linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Josh Poimboeuf <jpoimboe@redhat.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: tglx@linutronix.de, linux-kernel@vger.kernel.org, x86@kernel.org,
	mhiramat@kernel.org, mbenes@suse.cz
Subject: Re: [PATCH v4 01/13] objtool: Remove CFI save/restore special case
Date: Thu, 26 Mar 2020 20:00:01 -0500	[thread overview]
Message-ID: <20200327010001.i3kebxb4um422ycb@treble> (raw)
In-Reply-To: <20200326195718.GD2452@worktop.programming.kicks-ass.net>

On Thu, Mar 26, 2020 at 08:57:18PM +0100, Peter Zijlstra wrote:
> On Thu, Mar 26, 2020 at 04:49:38PM +0100, Peter Zijlstra wrote:
> > > The 'insn == first' check isn't ideal, but at least it works (I think?).
> > 
> > It works, yes, for exactly this one case.
>
> How's this? Ignore the ignore_cfi bits, that's a 'failed' experiment.

It still seems complex to me.

What do you think about this?  If we store save_insn in the state when
we see insn->save, the restore logic becomes a lot easier.  Then if we
get a restore without a save, we can just ignore the restore hint in
that path.  Later, when we see the restore insn again from the save
path, we can then compare the insn state with the saved state to make
sure they match.

This assumes no crazy save/restore scenarios.  It also means that the
restore path has to be reachable from the save path, for which I had to
make a change to make IRETQ *not* a dead end if there's a restore hint
immediately after it.

diff --git a/tools/objtool/check.c b/tools/objtool/check.c
index e637a4a38d2a..e9becd50f148 100644
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -1265,7 +1265,6 @@ static int read_unwind_hints(struct objtool_file *file)
 
 		} else if (hint->type == UNWIND_HINT_TYPE_RESTORE) {
 			insn->restore = true;
-			insn->hint = true;
 			continue;
 		}
 
@@ -2003,7 +2002,7 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
 			   struct instruction *first, struct insn_state state)
 {
 	struct alternative *alt;
-	struct instruction *insn, *next_insn;
+	struct instruction *insn, *next_insn, *save_insn = NULL;
 	struct section *sec;
 	u8 visited;
 	int ret;
@@ -2034,54 +2033,32 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
 
 		visited = 1 << state.uaccess;
 		if (insn->visited) {
-			if (!insn->hint && !insn_state_match(insn, &state))
+			if (!insn->hint && !insn->restore &&
+			    !insn_state_match(insn, &state)) {
 				return 1;
+			}
+
+			if (insn->restore && save_insn) {
+				if (!insn_state_match(insn, &save_insn->state))
+					return 1;
+				save_insn = NULL;
+			}
 
 			if (insn->visited & visited)
 				return 0;
 		}
 
-		if (insn->hint) {
-			if (insn->restore) {
-				struct instruction *save_insn, *i;
-
-				i = insn;
-				save_insn = NULL;
-				func_for_each_insn_continue_reverse(file, func, i) {
-					if (i->save) {
-						save_insn = i;
-						break;
-					}
-				}
-
-				if (!save_insn) {
-					WARN_FUNC("no corresponding CFI save for CFI restore",
-						  sec, insn->offset);
-					return 1;
-				}
-
-				if (!save_insn->visited) {
-					/*
-					 * Oops, no state to copy yet.
-					 * Hopefully we can reach this
-					 * instruction from another branch
-					 * after the save insn has been
-					 * visited.
-					 */
-					if (insn == first)
-						return 0;
-
-					WARN_FUNC("objtool isn't smart enough to handle this CFI save/restore combo",
-						  sec, insn->offset);
-					return 1;
-				}
+		if (insn->save)
+			save_insn = insn;
 
-				insn->state = save_insn->state;
-			}
+		if (insn->restore && save_insn) {
+			insn->state = save_insn->state;
+			save_insn = NULL;
+		}
 
+		if (insn->hint)
 			state = insn->state;
-
-		} else
+		else
 			insn->state = state;
 
 		insn->visited |= visited;
@@ -2191,12 +2168,17 @@ static int validate_branch(struct objtool_file *file, struct symbol *func,
 			break;
 
 		case INSN_CONTEXT_SWITCH:
-			if (func && (!next_insn || !next_insn->hint)) {
-				WARN_FUNC("unsupported instruction in callable function",
-					  sec, insn->offset);
-				return 1;
+			if (!next_insn || !next_insn->restore) {
+				if (func) {
+					WARN_FUNC("unsupported instruction in callable function",
+							sec, insn->offset);
+					return 1;
+				}
+
+				return 0;
 			}
-			return 0;
+
+			break;
 
 		case INSN_STACK:
 			if (update_insn_state(insn, &state))
@@ -2293,7 +2275,7 @@ static int validate_unwind_hints(struct objtool_file *file)
 	clear_insn_state(&state);
 
 	for_each_insn(file, insn) {
-		if (insn->hint && !insn->visited) {
+		if ((insn->hint || insn->save || insn->restore) && !insn->visited) {
 			ret = validate_branch(file, insn->func, insn, state);
 			if (ret && backtrace)
 				BT_FUNC("<=== (hint)", insn);


  reply	other threads:[~2020-03-27  1:00 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-25 17:45 [PATCH v4 00/13] objtool: vmlinux.o and moinstr validation Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 01/13] objtool: Remove CFI save/restore special case Peter Zijlstra
2020-03-26 11:30   ` Peter Zijlstra
2020-03-26 12:58     ` Peter Zijlstra
2020-03-26 13:44       ` Josh Poimboeuf
2020-03-26 15:38         ` Peter Zijlstra
2020-03-27  4:19           ` Josh Poimboeuf
2020-03-26 14:44       ` Miroslav Benes
2020-03-26 15:04         ` Miroslav Benes
2020-03-26 13:00     ` Peter Zijlstra
2020-03-26 13:56     ` Josh Poimboeuf
2020-03-26 15:49       ` Peter Zijlstra
2020-03-26 19:57         ` Peter Zijlstra
2020-03-27  1:00           ` Josh Poimboeuf [this message]
2020-03-30 17:02             ` Peter Zijlstra
2020-03-30 19:02               ` Josh Poimboeuf
2020-03-30 20:02                 ` Peter Zijlstra
2020-03-30 20:29                   ` Peter Zijlstra
2020-03-31 11:16                   ` [RFC][PATCH] objtool,ftrace: Implement UNWIND_HINT_RET_OFFSET Peter Zijlstra
2020-03-31 15:31                     ` Steven Rostedt
2020-03-31 16:06                       ` [RFC][PATCH] x86,ftrace: Shrink ftrace_regs_caller() by one byte Peter Zijlstra
2020-03-31 19:58                       ` [RFC][PATCH] objtool,ftrace: Implement UNWIND_HINT_RET_OFFSET Peter Zijlstra
2020-03-31 20:26                         ` Josh Poimboeuf
2020-03-31 20:23                     ` Josh Poimboeuf
2020-03-31 20:40                       ` Peter Zijlstra
2020-03-31 21:07                         ` Peter Zijlstra
2020-03-31 21:17                         ` Josh Poimboeuf
2020-03-31 21:20                           ` Josh Poimboeuf
2020-03-31 22:27                             ` [PATCH v2] " Peter Zijlstra
2020-04-01 14:14                               ` Josh Poimboeuf
2020-04-01 14:22                                 ` Peter Zijlstra
2020-04-01 14:39                                   ` Josh Poimboeuf
2020-04-01 15:38                                     ` Peter Zijlstra
2020-04-01 15:39                                     ` Steven Rostedt
2020-04-01 15:43                               ` Julien Thierry
2020-04-01 17:09                                 ` Peter Zijlstra
2020-04-01 17:33                                   ` Steven Rostedt
2020-04-01 17:45                                     ` Peter Zijlstra
2020-04-01 18:20                                       ` Steven Rostedt
2020-04-01 20:20                                         ` Peter Zijlstra
2020-04-01 17:37                                   ` Josh Poimboeuf
2020-04-02  6:41                                   ` Julien Thierry
2020-04-02  6:56                                     ` Julien Thierry
2020-04-02  7:50                                     ` Peter Zijlstra
2020-04-02  8:16                                       ` Julien Thierry
2020-04-02  8:17                                       ` Peter Zijlstra
2020-04-02  8:29                                         ` Julien Thierry
2020-04-02  8:58                                           ` Miroslav Benes
2020-03-25 17:45 ` [PATCH v4 02/13] objtool: Factor out CFI hints Peter Zijlstra
2020-03-25 18:26   ` Miroslav Benes
2020-03-25 19:41     ` Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 03/13] objtool: Rename struct cfi_state Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 04/13] objtool: Fix !CFI insn_state propagation Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 05/13] objtool: Implement noinstr validation Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 06/13] objtool: Optimize !vmlinux.o again Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 07/13] objtool: Use sec_offset_hash() for insn_hash Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 08/13] objtool: Detect loading function pointers across noinstr Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 09/13] kbuild/objtool: Add objtool-vmlinux.o pass Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 10/13] objtool: Avoid iterating !text section symbols Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 11/13] objtool: Rearrange validate_section() Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 12/13] objtool: Add STT_NOTYPE noinstr validation Peter Zijlstra
2020-03-25 17:45 ` [PATCH v4 13/13] objtool: Also consider .entry.text as noinstr Peter Zijlstra
2020-03-25 19:03 ` [PATCH v4 00/13] objtool: vmlinux.o and moinstr validation Miroslav Benes

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=20200327010001.i3kebxb4um422ycb@treble \
    --to=jpoimboe@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mbenes@suse.cz \
    --cc=mhiramat@kernel.org \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.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;
as well as URLs for NNTP newsgroup(s).