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>
In an image is loaded from memory we do not have a device path.
Do not install NULL as device path in this case.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When a UEFI payload just returns instead of calling the Exit() callback,
we handle that in efi_do_enter() and call Exit on its behalf, so that
the loaded_image->exit_status value is correct.
We were missing that logic in StartImage(). Call it there too.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Provide a better description for indent_string.
Fix a typo.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
EFI calls are usually defined as asmlinkage. That means we pass all parameters
to functions via the stack x86_32.
On x86_64 however, we need to also stick to the MS ABI calling conventions,
which the EFIAPI define conveniently handles for us. Most EFI functions were
also marked with EFIAPI, except for the entry call.
So this patch adjusts all entry calls to use EFIAPI instead of the manual
asmlinkage attribute.
While at it, we also change the prototype of the entry point to return
efi_status_t instead of ulong, as this is the correct prototype definition.
Signed-off-by: Alexander Graf <agraf@suse.de>
---
v1 -> v2:
- Use efi_status_t in all occurences
We have 2 users of the EFI headers: efi_loader and the EFI stub. Efi_loader
always expects that the bitness of the definitions it uses is identical to
the execution.
The EFI stub however allows to run x86_64 U-Boot on 32bit EFI and the other
way around, so it allows for different bitness of EFI definitions and U-Boot
environment.
This patch explicitly requests via Kconfig that efi_loader can only be enabled
if the bitness is identical. Because we can run efi_loader on x86_64 without
EFI stub enabled, it also ensures that this case propagates the correct ABI
constraints.
Signed-off-by: Alexander Graf <agraf@suse.de>
Create device path nodes for UCLASS_ETH udevices.
Create device path nodes of block device children of UCLASS_MMC udevices.
Consistently use debug for unsupported nodes.
Set the log level to error.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
[agraf: Fix build failure by adding #ifdef CONFIG_DM_ETH]
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch lets the implementation of ExitBootServices conform to
the UEFI standard.
The timer events must be disabled before calling the notification
functions of the exit boot services events.
The boot services must be disabled in the system table.
The handles in the system table should be defined as efi_handle_t.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
In ExitBootServices we need to signal events irrespective of the current
TPL level. A new parameter check_tpl is added to efi_signal_event().
Function efi_console_timer_notify() gets some comments.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
On a block device and its partitions the same protocols can be
installed. To tell the apart we can use the type of the last
node of the device path which is not the end node.
The patch provides a utility function to find this last node.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Add the revision constants.
Depending on the revision additional fields are needed in the
media descriptor.
Use efi_uintn_t for number of bytes to read or write.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Up to now we have been using efi_disk_create_partitions() to create
partitions for block devices that existed before starting an EFI
application.
We need to call it for block devices created by EFI
applications at run time. The EFI application will define the
handle for the block device and install a device path protocol
on it. We have to use this device path as stem for the partition
device paths.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Provide new function efi_dp_part_node() to create a device
node for a partition.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The GUID of the EFI_BLOCK_IO_PROTOCOL is needed in different code
parts. To avoid duplication make efi_block_io_guid a global symbol.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The interface type name can be used to look up the interface type.
Don't confound it with the driver name which may be different.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The calling convention for the entry point of an EFI image
is always 'asmlinkage'.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Check if the device tree and the SMBIOS table are available.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The category of memory allocated for an EFI image should depend on
its type (application, bootime service driver, runtime service driver).
Our helloworld.efi built on arm64 has an illegal image type. Treat it
like an EFI application.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Use %pD to print the device path instead of its address when
entering efi_load_image.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
In contrast to the description the code did not split the device
path into device part and file part.
The code should use the installed protocol and not refer to the
internal structure of the the disk object.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When installing the the simple file system protocol we have to path
the address of the structure and not the address of a pointer to the
structure.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
For the construction of device paths we need to call the
AllocatePool service. We should not ignore if it fails due to an
out of memory situation.
This patch changes the device path functions to return NULL if
the memory allocation fails.
Additional patches will be needed to fix the callers.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
We should consistently use the efi_handle_t typedef when
referring to handles.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
The UninstallProtocol boot service should first try to
disconnect controllers that have been connected with
EFI_OPEN_PROTOCOL_BY_DRIVER.
If the protocol is still opened by an agent, it should be
closed.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Unfortunately we need a forward declaration because both
OpenProtocol and CloseProtocol have to call DisconnectController.
And DisconnectController calls both OpenProtcol and CloseProtocol.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Handles should be passed as efi_handle_t and not as void *.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Implement the ConnectController boot service.
A unit test is supplied in a subsequent patch.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When a device path protocol is installed write the device
path to the console in debug mode.
For printing the new macro EFI_PRINT is used, which can be reused
for future diagnostic output.
Remove unused EFI_PRINT_GUID macro
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_open_protocol_information provides the agent and controller
handles as well as the attributes and open count of an protocol
on a handle.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
[agraf: fix counting error]
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_open_protocol and efi_close_protocol have to keep track of
opened protocols.
Check if the protocol was opened for the same agent and
controller.
Remove all open protocol information for this pair.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_open_protocol has to keep track of opened protocols.
OpenProtocol enters the agent and controller handle
information into this list.
A unit test is supplied with a subsequent patch.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Add a list of open protocol infos to each protocol of a handle.
Provide helper functions to access the list items.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Replace list_for_each_safe() and list_entry() by
list_for_each_entry_safe().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Avoid a failed assertion when an EFI app calls an EFI app.
Avoid that the indent level increases when calling 'bootefi hello'
repeatedly.
Avoid negative indent level when an EFI app calls an EFI app that
calls an EFI app (e.g. iPXE loads grub which starts the kernel).
Return the status code of a loaded image that returns without
calling the Exit boot service.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Without the patch a device path consisting only of an end node is
displayed as '/UNKNOWN(007f,00ff)'. It should be displayed as '/'.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
virt_size is of type unsigned long.
So it should be printed with %ul.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Add comments describing memory functions.
Fix the formatting of a function declaration.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
As the U-Boot is compiled with -fshort-wchar we can define
the firmware vendor constant as wide string.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Correctly create the device path for IDE and SCSI disks.
Support for SATA remains to be done in a future patch.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When creating the device path of a block device it has to
comprise the block device itself and should not end at
its parent.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When adding a partition, set the logical_partition member in the media
structure as mandated by the UEFI spec.
Signed-off-by: Emmanuel Vadot <manu@freebsd.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Add a description for dp_part_fill().
Reword a comment in the function.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When converting device nodes and paths to text we should
stick to the UEFI spec.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
According to the UEFI spec the numbering of partitions has to
start with 1.
Partion number 0 is reserved for the optional device path for
the complete block device.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The SD cards and eMMC devices have different device nodes.
The current coding interpretes all MMC devices as eMMC.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
If a failure occurs when trying to load an image, it is insufficient
to free() the EFI object. We must remove it from the object list,
too. Otherwise a use after free will occur the next time we
iterate over the object list.
Furthermore errors in setting up the image should be handled.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Provide a function to remove a handle from the object list
after removing all protocols.
To avoid forward declarations other functions have to move up
in the coding.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_setup_loaded_image() should return an error code indicating if
an error has occurred.
An error occurs if a protocol cannot be installed.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When calling efi_dp_find_obj(), we usually want to find the *exact* match
of an object for a given device path. However, I ran into a nasty corner case
where I had the following objects with paths available:
Handle 0x9feffa70
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]/USB(6,0)/EndEntire
Handle 0x9feffb58
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]/USB(6,0)/HD(1,800,32000,2de808cb00000000,1,1)/EndEntire
and was searching for
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]/USB(6,0)/HD(1,800,32000,2de808cb00000000,1,1)/EndEntire
But because our device path search looked for any substring match, it would
return
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]/USB(6,0)/EndEntire
because that path is a full substring of the path we were searching for.
So this patch adapts the device path search logic to always look for exact
matches first. The way we distinguish between those cases is by looking at
whether our caller actually deals with remainders.
As a side effect, the code as is from all I can tell now never does a
substring match anymore, because it always gets called with rem=NULL, so
we always only do exact matches now.
Reported-by: Jonathan Gray <jsg@jsg.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Commit bbf75dd934 ("efi_loader: output load options in helloworld")
introduced a const variable in efi_main() called loaded_image_guid which
got populated from a constant struct.
While you would usually expect a compiler to realize that this variable
should really just be a global pointer to .rodata, gcc disagrees and instead
puts it on the stack. Unfortunately in some implementations of gcc it does
so my calling memcpy() which we do not implement in our hello world
environment.
So let's explicitly move it to a global variable which in turn puts it in
.rodata reliably and gets rid of the memcpy().
Fixes: bbf75dd934 ("efi_loader: output load options in helloworld")
Reported-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Heinrich Schuchardt
Alexander Graf
Emmanuel Vadot