linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Alexander Popov <alex.popov@linux.com>
To: Kees Cook <keescook@chromium.org>,
	Emese Revfy <re.emese@gmail.com>,
	Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>,
	Masahiro Yamada <masahiroy@kernel.org>,
	Michal Marek <michal.lkml@markovi.net>,
	Andrew Morton <akpm@linux-foundation.org>,
	Masahiro Yamada <yamada.masahiro@socionext.com>,
	Thiago Jung Bauermann <bauerman@linux.ibm.com>,
	Luis Chamberlain <mcgrof@kernel.org>,
	Jessica Yu <jeyu@kernel.org>,
	Sven Schnelle <svens@stackframe.org>,
	Iurii Zaikin <yzaikin@google.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Vincenzo Frascino <vincenzo.frascino@arm.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Peter Collingbourne <pcc@google.com>,
	Naohiro Aota <naohiro.aota@wdc.com>,
	Alexander Monakov <amonakov@ispras.ru>,
	Mathias Krause <minipli@googlemail.com>,
	PaX Team <pageexec@freemail.hu>,
	Brad Spengler <spender@grsecurity.net>,
	Laura Abbott <labbott@redhat.com>,
	Florian Weimer <fweimer@redhat.com>,
	Alexander Popov <alex.popov@linux.com>,
	kernel-hardening@lists.openwall.com,
	linux-kbuild@vger.kernel.org, x86@kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, gcc@gcc.gnu.org
Cc: notify@kernel.org
Subject: [PATCH 1/5] gcc-plugins/stackleak: Exclude alloca() from the instrumentation logic
Date: Thu,  4 Jun 2020 16:49:53 +0300	[thread overview]
Message-ID: <20200604134957.505389-2-alex.popov@linux.com> (raw)
In-Reply-To: <20200604134957.505389-1-alex.popov@linux.com>

Some time ago Variable Length Arrays (VLA) were removed from the kernel.
The kernel is built with '-Wvla'. Let's exclude alloca() from the
instrumentation logic and make it simpler. The build-time assertion
against alloca() is added instead.

Unfortunately, for that assertion we can't simply check cfun->calls_alloca
during RTL phase. It turned out that gcc before version 7 called
allocate_dynamic_stack_space() from expand_stack_vars() for runtime
alignment of constant-sized stack variables. That caused cfun->calls_alloca
to be set for functions that don't use alloca().

Signed-off-by: Alexander Popov <alex.popov@linux.com>
---
 scripts/gcc-plugins/stackleak_plugin.c | 51 +++++++++++---------------
 1 file changed, 21 insertions(+), 30 deletions(-)

diff --git a/scripts/gcc-plugins/stackleak_plugin.c b/scripts/gcc-plugins/stackleak_plugin.c
index cc75eeba0be1..1ecfe50d0bf5 100644
--- a/scripts/gcc-plugins/stackleak_plugin.c
+++ b/scripts/gcc-plugins/stackleak_plugin.c
@@ -9,10 +9,9 @@
  * any of the gcc libraries
  *
  * This gcc plugin is needed for tracking the lowest border of the kernel stack.
- * It instruments the kernel code inserting stackleak_track_stack() calls:
- *  - after alloca();
- *  - for the functions with a stack frame size greater than or equal
- *     to the "track-min-size" plugin parameter.
+ * It instruments the kernel code inserting stackleak_track_stack() calls
+ * for the functions with a stack frame size greater than or equal to
+ * the "track-min-size" plugin parameter.
  *
  * This plugin is ported from grsecurity/PaX. For more information see:
  *   https://grsecurity.net/
@@ -46,7 +45,7 @@ static struct plugin_info stackleak_plugin_info = {
 		"disable\t\tdo not activate the plugin\n"
 };
 
-static void stackleak_add_track_stack(gimple_stmt_iterator *gsi, bool after)
+static void stackleak_add_track_stack(gimple_stmt_iterator *gsi)
 {
 	gimple stmt;
 	gcall *stackleak_track_stack;
@@ -56,12 +55,7 @@ static void stackleak_add_track_stack(gimple_stmt_iterator *gsi, bool after)
 	/* Insert call to void stackleak_track_stack(void) */
 	stmt = gimple_build_call(track_function_decl, 0);
 	stackleak_track_stack = as_a_gcall(stmt);
-	if (after) {
-		gsi_insert_after(gsi, stackleak_track_stack,
-						GSI_CONTINUE_LINKING);
-	} else {
-		gsi_insert_before(gsi, stackleak_track_stack, GSI_SAME_STMT);
-	}
+	gsi_insert_before(gsi, stackleak_track_stack, GSI_SAME_STMT);
 
 	/* Update the cgraph */
 	bb = gimple_bb(stackleak_track_stack);
@@ -87,14 +81,13 @@ static bool is_alloca(gimple stmt)
 
 /*
  * Work with the GIMPLE representation of the code. Insert the
- * stackleak_track_stack() call after alloca() and into the beginning
- * of the function if it is not instrumented.
+ * stackleak_track_stack() call into the beginning of the function.
  */
 static unsigned int stackleak_instrument_execute(void)
 {
 	basic_block bb, entry_bb;
-	bool prologue_instrumented = false, is_leaf = true;
-	gimple_stmt_iterator gsi;
+	bool is_leaf = true;
+	gimple_stmt_iterator gsi = { 0 };
 
 	/*
 	 * ENTRY_BLOCK_PTR is a basic block which represents possible entry
@@ -111,27 +104,17 @@ static unsigned int stackleak_instrument_execute(void)
 	 */
 	FOR_EACH_BB_FN(bb, cfun) {
 		for (gsi = gsi_start_bb(bb); !gsi_end_p(gsi); gsi_next(&gsi)) {
-			gimple stmt;
-
-			stmt = gsi_stmt(gsi);
+			gimple stmt = gsi_stmt(gsi);
 
 			/* Leaf function is a function which makes no calls */
 			if (is_gimple_call(stmt))
 				is_leaf = false;
 
-			if (!is_alloca(stmt))
-				continue;
-
-			/* Insert stackleak_track_stack() call after alloca() */
-			stackleak_add_track_stack(&gsi, true);
-			if (bb == entry_bb)
-				prologue_instrumented = true;
+			/* Variable Length Arrays are forbidden in the kernel */
+			gcc_assert(!is_alloca(stmt));
 		}
 	}
 
-	if (prologue_instrumented)
-		return 0;
-
 	/*
 	 * Special cases to skip the instrumentation.
 	 *
@@ -168,7 +151,7 @@ static unsigned int stackleak_instrument_execute(void)
 		bb = single_succ(ENTRY_BLOCK_PTR_FOR_FN(cfun));
 	}
 	gsi = gsi_after_labels(bb);
-	stackleak_add_track_stack(&gsi, false);
+	stackleak_add_track_stack(&gsi);
 
 	return 0;
 }
@@ -185,12 +168,20 @@ static bool large_stack_frame(void)
 /*
  * Work with the RTL representation of the code.
  * Remove the unneeded stackleak_track_stack() calls from the functions
- * which don't call alloca() and don't have a large enough stack frame size.
+ * that don't have a large enough stack frame size.
  */
 static unsigned int stackleak_cleanup_execute(void)
 {
 	rtx_insn *insn, *next;
 
+	/*
+	 * gcc before version 7 called allocate_dynamic_stack_space() from
+	 * expand_stack_vars() for runtime alignment of constant-sized stack
+	 * variables. That caused cfun->calls_alloca to be set for functions
+	 * that don't use alloca().
+	 * For more info see gcc commit 7072df0aae0c59ae437e.
+	 * Let's leave such functions instrumented.
+	 */
 	if (cfun->calls_alloca)
 		return 0;
 
-- 
2.25.2


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  reply	other threads:[~2020-06-04 13:50 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-04 13:49 [PATCH 0/5] Improvements of the stackleak gcc plugin Alexander Popov
2020-06-04 13:49 ` Alexander Popov [this message]
2020-06-04 14:01   ` [PATCH 1/5] gcc-plugins/stackleak: Exclude alloca() from the instrumentation logic Jann Horn
2020-06-04 15:23     ` Alexander Popov
2020-06-09 18:39       ` Kees Cook
2020-06-10 15:24         ` Alexander Popov
2020-06-04 13:49 ` [PATCH 2/5] gcc-plugins/stackleak: Use asm instrumentation to avoid useless register saving Alexander Popov
2020-06-04 15:05   ` Miguel Ojeda
2020-06-09 18:46   ` Kees Cook
2020-06-10 15:47     ` Alexander Popov
2020-06-10 20:03       ` Kees Cook
2020-06-11 23:45         ` Alexander Popov
2020-06-04 13:49 ` [PATCH 3/5] gcc-plugins/stackleak: Add 'verbose' plugin parameter Alexander Popov
2020-06-09 18:47   ` Kees Cook
2020-06-10 15:52     ` Alexander Popov
2020-06-10 20:04       ` Kees Cook
2020-06-04 13:49 ` [PATCH 4/5] gcc-plugins/stackleak: Don't instrument itself Alexander Popov
2020-06-09 18:48   ` Kees Cook
2020-06-04 13:49 ` [PATCH 5/5] gcc-plugins/stackleak: Don't instrument vgettimeofday.c in arm64 VDSO Alexander Popov
2020-06-04 13:58   ` Will Deacon
2020-06-04 14:14     ` Jann Horn
2020-06-04 14:20       ` Alexander Popov
2020-06-04 14:25         ` Jann Horn
2020-06-04 14:44           ` Alexander Popov
2020-06-09 19:09     ` Kees Cook
2020-06-10  7:30       ` Will Deacon
2020-06-10 15:18         ` Alexander Popov
2020-06-04 21:39 ` [PATCH 0/5] Improvements of the stackleak gcc plugin Kees Cook
2020-06-09 19:15 ` Kees Cook
2020-06-10 15:14   ` Alexander Popov

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=20200604134957.505389-2-alex.popov@linux.com \
    --to=alex.popov@linux.com \
    --cc=akpm@linux-foundation.org \
    --cc=amonakov@ispras.ru \
    --cc=bauerman@linux.ibm.com \
    --cc=catalin.marinas@arm.com \
    --cc=fweimer@redhat.com \
    --cc=gcc@gcc.gnu.org \
    --cc=jeyu@kernel.org \
    --cc=keescook@chromium.org \
    --cc=kernel-hardening@lists.openwall.com \
    --cc=labbott@redhat.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=masahiroy@kernel.org \
    --cc=mcgrof@kernel.org \
    --cc=michal.lkml@markovi.net \
    --cc=miguel.ojeda.sandonis@gmail.com \
    --cc=minipli@googlemail.com \
    --cc=naohiro.aota@wdc.com \
    --cc=notify@kernel.org \
    --cc=pageexec@freemail.hu \
    --cc=pcc@google.com \
    --cc=re.emese@gmail.com \
    --cc=spender@grsecurity.net \
    --cc=svens@stackframe.org \
    --cc=tglx@linutronix.de \
    --cc=vincenzo.frascino@arm.com \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    --cc=yamada.masahiro@socionext.com \
    --cc=yzaikin@google.com \
    /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).