Commit Graph

46 Commits

Author SHA1 Message Date
Peter Jones
c8811bfed2 lib/shell.c: minor cleanup
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
80b7937e17 lib/simple_file.c: minor cleanup
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
a55b4d6688 lib: Use EFI_ERROR() instead of comparing to EFI_SUCCESS everywhere.
Also consistently name our status variable "efi_status" unless there's a
good reason not to, such as already having another one of those.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
4816cd7533 lib: find_in_variable_esl(): Fix a tiny nitpick clang-analyze has.
clang-analyze believes the following:

311	EFI_STATUS
312	variable_enroll_hash(CHAR16 *var, EFI_GUID owner,
313			     UINT8 hash[SHA256_DIGEST_SIZE])
314	{
315		EFI_STATUS efi_status;
316
317		efi_status = find_in_variable_esl(var, owner, hash, SHA256_DIGEST_SIZE);
>               Calling 'find_in_variable_esl' →

260	EFI_STATUS
261	find_in_variable_esl(CHAR16* var, EFI_GUID owner, UINT8 *key, UINTN keylen)
262	{
263		UINTN DataSize;
264		UINT8 *Data;
>               ← 'Data' declared without an initial value →
265		EFI_STATUS efi_status;
266
267		efi_status = get_variable(var, &Data, &DataSize, owner);
>               ← Calling 'get_variable' →

237	EFI_STATUS
238	get_variable(CHAR16 *var, UINT8 **data, UINTN *len, EFI_GUID owner)
239	{
240		return get_variable_attr(var, data, len, owner, NULL);
>		← Calling 'get_variable_attr' →

213	EFI_STATUS
214	get_variable_attr(CHAR16 *var, UINT8 **data, UINTN *len, EFI_GUID owner,
215			  UINT32 *attributes)
216	{
217		EFI_STATUS efi_status;
218
219		*len = 0;
220
221		efi_status = GetVariable(var, &owner, NULL, len, NULL);
>		← Calling 'GetVariable' →
>		← Returning from 'GetVariable' →
222		if (efi_status != EFI_BUFFER_TOO_SMALL)
>		← Assuming the condition is true →
>		← Taking true branch →
223			return efi_status;
224
225		*data = AllocateZeroPool(*len);
226		if (!*data)
227			return EFI_OUT_OF_RESOURCES;
228
229		efi_status = GetVariable(var, &owner, attributes, len, *data);
230		if (EFI_ERROR(efi_status)) {
231			FreePool(*data);
232			*data = NULL;
233		}
234		return efi_status;
235	}

And it can't figure out that the first GetVariable() call will, in fact,
always return EFI_BUFFER_TOO_SMALL, and that AllocateZeroPool() will
then *correctly* clobber the two variables we never assigned the value
from.  It also then believes that efi_status might have been returned
/without/ being an error, and thinks that means we'll use the
uninitialized pointer.

This won't happen, but hey, let's make the code better express to the
checker what is intended.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
0d17c49219 lib: simple_file_selector(): remove some dead code.
clang-analyzer correctly believes this:

465                             int i;
466
467				i = StrLen(name) - 1;

                                ^ Value stored to 'i' is never read

468
469				for (i = StrLen(name); i > 0; --i) {
470					if (name[i] == '\\')
471						break;
472				}

And it's right; that's completely dead code.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
70a4c4a395 lib: simple_file_selector(): simplify the error path to confuse covscan less.
Because they don't believe code should be defensive against future
changes, covscan believes:

520 out_free:
521        FreePool(dmp);
   CID 182824 (#1 of 1): Dereference before null check
   (REVERSE_INULL)check_after_deref: Null-checking entries suggests that
   it may be null, but it has already been dereferenced on all paths
   leading to the check.
522        if (entries) {
523                free_entries(entries, count);
524                FreePool(entries);
525        }
526 out_free_name:
527        FreePool(name);
528}

Which is technically correct, but still kind of dumb.  So this patch
combines the two error out paths into just being out_free, so that the
first path there is before entries is allocated.  (It also initializes
dmp to NULL and checks that before freeing it.)

I also Lindent-ed that function.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
7ee19bdc41 Use gcc's offsetof() instead of hacking out our own.
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
54f8f1f9ed Get rid of all the places we cast to (CHAR16 *[])
Lindent gets confused by these, and they're hard to read anyway.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
9facc22ebe Fix some "if (x < 0)" tests where x is UINTN.
Obviously, these are not correct.  Most of them are just useless; one
can be changed to a more useful test.

Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
b953468e91 Don't have tons of local guid definitions for no reason at all.
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
4d70f10481 lib/variables.c: reformat CreateTimeBasedPayload()
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
dc62a3c4dc Move includes around to clean the source tree up a bit.
Signed-off-by: Peter Jones <pjones@redhat.com>
2018-03-12 16:21:43 -04:00
Peter Jones
97a3f6cf94 "in_protocol" is used in more than shim.o; make it not static.
Signed-off-by: Peter Jones <pjones@redhat.com>
2017-12-19 16:52:01 -05:00
Peter Jones
96b347c200 Add a mechanism to print openssl errors
Signed-off-by: Peter Jones <pjones@redhat.com>
2017-08-31 15:13:45 -04:00
Peter Jones
78f6b007e7 Make msleep() be a thing
Signed-off-by: Peter Jones <pjones@redhat.com>
2017-08-31 15:13:34 -04:00
Peter Jones
8a6d270d70 Make it possible to build in a subdirectory.
This lets you do:

mkdir build-x64 build-ia32
cd build-x64
make TOPDIR=.. -f ../Makefile
cd ../build-ia32
setarch i686 -B make ARCH=ia32 TOPDIR=.. -f ../Makefile

And not worry about generated sources and headers mixing and matching.

Signed-off-by: Peter Jones <pjones@redhat.com>
2017-07-24 20:11:28 -04:00
Lans Zhang
3537d91c61 become more friendly for the cross compilation
Allow to overwrite certain settings from the external.

Signed-off-by: Lans Zhang <jia.zhang@windriver.com>
2017-06-15 11:30:05 -04:00
Max Resch
047242c624 fix compilerer errors in security policy
When compiling with -Werror security policy can't be compiled because of
type errors. This commit fixes this problem.
2017-06-15 11:21:04 -04:00
Peter Jones
abe6ed8b57 More incorrect unsigned vs signed fixups from yours truly.
Woops.

Signed-off-by: Peter Jones <pjones@redhat.com>
2015-06-29 14:41:21 -04:00
Peter Jones
7bb0d6f71d Fix console_print_box*() parameters.
When we made lib build with the correct CFLAGS, it inherited
-Werror=sign-compare, and I fixed up some parameters on
console_print_box() and console_print_box_at() to avoid sign comparison
errors.

The fixups were *completely wrong*, as some behavior relies on negative
values.  So this fixes them in a completely different way, by casting
appropriately to signed types where we're doing comparisons.

Signed-off-by: Peter Jones <pjones@redhat.com>
2015-06-16 11:41:32 -04:00
Peter Jones
605be9f179 Make lib/ use the right CFLAGS.
Signed-off-by: Peter Jones <pjones@redhat.com>
2015-04-13 19:55:25 -04:00
Peter Jones
7fdbd9d48a Make lib/ build right with the cflags it should be using...
... but isn't.

Signed-off-by: Peter Jones <pjones@redhat.com>
2015-04-13 19:55:25 -04:00
Gary Ching-Pang Lin
361716dd4a Add nostdinc to the CFLAGS for lib
We don't need the headers from the standard include path.

Signed-off-by: Gary Ching-Pang Lin <glin@suse.com>
2014-12-11 09:48:50 -05:00
Ard Biesheuvel
99d7b5e858 Factor out x86-isms and add cross compile support
This patch cleans up and refactors the Makefiles to better allow new
architectures to be added:
- remove unused Makefile definitions
- import Makefile definitions from top level rather than redefining
- move x86 specific CFLAGS to inside ifeq() blocks
- remove x86 inline asm
- allow $(FORMAT) to be overridden: this is necessary as there exists no
  EFI or PE/COFF aware objcopy for ARM

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
2014-08-12 10:54:05 -04:00
Peter Jones
eb4cb6a509 Make sure we default to assuming we're locked down.
If "SecureBoot" exists but "SetupMode" does not, assume "SetupMode" says
we're not in Setup Mode.

Signed-off-by: Peter Jones <pjones@redhat.com>
2014-06-25 10:55:56 -04:00
Gary Ching-Pang Lin
868b372115 Check the secure variables with the lib functions
There are functions defined in lib to check the secure variables.
Use the functions to shun the duplicate code.

Signed-off-by: Gary Ching-Pang Lin <glin@suse.com>

Conflicts:
	shim.c
2014-06-25 10:55:12 -04:00
Gary Ching-Pang Lin
8bd7bad4b7 Remove the duplicate calls in lib/console.c
Signed-off-by: Gary Ching-Pang Lin <glin@suse.com>
2014-06-25 10:15:31 -04:00
Gary Ching-Pang Lin
22254e2633 MokManager: handle the error status from ReadKeyStroke
On some machines, even though the key event was signaled, ReadKeyStroke
still got EFI_NOT_READY. This commit handles the error status to avoid
console_get_keystroke from returning unexpected keys.

Signed-off-by: Gary Ching-Pang Lin <glin@suse.com>

Conflicts:
	MokManager.c
2014-06-25 10:02:18 -04:00
Peter Jones
293f28d1fe Error check the right thing in get_variable_attr() when allocating.
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-11-21 11:48:24 -05:00
Peter Jones
4dbef508ab Initialize entries before we pass it to another function.
Coverity scan noticed that entries is uninitialized when we pass its
location to another function.

Signed-off-by: Peter Jones <pjones@redhat.com>
2013-11-21 11:48:24 -05:00
Peter Jones
af25679e16 Fix wrong sizeof().
CHAR16* vs CHAR16**, so the result is the same on all platforms.

Detected by coverity.

Signed-off-by: Peter Jones <pjones@redhat.com>
2013-11-21 11:48:24 -05:00
Andrew Boie
9712a7e77d allow 32-bit compilation with 64-bit compiler
Also removed unused LIB_PATH from some Makefiles.

Change-Id: I7d28d18f7531b51b6121a2ffb88bcaedec57c467
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
2013-11-12 10:31:59 -05:00
Peter Jones
3c0ee51ad8 Put SHIM_VERBOSE under shim's guid, not global.
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-04 13:54:35 -04:00
Peter Jones
880f9de412 Try to actually make debug printing look reasonable.
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-04 11:51:09 -04:00
Peter Jones
fc986307fb Add ident-like blobs to shim.efi for version checking.
I feel dirty.
2013-10-03 11:11:09 -04:00
Peter Jones
aec345c0b2 CompareMem expects void * and gcc complains.
Sorry about that.

Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-02 10:02:01 -04:00
Peter Jones
6e5d86e565 Clean up warnings.
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-01 14:38:09 -04:00
Peter Jones
f330528786 Conditionalize overriding the security policy.
Make OVERRIDE_SECURITY_POLICY a build option.

Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-01 14:03:16 -04:00
Peter Jones
4537217422 Merge console_control.h and console.h
Since these are topically the same thing, they can live together.

Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-01 14:03:16 -04:00
Peter Jones
09a37bbc69 Make verbose stuff use console_notify
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-10-01 14:03:16 -04:00
Gary Ching-Pang Lin
3508c40c39 integrate security override 2013-09-26 11:58:03 -04:00
Gary Ching-Pang Lin
7d602e843c Merge variable retrieving functions 2013-09-26 11:58:02 -04:00
Gary Ching-Pang Lin
79424b09ca Merge signature.h into efiauthenticated.h and guid.h
Conflicts:
	shim.c
2013-09-26 11:58:02 -04:00
Peter Jones
4a7f9bd4a6 Make EFI_PATH easily resettable from the build command line.
Signed-off-by: Peter Jones <pjones@redhat.com>
2013-09-26 11:58:01 -04:00
Gary Ching-Pang Lin
b82d6d7cb1 simple_file: Allocate buffers for file entries
The dir filter appends L'/' to the directory entries without
allocating a new buffer, and this could crash the whole program.
2013-09-26 11:58:01 -04:00
Matthew Garrett
d359712e1b Port MokManager to Linux Foundation loader UI code
This is the first stage of porting the MokManager UI to the UI code used
by the Linux Foundation UEFI loader.
2013-09-26 11:57:59 -04:00