Generating the canary at build time allows the canary to be different for
every build which could limit the effectiveness of certain exploits.
Fallback to the statically generated random bytes if /dev/urandom is not
readable, e.g. Windows.
On 32-bit architectures, which use a 32-bit canary, reduce the canary to
4 bytes with one byte being NUL to filter out string buffer overflow attacks.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The canary, __stack_chk_guard, is in the BSS and so will get initialized to
zero if it is not explicitly initialized. If the UEFI firmware does not
support the RNG protocol, then the canary will not be randomized and will
be zero. This seems like a possibly easier value to write by an attacker.
Initialize canary to static random bytes, so that it is still random when
there is no RNG protocol. Set at least one byte to NUL to protect against
string buffer overflow attacks [1]. Code that writes NUL terminated strings
will terminate when a NUL is encountered in the input byte stream. So the
attacker will not be able to forge the canary by including it in the input
stream without terminating the string operation and thus limiting the
stack corruption.
[1] https://www.sans.org/blog/stack-canaries-gingerly-sidestepping-the-cage/
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GCC is electing to instrument grub_efi_init() to give it stack smashing
protection when configuring with --enable-stack-protector on the x86_64-efi
target. In the function prologue, the canary at the top of the stack frame
is set to the value of the stack guard. And in the epilogue, the canary is
checked to verify if it is equal to the guard and if not to call the stack
check fail function. The issue is that grub_efi_init() sets up the guard
by initializing it with random bytes, if the firmware supports the RNG
protocol. So in its prologue the canary will be set with the value of the
uninitialized guard, likely NUL bytes. Then the guard is initialized, and
finally the epilogue checks the canary against the guard, which will almost
certainly be different. This causes the code path for a smashed stack to be
taken, causing the machine to print out a message that stack smashing was
detected, wait 5 seconds, and then reboot. Disable grub_efi_init()
instrumentation so there is no stack smashing false positive generated.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
EFI firmware determines where to load the GRUB EFI at runtime, and so the
addresses of debug symbols are not known ahead of time. There is a command
defined in the gdb_grub script which will load the debug symbols at the
appropriate addresses, if given the application load address for GRUB.
So add a command named "gdbinfo" to allow the user to print this GDB command
string with the application load address on-demand. For the outputted GDB
command to have any effect when entered into a GDB session, GDB should have
been started with the script as an argument to the -x option or sourced into
an active GDB session before running the outputted command.
Documentation for the gdbinfo command is also added.
Co-developed-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are 3 implementations of a GUID in GRUB. Replace them with
a common one, placed in types.h.
It uses the "packed" flavor of the GUID structs, the alignment attribute
is dropped, since it is not required.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Now that GCC can generate function calls using the correct calling
convention for us, we can stop using the efi_call_XX() wrappers, and
just dereference the function pointers directly.
This avoids the untyped variadic wrapper routines, which means better
type checking for the method calls.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The initial implementation of the stack protector just busy looped
in __stack_chk_fail in order to reduce the amount of code being
executed after the stack has been compromised because of a lack of
firmware memory protections. With future firmware implementations
incorporating memory protections such as W^X, call in to boot services
when an error occurs in order to log a message to the console before
automatically rebooting the machine.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It works only on UEFI platforms but can be quite easily extended to
others architectures and platforms if needed.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Marco A Benatto <mbenatto@redhat.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
If the UEFI Secure Boot is enabled then the GRUB must be locked down
to prevent executing code that can potentially be used to subvert its
verification mechanisms.
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Move the shim_lock verifier from its own module into the core image. The
Secure Boot lockdown mechanism has the intent to prevent the load of any
unsigned code or binary when Secure Boot is enabled.
The reason is that GRUB must be able to prevent executing untrusted code
if UEFI Secure Boot is enabled, without depending on external modules.
Signed-off-by: Marco A Benatto <mbenatto@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
commit 92bfc33db984 ("efi: Free malloc regions on exit")
introduced memory freeing in grub_efi_fini(), which is
used not only by exit path but by halt/reboot one as well.
As result of memory freeing, code and data regions used by
modules, such as halt, reboot, acpi (used by halt) also got
freed. After return to module code, CPU executes, filled
by UEFI firmware (tested with edk2), 0xAFAFAFAF pattern as
a code. Which leads to #UD exception later.
grub> halt
!!!! X64 Exception Type - 06(#UD - Invalid Opcode) CPU Apic ID - 00000000 !!!!
RIP - 0000000003F4EC28, CS - 0000000000000038, RFLAGS - 0000000000200246
RAX - 0000000000000000, RCX - 00000000061DA188, RDX - 0A74C0854DC35D41
RBX - 0000000003E10E08, RSP - 0000000007F0F860, RBP - 0000000000000000
RSI - 00000000064DB768, RDI - 000000000832C5C3
R8 - 0000000000000002, R9 - 0000000000000000, R10 - 00000000061E2E52
R11 - 0000000000000020, R12 - 0000000003EE5C1F, R13 - 00000000061E0FF4
R14 - 0000000003E10D80, R15 - 00000000061E2F60
DS - 0000000000000030, ES - 0000000000000030, FS - 0000000000000030
GS - 0000000000000030, SS - 0000000000000030
CR0 - 0000000080010033, CR2 - 0000000000000000, CR3 - 0000000007C01000
CR4 - 0000000000000668, CR8 - 0000000000000000
DR0 - 0000000000000000, DR1 - 0000000000000000, DR2 - 0000000000000000
DR3 - 0000000000000000, DR6 - 00000000FFFF0FF0, DR7 - 0000000000000400
GDTR - 00000000079EEA98 0000000000000047, LDTR - 0000000000000000
IDTR - 0000000007598018 0000000000000FFF, TR - 0000000000000000
FXSAVE_STATE - 0000000007F0F4C0
Proposal here is to continue to free allocated memory for
exit boot services path but keep it for halt/reboot path
as it won't be much security concern here.
Introduced GRUB_LOADER_FLAG_EFI_KEEP_ALLOCATED_MEMORY
loader flag to be used by efi halt/reboot path.
Signed-off-by: Alexey Makhalov <amakhalov@vmware.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When we exit grub, we don't free all the memory that we allocated earlier
for our heap region. This can cause problems with setups where you try
to descend the boot order using "exit" entries, such as PXE -> HD boot
scenarios.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The path returned by grub_efi_net_config has already been stripped for the
directory part extracted from cached bootp packet. We should just return the
result to avoild it be stripped again.
It fixed the problem that grub.efi as NBP image always looking for grub.cfg and
platform directory in upper folder rather than current one it gets loaded while
$prefix is empty. The behavior is inconsistent with other architecture and how
we would expect empty $prefix going to be in general.
The only exception to the general rule of empty $prefix is that when loaded
from platform directory itself, the platform part is stripped thus upper folder
is used for looking up files. It meets the case for how grub-mknetdir lay out
the files under tftp root directory, but also hide away this issue to be
identified as it appears to be just works.
Also fix possible memory leak by moving grub_efi_get_filename() call after
grub_efi_net_config().
We only support FDT files with EFI on arm and arm64 systems, not
on x86. So move the helper that finds a prepopulated FDT UUID
into its own file and only build it for architectures where it
also gets called.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Searching for a device tree that EFI passes to us via configuration tables
is nothing architecture specific. Move it into generic code.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI version of grub_machine_get_bootlocation crops the boot image
name back to the last / in order to get a directory path. However, it
does not check that *name is actually set before calling grub_strrchr
to do this, and neither does grub_strrchr before dereferencing a NULL
pointer.
Parent function, grub_set_prefix_and_root, does check the pointer
before using.
starts with "(,", fill the drive containing the loaded image in
between those two characters, but expect that a full partition
specification including partition map names will follow.
