linux-kernel - Re: [PATCH 00/24] objtool: retpoline and asm-goto validation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 23 Jan 2018 16:42:36 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     David Woodhouse <dwmw2@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Josh Poimboeuf <jpoimboe@...hat.com>
Cc:     linux-kernel@...r.kernel.org, Dave Hansen <dave.hansen@...el.com>,
        Ashok Raj <ashok.raj@...el.com>,
        Tim Chen <tim.c.chen@...ux.intel.com>,
        Andy Lutomirski <luto@...nel.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Greg KH <gregkh@...uxfoundation.org>,
        Andrea Arcangeli <aarcange@...hat.com>,
        Andi Kleen <ak@...ux.intel.com>,
        Arjan Van De Ven <arjan.van.de.ven@...el.com>,
        Dan Williams <dan.j.williams@...el.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Jun Nakajima <jun.nakajima@...el.com>,
        Asit Mallick <asit.k.mallick@...el.com>,
        Jason Baron <jbaron@...mai.com>
Subject: Re: [PATCH 00/24] objtool: retpoline and asm-goto validation

On Tue, Jan 23, 2018 at 04:25:39PM +0100, Peter Zijlstra wrote:
> This series adds two features to objtool:
> 
>  - retpoline validation; it verifies no indirect jumps/calls are left in
>    CONFIG_RETPOLINE=y builds.
> 
>    includes fixups for kvm and others
> 
>  - asm-goto validation; allows asserting that a block of code is not reachable
>    through a dynamic branch.
> 
> After this it forces an asm-goto capable compiler to build arch/x86 and removes
> the fallback code for static_cpu_has(). After which we can use static_cpu_has()
> to guarantee a lack of dynamic branches, which is a requested feature for
> various spectre related things.

x86_64-allmodconfig -KCOV -KASAN +patch gives:

drivers/infiniband/hw/bnxt_re/.tmp_qplib_fp.o: warning: objtool: bnxt_qplib_poll_cq()+0xf8: sibling call from callable instruction with modified stack frame
drivers/watchdog/.tmp_hpwdt.o: warning: objtool: .text+0x24: indirect call found in RETPOLINE build


---
Subject: jump_label: Add static assertion to every static_branch
From: Peter Zijlstra <peterz@...radead.org>
Date: Tue Jan 16 15:27:36 CET 2018

For testing only.. can't use in general because people do:

  static_branch() || more-cond

which confuses the living daylight out of objtool validation and
really isn't fixable.

DO NOT MERGE

Not-signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
---
 arch/x86/include/asm/jump_label.h |    1 +
 include/linux/cgroup.h            |    2 +-
 include/linux/jump_label.h        |   15 +++++++++++++--
 3 files changed, 15 insertions(+), 3 deletions(-)

--- a/arch/x86/include/asm/jump_label.h
+++ b/arch/x86/include/asm/jump_label.h
@@ -76,6 +76,7 @@ static __always_inline bool arch_static_
  *
  * Also works with static_cpu_has().
  */
+#define arch_static_assert arch_static_assert
 static __always_inline void arch_static_assert(void)
 {
 	asm volatile ("1:\n\t"
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -88,7 +88,7 @@ extern struct css_set init_css_set;
  * @ss: subsystem in question
  */
 #define cgroup_subsys_on_dfl(ss)						\
-	static_branch_likely(&ss ## _on_dfl_key)
+	__static_branch_likely(&ss ## _on_dfl_key)
 
 bool css_has_online_children(struct cgroup_subsys_state *css);
 struct cgroup_subsys_state *css_from_id(int id, struct cgroup_subsys *ss);
--- a/include/linux/jump_label.h
+++ b/include/linux/jump_label.h
@@ -323,6 +323,10 @@ extern bool ____wrong_branch_error(void)
 
 #ifdef HAVE_JUMP_LABEL
 
+#ifndef arch_static_assert
+#define arch_static_assert (void)
+#endif
+
 /*
  * Combine the right initial value (type) with the right branch order
  * to generate the desired result.
@@ -379,7 +383,7 @@ extern bool ____wrong_branch_error(void)
  * See jump_label_type() / jump_label_init_type().
  */
 
-#define static_branch_likely(x)							\
+#define __static_branch_likely(x)						\
 ({										\
 	bool branch;								\
 	if (__builtin_types_compatible_p(typeof(*x), struct static_key_true))	\
@@ -391,7 +395,9 @@ extern bool ____wrong_branch_error(void)
 	likely(branch);								\
 })
 
-#define static_branch_unlikely(x)						\
+#define static_branch_likely(x) (__static_branch_likely(x) && (arch_static_assert(), true))
+
+#define __static_branch_unlikely(x)						\
 ({										\
 	bool branch;								\
 	if (__builtin_types_compatible_p(typeof(*x), struct static_key_true))	\
@@ -403,11 +409,16 @@ extern bool ____wrong_branch_error(void)
 	unlikely(branch);							\
 })
 
+#define static_branch_unlikely(x) (__static_branch_unlikely(x) && (arch_static_assert(), true))
+
 #else /* !HAVE_JUMP_LABEL */
 
 #define static_branch_likely(x)		likely(static_key_enabled(&(x)->key))
 #define static_branch_unlikely(x)	unlikely(static_key_enabled(&(x)->key))
 
+#define __static_branch_likely(x)		likely(static_key_enabled(&(x)->key))
+#define __static_branch_unlikely(x)	unlikely(static_key_enabled(&(x)->key))
+
 #endif /* HAVE_JUMP_LABEL */
 
 /*