From: Peter Zijlstra <peterz@infradead.org>
From: Peter Zijlstra <peterz@infradead.org>
To: David Woodhouse <dwmw2@infradead.org>,
Thomas Gleixner <tglx@linutronix.de>,
Josh Poimboeuf <jpoimboe@redhat.com>
Cc: linux-kernel@vger.kernel.org, Dave Hansen <dave.hansen@intel.com>,
Ashok Raj <ashok.raj@intel.com>,
Tim Chen <tim.c.chen@linux.intel.com>,
Andy Lutomirski <luto@kernel.org>,
Linus Torvalds <torvalds@linux-foundation.org>,
Greg KH <gregkh@linuxfoundation.org>,
Andrea Arcangeli <aarcange@redhat.com>,
Andi Kleen <ak@linux.intel.com>,
Arjan Van De Ven <arjan.van.de.ven@intel.com>,
Dan Williams <dan.j.williams@intel.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Jun Nakajima <jun.nakajima@intel.com>,
Asit Mallick <asit.k.mallick@intel.com>,
Jason Baron <jbaron@akamai.com>,
Peter Zijlstra <peterz@infradead.org>
Subject: [PATCH 17/35] objtool: Even more complex static block checks
Date: Thu, 18 Jan 2018 14:48:17 +0100 [thread overview]
Message-ID: <20180118140152.477421083@infradead.org> (raw)
In-Reply-To: <20180118134800.711245485@infradead.org>
[-- Attachment #1: peterz-objtool-more-clever.patch --]
[-- Type: text/plain, Size: 3465 bytes --]
I've observed GCC transform:
f()
{
if (!static_branch_unlikely())
return;
static_assert();
A;
}
g()
{
f();
}
Into:
f()
{
static_assert();
A;
}
g()
{
if (static_branch_unlikely())
f();
}
Which results in the assertion landing at f+0. The transformation is
valid and useful; it avoids a pointless CALL+RET sequence, so we'll
have to teach objtool how to deal with this.
Do this by marking all CALL destinations with static_call when called
from a static_block and non_static_call when called outside a
static_block. This allows us to identify functions called exclusively
from a static_block and start them with a static_block.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
tools/objtool/check.c | 81 +++++++++++++++++++++++++++++++++++++-------------
tools/objtool/elf.h | 1
2 files changed, 61 insertions(+), 21 deletions(-)
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -1207,36 +1207,75 @@ static int read_retpoline_hints(struct o
return 0;
}
+static bool __grow_static_block(struct objtool_file *file,
+ struct instruction *insn, bool ign_bt)
+{
+ /* if a jump can come in here, terminate */
+ if (insn->branch_target && !ign_bt)
+ return false;
+
+ switch (insn->type) {
+ case INSN_JUMP_UNCONDITIONAL:
+ /* mark this instruction, terminate this section */
+ insn->static_jump_dest = true;
+ return false;
+
+ /* these disturb unconditional code flow, terminate */
+ case INSN_JUMP_CONDITIONAL:
+ case INSN_JUMP_DYNAMIC:
+ case INSN_RETURN:
+ case INSN_BUG:
+ return false;
+
+ /* these return right back and don't disturb the code flow */
+ case INSN_CALL:
+ case INSN_CALL_DYNAMIC:
+ break;
+ }
+
+ /* mark this insn, and continue the section */
+ insn->static_jump_dest = true;
+ return true;
+}
+
static int grow_static_blocks(struct objtool_file *file)
{
- struct instruction *insn;
bool static_block = false;
+ struct symbol *func, *tmp;
+ struct instruction *insn;
+ struct section *sec;
for_each_insn(file, insn) {
- if (!static_block && !insn->static_jump_dest)
- continue;
+ if (static_block || insn->static_jump_dest)
+ static_block = __grow_static_block(file, insn, !static_block);
- if (insn->static_jump_dest) {
- static_block = true;
- continue;
+ if (insn->type == INSN_CALL) {
+ func = insn->call_dest;
+ if (!func)
+ continue;
+
+ if (static_block)
+ func->static_call = true;
+ else
+ func->non_static_call = true;
}
+ }
- if (insn->branch_target) {
- static_block = false;
- continue;
- } else switch (insn->type) {
- case INSN_JUMP_CONDITIONAL:
- case INSN_JUMP_UNCONDITIONAL:
- case INSN_JUMP_DYNAMIC:
- case INSN_CALL:
- case INSN_CALL_DYNAMIC:
- case INSN_RETURN:
- case INSN_BUG:
- static_block = false;
- continue;
+ for_each_sec(file, sec) {
+ list_for_each_entry_safe(func, tmp, &sec->symbol_list, list) {
+ if (!func->static_call)
+ continue;
+
+ if (func->non_static_call)
+ continue;
+
+ /* static && !non_static -- only static callers */
+
+ func_for_each_insn(file, func, insn) {
+ if (!__grow_static_block(file, insn, false))
+ break;
+ }
}
-
- insn->static_jump_dest = static_block;
}
return 0;
--- a/tools/objtool/elf.h
+++ b/tools/objtool/elf.h
@@ -61,6 +61,7 @@ struct symbol {
unsigned char bind, type;
unsigned long offset;
unsigned int len;
+ bool static_call, non_static_call;
};
struct rela {
next prev parent reply other threads:[~2018-01-18 15:04 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-01-18 13:48 [PATCH 00/35] jump_label, objtool, IBRS and IBPB Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 01/35] jump_label: Add branch hints to static_branch_{un,}likely() Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 02/35] sched: Optimize ttwu_stat() Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 03/35] x86: Reindent _static_cpu_has Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 04/35] x86: Update _static_cpu_has to use all named variables Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 05/35] x86: Add a type field to alt_instr Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 06/35] objtool: Implement base jump_assert support Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 07/35] x86: Annotate static_cpu_has alternative Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 08/35] objtool: Implement jump_assert for _static_cpu_has() Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 09/35] objtool: Introduce special_type Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 10/35] x86/jump_label: Implement arch_static_assert() Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 11/35] objtool: Add retpoline validation Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 12/35] x86/paravirt: Annotate indirect calls Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 13/35] x86,nospec: Annotate indirect calls/jumps Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 14/35] x86: Annotate indirect jump in head_64.S Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 15/35] objtool: More complex static jump implementation Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 16/35] objtool: Use existing global variables for options Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` Peter Zijlstra, Peter Zijlstra [this message]
2018-01-18 13:48 ` [PATCH 18/35] objtool: Another static block fail Peter Zijlstra, Peter Zijlstra
2018-01-19 16:42 ` Peter Zijlstra
2018-01-29 18:01 ` Josh Poimboeuf
2018-01-29 18:24 ` Peter Zijlstra
2018-01-18 13:48 ` [PATCH 19/35] objtool: Skip static assert when KCOV/KASAN Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 20/35] x86: Force asm-goto Peter Zijlstra, Peter Zijlstra
2018-01-18 16:25 ` David Woodhouse
2018-01-18 13:48 ` [PATCH 21/35] x86: Remove FAST_FEATURE_TESTS Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 22/35] x86/cpufeatures: Detect Speculation control feature Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 23/35] x86/speculation: Add basic speculation control code Peter Zijlstra, Peter Zijlstra
2018-01-18 16:37 ` Josh Poimboeuf
2018-01-18 17:08 ` Dave Hansen
2018-01-18 17:12 ` Paolo Bonzini
2018-01-18 18:24 ` Josh Poimboeuf
2018-01-18 19:08 ` Andrea Arcangeli
2018-01-18 23:25 ` Andy Lutomirski
2018-01-18 23:35 ` Andrew Cooper
2018-01-19 1:41 ` Andrea Arcangeli
2018-01-19 4:10 ` Andy Lutomirski
2018-01-19 4:15 ` Van De Ven, Arjan
2018-01-19 15:47 ` Andrea Arcangeli
2018-01-18 13:48 ` [PATCH 24/35] x86/msr: Move native_*msr macros out of microcode.h Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 25/35] x86/speculation: Add inlines to control Indirect Branch Speculation Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 26/35] x86/enter: Create macros to stop/restart " Peter Zijlstra, Peter Zijlstra
2018-01-18 19:44 ` Tim Chen
2018-01-18 13:48 ` [PATCH 27/35] x86/enter: Use IBRS on syscall and interrupts Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 28/35] x86/idle: Control Indirect Branch Speculation in idle Peter Zijlstra, Peter Zijlstra
2018-01-18 19:52 ` Andrew Cooper
2018-01-18 13:48 ` [PATCH 29/35] x86/speculation: Add IPBP support Peter Zijlstra, Peter Zijlstra
2018-01-18 16:22 ` Josh Poimboeuf
2018-01-18 18:31 ` Borislav Petkov
2018-01-18 18:35 ` Josh Poimboeuf
2018-01-18 18:46 ` Borislav Petkov
2018-01-18 13:48 ` [PATCH 30/35] x86/speculation: Use Indirect Branch Prediction Barrier in context switch Peter Zijlstra, Peter Zijlstra
2018-01-19 0:38 ` Tim Chen
2018-01-19 4:03 ` Kevin Easton
2018-01-19 20:26 ` Tim Chen
2018-01-18 13:48 ` [PATCH 31/35] x86/ibrs: Add new helper macros to save/restore MSR_IA32_SPEC_CTRL Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 32/35] x86/vmx: Direct access to MSR_IA32_SPEC_CTRL Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 33/35] x86/svm: " Peter Zijlstra, Peter Zijlstra
2018-01-18 13:48 ` [PATCH 34/35] x86/kvm: Add IBPB support Peter Zijlstra, Peter Zijlstra
2018-01-18 15:32 ` Paolo Bonzini
2018-01-19 15:25 ` Paolo Bonzini
2018-01-19 16:08 ` David Woodhouse
2018-01-19 16:27 ` Andy Lutomirski
2018-01-19 16:48 ` Paolo Bonzini
2018-01-18 13:48 ` [PATCH 35/35] x86/nospec: Add static assertions Peter Zijlstra, Peter Zijlstra
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180118140152.477421083@infradead.org \
--to=peterz@infradead.org \
--cc=aarcange@redhat.com \
--cc=ak@linux.intel.com \
--cc=arjan.van.de.ven@intel.com \
--cc=ashok.raj@intel.com \
--cc=asit.k.mallick@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.hansen@intel.com \
--cc=dwmw2@infradead.org \
--cc=gregkh@linuxfoundation.org \
--cc=jbaron@akamai.com \
--cc=jpoimboe@redhat.com \
--cc=jun.nakajima@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=pbonzini@redhat.com \
--cc=tglx@linutronix.de \
--cc=tim.c.chen@linux.intel.com \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome