summaryrefslogtreecommitdiff
path: root/tools/objtool
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2024-03-12 14:49:30 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2024-03-12 14:49:30 -0700
commit216532e147b2fee6ee830f4a844bbc3cbb9137af (patch)
tree0177c6fe41fcd036a8e20fe50ab03a982553fe97 /tools/objtool
parentb32273ee89a866b01b316b9a8de407efde01090c (diff)
parent3fe1eb4dd2e4b872ffb7b9b081b34ffcfa934ba7 (diff)
Merge tag 'hardening-v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
Pull hardening updates from Kees Cook: "As is pretty normal for this tree, there are changes all over the place, especially for small fixes, selftest improvements, and improved macro usability. Some header changes ended up landing via this tree as they depended on the string header cleanups. Also, a notable set of changes is the work for the reintroduction of the UBSAN signed integer overflow sanitizer so that we can continue to make improvements on the compiler side to make this sanitizer a more viable future security hardening option. Summary: - string.h and related header cleanups (Tanzir Hasan, Andy Shevchenko) - VMCI memcpy() usage and struct_size() cleanups (Vasiliy Kovalev, Harshit Mogalapalli) - selftests/powerpc: Fix load_unaligned_zeropad build failure (Michael Ellerman) - hardened Kconfig fragment updates (Marco Elver, Lukas Bulwahn) - Handle tail call optimization better in LKDTM (Douglas Anderson) - Use long form types in overflow.h (Andy Shevchenko) - Add flags param to string_get_size() (Andy Shevchenko) - Add Coccinelle script for potential struct_size() use (Jacob Keller) - Fix objtool corner case under KCFI (Josh Poimboeuf) - Drop 13 year old backward compat CAP_SYS_ADMIN check (Jingzi Meng) - Add str_plural() helper (Michal Wajdeczko, Kees Cook) - Ignore relocations in .notes section - Add comments to explain how __is_constexpr() works - Fix m68k stack alignment expectations in stackinit Kunit test - Convert string selftests to KUnit - Add KUnit tests for fortified string functions - Improve reporting during fortified string warnings - Allow non-type arg to type_max() and type_min() - Allow strscpy() to be called with only 2 arguments - Add binary mode to leaking_addresses scanner - Various small cleanups to leaking_addresses scanner - Adding wrapping_*() arithmetic helper - Annotate initial signed integer wrap-around in refcount_t - Add explicit UBSAN section to MAINTAINERS - Fix UBSAN self-test warnings - Simplify UBSAN build via removal of CONFIG_UBSAN_SANITIZE_ALL - Reintroduce UBSAN's signed overflow sanitizer" * tag 'hardening-v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux: (51 commits) selftests/powerpc: Fix load_unaligned_zeropad build failure string: Convert helpers selftest to KUnit string: Convert selftest to KUnit sh: Fix build with CONFIG_UBSAN=y compiler.h: Explain how __is_constexpr() works overflow: Allow non-type arg to type_max() and type_min() VMCI: Fix possible memcpy() run-time warning in vmci_datagram_invoke_guest_handler() lib/string_helpers: Add flags param to string_get_size() x86, relocs: Ignore relocations in .notes section objtool: Fix UNWIND_HINT_{SAVE,RESTORE} across basic blocks overflow: Use POD in check_shl_overflow() lib: stackinit: Adjust target string to 8 bytes for m68k sparc: vdso: Disable UBSAN instrumentation kernel.h: Move lib/cmdline.c prototypes to string.h leaking_addresses: Provide mechanism to scan binary files leaking_addresses: Ignore input device status lines leaking_addresses: Use File::Temp for /tmp files MAINTAINERS: Update LEAKING_ADDRESSES details fortify: Improve buffer overflow reporting fortify: Add KUnit tests for runtime overflows ...
Diffstat (limited to 'tools/objtool')
-rw-r--r--tools/objtool/check.c12
-rw-r--r--tools/objtool/noreturns.h2
2 files changed, 13 insertions, 1 deletions
diff --git a/tools/objtool/check.c b/tools/objtool/check.c
index 8440b7bb343c..eb7e12ebc1d0 100644
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -3620,6 +3620,18 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
}
if (!save_insn->visited) {
+ /*
+ * If the restore hint insn is at the
+ * beginning of a basic block and was
+ * branched to from elsewhere, and the
+ * save insn hasn't been visited yet,
+ * defer following this branch for now.
+ * It will be seen later via the
+ * straight-line path.
+ */
+ if (!prev_insn)
+ return 0;
+
WARN_INSN(insn, "objtool isn't smart enough to handle this CFI save/restore combo");
return 1;
}
diff --git a/tools/objtool/noreturns.h b/tools/objtool/noreturns.h
index 1685d7ea6a9f..3a301696f005 100644
--- a/tools/objtool/noreturns.h
+++ b/tools/objtool/noreturns.h
@@ -6,6 +6,7 @@
*
* Yes, this is unfortunate. A better solution is in the works.
*/
+NORETURN(__fortify_panic)
NORETURN(__kunit_abort)
NORETURN(__module_put_and_kthread_exit)
NORETURN(__reiserfs_panic)
@@ -22,7 +23,6 @@ NORETURN(do_exit)
NORETURN(do_group_exit)
NORETURN(do_task_dead)
NORETURN(ex_handler_msr_mce)
-NORETURN(fortify_panic)
NORETURN(hlt_play_dead)
NORETURN(hv_ghcb_terminate)
NORETURN(kthread_complete_and_exit)