Diffstat (limited to 'tools/objtool/Documentation/stack-validation.txt')
1 files changed, 17 insertions, 39 deletions
diff --git a/tools/objtool/Documentation/stack-validation.txt b/tools/objtool/Documentation/stack-validation.txt
index 17c1195f11f4..6a1af43862df 100644
@@ -11,9 +11,6 @@ analyzes every .o file and ensures the validity of its stack metadata.
It enforces a set of rules on asm code and C inline assembly code so
that stack traces can be reliable.
-Currently it only checks frame pointer usage, but there are plans to add
-CFI validation for C files and CFI generation for asm files.
For each function, it recursively follows all possible code paths and
validates the correct frame pointer state at each instruction.
@@ -23,6 +20,10 @@ alternative execution paths to a given instruction (or set of
instructions). Similarly, it knows how to follow switch statements, for
which gcc sometimes uses jump tables.
+(Objtool also has an 'orc generate' subcommand which generates debuginfo
+for the ORC unwinder. See Documentation/x86/orc-unwinder.txt in the
+kernel tree for more details.)
Why do we need stack metadata validation?
@@ -93,37 +94,14 @@ a) More reliable stack traces for frame pointer enabled kernels
or at the very end of the function after the stack frame has been
destroyed. This is an inherent limitation of frame pointers.
-b) 100% reliable stack traces for DWARF enabled kernels
- (NOTE: This is not yet implemented)
- As an alternative to frame pointers, DWARF Call Frame Information
- (CFI) metadata can be used to walk the stack. Unlike frame pointers,
- CFI metadata is out of band. So it doesn't affect runtime
- performance and it can be reliable even when interrupts or exceptions
- are involved.
- For C code, gcc automatically generates DWARF CFI metadata. But for
- asm code, generating CFI is a tedious manual approach which requires
- manually placed .cfi assembler macros to be scattered throughout the
- code. It's clumsy and very easy to get wrong, and it makes the real
- code harder to read.
- Stacktool will improve this situation in several ways. For code
- which already has CFI annotations, it will validate them. For code
- which doesn't have CFI annotations, it will generate them. So an
- architecture can opt to strip out all the manual .cfi annotations
- from their asm code and have objtool generate them instead.
+b) ORC (Oops Rewind Capability) unwind table generation
- We might also add a runtime stack validation debug option where we
- periodically walk the stack from schedule() and/or an NMI to ensure
- that the stack metadata is sane and that we reach the bottom of the
+ An alternative to frame pointers and DWARF, ORC unwind data can be
+ used to walk the stack. Unlike frame pointers, ORC data is out of
+ band. So it doesn't affect runtime performance and it can be
+ reliable even when interrupts or exceptions are involved.
- So the benefit of objtool here will be that external tooling should
- always show perfect stack traces. And the same will be true for
- kernel warning/oops traces if the architecture has a runtime DWARF
+ For more details, see Documentation/x86/orc-unwinder.txt.
c) Higher live patching compatibility rate
@@ -211,7 +189,7 @@ they mean, and suggestions for how to fix them.
function, add proper frame pointer logic using the FRAME_BEGIN and
FRAME_END macros. Otherwise, if it's not a callable function, remove
its ELF function annotation by changing ENDPROC to END, and instead
- use the manual CFI hint macros in asm/undwarf.h.
+ use the manual unwind hint macros in asm/unwind_hints.h.
If it's a GCC-compiled .c file, the error may be because the function
uses an inline asm() statement which has a "call" instruction. An
@@ -231,8 +209,8 @@ they mean, and suggestions for how to fix them.
If the error is for an asm file, and the instruction is inside (or
reachable from) a callable function, the function should be annotated
with the ENTRY/ENDPROC macros (ENDPROC is the important one).
- Otherwise, the code should probably be annotated with the CFI hint
- macros in asm/undwarf.h so objtool and the unwinder can know the
+ Otherwise, the code should probably be annotated with the unwind hint
+ macros in asm/unwind_hints.h so objtool and the unwinder can know the
stack state associated with the code.
If you're 100% sure the code won't affect stack traces, or if you're
@@ -258,7 +236,7 @@ they mean, and suggestions for how to fix them.
instructions aren't allowed in a callable function, and are most
likely part of the kernel entry code. They should usually not have
the callable function annotation (ENDPROC) and should always be
- annotated with the CFI hint macros in asm/undwarf.h.
+ annotated with the unwind hint macros in asm/unwind_hints.h.
6. file.o: warning: objtool: func()+0x26: sibling call from callable instruction with modified stack frame
@@ -272,7 +250,7 @@ they mean, and suggestions for how to fix them.
If the instruction is not actually in a callable function (e.g.
kernel entry code), change ENDPROC to END and annotate manually with
- the CFI hint macros in asm/undwarf.h.
+ the unwind hint macros in asm/unwind_hints.h.
7. file: warning: objtool: func()+0x5c: stack state mismatch
@@ -288,8 +266,8 @@ they mean, and suggestions for how to fix them.
Another possibility is that the code has some asm or inline asm which
does some unusual things to the stack or the frame pointer. In such
- cases it's probably appropriate to use the CFI hint macros in
+ cases it's probably appropriate to use the unwind hint macros in
8. file.o: warning: objtool: funcA() falls through to next function funcB()