This commit fixes 2 issues with the TPM support code:
1) Remove "REQUIRE_TPM ?=" line from the Makefile, further down the Makefile
checks if REQUIRE_TPM is undefined, but the above line sets it to an empty
string, which is not the same as undefined. Without this handle_image fails
after the tpm_log_pe() call even if REQUIRE_TPM=1 once was not set when
building the shim
2) When secure-boot is disabled then shim_verify() would exit with the
status of tpm_log_pe(), which on systems with a TPM is an error. Combined
with the recent change to always install the shim protocols, this causes
grub to refuse to boot any kernel since the verify() call now always fails.
This commit fixes this by explicitly setting status = EFI_SUCCESS when
secure-boot is disabled.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Currently the shim_lock protocol is only installed when SecureBoot is enabled.
However, having Verify just measure into the TPM without SecureBoot is a useful
feature.
Signed-off-by: Tamas K Lengyel <lengyelt@ainfosec.com>
Currently the only measurement the shim logs in the TPM is that of the EFI
application it directly loads. However, there are no measurements being taken
of application that are being verified through the shim_lock protocol. In this
patch we extend PCR4 for any binary for which Verify is being called through
the shim_lock protocol.
Signed-off-by: Tamas K Lengyel <lengyelt@ainfosec.com>
Originally, we check if the last 2 nodes in the device path are
IPv4()/Uri() or IPv6()/Uri() to determine whether httpboot is used or
not. However, since UEFI 2.7, the DNS node will be inserted between the
IP node and the URI node if the server provides the DNS server address.
This commit changes the matching rule to search IP node and URI node
and ignore any node between those two nodes.
Signed-off-by: Gary Lin <glin@suse.com>
Make sure if we chainload things, a chainloaded bootloader will be able to use
the latest systab replacements and protocols. They need to match for things
to validate correctly.
Signed-off-by: Mathieu Trudel-Lapierre <mathieu.trudel-lapierre@canonical.com>
shim 13:
- OpenSSL reverted to 1.0.2k to make the cert chaining of existing deployments stay working
- Better PCR usage for TPM
- TPM documentation in README.tpm
- More configurable build via make variables:
ENABLE_SHIM_CERT
ENABLE_SHIM_HASH
ENABLE_SBSIGN
LIBDIR
EFIDIR
VENDOR_CERT_FILE
VENDOR_DB_FILE
- Better MoK documentation in MokVars.txt
- Better debuginfo generation
- Lots of minor bug fixes.
Signed-off-by: Peter Jones <pjones@redhat.com>
If you build with ENABLE_SHIM_CERT=1, the include chain right now winds
up meaning shim_cert is defined in a header that gets included in
netboot.c as well, which never uses it:
In file included from shim.h:125:0,
from netboot.c:36:
shim_cert.h:1:14: error: ‘shim_cert’ defined but not used [-Werror=unused-variable]
static UINT8 shim_cert[] = {
^~~~~~~~~
cc1: all warnings being treated as errors
So make that okay by adding __attribute__((__unused__)) to the variable
decl.
Signed-off-by: Peter Jones <pjones@redhat.com>
Cyphermox discovered that when you run this:
( printf "\xff\x00\xfe\x00" ; echo "shimx64.efi,foo,,This is the boot entry for foo" ) | sed -z 's/./&\x00/g'
on some debian machines, printf(1) doesn't interpret the \x.. characters,
and that results in this being the encoded text:
00000000 5c 78 66 66 5c 78 66 65 73 00 68 00 69 00 6d 00 |\xff\xfes.h.i.m.|
00000010 78 00 36 00 34 00 2e 00 65 00 66 00 69 00 2c 00 |x.6.4...e.f.i.,.|
00000020 66 00 6f 00 6f 00 2c 00 2c 00 54 00 68 00 69 00 |f.o.o.,.,.T.h.i.|
which... yeah, that's wrong. So instead, use iconv instead of
printf+sed to encode it in UCS-2. Unfortunately, that means we don't
get endian markers, because for some reason iconv(1) doesn't have any way
to say it should include them. But that's okay; fallback already
handles not having them and just assumes the second byte being \x00
means UCS-2LE.
Signed-off-by: Peter Jones <pjones@redhat.com>
Commit 1e71734992 inadvertantly switched ARM's
LDFLAGS+=--defsym=EFI_SUBSYSTEM=$(SUBSYSTEM) to be before LDFLAGS is set,
and so it got clobbered away.
Signed-off-by: Peter Jones <pjones@redhat.com>
shim_cert.h is required by other pieces (such as netboot.o, cert.o) and
might not be built by the time these targets are reached. In that case the
build would fail as it can't find a required header.
Signed-off-by: Mathieu Trudel-Lapierre <mathieu.trudel-lapierre@canonical.com>
I'm still having some trouble with the offsetof() definition, so just
nerf it to what stddef.h would say anyway.
Signed-off-by: Peter Jones <pjones@redhat.com>
Disable DES completely since it's already old and insecure.
This makes MokManager not support the DES based password hash but
probably no one is using it.
Signed-off-by: Gary Lin <glin@suse.com>
- Declare some functions in the proper headers
+ We missed them for a long time...
- Cast offsetof to UINTN
+ The original casting triggers the gcc warning since int can not
present the offset for the 64bit machines.
- Cast the "char" array to "CHAR8 *" to avoid the gcc warnings
- Implement atoi correctly
Signed-off-by: Gary Lin <glin@suse.com>
strcmp() and strcasecmp() are widely used in openssl. Implement those
two functions to eliminate the gcc warnings and the potential crash.
Signed-off-by: Gary Lin <glin@suse.com>
The changes in the openssl headers cause the inclusion of
CrtLibSupport.h eariler than the inclusion of stddef.h, so "offsetof"
was defined twice and this caused the followling build error:
In file included from Cryptlib/Include/openssl/buffer.h:23:0,
from Cryptlib/Include/openssl/x509.h:22,
from shim.c:56:
/usr/lib64/gcc/x86_64-suse-linux/6/include/stddef.h:417:0: error: "offsetof" redefined [-Werror]
#define offsetof(TYPE, MEMBER) __builtin_offsetof (TYPE, MEMBER)
In file included from Cryptlib/Include/limits.h:15:0,
from Cryptlib/Include/openssl/ossl_typ.h:13,
from Cryptlib/Include/openssl/x509.h:20,
from shim.c:56:
Cryptlib/Include/CrtLibSupport.h:192:0: note: this is the location of the previous definition
#define offsetof(type, member) ( (int) & ((type*)0) -> member )
We can lower the priority of the gcc include path or just remove the
path, but this might cause problem since the path was introduced on
purpose(*). Instead, including stddef.h first is more feasible.
(*) d51739a416
Signed-off-by: Gary Lin <glin@suse.com>
OpenSSL changes quite a bit of the key validation, and most of the keys
I can find in the wild aren't marked as trusted by the new checker.
Intel noticed this too: https://github.com/vathpela/edk2/commit/f536d7c3ed
but instead of fixing the compatibility error, they switched their test
data to match the bug.
So that's pretty broken.
For now, I'm reverting OpenSSL 1.1.0e, because we need those certs in
the wild to work.
This reverts commit 513cbe2aea.
This reverts commit e9cc33d6f2.
This reverts commit 80d49f758e.
This reverts commit 9bc647e2b2.
This reverts commit ae75df6232.
This reverts commit e883479f35.
This reverts commit 97469449fd.
This reverts commit e39692647f.
This reverts commit 0f3dfc01e2.
This reverts commit 4da6ac8195.
This reverts commit d064bd7eef.
This reverts commit 9bc86cfd6f.
This reverts commit ab9a05a10f.
Signed-off-by: Peter Jones <pjones@redhat.com>
Rather than looping once through the possible actions (MokNew, MokDel, etc.),
revise the logic so that instead of rebooting immediately we get back to the
main menu setting a flag to replace "Continue booting" with a proper reboot.
Getting back to the menu means we can go make other changes before rebooting.
For instance, you might want to enable validation, but beforehand you also
need to enroll a MOK. You can already do so from userland; except the requests
were cleared as soon as one of them was processed.
This involves some extra cleanup of the states to avoid running the same
request more than once, removing the option from the menu once it's done, and
changing prompting functions to return an EFI_STATUS so we can better track
whether the process has succeeded.
Signed-off-by: Mathieu Trudel-Lapierre <mathieu.trudel-lapierre@canonical.com>
