchi@hpc.bs1.fc.nec.co.jp>
	* added support for initrd (-i option) original ELI code from Bill Nottingham <notting@redhat.com>)
	* lots of cleanups 
	* got rid of #ifdef LILO_DEBUG and uses macro instead
	* fix a few extra memory leaks in create_boot_params()
	* added exit capability just before starting the kernel

2000-06-22  David Mosberger  <davidm@hpl.hp.com>

	* gnuefi/elf_ia64_efi.lds: Add .srodata, .ctors, .IA64.unwind,
	.IA64.unwind_info to .data section and .rela.ctors to .rela
	section.

2000-04-03  David Mosberger  <davidm@hpl.hp.com>

	* lilo/lilo.c (LILO_VERSION): Up version number to 0.9.

	* gnuefi/elf_ia64_efi.lds: Include .IA_64.unwind and
	.IA_64.unwind_info in .data segment to avoid EFI load error
	"ImageAddress: pointer outside of image" error due to the .dynsym
	relocations against these sections.

	* ChangeLog: Moved from lilo/ChangeLogs.

	* gnuefi/reloc_ia64.S: fixed typo: .space directive had constant
	100 hardcoded instead of using MAX_FUNCTION_DESCRIPTORS
	macro. Duh.

Fri Mar 17 15:19:18 PST 2000 Stephane Eranian <eranian@hpl.hp.com>

	* Released 0.8
	* replace the  getopt.c with new version free with better license
	* created a documentation file
	* fix a couple of memory leaks
	* code cleanups
	* created a separate directory for lilo in the gnu-efi package.
	* added support for the BOOT_IMAGE argument to kernel
	* default is to build natively now