We need to check that device paths provided via UEFI variables are not
malformed.
Provide function efi_dp_check_length() to check if a device path has an
end node within a given number of bytes.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Provide device path nodes for SATA devices.
This avoids creation of two handles with the same device path indicating
our root node.
This is what the device paths for a SATA drive with four partitions could
like:
/VenHw(..)/Sata(0x0,0xffff,0x0)
/VenHw(..)/Sata(0x0,0xffff,0x0)/HD(1,MBR,0x81ea591f,0x800,0x63ff830)
/VenHw(..)/Sata(0x0,0xffff,0x0)/HD(2,MBR,0x81ea591f,0x6400800,0x9ff830)
/VenHw(..)/Sata(0x0,0xffff,0x0)/HD(3,MBR,0x81ea591f,0x6e00800,0x16ef2ab0)
/VenHw(..)/Sata(0x0,0xffff,0x0)/HD(4,MBR,0x81ea591f,0x1dcf3800,0x1dcedab0)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The UEFI specification does not define a device sub-type for virtio.
Let's use a vendor hardware node here.
This avoids creation of two handles with the same device path indicating
our root node.
Reported-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
We should not be using typedefs and these make it harder to use
forward declarations (to reduce header file inclusions). Drop the typedef.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.
Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.
Signed-off-by: Simon Glass <sjg@chromium.org>
Correctly check the return value of efi_dp_from_file().
If we can determine device path for the file, should not depend on the
device path for the device being requested.
Provide a function description for efi_dp_from_name().
Reported-by: Coverity CID 273159, CID 273158
Fixes: 08c51fff30 ("efi_loader: device_path: check against file path length")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
device_path strcuture has 2 bytes of "length" field, and so
file path length should not exceed this limit, 65535.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This allows our EFI API to create a device path node for NVMe
devices. It adds the necessary device path struct, uses the
nvme namespace accessor to retrieve the id and eui64, and also
provides support for the device path text protocol.
Signed-off-by: Patrick Wildt <patrick@blueri.se>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Sandbox's "host" devices are currently described as UCLASS_ROOT udevice
with DEV_IF_HOST block device. As the current implementation of
efi_device_path doesn't support such a type, any "host" device
on sandbox cannot be seen as a distinct object.
For example,
=> host bind 0 /foo/disk.img
=> efi devices
Scanning disk host0...
Found 1 disks
Device Device Path
================ ====================
0000000015c19970 /VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)
0000000015c19d70 /VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)
=> efi dh
Handle Protocols
================ ====================
0000000015c19970 Device Path, Device Path To Text, Device Path Utilities, Unicode Collation 2, HII String, HII Database, HII Config Routing
0000000015c19ba0 Driver Binding
0000000015c19c10 Simple Text Output
0000000015c19c80 Simple Text Input, Simple Text Input Ex
0000000015c19d70 Block IO, Device Path, Simple File System
As you can see here, efi_root (0x0000000015c19970) and host0 device
(0x0000000015c19d70) have the same representation of device path.
This is not only inconvenient, but also confusing since two different
efi objects are associated with the same device path and
efi_dp_find_obj() will possibly return a wrong result.
Solution:
Each "host" device should be given an additional device path node
of "vendor device path" to make it distinguishable.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Correct the name of the partition size component in struct
efi_device_path_cdrom_path.
Render entry, start, and size when converting a CD-ROM device path node to
text.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When creating a device path node ensure that the size of the allocated
memory at lest matches the size of the node header.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This is a preparatory patch.
efi_dp_split_file_path() is used to create device_path and file_path
from file_path for efi_setup_loaded_image().
In a special case, however, of HARDWARE_DEVICE/MEMORY, it doesn't
work expectedly since this path doesn't contain any FILE_PATH sub-type.
This patch makes a workaround.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Adjust the logic such that for all paths that do no end on a media file
path we return NULL as file_path.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Check the value of block_dev before to use this pointer.
This patch solves problem for the command "load" when ubifs
is previously mounted: in this case the function
blk_get_device_part_str("ubi 0") don't return error but return
block_dev = NULL and then data abort.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The statement '#define LOG_CATEGORY LOGL_ERR' makes not sense. LOGL_ERR is
not a LOG_CATEGORY. Remove the statement.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In the `efidebug boot add` command we do not want an unsolicited leading
backslash added to the file name.
There is no good reason to mark a loaded file with a backslash as absolute.
Anyway when reading files the file name will be interpreted as relative to
root directory of the device.
So let's get rid of this backslash.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
If the path passed to efi_dp_split_file_path() does not contain a
reference to a file it returns EFI_OUT_OF_RESOURCES. This does not properly
indicate the kind of the problem that occurred. Return
EFI_INVALID_PARAMETER instead.
Update function description.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This is a preparatory patch for use in efi_serialize_load_option()
as a load option's file_path should have both a device path and
a file path.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Factor out efi_set_bootdev() and extract efi_dp_from_name().
This function will be used to set a boot device in efishell command.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
A pointer to a struct efi_object is a handle. We do not need any handle
member in this structure. Let's eliminate it.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Remove a superfluous statement in efi_dp_get_next_instance().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently we assign a lot of protocols to loaded images though
these protocols are not related to them. Instead they should be
installed on a separate handle. Via the device path it is the
parent to the devices like the network adapter.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
In int-ll64.h, we always use the following typedefs:
typedef unsigned int u32;
typedef unsigned long uintptr_t;
typedef unsigned long long u64;
This does not need to match to the compiler's <inttypes.h>.
Do not include it.
The use of PRI* makes the code super-ugly. You can simply use
"l" for printing uintptr_t, "ll" for u64, and no modifier for u32.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
The missing services of the EFI_DEVICE_PATH_UTILITIES_PROTOCOL are
implemented.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Device paths may consist of multiple instances. Up to now we have only
considered the size of the first instance. For the services of the
EFI_DEVICE_PATH_UTILITIES_PROTOCOL in most cases the total length of the
device path is relevant.
So let's rename efi_dp_size() to efi_dp_instance_size() and create a new
function efi_dp_size() that calculates the total device path length.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_dp_size() is meant to return the device path length without the end
node.
The length of a device path containing only an end node was incorrectly
reported as 4.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The logic of the AppendDevicePath service of the
EFI_DEVICE_PATH_UTILITIES_PROTOCOL is incorrectly implemented:
* if both paths are NULL an end node has to be returned
* if both paths are not NULL the end node of the second device path has to
be kept
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Implement the CreateDeviceNode service of the device path utility protocol.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
This reverts the parts of commit 3b3ea2c56e
where it changed the EFI dependency on NET.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Duncan Hare <dh@synoia.com>
Since the backing memory for a new device path can contain stale
data we have to make sure that we zero the buffer. Otherwise some
code paths that don't set all fields in a structure backed by this
device path might contain unwanted stale data.
Signed-off-by: Patrick Wildt <patrick@blueri.se>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
When CONFIG_DM_ETH is set, we assemble the device path properly with a
full device hierarchy. Our helper function dp_fill() even put the MAC
node itself in it for us.
However, for non-DM compatibility we also have code in that added the
MAC node manually. That code now runs on top of the existing MAC node:
Handle 0x3db2f6b0
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]
/USBClass(0,0,9,0,0)/USBClass(424,9514,9,0,2)/MacAddr(b8:27:eb:e1:81:47,1)
/MacAddr(b8:27:eb:e1:81:47,57)/EndEntire
We obviously don't need the additional node and in fact, grub chokes on
it and fails to match the DP against the ethernet device node. So this
patch moves the additional MAC node into the non-DM code path:
Handle 0x3db3fde0
/HardwareVendor(e61d73b9-a384-4acc-aeab-82e828f3628b)[0: ]
/USBClass(0,0,9,0,0)/USBClass(424,9514,9,0,2)/MacAddr(b8:27:eb:e1:81:47,1)
/EndEntire
While at it, we also mark the non-DM MAC node as ethernet.
Fixes: b66c60dde9 ("efi_loader: add device-path utils")
Signed-off-by: Alexander Graf <agraf@suse.de>
Enable networking command only when NET is enabled.
And remove selecting NET for CMD_NET
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
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>
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>
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>
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>
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>
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>
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>
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>
Use function efi_search_protocol().
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Zero partition_signature in the efi_device_path_hard_drive_path
structure when signature_type is 0 (no signature) as required by the
UEFI specification.
This is required so that efi_dp_match() will work as expected
when doing memcmp() comparisons. Previously uninitialised memory
would cause it not match nodes when it should have when the signature
type was not GUID.
Corrects a problem where the loaded image protocol would not return a
device path with MEDIA_DEVICE causing the OpenBSD bootloader to fail
on rpi_3 and other targets.
v2: Also handle signature_type 1 (MBR) as described in the specification
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Tested-by: Artturi Alm <artturi.alm@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
efi_dp_match does not change its arguments.
So they should be marked as const.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
