Go to file
Peter Jones a7249a65af Actually refer to the base relocation table of our loaded image.
Currently when we process base relocations, we get the correct Data
Directory pointer from the headers (context->RelocDir), and that header
has been copied into our pristine allocated image when we copied up to
SizeOfHeaders.  But the data it points to has not been mirrored in to
the new image, so it is whatever data AllocPool() gave us.

This patch changes relocate_coff() to refer to the base relocation table
from the image we loaded from disk, but apply the fixups to the new
copy.

I have no idea how x86_64 worked without this, but I can't make aarch64
work without it.  I also don't know how Ard or Leif have seen aarch64
work.  Maybe they haven't?  Leif indicated on irc that they may have
only tested shim with simple "hello world" applications from gnu-efi;
they are certainly much less complex than grub.efi, and are generated
through a different linking process.

My only theory is that we're getting recycled data there pretty reliably
that just makes us /not/ process any relocations, but since our
ImageBase is 0, and I don't think we ever load grub with 0 as its base
virtual address, that doesn't follow.  I'm open to any other ideas
anybody has.

I do know that on x86_64 (and presumably aarch64 as well), we don't
actually start seeing *symptoms* of this bug until the first chunk[0] of
94c9a77f is applied[1].  Once that is applied, relocate_coff() starts
seeing zero[2] for both RelocBase->VirtualAddress and
RelocBase->SizeOfBlock, because RelocBase is a (generated, relative)
pointer that only makes sense in the context of the original binary, not
our partial copy.  Since RelocBase->SizeOfBlock is tested first,
relocate_base() gives us "Reloc block size is invalid"[3] and returns
EFI_UNSUPPORTED.  At that point shim exits with an error.

[0] The second chunk of 94c9a77f patch makes no difference on this
    issue.
[1] I don't see why at all.
[2] Which could really be any value since it's AllocatePool() and not
    AllocateZeroPool() results, but 0 is all I've observed; I think
    AllocatePool() has simply never recycled any memory in my test
    cases.
[3] which is silent because perror() tries to avoid talking because that
    has caused much crashing in the past; work needs to go in to 0.9 for
    this.

Signed-off-by: Peter Jones <pjones@redhat.com>
2014-09-19 09:30:26 -04:00
Cryptlib Update openssl to 0.9.8zb 2014-08-19 14:20:23 -04:00
include Make sure we don't try to load a binary from a different arch. 2014-08-27 16:40:57 -04:00
lib Factor out x86-isms and add cross compile support 2014-08-12 10:54:05 -04:00
.gitignore Add ident-like blobs to shim.efi for version checking. 2013-10-03 11:11:09 -04:00
cert.S Add support for 32-bit ARM 2014-08-12 10:54:05 -04:00
COPYRIGHT Add copyright file 2012-07-09 11:03:12 -04:00
crypt_blowfish.c MokManager: support blowfish-based crypt() hash 2013-09-26 11:58:01 -04:00
crypt_blowfish.h MokManager: support blowfish-based crypt() hash 2013-09-26 11:58:01 -04:00
elf_aarch64_efi.lds Add support for 64-bit ARM (AArch64) 2014-08-12 10:54:05 -04:00
elf_arm_efi.lds Fix typo from Ard's old tree 32-bit ARM patch. 2014-08-27 11:49:39 -04:00
elf_ia32_efi.lds Move embedded certificates to their own section. 2013-06-10 17:35:33 -04:00
elf_ia64_efi.lds Move embedded certificates to their own section. 2013-06-10 17:35:33 -04:00
elf_x86_64_efi.lds Move embedded certificates to their own section. 2013-06-10 17:35:33 -04:00
fallback.c [fallback] Try to boot the first boot option anyway 2014-05-13 13:30:07 -04:00
make-certs Sign MokManager with a locally-generated key 2012-11-26 13:43:50 -05:00
Makefile Add support for 32-bit ARM 2014-08-12 10:54:05 -04:00
MokManager.c MokManager: handle the error status from ReadKeyStroke 2014-06-25 10:02:18 -04:00
MokVars.txt Add support for disabling db for verification 2013-10-02 11:29:34 -04:00
netboot.c Factor out x86-isms and add cross compile support 2014-08-12 10:54:05 -04:00
netboot.h netboot.h: fix build error on 32-bit systems 2013-11-12 10:25:40 -05:00
PasswordCrypt.c additional bounds-checking on section sizes 2014-04-11 14:41:22 -04:00
PasswordCrypt.h MokManager: support Tradition DES hash 2013-09-26 11:58:01 -04:00
README Replace build instructions in README with something not completely wrong. 2014-07-21 16:15:07 -04:00
replacements.c Don't name something exit(). 2014-08-27 16:40:57 -04:00
replacements.h Allow fallback to use the system's LoadImage/StartImage . 2014-02-14 17:48:01 -05:00
shim.c Actually refer to the base relocation table of our loaded image. 2014-09-19 09:30:26 -04:00
shim.h Make SHIM_LOCK_GUID a first-class object with a symbol. 2013-09-23 10:40:49 -04:00
testplan.txt Add a failure case to the test plan and fix an ordering error. 2014-02-14 17:48:01 -05:00
TODO Update for Josh's changes. 2013-10-02 13:33:52 -04:00
ucs2.h Add StrCSpn() 2013-04-30 09:46:22 -04:00
version.c.in Add ident-like blobs to shim.efi for version checking. 2013-10-03 11:11:09 -04:00
version.h Add ident-like blobs to shim.efi for version checking. 2013-10-03 11:11:09 -04:00

shim is a trivial EFI application that, when run, attempts to open and
execute another application. It will initially attempt to do this via the
standard EFI LoadImage() and StartImage() calls. If these fail (because secure
boot is enabled and the binary is not signed with an appropriate key, for
instance) it will then validate the binary against a built-in certificate. If
this succeeds and if the binary or signing key are not blacklisted then shim
will relocate and execute the binary.

shim will also install a protocol which permits the second-stage bootloader
to perform similar binary validation. This protocol has a GUID as described
in the shim.h header file and provides a single entry point. On 64-bit systems
this entry point expects to be called with SysV ABI rather than MSABI, and
so calls to it should not be wrapped.

To use shim, simply place a DER-encoded public certificate in a file such as
pub.cer and build with "make VENDOR_CERT_FILE=pub.cer".