Allow booting from a short form device-path without file path, e.g.
/HD(1,GPT,5ef79931-a1aa-4c70-9d67-611e8f69eafd,0x800,0x1000)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Booting from a short-form device path which starts with the first element
being a File Path Media Device Path failed because it doesn't contain
any valid device with simple file system protocol and efi_dp_find_obj()
in efi_load_image_from_path() will return NULL.
For instance,
/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)/Scsi(0,0)/\helloworld.efi
-> shortened version: /\helloworld.efi
With this patch applied, all the media devices with simple file system
protocol are enumerated and the boot manager attempts to boot temporarily
generated device paths one-by-one.
This new implementation is still a bit incompatible with the UEFI
specification in terms of:
* not creating real boot options
* not try
"If a device does not support the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL, but
supports the EFI_BLOCK_IO_PROTOCOL protocol, then the EFI Boot Service
ConnectController must be called for this device with DriverImageHandle
and RemainingDevicePath set to NULL and the Recursive flag is set to TRUE."
(See section 3.1.2 "Load Option Processing".)
But it still gives us a closer and better solution than the current.
Fixes: commit 9cdf470274 ("efi_loader: support booting via short-form device-path")
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Under the current implementation, booting from removable media using
a architecture-specific default image name, say BOOTAA64.EFI, is
supported only in distro_bootcmd script. See the commit 74522c898b
("efi_loader: Add distro boot script for removable media").
This is, however, half-baked implementation because
1) UEFI specification requires this feature to be implemented as part
of Boot Manager's responsibility:
3 - Boot Manager
3.5.1 Boot via the Simple File Protocol
When booting via the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL, the FilePath will
start with a device path that points to the device that implements the
EFI_SIMPLE_FILE_SYSTEM_PROTOCOL or the EFI_BLOCK_IO_PROTOCOL. The next
part of the FilePath may point to the file name, including
subdirectories, which contain the bootable image. If the file name is
a null device path, the file name must be generated from the rules
defined below.
...
3.5.1.1 Removable Media Boot Behavior
To generate a file name when none is present in the FilePath, the
firmware must append a default file name in the form
\EFI\BOOT\BOOT{machine type short-name}.EFI ...
2) So (1) entails the hehavior that the user's preference of boot media
order should be determined by Boot#### and BootOrder variables.
With this patch, the semantics mentioned above is fully implemented.
For example, if you want to boot the system from USB and SCSI in this
order,
* define Boot0001 which contains only a device path to the USB device
(without any file path/name)
* define Boot0002 which contains only a device path to the SCSI device,
and
* set BootOrder to Boot0001:Boot0002
To avoid build error for sandbox, default file name "BOOTSANDBOX.efi"
is defined even if it is out of scope of UEFI specification.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
On sandbox use binary name corresponding to host architecture.
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
The log functions print file name, line number, and function name if
selected via the log command or customizing. Don't print the function
name twice.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
At present we use wide characters for unicode but this is not necessary.
Change the code to use the 'u' literal instead. This helps to fix build
warnings for sandbox on rpi.
Signed-off-by: Simon Glass <sjg@chromium.org>
Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
get_var() is defined statically in efi_bootmgr.c and doesn't properly
check a buffer allocation. Remove it completely and use the exported
function from efi_var_common.c that does the same thing
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Up to now we install EFI_LOAD_FILE2_PROTOCOL to load an initrd
unconditionally. Although we correctly return various EFI exit codes
depending on the file status (i.e EFI_NO_MEDIA, EFI_NOT_FOUND etc), the
kernel loader, only falls back to the cmdline interpreted initrd if the
protocol is not installed.
This creates a problem for EFI installers, since they won't be able to
load their own initrd and continue the installation. It also makes the
feature hard to use, since we can either have a single initrd or we have
to recompile u-boot if the filename changes.
So let's introduce a different logic that will decouple the initrd
path from the config option we currently have.
When defining a UEFI BootXXXX we can use the filepathlist and store
a file path pointing to our initrd. Specifically the EFI spec describes:
"The first element of the array is a device path that describes the device
and location of the Image for this load option. Other device paths may
optionally exist in the FilePathList, but their usage is OSV specific"
When the EFI application is launched through the bootmgr, we'll try to
interpret the extra device path. If that points to a file that exists on
our disk, we'll now install the load_file2 and the efi-stub will be able
to use it.
This opens up another path using U-Boot and defines a new boot flow.
A user will be able to control the kernel/initrd pairs without explicit
cmdline args or GRUB.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Move all load options related functions to a new module. So that they can
be compiled independently.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
load_options passed from do_efibootmgr() to do_bootefi_exec() may contain
invalid data from the stack which will lead to an invalid free().
Fixes: 0ad64007fe ("efi_loader: set load options in boot manager")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Bootxxxx variables are provided by the user and therefore cannot be
trusted. We have to validate them before usage.
A device path provided by a Bootxxxx variable must have an end node within
the indicated device path length.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Up to now we used the value of the bootargs environment variable as load
options in the boot manager. This is not correct. The data has to be taken
from the Boot#### variable.
Let the boot manager copy the optional data of the EFI_LOAD_OPTION as load
options to the loaded image protocol.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The bootefi bootmgr command has to set the load options for a loaded image
from the value of BootXXXX variable. If the boot manager is not used, the
value is set from the environment variable bootargs (or efi_selftest).
Factor out a common function efi_set_load_options().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Simplify the implementation of the UEFI boot manager:
* avoid EFI_CALL for SetVariable() and GetVariable()
* remove unnecessary type conversions
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When booting via the boot manager use log function for user messages
instead of printf() and debug().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
For passing the optional data of the load option to the loaded imaged
protocol we need its size.
efi_deserialize_load_option() is changed to return the size of the optional
data.
As a by-product we get a partial validation of the load option.
Checking the length of the device path remains to be implemented.
Some Coverity defects identified the load options as user input because
get_unaligned_le32() and get_unaligned_le16() is called. But non of these
Coverity defects can be resolved without marking functions with Coverity
specific tags.
Reported-by: Coverity (CID 303760)
Reported-by: Coverity (CID 303768)
Reported-by: Coverity (CID 303776)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
If a user defines BootNext but not BootOrder and loading from BootNext
fails, you will see only a message like this:
BootOrder not defined
This may confuse a user. Adding an error message will be helpful.
An example output looks like this:
=> efidebug boot add 0001 label1 scsi 0:1 "\path1\file1.efi" "--option foo"
=> efidebug boot add 0002 label2 scsi 0:1 "\path2\file2.efi" "--option bar"
=> efidebug boot add 0003 label3 scsi 0:1 "\path3\file3.efi" "--option no"
=> efidebug boot order 0001 0002
=> efidebug boot next 0003
=> bootefi bootmgr
Loading from Boot0003 'label3' failed
Loading from BootNext failed, falling back to BootOrder
Loading from Boot0001 'label1' failed
Loading from Boot0002 'label2' failed
EFI boot manager: Cannot load any image
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Adjust messages.
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The field boot OptionalData in structure _EFI_LOAD_OPTIONS is for binary
data.
When we use `efidebug boot add` we should convert the 5th argument from
UTF-8 to UTF-16 before putting it into the BootXXXX variable.
When printing boot variables with `efidebug boot dump` we should support
the OptionalData being arbitrary binary data. So let's dump the data as
hexadecimal values.
Here is an example session protocol:
=> efidebug boot add 00a1 label1 scsi 0:1 doit1 'my option'
=> efidebug boot add 00a2 label2 scsi 0:1 doit2
=> efidebug boot dump
Boot00A0:
attributes: A-- (0x00000001)
label: label1
file_path: .../HD(1,MBR,0xeac4e18b,0x800,0x3fffe)/doit1
data:
00000000: 6d 00 79 00 20 00 6f 00 70 00 74 00 69 00 6f 00 m.y. .o.p.t.i.o.
00000010: 6e 00 00 00 n...
Boot00A1:
attributes: A-- (0x00000001)
label: label2
file_path: .../HD(1,MBR,0xeac4e18b,0x800,0x3fffe)/doit2
data:
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In the current implementation, bootefi command and EFI boot manager
don't use load_image API, instead, use more primitive and internal
functions. This will introduce duplicated code and potentially
unknown bugs as well as inconsistent behaviours.
With this patch, do_efibootmgr() and do_boot_efi() are completely
overhauled and re-implemented using load_image API.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Use efi_root as parent handle for the loaded image.
LoadImage() should be called with BootPolicy = true by the boot manager.
Avoid duplicate free_pool().
Eliminate variable memdp which is not needed after anymore due to
"efi_loader: correctly split device path of loaded image".
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Since commit 914df75b0c ("efi_loader: fix EFI entry counting")
entry_count is already set to 1 before efi_bootmgr_load() is called. So we
should not increment it when entering the function.
Without the patch an assert error occurs in efi_get_variable() if DEBUG is
defined.
Fixes: 914df75b0c ("efi_loader: fix EFI entry counting")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
For booting via `bootefi bootmgr` it is necessary that the EFI variable
BootOrder is defined. Provide a diagnostic message if the variable is
missing.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
If we want to properly unload images in Exit() the memory should always be
allocated in the same way. As we allocate memory when reading from file we
should do the same when the original image is in memory.
A further patch will be needed to free the memory when Exit() is called.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In this patch, helper functions for an load option variable (BootXXXX)
are added:
* efi_deserialize_load_option(): parse a string into load_option data
(renamed from parse_load_option and exported)
* efi_serialize_load_option(): convert load_option data into a string
Those functions will be used to implement efishell command.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
The function names utf16_strlen() and utf16_strnlen() are misnomers.
The functions do not count utf-16 characters but non-zero words.
So let's rename them to u16_strlen and u16_strnlen().
In utf16_dup() avoid assignment in if clause.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The definitons of the variable services are adjusted:
- use efi_uintn_t instead of unsigned long
- use u16 * instead of s16 * for Unicode strings
- correct definition of QueryVariableInfo
- rename efi_get_next_variable to efi_get_next_variable_name
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
As part of the main conversion a few files were missed. These files had
additional whitespace after the '*' and before the SPDX tag and my
previous regex was too strict. This time I did a grep for all SPDX tags
and then filtered out anything that matched the correct styles.
Fixes: 83d290c56f ("SPDX: Convert all of our single license tags to Linux Kernel style")
Reported-by: Heinrich Schuchardt <xypron.debian@gmx.de>
Signed-off-by: Tom Rini <trini@konsulko.com>
Now that we have %pD support in vsprintf we should avoid separate
logic for printing device paths in other places.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Similar to a "real" UEFI implementation, the bootmgr looks at the
BootOrder and BootXXXX variables to try to find an EFI payload to load
and boot. This is added as a sub-command of bootefi.
The idea is that the distro bootcmd would first try loading a payload
via the bootmgr, and then if that fails (ie. first boot or corrupted
EFI variables) it would fallback to loading bootaa64.efi. (Which
would then load fallback.efi which would look for \EFI\*\boot.csv and
populate BootOrder and BootXXXX based on what it found.)
Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
