Android: inital porting of efibootmgr

The libefivar porting is also required.

Signed-off-by: Chih-Wei Huang <cwhuang@linux.org.tw>

remove extra decl

Signed-off-by: Peter Jones <pjones@redhat.com>

Bump version to 17

Signed-off-by: Peter Jones <pjones@redhat.com>

Add a bit more logging

Signed-off-by: Peter Jones <pjones@redhat.com>

Propogate verbosity to libefivar 36's internal logging facility

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix an invalid free()

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr -e: improve parsing for efivar-36 compat

Signed-off-by: Peter Jones <pjones@redhat.com>

Make it easier to build with a devel branch of efivar

Signed-off-by: Peter Jones <pjones@redhat.com>

Don't set LIBEFIBOOT_REPORT_GPT_ERRORS=1

Signed-off-by: Peter Jones <pjones@redhat.com>

Change the default partition choice.

This requires efivar >= 36 to actually make sense, unfortunately.

Signed-off-by: Peter Jones <pjones@redhat.com>

make_linux_load_option(): add some more efi_error() calls

Signed-off-by: Peter Jones <pjones@redhat.com>

Make.defaults: fix pkg-config invocation for LDFLAGS

Signed-off-by: Peter Jones <pjones@redhat.com>

Work around coverity being a pretty lazy and poor tool.

I mean, come on, at least dynamically import the types you've missed
from the version of GCC that's installed on the system and you're
pretending to emulate.

Signed-off-by: Peter Jones <pjones@redhat.com>

Update travis to use some better build scripts

Signed-off-by: Peter Jones <pjones@redhat.com>

travis: Move to common efi-ci-rawhide docker image

Signed-off-by: Peter Jones <pjones@redhat.com>

Bump version to 16

Signed-off-by: Peter Jones <pjones@redhat.com>

Add bumpver in 'make archive'

Signed-off-by: Peter Jones <pjones@redhat.com>

travis: build efivar first...

Signed-off-by: Peter Jones <pjones@redhat.com>

travis: Use the right docker image.

Signed-off-by: Peter Jones <pjones@redhat.com>

Update travis/docker stuff.

Signed-off-by: Peter Jones <pjones@redhat.com>

Work around some minor clang analyzer complaints.

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix uninitialized values for mirror status.

In the case where there's a request for mirroring but it isn't enacted
yet, we'll have above4g and below4g uninitialized, and the comparison to
rabove4g and rbelow4g will be meaningless.  So make them 0.

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix minor scan-build complaints

clang analyzer doesn't like when we set variables and don't use the
value, because it doesn't understand software maintenance as an ongoing
process.

Signed-off-by: Peter Jones <pjones@redhat.com>

Better scan makefiles

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: fix error check on add_to_boot_order()

Previously the error was checked whether or not we actually /call/
add_to_boot_order(), so sometimes we'll get a failure even though
nothing was even done.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add coverity make rules

Signed-off-by: Peter Jones <pjones@redhat.com>

Update docs for old URLs and general formatting.

Fix GitHub URL

make_linux_load_option(): check data_size correctly.

Coverity noticed that we're not checking data_size for NULLness before
dereferencing it to pass the value to efi_loadopt_create().  So now we
pass a temporary variable and assign it only if it's non-NULL.

Signed-off-by: Peter Jones <pjones@redhat.com>

unparse_path: remove unneeded include

This fixes an invalid redefinition of wchar_t when compiling with hardened gcc-musl on gentoo.

Bump version to 15

Make conditional default loader path use EFIDIR like fwupdate does.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make default '--loader' build-time configurable.

Each distribution uses a specific 'vendor'-directory in the
'efi' directory on the EFI System Partition to store their
EFI loader.  There's little use in hardcoding the value
for just one in 'efibootmgr', which is displayed in '--help'
and used as default without '--loader'.

Simply use
  make OS_VENDOR=<vendor> EFI_LOADER=<loader>.efi
to get the desired value.  :)

Signed-off-by: Raymund Will <rw@suse.com>

efibootmgr: sanitize `set/get_mirror()`

get_mirror()
- don't short-circuit the assignment of values in case of version mismatch
  as `show_mirror()` happily tries to use them nevertheless
- move redundant warning (from `show_mirror()`s PoV) to `set_mirror()`
- add missing `free(data)`

set_mirror()
- skip obsolete second assignment of `data`
- avoid potentially "uninitialized" access to `above/below4g` by protecting
  it with `opts.set_mirror_hi/_lo` respectively, thus simplifying the code
  down the line
Note: this will prevent unnecessary write-operations!

Signed-off-by: Raymund Will <rw@suse.com>

Add support for parsing optional data as ucs2

Remove extra const keywords gcc 7 gripes about.

Signed-off-by: Peter Jones <pjones@redhat.com>

Don't use -fshort-wchar when building (#63)

It's not needed and is causing build failures with gcc 6. Closes
Debian bug #849651

Signed-off-by: Steve McIntyre <steve@einval.com>

Bump version to 14

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootdump: Get rid of extra debug print

Signed-off-by: Peter Jones <pjones@redhat.com>

Move -flto from cflags to CFLAGS

LTO is an optimization that isn't always desired. This change allows it
to be disabled easily.

Don't bother with --default-symver since we've got no library code here.

ld.gold breaks with it, for basically no reason, but we're not really
using it anyway.

Resolves github issue #56.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make compiler detection slightly better.

Resolves github issue #57.

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix EFIBOOTMGR_VERSION setting and -V output

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: fixes in man page synopsis and README

- more --delete-bootnum description fixes
- make -r | -y bold
- delete -H and -U (removed in 83f34d39299)

Install efibootdump with make

Create README.md

Fix a typo in popt options.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: remove "(hex)" from description of --delete-bootnum

Boot entry selected with --bootnum

efibootmgr 13

- Add support for --sysprep and --driver to support UEFI System Prep
  Applications and UEFI Drivers.
- use efivar's error reporting facility, and show error traces when
  "-v -v" is used.
- Still yet better error codes returned on failures.
- Add -m and -M to support Memory Address Range Mirroring.
- Add efibootdump, to examine Boot* variables found in tarballs in bug
  reports and similar.
- miscellaneous bugfixes.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootdump: give it the error.h treatment

Signed-off-by: Peter Jones <pjones@redhat.com>

Add support for --sysprep and --driver

Maybe this fixes issue rhinstaller/efibootmgr#40 .

Signed-off-by: Peter Jones <pjones@redhat.com>

.spec: fix requirements for newer efivar and man page path.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: use efivar's error facility.

Stop preceeding our version number with a "0.".

It's useless and it just makes people whine about it not being >1.0.

Signed-off-by: Peter Jones <pjones@redhat.com>

README: update github link

README: fix netboot command line since we have acpi bits now.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: fix usage of efi_loadopt_path() /again/.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: fix up a lot of error result codes.

Signed-off-by: Peter Jones <pjones@redhat.com>

Handle "-@ -" correctly again.

This should fix issue #39.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add lacking description of man for -m and -M option

Currently there's no description of efibootmgr man page
for -m and -M option.

This patch adds the lacking description for -m and -M
option.

Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>

Add lacking description of man for -m and -M option

Currently there's no description of efibootmgr man page
for -m and -M option.

This patch adds the lacking description for -m and -M
option.

Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>

Explicitly pass a mode to efi_set_variable() in all cases.

We'll have to do this for efivar-0.24 .

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr.spec: allow building on arm

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr.spec: require popt-devel

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: fix some types the compiler doesn't like.

Basically uint8_t * vs char *.

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootdump: add efibootdump

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootnext: add make targets (but not default)

Signed-off-by: Peter Jones <pjones@redhat.com>

efoconman: remove from default make targets for now

Signed-off-by: Peter Jones <pjones@redhat.com>

efibootmgr: update for efivar 0.24

As of 0.24 efi_loadopt_pathlen() and efi_loadopt_path() require a limit
argument to avoid overruns of the input data.

This patch adds reasonable limits to those calls.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add -Wextra, because it finds a lot of bugs in GCC6

Signed-off-by: Peter Jones <pjones@redhat.com>

eficonman: intialize poptOptions with named initializers.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add efibootnext beginnings

efibootnext will (eventually) attempt to find any suitable boot device
and set it as BootNext.  To be used when other variable writes have
failed, in case it will use less space.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add eficonman beginnings

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix insusufficient validation check of -M option

The acceptable range of -M (--mirror-above-4G) option should
be from 0 to 50. So the negative value should be rejected.
Otherwise unintended value can be specified.

e.g.

 > efibootmgr -M -3
...
MirroredPercentageAbove4G: 25.00
MirrorMemoryBelow4GB: true
RequestMirroredPercentageAbove4G: 652.36
RequestMirrorMemoryBelow4GB: true

 652.36 is invalid value for MirroredPercentageAbove4G and should
 not be specified.

Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>

Add LOCALEDIR make goo

Signed-off-by: Peter Jones <pjones@redhat.com>

Simplify directory structure.

We're making a lot less stuff than we did before libefivar.so existed,
and there's not a lot of reason for it.  Collapse the directory
structure entirely, and move to more modern makefiles.

Signed-off-by: Peter Jones <pjones@redhat.com>

Set pkg-config binary as variable

Allow the location of the pkg-config binary to be set using the
parameter PKG_CONFIG. Needed to support cross compilation toolchains.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make it a little clearer what's going on when efi_del_variable() fails.

Right now if efi_del_variable() fails we try a couple of things for
compatibility reasons and we have too many different error messages.
Hopefully this makes it clearer.

Signed-off-by: Peter Jones <pjones@redhat.com>

INSTALL: add note about missing efivar.h error

Signed-off-by: Peter Jones <pjones@redhat.com>

Set guid when reading boot entries in read_vars.

set_active_state use boot->guid, so set it to EFI_GLOBAL_GUID.

Delete BootOrder only when necessary.

Signed-off-by: Shih-Yuan Lee (FourDollars) <sylee@ubuntu.com>

Add -m and -M options to support address range mirror

BIOS and OS pass messages for address range mirror using a pair of
UEFI variables:

   MirrorCurrent: Written by BIOS to tell the OS that the platform supports
address range mirror and to say whether mirroring was set up using
the desired parameters
   MirrorRequest: Written by OS to tell BIOS new paramters to use on next boot

This is documented in the RAS validation guide for memory address range
mirroring for the Xeon E7-v3 family:

  https://software.intel.com/sites/default/files/managed/43/6a/Memory%20Address%20Range%20Mirroring%20Validation%20Guide.pdf

Signed-off-by: Tony Luck <tony.luck@intel.com>

Update version to 0.12

- use -lefivar and -lefiboot to make device paths and load options.

Signed-off-by: Peter Jones <pjones@redhat.com>

Update build deps in the spec to match what libraries we actually use.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make the release procedure a bit more streamlined.

Signed-off-by: Peter Jones <pjones@redhat.com>

Add some language about efivar in INSTALL

Signed-off-by: Peter Jones <pjones@redhat.com>

Get rid of pci-utils and zlib deps; we don't need them any more.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make loadopt function names use efivar-0.19 names.

Signed-off-by: Peter Jones <pjones@redhat.com>

Get us back to where we were before with network entries.

Signed-off-by: Peter Jones <pjones@redhat.com>

We don't need the ucs2/utf8 code either.

Signed-off-by: Peter Jones <pjones@redhat.com>

Get rid of efichar.h; it's no longer used.

Signed-off-by: Peter Jones <pjones@redhat.com>

Fix the github url once more.

This currently lives under the rhinstaller github account.

Signed-off-by: Peter Jones <pjones@redhat.com>

Get rid of manually specified ACPI arguments; we don't need them.

Since we have pretty good detection of acpi hid/uid now, we don't
actually need them to be manually specified anywhere.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make -e 1 and -e 3 work again after the libefiboot shuffle.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make file boot option creation use libefiboot to create the device path.

(right now network options are horribly broken, but they never really
worked anyway, so... yeah, I'll fix them in a subsequent commit.)

Signed-off-by: Peter Jones <pjones@redhat.com>

Add "efibootmgr -C"; create only without appending to BootOrder

Signed-off-by: Peter Jones <pjones@redhat.com>

Get rid of a lot of DP stuff that efivar does now.

Get run of unparse_path() and unparse_raw()

We don't use these any more.

Signed-off-by: Peter Jones <pjones@redhat.com>

Make show_boot_vars use efivars DP formatting code.

Signed-off-by: Peter Jones <pjones@redhat.com>

Don't depend on vendor-defined calloc(n, 0) behavior.

If we don't have any boot variables, we know that 0 is free, so just
return that.

Signed-off-by: Peter Jones <pjones@redhat.com>