提交 2d21889f 编写于 作者: J Jean-Philippe Brucker 提交者: Will Deacon

arm64: Don't insert a BTI instruction at inner labels

Some ftrace features are broken since commit 714a8d02 ("arm64: asm:
Override SYM_FUNC_START when building the kernel with BTI"). For example
the function_graph tracer:

$ echo function_graph > /sys/kernel/debug/tracing/current_tracer
[   36.107016] WARNING: CPU: 0 PID: 115 at kernel/trace/ftrace.c:2691 ftrace_modify_all_code+0xc8/0x14c

When ftrace_modify_graph_caller() attempts to write a branch at
ftrace_graph_call, it finds the "BTI J" instruction inserted by
SYM_INNER_LABEL() instead of a NOP, and aborts.

It turns out we don't currently need the BTI landing pads inserted by
SYM_INNER_LABEL:

* ftrace_call and ftrace_graph_call are only used for runtime patching
  of the active tracer. The patched code is not reached from a branch.
* install_el2_stub is reached from a CBZ instruction, which doesn't
  change PSTATE.BTYPE.
* __guest_exit is reached from B instructions in the hyp-entry vectors,
  which aren't subject to BTI checks either.

Remove the BTI annotation from SYM_INNER_LABEL.

Fixes: 714a8d02 ("arm64: asm: Override SYM_FUNC_START when building the kernel with BTI")
Signed-off-by: NJean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: NMark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20200624112253.1602786-1-jean-philippe@linaro.orgSigned-off-by: NWill Deacon <will@kernel.org>
上级 e56404e8
......@@ -12,7 +12,6 @@
* instead.
*/
#define BTI_C hint 34 ;
#define BTI_J hint 36 ;
/*
* When using in-kernel BTI we need to ensure that PCS-conformant assembly
......@@ -43,11 +42,6 @@
SYM_START(name, SYM_L_WEAK, SYM_A_NONE) \
BTI_C
#define SYM_INNER_LABEL(name, linkage) \
.type name SYM_T_NONE ASM_NL \
SYM_ENTRY(name, linkage, SYM_A_NONE) \
BTI_J
#endif
/*
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册