In addition to what was already there, Gnulib's <intprops.h> needs SCHAR_MIN,
SCHAR_MAX, SHRT_MIN, INT_MIN, LONG_MIN, and LONG_MAX. Fixes build on CentOS 7.
Reported-by: "Chen, Farrah" <farrah.chen@intel.com>
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Mirror behaviour of ELF loader in libxc: first look for Xen notes in
PT_NOTE segment, then in SHT_NOTE section and only then fallback to
a section with __xen_guest name. This fixes loading PV kernels that
Xen note have outside of PT_NOTE. While this may be result of a buggy
linker script, loading such kernel directly works fine, so make it work
with GRUB too. Specifically, this applies to binaries built from Unikraft.
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Various GRUB utilities fail if the current directory doesn't exist,
because grub_find_device() chdirs to a different directory and then
fails when trying to chdir back. Gnulib's save-cwd module uses fchdir()
instead when it can, avoiding this category of problem.
Fixes Debian bug #918700.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Mostly for cosmetic reasons, we add a "net_dhcp" command, which is (at the
moment) identical to the existing "net_bootp" command. Both actually trigger
a DHCP handshake now, and both should be able to deal with pure BOOTP servers.
We could think about dropping the DHCP options from the initial DISCOVER packet
when the user issues the net_bootp command, but it's unclear whether this is
really useful, as both protocols should be able to coexist.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Even though we were parsing some DHCP options sent by the server, so far
we are only using the BOOTP 2-way handshake, even when talking to a DHCP
server.
Change this by actually sending out DHCP DISCOVER packets instead of the
generic (mostly empty) BOOTP BOOTREQUEST packets.
A pure BOOTP server would ignore the extra DHCP options in the DISCOVER
packet and would just reply with a BOOTREPLY packet, which we also
handle in the code.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In respone to a BOOTREQUEST packet a BOOTP server would answer with a BOOTREPLY
packet, which ends the conversation for good. DHCP uses a 4-way handshake,
where the initial server respone is an OFFER, which has to be answered with
REQUEST by the client again, only to be completed by an ACKNOWLEDGE packet
from the server.
Teach the grub_net_process_dhcp() function to deal with OFFER packets,
and treat ACK packets the same es BOOTREPLY packets.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The BOOTP RFC describes the boot file name and the server name as being part
of the integral BOOTP data structure, with some limits on the size of them.
DHCP extends this by allowing them to be separate DHCP options, which is more
flexible.
Teach the code dealing with those fields to check for those DHCP options first
and use this information, if provided. We fall back to using the BOOTP
information if those options are not used.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently we have a global timeout for all network cards in the BOOTP/DHCP
discovery process.
Make this timeout a per-interface one, so better accommodate the upcoming
4-way DHCP handshake and to also cover the lease time limit a DHCP offer
will come with.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Change the interface of the function dealing with incoming BOOTP packets
to take an interface instead of a card, to allow more fine per-interface
state (timeout, handshake state) later on.
Use the opportunity to clean up the code a bit.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In contrast to BOOTP, DHCP uses a 4-way handshake, so requires to send
packets more often.
Refactor the generation and sending of the BOOTREQUEST packet into
a separate function, so that future code can more easily reuse this.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
DHCP specifies a special dummy option OVERLOAD, to allow DHCP options to
spill over into the (legacy) BOOTFILE and SNAME fields.
Parse and handle this option properly.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For proper DHCP support we will need to parse DHCP options from a packet
more often and at various places.
Refactor the option parsing into a new function, which will scan a packet to
find *a particular* option field. Use that new function in places where we
were dealing with DHCP options before.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The comment is right, the "giaddr" fields holds the IP address of the BOOTP
relay, not a general purpose router address. Just remove the commented code,
archeologists can find it in the git history.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In order to be able to read from and write to model-specific registers,
two new modules are added. They are i386 specific, as the cpuid module.
rdmsr module registers the command rdmsr that allows reading from a MSR.
wrmsr module registers the command wrmsr that allows writing to a MSR.
wrmsr module is disabled if UEFI secure boot is enabled.
Please note that on SMP systems, interacting with a MSR that has a scope
per hardware thread, implies that the value only applies to the
particular cpu/core/thread that ran the command.
Also, if you specify a reserved or unimplemented MSR address, it will
cause a general protection exception (which is not currently being
handled) and the system will reboot.
Signed-off-by: Jesús Diéguez Fernández <jesusdf@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In order to maintain the coding style consistency, it was requested to
replace the methods that use "__asm__ __volatile__" with "asm volatile".
Signed-off-by: Jesús Diéguez Fernández <jesusdf@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add BIOS Boot Partition support for sparc64 platforms. This will work a
little different than x86. With GPT, both the OBP "load" and "boot" commands
are partition aware and neither command can see the partition table. Therefore
the entire boot-loader is stored within the BIOS Boot Partition and nothing
is stored within the bootstrap code area of MBR.
To use it, the end user will issue the boot command with the path pointing to
the BIOS Boot Partition.
For example with the disk below:
Model: Unknown (unknown)
Disk /dev/nvme1n1: 1600GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Number Start End Size File system Name Flags
1 1049kB 1075MB 1074MB ext3
2 1075MB 1076MB 1049kB bios_grub
3 1076MB 1600GB 1599GB lvm
To boot grub2 from OBP, you would use:
boot /pci@302/pci@1/pci@0/pci@13/nvme@0/disk@1:b
Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a new disk driver called obdisk for IEEE1275 platforms. Currently
the only platform using this disk driver is SPARC, however other IEEE1275
platforms could start using it if they so choose. While the functionality
within the current IEEE1275 ofdisk driver may be suitable for PPC and x86, it
presented too many problems on SPARC hardware.
Within the old ofdisk, there is not a way to determine the true canonical
name for the disk. Within Open Boot, the same disk can have multiple names
but all reference the same disk. For example the same disk can be referenced
by its SAS WWN, using this form:
/pci@302/pci@2/pci@0/pci@17/LSI,sas@0/disk@w5000cca02f037d6d,0
It can also be referenced by its PHY identifier using this form:
/pci@302/pci@2/pci@0/pci@17/LSI,sas@0/disk@p0
It can also be referenced by its Target identifier using this form:
/pci@302/pci@2/pci@0/pci@17/LSI,sas@0/disk@0
Also, when the LUN=0, it is legal to omit the ,0 from the device name. So with
the disk above, before taking into account the device aliases, there are 6 ways
to reference the same disk.
Then it is possible to have 0 .. n device aliases all representing the same disk.
Within this new driver the true canonical name is determined using the the
IEEE1275 encode-unit and decode-unit commands when address_cells == 4. This
will determine the true single canonical name for the device so multiple ihandles
are not opened for the same device. This is what frequently happens with the old
ofdisk driver. With some devices when they are opened multiple times it causes
the entire system to hang.
Another problem solved with this driver is devices that do not have a device
alias can be booted and used within GRUB. Within the old ofdisk, this was not
possible, unless it was the original boot device. All devices behind a SAS
or SCSI parent can be found. Within the old ofdisk, finding these disks
relied on there being an alias defined. The alias requirement is not
necessary with this new driver. It can also find devices behind a parent
after they have been hot-plugged. This is something that is not possible
with the old ofdisk driver.
The old ofdisk driver also incorrectly assumes that the device pointing to by a
device alias is in its true canonical form. This assumption is never made with
this new driver.
Another issue solved with this driver is that it properly caches the ihandle
for all open devices. The old ofdisk tries to do this by caching the last
opened ihandle. However this does not work properly because the layer above
does not use a consistent device name for the same disk when calling into the
driver. This is because the upper layer uses the bootpath value returned within
/chosen, other times it uses the device alias, and other times it uses the
value within grub.cfg. It does not have a way to figure out that these devices
are the same disk. This is not a problem with this new driver.
Due to the way GRUB repeatedly opens and closes the same disk. Caching the
ihandle is important on SPARC. Without caching, some SAS devices can take
15 - 20 minutes to get to the GRUB menu. This ihandle caching is not possible
without correctly having the canonical disk name.
When available, this driver also tries to use the deblocker #blocks and
a way of determining the disk size.
Finally and probably most importantly, this new driver is also capable of
seeing all partitions on a GPT disk. With the old driver, the GPT
partition table can not be read and only the first partition on the disk
can be seen.
Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
By default all file system modules are added to the GRUB coreboot
payload `default_payload.elf`. This makes the image quite big,
especially as often not all modules are needed.
Introduce the variable `FS_PAYLOAD_MODULES`, which can be used to
explicitly set file systems modules to be added.
$ make default_payload.elf
test -f default_payload.elf && rm default_payload.elf || true
pkgdatadir=. ./grub-mkstandalone --grub-mkimage=./grub-mkimage -O i386-coreboot -o default_payload.elf --modules='ahci pata ehci uhci ohci usb_keyboard usbms part_msdos ext2 fat at_keyboard part_gpt usbserial_usbdebug cbfs' --install-modules='ls linux search configfile normal cbtime cbls memrw iorw minicmd lsmmap lspci halt reboot hexdump pcidump regexp setpci lsacpi chain test serial multiboot cbmemc linux16 gzio echo help syslinuxcfg xnu affs afs bfs btrfs cbfs cpio cpio_be exfat ext2 f2fs fat hfs hfsplus iso9660 jfs minix minix2 minix2_be minix3 minix3_be minix_be newc nilfs2 ntfs odc procfs reiserfs romfs sfs squash4 tar udf ufs1 ufs1_be ufs2 xfs zfs password_pbkdf2 ' --fonts= --themes= --locales= -d grub-core/ /boot/grub/grub.cfg=./coreboot.cfg
$ ls -l default_payload.elf
-rw-rw---- 1 joey joey 1199568 Mar 6 13:58 default_payload.elf
$ make default_payload.elf FS_PAYLOAD_MODULES="" # ext2 already in `--modules`
test -f default_payload.elf && rm default_payload.elf || true
pkgdatadir=. ./grub-mkstandalone --grub-mkimage=./grub-mkimage -O i386-coreboot -o default_payload.elf --modules='ahci pata ehci uhci ohci usb_keyboard usbms part_msdos ext2 fat at_keyboard part_gpt usbserial_usbdebug cbfs' --install-modules='ls linux search configfile normal cbtime cbls memrw iorw minicmd lsmmap lspci halt reboot hexdump pcidump regexp setpci lsacpi chain test serial multiboot cbmemc linux16 gzio echo help syslinuxcfg xnu password_pbkdf2 ' --fonts= --themes= --locales= -d grub-core/ /boot/grub/grub.cfg=./coreboot.cfg
$ ls -l default_payload.elf
-rw-rw---- 1 joey joey 832976 Mar 7 12:13 default_payload.elf
So, the resulting payload size is around 370 kB smaller. (Adding it to
the CBFS, it will be compressed, so the effective size difference will
be smaller.)
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Upgrade Gnulib files to 20190105.
It's much easier to maintain GRUB's use of portability support files
from Gnulib when the process is automatic and driven by a single
configuration file, rather than by maintainers occasionally running
gnulib-tool and committing the result. Removing these
automatically-copied files from revision control also removes the
temptation to hack the output in ways that are difficult for future
maintainers to follow. Gnulib includes a "bootstrap" program which is
designed for this.
The canonical way to bootstrap GRUB from revision control is now
"./bootstrap", but "./autogen.sh" is still useful if you just want to
generate the GRUB-specific parts of the build system.
GRUB now requires Autoconf >= 2.63 and Automake >= 1.11, in line with
Gnulib.
Gnulib source code is now placed in grub-core/lib/gnulib/ (which should
not be edited directly), and GRUB's patches are in
grub-core/lib/gnulib-patches/. I've added a few notes to the developer
manual on how to maintain this.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
syslinux_parse simplifies some filenames by removing things like ".."
segments, but the tests assumed that @abs_top_srcdir@ would be
untouched, which is not true in the case of out-of-tree builds where
@abs_top_srcdir@ may contain ".." segments.
Performing the substitution requires some awkwardness in Makefile.am due
to details of how config.status works.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Many of GRUB's utilities don't check anywhere near all the possible
write errors. For example, if grub-install runs out of space when
copying a file, it won't notice. There were missing checks for the
return values of write, fflush, fsync, and close (or the equivalents on
other OSes), all of which must be checked.
I tried to be consistent with the existing logging practices of the
various hostdisk implementations, but they weren't entirely consistent
to start with so I used my judgement. The result at least looks
reasonable on GNU/Linux when I provoke a write error:
Installing for x86_64-efi platform.
grub-install: error: cannot copy `/usr/lib/grub/x86_64-efi-signed/grubx64.efi.signed' to `/boot/efi/EFI/debian/grubx64.efi': No space left on device.
There are more missing checks in other utilities, but this should fix
the most critical ones.
Fixes Debian bug #922741.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Steve McIntyre <93sam@debian.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For EBR partitions, "start" is the relative starting sector of the EBR
header itself, whereas "offset" is the relative starting byte of the
partition's contents, excluding the EBR header and any padding. Thus we
must use "offset", and divide by the sector size to convert to sectors.
Fixes Debian bug #923253.
Signed-off-by: James Clarke <jrtc27@jrtc27.com>
Reviewed-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Much like on x86, we can work out if the system is running on top of EFI
firmware. If so, return "arm-efi". If not, fall back to "arm-uboot" as
previously.
Split out the code to (maybe) load the efivar module and check for
/sys/firmware/efi into a common helper routine is_efi_system().
Signed-off-by: Steve McIntyre <93sam@debian.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This reverts commit 082fd84d52.
Incorrect version of the patch was pushed into the git repo.
Reported-by: Leif Lindholm <leif.lindholm@linaro.org>
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
There is a really convenient service for open source project from Travis
CI: They allow for free CI testing using their infrastructure.
GRUB has had issues with broken builds for various targets for a long time
already. The main reason is a lack of CI to just do smoke tests on whether
all targets still at least compile.
This patch adds a Travis config file which builds (almost) all currently
available targets.
On top of that, this Travis config also runs a small execution test on the
x86_64-efi target.
All of this config file can easily be extended further on. It probably
makes sense to do something similar to the u-boot test infrastructure
that communicates with the payload properly. Going forward, we also will
want to do more QEMU runtime checks for other targets.
Currently, with this config alone, I already see about half of the available
targets as broken. So it's definitely desperately needed :).
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Much like on x86, we can work out if the system is running on top
of EFI firmware. If so, return "arm-efi". If not, fall back to
"arm-uboot" as previously.
Heavily inspired by the existing code for x86.
Signed-off-by: Steve McIntyre <93sam@debian.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_efi_get_ram_base() looks for the lowest available RAM address by
traversing the memory map, comparing lowest address found so far.
Due to a brain glitch, that "so far" was initialized to GRUB_UINT_MAX -
completely preventing boot on systems without RAM below 4GB.
Change the initial value to GRUB_EFI_MAX_USABLE_ADDRESS, as originally
intended.
Reported-by: Steve McIntyre <93sam@debian.org>
Signed-off-by: Leif Lindholm <leif.lindholm@linaro.org>
Tested-by: Steve McIntyre <93sam@debian.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Some terminals, like `grub-core/term/at_keyboard.c`, return `-1` in case
they are not ready yet.
if (! KEYBOARD_ISREADY (grub_inb (KEYBOARD_REG_STATUS)))
return -1;
Currently, that is treated as a key press, and the menu time-out is
cancelled/cleared. This is unwanted, as the boot is stopped and the user
manually has to select a menu entry. Therefore, adapt the condition to
require the key value also to be greater than 0.
`GRUB_TERM_NO_KEY` is defined as 0, so the condition could be collapsed
to greater or equal than (≥) 0, but the compiler will probably do that
for us anyway, so keep the cases separate for clarity.
This is tested with coreboot, the GRUB default payload, and the
configuration file `grub.cfg` below.
For GRUB:
$ ./autogen.sh
$ ./configure --with-platform=coreboot
$ make -j`nproc`
$ make default_payload.elf
For coreboot:
$ more grub.cfg
serial --unit 0 --speed 115200
set timeout=5
menuentry 'halt' {
halt
}
$ build/cbfstool build/coreboot.rom add-payload \
-f /dev/shm/grub/default_payload.elf -n fallback/payload -c lzma
$ build/cbfstool build/coreboot.rom add -f grub.cfg -n etc/grub.cfg -t raw
$ qemu-system-x86_64 --version
QEMU emulator version 3.1.0 (Debian 1:3.1+dfsg-2+b1)
Copyright (c) 2003-2018 Fabrice Bellard and the QEMU Project developers
$ qemu-system-x86_64 -M pc -bios build/coreboot.rom -serial stdio -nic none
Currently, the time-out is cancelled/cleared. With the commit, it is not.
With a small GRUB payload, this the problem is also reproducible on the
ASRock E350M1.
Link: http://lists.gnu.org/archive/html/grub-devel/2019-01/msg00037.html
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We now have signature check logic in grub which allows us to treat
files differently depending on their file type.
Treat a loaded device tree like an overlayed ACPI table.
Both describe hardware, so I suppose their threat level is the same.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
This patch adds support for RISC-V to the grub build system. With this
patch, I can successfully build grub on RISC-V as a UEFI application.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Gcc may decide it wants to call helper functions to execute clz. Provide
them in our own copy of libgcc.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
To support a new architecture we need to provide a few helper functions
for memory, cache, timer, etc support.
This patch adds the remainders of those. Some bits are still disabled,
as I couldn't guarantee that we're always running on models / in modes
where the respective hardware is available.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds awareness of RISC-V relocations throughout the grub tools
as well as dynamic linkage and elf->PE relocation conversion support.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We currently only support to run grub on RISC-V as UEFI payload. Ideally,
we also only want to support running Linux underneath as UEFI payload.
Prepare that with some Linux boot stub code. Once the arm64 target is
generalized, we can hook into that one and gain boot functionality.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On entry, we need to save the system table pointer as well as our image
handle. Add an early startup file that saves them and then brings us
into our main function.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds a 32/64 capable setjmp implementation for RISC-V.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The RISC-V ABI document outlines ELF header structure and relocation
information. Pull the respective magic numbers into our elf header
so we can make use of them.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The PE format defines magic numbers as well as relocation identifiers for
RISC-V. Add them to our include file, so we can make use of them.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Some architectures want to boot Linux as plain UEFI binary. Today that
really only encompasses ARM and AArch64, but going forward more
architectures may adopt that model.
So rename our internal API accordingly.
Signed-off-by: Alexander Graf <agraf@suse.de>
Acked-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are a few spots in the PE generation code for EFI binaries that uses
the section alignment rather than file alignment, even though the alignment
is really only file bound.
Replace those cases with the file alignment constant instead.
Reported-by: Daniel Kiper <dkiper@net-space.pl>
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Julien ROBIN <julien.robin28@free.fr>
There is UEFI firmware popping up in the wild now that implements stricter
permission checks using NX and write protect page table entry bits.
This means that firmware now may fail to load binaries if its individual
sections are not page aligned, as otherwise it can not ensure permission
boundaries.
So let's bump all efi section alignments up to 4k (EFI page size). That way
we will stay compatible going forward.
Unfortunately our internals can't deal very well with a mismatch of alignment
between the virtual and file offsets, so we have to also pad our target
binary a bit.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Julien ROBIN <julien.robin28@free.fr>
The efi-arm case was defining its own header size calculation, even though it's
100% identical to the common EFI32_HEADER_SIZE definition.
So let's clean it up to use the common define.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Julien ROBIN <julien.robin28@free.fr>
This patch allows to have bigger kernels. If the kernel grows, then it will
overwrite the initrd when it is extracted.
Signed-off-by: Guillaume GARDET <guillaume.gardet@arm.com>
Acked-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A certain amount of dynamic space is required for the handover from
GRUB/Linux-EFI-stub. This entails things like initrd addresses,
address-cells entries and associated strings.
But move this into a proper centralised #define rather than live-code
it in the loader.
Signed-off-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
uboot_disk_write() is currently lacking the write support
to storage devices because, historically, those devices did not
implement block_write() in U-Boot.
The solution has been tested using a patched U-Boot loading
and booting GRUB in a QEMU vexpress-a9 environment.
The disk write operations were triggered with GRUB's save_env
command.
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The value of tpm_handle changes between successive calls to grub_tpm_handle_find(),
as instead of simply copying the stored pointer we end up taking the address of
said pointer when using the cached value of grub_tpm_handle.
This causes grub_efi_open_protocol() to return a nullptr in grub_tpm2_execute()
and grub_tpm2_log_event(). Said nullptr goes unchecked and
efi_call_5(tpm->hash_log_extend_event,...) ends up jumping to 0x0, Qemu crashes
once video ROM is reached at 0xb0000.
This patch seems to do the trick of fixing that bug, but we should also ensure
that all calls to grub_efi_open_protocol() are checked so that we don't start
executing low memory.
Signed-off-by: Max Tottenham <mtottenh@akamai.com>
Reviewed-by: Matthew Garrett <mjg59@google.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit b07feb8746 (verifiers: Rename
verify module to pgp module) renamed the "verify" module to "pgp", but
the GRUB_MOD_INIT and GRUB_MOD_FINI macros were left as "verify", which
broke the emu target build; and file_filter_test still referred to the
now non-existent "verify" module. Fix both of these.
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add support for multiple, shared, early initrd images. These early
images will be loaded in the order declared, and all will be loaded
before the initrd image.
While many classes of data can be provided by early images, the
immediate use case would be for distributions to provide CPU
microcode to mitigate the Meltdown and Spectre vulnerabilities.
Xen has also support to load microcode updates provided as additional
modules by the bootloader.
There are two environment variables provided for declaring the early
images.
* GRUB_EARLY_INITRD_LINUX_STOCK is for the distribution declare
images that are provided by the distribution or installed packages.
If undeclared, this will default to a set of common microcode image
names.
* GRUB_EARLY_INITRD_LINUX_CUSTOM is for user created images. User
images will be loaded after the stock images.
These separate configurations allow the distribution and user to
declare different image sets without clobbering each other.
This also makes a minor update to ensure that UUID partition labels
stay disabled when no initrd image is found, even if early images are
present.
This is basically a copy of a698240d "grub-mkconfig/10_linux: Support
multiple early initrd images" by Matthew S. Turnbull.
Signed-off-by: Peter Große <pegro@friiks.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We should not try to copy any memory area which is outside of the original
fdt. If this extra memory is controlled by a hypervisor this might end
with a crash.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
