We bumped into the build error while testing gcc-10 pre-release.
../../grub-core/disk/mdraid1x_linux.c: In function 'grub_mdraid_detect':
../../grub-core/disk/mdraid1x_linux.c:181:15: error: array subscript <unknown> is outside array bounds of 'grub_uint16_t[0]' {aka 'short unsigned int[0]'} [-Werror=array-bounds]
181 | (char *) &sb.dev_roles[grub_le_to_cpu32 (sb.dev_number)]
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../grub-core/disk/mdraid1x_linux.c:98:17: note: while referencing 'dev_roles'
98 | grub_uint16_t dev_roles[0]; /* Role in array, or 0xffff for a spare, or 0xfffe for faulty. */
| ^~~~~~~~~
../../grub-core/disk/mdraid1x_linux.c:127:33: note: defined here 'sb'
127 | struct grub_raid_super_1x sb;
| ^~
cc1: all warnings being treated as errors
Apparently gcc issues the warning when trying to access sb.dev_roles
array's member, since it is a zero length array as the last element of
struct grub_raid_super_1x that is allocated sparsely without extra
chunks for the trailing bits, so the warning looks legitimate in this
regard.
As the whole thing here is doing offset computation, it is undue to use
syntax that would imply array member access then take address from it
later. Instead we could accomplish the same thing through basic array
pointer arithmetic to pacify the warning.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
With recent versions of gcc on Ubuntu a very large lzma_decompress.img file is
output. (e.g. 134479600 bytes instead of 2864.) This causes grub-mkimage to
fail with: "error: Decompressor is too big."
This seems to be caused by a section .note.gnu.property that is placed at an
offset such that objcopy needs to pad the img file with zeros.
This issue is present on:
Ubuntu 19.10 with gcc (Ubuntu 8.3.0-26ubuntu1~19.10) 8.3.0
Ubuntu 19.10 with gcc (Ubuntu 9.2.1-9ubuntu2) 9.2.1 20191008
This issue is not present on:
Ubuntu 19.10 with gcc (Ubuntu 7.5.0-3ubuntu1~19.10) 7.5.0
RHEL 8.0 with gcc 8.3.1 20190507 (Red Hat 8.3.1-4)
The issue can be fixed by removing the section using objcopy as shown in
this patch.
Signed-off-by: Simon Hardy <simon.hardy@itdev.co.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The memory requested for the event is not released here,
causing memory leaks. This patch fixes this problem.
Signed-off-by: Jia Zhang <zhang.jia@linux.alibaba.com>
Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add notes on LVM cache booting to the GRUB manual to help user understanding
the outstanding issue and status.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The LVM cache logical volume is the logical volume consisting of the original
and the cache pool logical volume. The original is usually on a larger and
slower storage device while the cache pool is on a smaller and faster one. The
performance of the original volume can be improved by storing the frequently
used data on the cache pool to utilize the greater performance of faster
device.
The default cache mode "writethrough" ensures that any data written will be
stored both in the cache and on the origin LV, therefore grub can be straight
to read the original lv as no data loss is guarenteed.
The second cache mode is "writeback", which delays writing from the cache pool
back to the origin LV to have increased performance. The drawback is potential
data loss if losing the associated cache device.
During the boot time grub reads the LVM offline i.e. LVM volumes are not
activated and mounted, hence it should be fine to read directly from original
lv since all cached data should have been flushed back in the process of taking
it offline.
It is also not much helpful to the situation by adding fsync calls to the
install code. The fsync did not force to write back dirty cache to the original
device and rather it would update associated cache metadata to complete the
write transaction with the cache device. IOW the writes to cached blocks still
go only to the cache device.
To write back dirty cache, as LVM cache did not support dirty cache flush per
block range, there'no way to do it for file. On the other hand the "cleaner"
policy is implemented and can be used to write back "all" dirty blocks in a
cache, which effectively drain all dirty cache gradually to attain and last in
the "clean" state, which can be useful for shrinking or decommissioning a
cache. The result and effect is not what we are looking for here.
In conclusion, as it seems no way to enforce file writes to the original
device, grub may suffer from power failure as it cannot assemble the cache
device and read the dirty data from it. However since the case is only
applicable to writeback mode which is sensitive to data lost in nature, I'd
still like to propose my (relatively simple) patch and treat reading dirty
cache as improvement.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When building GRUB with memory management debugging enabled, then the
build fails because of `grub_debug_malloc()` and `grub_debug_free()`
being undefined in the luks2 module. The cause is that we patch
"base64.h" to unconditionaly include "config-util.h", which shouldn't be
included for modules at all. As a result, `MM_DEBUG` is defined when
building the module, causing it to use the debug memory allocation
functions. As these are not built into modules, we end up with a linker
error.
Fix the issue by removing the <config-util.h> include altogether. The
sole reason it was included was for the `_GL_ATTRIBUTE_CONST` macro,
which we can simply define as empty in case it's not set.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The memory management system supports a debug mode that can be enabled
at build time by passing "--enable-mm-debug" to the configure script.
Passing the option will cause us define MM_DEBUG as expected, but in
fact the reverse option "--disable-mm-debug" will do the exact same
thing and also set up the define. This currently causes the build of
"lib/gnulib/base64.c" to fail as it tries to use `grub_debug_malloc()`
and `grub_debug_free()` even though both symbols aren't defined.
Seemingly, `AC_ARG_ENABLE()` will always execute the third argument if
either the positive or negative option was passed. Let's thus fix the
issue by moving the call to`AC_DEFINE()` into an explicit `if test
$xenable_mm_debug` block, similar to how other defines work.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
This allows comparing file ages on EFI system partitions.
Signed-off-by: David Michael <fedora.dm0@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This provides the node's attributes outside the iterator function
so the file modification time can be accessed and reported.
Signed-off-by: David Michael <fedora.dm0@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fixes a build failure:
grub-core/commands/date.c:49: undefined reference to `grub_get_weekday_name'
grub-core/commands/ls.c:155: undefined reference to `grub_unixtime2datetime'
Bug: https://bugs.gentoo.org/711512
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Tested-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While GRUB has no platform support for SuperH (sh4) yet, this change
adds the target-specific handling of soft-floats such that the GRUB
utilities can be built on this target.
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently, in some builds with some checkers, we see:
1. grub-core/disk/efi/efidisk.c:601: error[shiftTooManyBitsSigned]: Shifting signed 64-bit value by 63 bits is undefined behaviour
This is because grub_efi_status_t is defined as grub_efi_intn_t, which is
signed, and shifting into the sign bit is not defined behavior. UEFI fixed
this in the spec in 2.3:
2.3 | Change the defined type of EFI_STATUS from INTN to UINTN | May 7, 2009
And the current EDK2 code has:
MdePkg/Include/Base.h-//
MdePkg/Include/Base.h-// Status codes common to all execution phases
MdePkg/Include/Base.h-//
MdePkg/Include/Base.h:typedef UINTN RETURN_STATUS;
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-/**
MdePkg/Include/Base.h- Produces a RETURN_STATUS code with the highest bit set.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @param StatusCode The status code value to convert into a warning code.
MdePkg/Include/Base.h- StatusCode must be in the range 0x00000000..0x7FFFFFFF.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @return The value specified by StatusCode with the highest bit set.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-**/
MdePkg/Include/Base.h-#define ENCODE_ERROR(StatusCode) ((RETURN_STATUS)(MAX_BIT | (StatusCode)))
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-/**
MdePkg/Include/Base.h- Produces a RETURN_STATUS code with the highest bit clear.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @param StatusCode The status code value to convert into a warning code.
MdePkg/Include/Base.h- StatusCode must be in the range 0x00000000..0x7FFFFFFF.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @return The value specified by StatusCode with the highest bit clear.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-**/
MdePkg/Include/Base.h-#define ENCODE_WARNING(StatusCode) ((RETURN_STATUS)(StatusCode))
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-/**
MdePkg/Include/Base.h- Returns TRUE if a specified RETURN_STATUS code is an error code.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- This function returns TRUE if StatusCode has the high bit set. Otherwise, FALSE is returned.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @param StatusCode The status code value to evaluate.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h- @retval TRUE The high bit of StatusCode is set.
MdePkg/Include/Base.h- @retval FALSE The high bit of StatusCode is clear.
MdePkg/Include/Base.h-
MdePkg/Include/Base.h-**/
MdePkg/Include/Base.h-#define RETURN_ERROR(StatusCode) (((INTN)(RETURN_STATUS)(StatusCode)) < 0)
...
Uefi/UefiBaseType.h:typedef RETURN_STATUS EFI_STATUS;
This patch makes grub's implementation match the Edk2 declaration with regards
to the signedness of the type.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add debug information to EFI GOP video driver probing function.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
All other video drivers use "video" as the debug condition instead of "fb"
so change this in the efi/uga driver to make it consistent with the others.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
No messages were printed in this function, add some to ease debugging.
Also, the function returns a void * pointer so return NULL instead of
0 to make the code more readable.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We get 64 bit from PCI BAR but then truncate by assigning to 32 bit.
Make sure to check that pointer does not overflow on 32 bit platform.
Closes: 50931
Signed-off-by: Andrei Borzenkov <arvidjaar@gmail.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
EFI GOP has support for multiple different bitness types of frame buffers
and for a special "BLT only" type which is always defined to be RGBx.
Because grub2 doesn't ever directly access the frame buffer but instead
only renders graphics via the BLT interface anyway, we can easily support
these adapters.
The reason this has come up now is the emerging support for virtio-gpu
in OVMF. That adapter does not have the notion of a memory mapped frame
buffer and thus is BLT only.
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity Scan reports that the grub_strrchr() function can return NULL if
the character is not found. Check if that's the case for dirfile pointer.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a grub_debug_enabled() helper function instead of open coding it.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The function that searches the mods section base address does not have
any debug information. Add some debugging outputs that could be useful.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Micron PCIe SSDs Linux driver (mtip32xx) exposes block devices
as /dev/rssd[a-z]+[0-9]*. Add support for these rssd device names.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Linux creates modalias strings by filtering out non-ASCII, space,
and colon characters. Provide an option that does the same filtering
so people can create a modalias string in GRUB, and then match their
modalias patterns against it.
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When running make dist, I hit this error:
rm -f en@arabic.gmo && /usr/bin/gmsgfmt -c --statistics --verbose -o en@arabic.gmoen@arabic.po
en@arabic.po:5312: 'msgstr' is not a valid C format string, unlike 'msgid'.
Reason: The character that terminates the directive number 3 is not a valid conversion specifier.
/usr/bin/gmsgfmt: found 1 fatal error
This was caused by "%m" being replaced with foreign Unicode characters.
For example:
msgid "cannot rename the file %s to %s: %m"
msgstr "ﺹﺎﻨﻧﻮﺗ ﺮﻌﻧﺎﻤﻋ ﺖﻬﻋ ﻒִﻴﻠﻋ %s ﺕﻭ %s: %ﻡ"
Mimic the workaround used for "%s" by reversing the replacement of "%m" at
the end of the sed programs.
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These were inadvertently lost during the conversion to Gnulib (gnulib:
Upgrade Gnulib and switch to bootstrap tool; commit 35b909062). The
files in po/gettext-patches/ can be imported using "git am" on top of
the gettext tag corresponding to AM_GNU_GETTEXT_VERSION in configure.ac
(currently 0.18.3). They handle translation of messages in shell files,
make msgfmt output in little-endian format, and arrange to use @SHELL@
rather than /bin/sh.
There were some changes solely for the purpose of distributing extra
files; for ease of maintenance, I've added these to
conf/Makefile.extra-dist instead.
Fixes: https://savannah.gnu.org/bugs/?57298
Signed-off-by: Colin Watson <cjwatson@ubuntu.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently the string functions grub_strtol(), grub_strtoul(), and
grub_strtoull() don't declare the "end" pointer in such a way as to
require the pointer itself or the character array to be immutable to the
implementation, nor does the C standard do so in its similar functions,
though it does require us not to change any of it.
The typical declarations of these functions follow this pattern:
long
strtol(const char * restrict nptr, char ** restrict endptr, int base);
Much of the reason for this is historic, and a discussion of that
follows below, after the explanation of this change. (GRUB currently
does not include the "restrict" qualifiers, and we name the arguments a
bit differently.)
The implementation is semantically required to treat the character array
as immutable, but such accidental modifications aren't stopped by the
compiler, and the semantics for both the callers and the implementation
of these functions are sometimes also helped by adding that requirement.
This patch changes these declarations to follow this pattern instead:
long
strtol(const char * restrict nptr,
const char ** const restrict endptr,
int base);
This means that if any modification to these functions accidentally
introduces either an errant modification to the underlying character
array, or an accidental assignment to endptr rather than *endptr, the
compiler should generate an error. (The two uses of "restrict" in this
case basically mean strtol() isn't allowed to modify the character array
by going through *endptr, and endptr isn't allowed to point inside the
array.)
It also means the typical use case changes to:
char *s = ...;
const char *end;
long l;
l = strtol(s, &end, 10);
Or even:
const char *p = str;
while (p && *p) {
long l = strtol(p, &p, 10);
...
}
This fixes 26 places where we discard our attempts at treating the data
safely by doing:
const char *p = str;
long l;
l = strtol(p, (char **)&ptr, 10);
It also adds 5 places where we do:
char *p = str;
while (p && *p) {
long l = strtol(p, (const char ** const)&p, 10);
...
/* more calls that need p not to be pointer-to-const */
}
While moderately distasteful, this is a better problem to have.
With one minor exception, I have tested that all of this compiles
without relevant warnings or errors, and that /much/ of it behaves
correctly, with gcc 9 using 'gcc -W -Wall -Wextra'. The one exception
is the changes in grub-core/osdep/aros/hostdisk.c , which I have no idea
how to build.
Because the C standard defined type-qualifiers in a way that can be
confusing, in the past there's been a slow but fairly regular stream of
churn within our patches, which add and remove the const qualifier in many
of the users of these functions. This change should help avoid that in
the future, and in order to help ensure this, I've added an explanation
in misc.h so that when someone does get a compiler warning about a type
error, they have the fix at hand.
The reason we don't have "const" in these calls in the standard is
purely anachronistic: C78 (de facto) did not have type qualifiers in the
syntax, and the "const" type qualifier was added for C89 (I think; it
may have been later). strtol() appears to date from 4.3BSD in 1986,
which means it could not be added to those functions in the standard
without breaking compatibility, which is usually avoided.
The syntax chosen for type qualifiers is what has led to the churn
regarding usage of const, and is especially confusing on string
functions due to the lack of a string type. Quoting from C99, the
syntax is:
declarator:
pointer[opt] direct-declarator
direct-declarator:
identifier
( declarator )
direct-declarator [ type-qualifier-list[opt] assignment-expression[opt] ]
...
direct-declarator [ type-qualifier-list[opt] * ]
...
pointer:
* type-qualifier-list[opt]
* type-qualifier-list[opt] pointer
type-qualifier-list:
type-qualifier
type-qualifier-list type-qualifier
...
type-qualifier:
const
restrict
volatile
So the examples go like:
const char foo; // immutable object
const char *foo; // mutable pointer to object
char * const foo; // immutable pointer to mutable object
const char * const foo; // immutable pointer to immutable object
const char const * const foo; // XXX extra const keyword in the middle
const char * const * const foo; // immutable pointer to immutable
// pointer to immutable object
const char ** const foo; // immutable pointer to mutable pointer
// to immutable object
Making const left-associative for * and right-associative for everything
else may not have been the best choice ever, but here we are, and the
inevitable result is people using trying to use const (as they should!),
putting it at the wrong place, fighting with the compiler for a bit, and
then either removing it or typecasting something in a bad way. I won't
go into describing restrict, but its syntax has exactly the same issue
as with const.
Anyway, the last example above actually represents the *behavior* that's
required of strtol()-like functions, so that's our choice for the "end"
pointer.
Signed-off-by: Peter Jones <pjones@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
PIE should be disabled in assembly sources as well, or else GRUB will
fail to boot.
Bug: https://bugs.gentoo.org/667852
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
Signed-off-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
On a 32-bit SPARC userland, configure fails to compile assembly and the
build fails:
checking for options to compile assembly... configure: error: could not compile assembly
config.log shows:
asm-tests/sparc64.S: Assembler messages:
asm-tests/sparc64.S:5: Error: Architecture mismatch on "lduw [%o4+4],%o4".
asm-tests/sparc64.S:5: (Requires v9|v9a|v9b|v9c|v9d|v9e|v9v|v9m|m8; requested architecture is sparclite.)
asm-tests/sparc64.S:7: Error: Architecture mismatch on "stw %o5,[%o3]".
asm-tests/sparc64.S:7: (Requires v9|v9a|v9b|v9c|v9d|v9e|v9v|v9m|m8; requested architecture is sparclite.)
asm-tests/sparc64.S:8: Error: Architecture mismatch on "bne,pt %icc,1b ,pt %icc,1b".
asm-tests/sparc64.S:8: (Requires v9|v9a|v9b|v9c|v9d|v9e|v9v|v9m|m8; requested architecture is sparclite.)
Simply moving these blocks earlier in configure.ac is sufficient to
ensure that the tests are executed with the appropriate flags
(specifically -m64 in this case).
Bug: https://bugs.gentoo.org/667850
Signed-off-by: Mike Gilbert <floppym@gentoo.org>
Signed-off-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
The debug message printed when decryption with a keyslot fails is
missing its trailing newline. Add it to avoid mangling it with
subsequent output.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The necessary check for NULL before use of function ver->close is not
taking place in the failure path. This patch simply adds the missing
check and fixes the problem that GRUB hangs indefinitely after booting
rogue image without valid signature if secure boot is turned on.
Now it displays like this for booting rogue UEFI image:
error: bad shim signature
error: you need to load the kernel first
Press any key to continue...
and then you can go back to boot menu by pressing any key or after a few
seconds expired.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub-editenv create command will wrongly overwrite /boot/grub2/grubenv
with a regular file if grubenv is a symbolic link. But instead, it should
create a new file in the path the symlink points to.
This lets /boot/grub2/grubenv be a symlink to /boot/efi/EFI/fedora/grubenv
even when they're different mount points, which allows grub2-editenv to be
the same across platforms (i.e. UEFI vs BIOS).
For example, in Fedora the GRUB EFI builds have prefix set to /EFI/fedora
(on the EFI System Partition), but for BIOS machine it'll be /boot/grub2
(which may or may not be its own mountpoint).
With this patch, on EFI machines we can make /boot/grub2/grubenv a symlink
to /boot/efi/EFI/fedora/grubenv, and the same copy of grub-set-default will
work on both kinds of systems.
Windows doesn't implement a readlink primitive, so the current behaviour is
maintained for this operating system.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jonathan Lebon <jlebon@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently grub-editenv and related tools are not able to follow symbolic
links when finding their config file. For example the grub-editenv create
command will wrongly overwrite a symlink in /boot/grub2/grubenv with a new
regular file, instead of creating a file in the path the symlink points to.
A following patch will change that and add support in grub-editenv to
follow symbolic links when finding the grub environment variables file.
Add a grub_util_readlink() helper function that is just a wrapper around
the platform specific function to read the value of a symbolic link. This
helper function will be used by the following patch for grub-editenv.
The helper function is not added for Windows, since this operating system
doesn't have a primitive to read the contents of a symbolic link.
Signed-off-by: Peter Jones <pjones@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add documentation to the GRUB manual that specifies the order netboot
clients use to select a GRUB configuration file.
Also explain that the feature is enabled by default but can be disabled
by setting the "feature_net_search_cfg" environment variable to "n" in
an embedded configuration file.
Signed-off-by: Robert Marshall <rmarshall@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch implements a search for a specific configuration when the config
file is on a remoteserver. It uses the following order:
1) DHCP client UUID option.
2) MAC address (in lower case hexadecimal with dash separators);
3) IP (in upper case hexadecimal) or IPv6;
4) The original grub.cfg file.
This procedure is similar to what is used by pxelinux and yaboot:
http://www.syslinux.org/wiki/index.php/PXELINUX#config
It is enabled by default but can be disabled by setting the environment
variable "feature_net_search_cfg" to "n" in an embedded configuration.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=873406
Signed-off-by: Paulo Flabiano Smorigo <pfsmorigo@br.ibm.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch sets a net_<interface>_clientid and net_<interface>_clientuuid
GRUB environment variables, using the DHCP client ID and UUID options if
these are found.
In the same way than net_<interface>_<option> variables are set for other
options such domain name, boot file, next server, etc.
Signed-off-by: Paulo Flabiano Smorigo <pfsmorigo@br.ibm.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The DHCP Options and BOOTP Vendor Extensions enum values are a mixture of
decimal and hexadecimal numbers. Change this to consistently use decimal
numbers for all since that is how these values are defined by RFC 2132.
Suggested-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The printf(3) function has support for the %X format specifier, to output
an unsigned hexadecimal integer in uppercase.
This can be achived in GRUB using the %x format specifier in grub_printf()
and calling grub_toupper(), but it is more convenient if there is support
for %X in grub_printf().
Signed-off-by: Paulo Flabiano Smorigo <pfsmorigo@br.ibm.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The common datetime helper functions are currently included in the normal
module, but this makes any other module that calls these functions to have
a dependency with the normal module only for this reason.
Since the normal module does a lot of stuff, it calls functions from other
modules. But since other modules may depend on it for calling the datetime
helpers, this could lead to circular dependencies between modules.
As an example, when platform == xen the grub_get_datetime() function from
the datetime module calls to the grub_unixtime2datetime() helper function
from the normal module. Which leads to the following module dependency:
datetime -> normal
and send_dhcp_packet() from the net module calls the grub_get_datetime()
function, which leads to the following module dependency:
net -> datetime -> normal
but that means that the normal module is not allowed to depend on net or
any other module that depends on it due the transitive dependency caused
by datetime. A recent patch attempted to add support to fetch the config
file over the network, which leads to the following circular dependency:
normal -> net -> datetime -> normal
So having the datetime helpers in the normal module makes it quite fragile
and easy to add circular dependencies like these, that break the build due
the genmoddep.awk script catching the issues.
Fix this by taking the datetime helper functions out of the normal module
and instead add them to the datetime module itself. Besides fixing these
issues, it makes more sense to have these helper functions there anyways.
Reported-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch updates the miniLZO library to a newer version, which among other
things fixes "CVE-2014-4607 - lzo: lzo1x_decompress_safe() integer overflow"
that is present in the current used in GRUB.
It also updates the "GRUB Developers Manual", to mention that the library is
used and describes the process to update it to a newer release when needed.
Resolves: http://savannah.gnu.org/bugs/?42635
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
gcc says:
grub-core/fs/squash4.c: In function ‘direct_read’:
grub-core/fs/squash4.c:868:10: error: ‘err’ may be used uninitialized in
this function [-Werror=maybe-uninitialized]
868 | if (err)
| ^
cc1: all warnings being treated as errors
This patch initializes it to GRUB_ERR_NONE.
Signed-off-by: Peter Jones <pjones@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While testing the 86-DOS lDebug [1] booting from GRUB2, newer versions of the
debugger would fail to load when booted using GRUB's freedos command. The
behaviour observed in a qemu i386 machine was that the ROM-BIOS's boot load
would start anew, instead of loading the selected debugger as kernel.
It came to light that there was a size limit: Kernel files that were 58880
bytes (E600h) long or shorter succeeded to boot, while files that were 64000
bytes or longer failed in the manner described.
Eventually it turned out that the relocator16 stub succeeded whenever it was
placed completely within the first 64 KiB of the Low Memory Area. The chunk
for the relocator is allocated with a minimum address of 0x8010 and a maximum
address just below 0xA0000 [2]. That means if the kernel is, for instance,
E600h bytes long, then the kernel will be allocated memory starting at 00600h
(the fixed FreeDOS kernel load address) up to E600h + 00600h = 0EC00h, which
leaves 1400h (5120) bytes for the relocator to stay in the first 64 KiB.
If the kernel is 64000 bytes (FA00h) long, then the relocator must go to
FA00h + 00600h = 10000h at least which is outside the first 64 KiB.
The problem is that the relocator16 initialises the DS register with a
"pseudo real mode" descriptor, which is defined with a segment limit of
64 KiB and a segment base of zero. After that, the relocator addressed
parts of itself (implicitly) using the DS register, with an offset from
ESI, which holds the linear address of the relocator's base [3]. With the
larger kernel files this would lead to accessing data beyond the 64 KiB
segment limit, presumably leading to a fault and perhaps a subsequent
triple-fault or such.
This patch fixes the relocator to set the segment base of the descriptors
to the base address of the relocator; then, the subsequent accesses to
the relocator's variables are done without the ESI register as an index.
This does not interfere with the relocator's or its target's normal
operation; the segment limits are still loaded with 64 KiB and all the
segment bases are subsequently reset by the relocator anyway.
Current versions of the debugger to test are uploaded to [4]. The file
ldebugnh.com (LZ4-compressed and built with -D_EXTHELP=0) at 58368 bytes
loads successfully, whereas ldebug.com at 64000 bytes fails. Loading one
of these files requires setting root to a FAT FS partition and using the
freedos command to specify the file as kernel:
set root='(hd0,msdos1)'
freedos /ldebug.com
boot
Booting the file using the multiboot command (which uses a WIP entrypoint
of the debugger) works, as it does not use GRUB's relocator16 but instead
includes a loader in the kernel itself, which drops it back to 86 Mode.
[1]: https://hg.ulukai.org/ecm/ldebug
[2]: http://git.savannah.gnu.org/cgit/grub.git/tree/grub-core/lib/i386/relocator.c?id=495781f5ed1b48bf27f16c53940d6700c181c74c#n127
[3]: http://git.savannah.gnu.org/cgit/grub.git/tree/grub-core/lib/i386/relocator16.S?id=495781f5ed1b48bf27f16c53940d6700c181c74c#n97
[4]: https://ulukai.org/ecm/lDebug-5479a7988d21-nohelp.zip
Signed-off-by: C. Masloch <pushbx@ulukai.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
With cryptsetup 2.0, a new version of LUKS was introduced that breaks
compatibility with the previous version due to various reasons. GRUB
currently lacks any support for LUKS2, making it impossible to decrypt
disks encrypted with that version. This commit implements support for
this new format.
Note that LUKS1 and LUKS2 are quite different data formats. While they
do share the same disk signature in the first few bytes, representation
of encryption parameters is completely different between both versions.
While the former version one relied on a single binary header, only,
LUKS2 uses the binary header only in order to locate the actual metadata
which is encoded in JSON. Furthermore, the new data format is a lot more
complex to allow for more flexible setups, like e.g. having multiple
encrypted segments and other features that weren't previously possible.
Because of this, it was decided that it doesn't make sense to keep both
LUKS1 and LUKS2 support in the same module and instead to implement it
in two different modules luks and luks2.
The proposed support for LUKS2 is able to make use of the metadata to
decrypt such disks. Note though that in the current version, only the
PBKDF2 key derival function is supported. This can mostly attributed to
the fact that the libgcrypt library currently has no support for either
Argon2i or Argon2id, which are the remaining KDFs supported by LUKS2. It
wouldn't have been much of a problem to bundle those algorithms with
GRUB itself, but it was decided against that in order to keep down the
number of patches required for initial LUKS2 support. Adding it in the
future would be trivial, given that the code structure is already in
place.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The luks module contains quite a lot of logic to parse cipher and
cipher-mode strings like aes-xts-plain64 into constants to apply them
to the grub_cryptodisk_t structure. This code will be required by the
upcoming luks2 module, as well, which is why this commit moves it into
its own function grub_cryptodisk_setcipher in the cryptodisk module.
While the strings are probably rather specific to the LUKS modules, it
certainly does make sense that the cryptodisk module houses code to set
up its own internal ciphers instead of hosting that code in the luks
module.
Except for necessary adjustments around error handling, this commit does
an exact move of the cipher configuration logic from luks.c to
cryptodisk.c. Any behavior changes are unintentional.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While the AFSplitter code is currently used only by the luks module,
upcoming support for luks2 will add a second module that depends on it.
To avoid any linker errors when adding the code to both modules because
of duplicated symbols, this commit moves it into its own standalone
module afsplitter as a preparatory step.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The upcoming support for LUKS2 disc encryption requires us to include a
parser for base64-encoded data, as it is used to represent salts and
digests. As gnulib already has code to decode such data, we can just
add it to the boostrapping configuration in order to make it available
in GRUB.
The gnulib module makes use of booleans via the <stdbool.h> header. As
GRUB does not provide any POSIX wrapper header for this, but instead
implements support for bool in <sys/types.h>, we need to patch
base64.h to not use <stdbool.h> anymore. We unfortunately cannot include
<sys/types.h> instead, as it would then use gnulib's internal header
while compiling the gnulib object but our own <sys/types.h> when
including it in a GRUB module. Because of this, the patch replaces the
include with a direct typedef.
A second fix is required to make available _GL_ATTRIBUTE_CONST, which
is provided by the configure script. As base64.h does not include
<config.h>, it is thus not available and results in a compile error.
This is fixed by adding an include of <config-util.h>.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While the newly added jsmn library provides the parsing interface, it
does not provide any kind of interface to act on parsed tokens. Instead,
the caller is expected to handle pointer arithmetics inside of the token
array in order to extract required information. While simple, this
requires users to know some of the inner workings of the library and is
thus quite an unintuitive interface.
This commit adds a new interface on top of the jsmn parser that provides
convenience functions to retrieve values from the parsed json type, grub_json_t.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The upcoming support for LUKS2 encryption will require a JSON parser to
decode all parameters required for decryption of a drive. As there is
currently no other tool that requires JSON, and as gnulib does not
provide a parser, we need to introduce a new one into the code base. The
backend for the JSON implementation is going to be the jsmn library [1].
It has several benefits that make it a very good fit for inclusion in
GRUB:
- It is licensed under MIT.
- It is written in C89.
- It has no dependencies, not even libc.
- It is small with only about 500 lines of code.
- It doesn't do any dynamic memory allocation.
- It is testen on x86, amd64, ARM and AVR.
The library itself comes as a single header, only, that contains both
declarations and definitions. The exposed interface is kind of
simplistic, though, and does not provide any convenience features
whatsoever. Thus there will be a separate interface provided by GRUB
around this parser that is going to be implemented in the following
commit. This change only imports jsmn.h from tag v1.1.0 and adds it
unmodified to a new json module with the following command:
curl -L https://raw.githubusercontent.com/zserge/jsmn/v1.1.0/jsmn.h \
-o grub-core/lib/json/jsmn.h
Upstream jsmn commit hash: fdcef3ebf886fa210d14956d3c068a653e76a24e
Upstream jsmn commit name: Modernize (#149), 2019-04-20
[1]: https://github.com/zserge/jsmn
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In some cases GRUB2 allocates multiboot2 structure at 0 address, that is
a confusing behavior. Consumers of that structure can have internal NULL-checks
that will throw an error when get a pointer to data allocated at address 0.
To prevent that, define min address for mbi allocation on x86 and x86_64
platforms.
Signed-off-by: Lukasz Hawrylko <lukasz.hawrylko@linux.intel.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Recent work around x86 Linux kernel loader revealed an underflow in the
setup_header length calculation and another related issue. Both lead to
the memory overwrite and later machine crash.
Currently when the GRUB copies the setup_header into the linux_params
(struct boot_params, traditionally known as "zero page") it assumes the
setup_header size as sizeof(linux_i386_kernel_header/lh). This is
incorrect. It should use the value calculated accordingly to the Linux
kernel boot protocol. Otherwise in case of pretty old kernel, to be
exact Linux kernel boot protocol, the GRUB may write more into
linux_params than it was expected to. Fortunately this is not very big
issue. Though it has to be fixed. However, there is also an underflow
which is grave. It happens when
sizeof(linux_i386_kernel_header/lh) > "real size of the setup_header".
Then len value wraps around and grub_file_read() reads whole kernel into
the linux_params overwriting memory past it. This leads to the GRUB
memory allocator breakage and finally to its crash during boot.
The patch fixes both issues. Additionally, it moves the code not related to
grub_memset(linux_params)/grub_memcpy(linux_params)/grub_file_read(linux_params)
section outside of it to not confuse the reader.
Fixes: e683cfb0cf (loader/i386/linux: Calculate the setup_header length)
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Krystian Hebel <krystian.hebel@3mdeb.com>
