linux-kernel - [PATCH 3.16 205/294] modpost: don't emit section mismatch warnings for compiler optimizations

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]

Message-ID: <lsq.1510009382.8692801@decadent.org.uk>
Date:   Mon, 06 Nov 2017 23:03:02 +0000
From:   Ben Hutchings <ben@...adent.org.uk>
To:     linux-kernel@...r.kernel.org, stable@...r.kernel.org
CC:     akpm@...ux-foundation.org,
        "Paul Gortmaker" <paul.gortmaker@...driver.com>,
        "Rusty Russell" <rusty@...tcorp.com.au>,
        "Arnd Bergmann" <arnd@...db.de>
Subject: [PATCH 3.16 205/294] modpost: don't emit section mismatch
 warnings for compiler optimizations

3.16.50-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Paul Gortmaker <paul.gortmaker@...driver.com>

commit 4a3893d069b788f3570c19c12d9e986e8e15870f upstream.

Currently an allyesconfig build [gcc-4.9.1] can generate the following:

   WARNING: vmlinux.o(.text.unlikely+0x3864): Section mismatch in
   reference from the function cpumask_empty.constprop.3() to the
   variable .init.data:nmi_ipi_mask

which comes from the cpumask_empty usage in arch/x86/kernel/nmi_selftest.c.

Normally we would not see a symbol entry for cpumask_empty since it is:

	static inline bool cpumask_empty(const struct cpumask *srcp)

however in this case, the variant of the symbol gets emitted when GCC does
constant propagation optimization.

Fix things up so that any locally optimized constprop variants don't warn
when accessing variables that live in the __init sections.

[arnd: adapted text_sections definition to 3.18]

Signed-off-by: Paul Gortmaker <paul.gortmaker@...driver.com>
Signed-off-by: Rusty Russell <rusty@...tcorp.com.au>
Signed-off-by: Arnd Bergmann <arnd@...db.de>
Signed-off-by: Ben Hutchings <ben@...adent.org.uk>
---
 scripts/mod/modpost.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -930,6 +930,10 @@ static const char *init_sections[] = { A
 static const char *init_exit_sections[] =
 	{ALL_INIT_SECTIONS, ALL_EXIT_SECTIONS, NULL };
 
+/* all text sections */
+static const char *const text_sections[] = { ALL_INIT_TEXT_SECTIONS,
+				ALL_EXIT_TEXT_SECTIONS, TEXT_SECTIONS, NULL };
+
 /* data section */
 static const char *data_sections[] = { DATA_SECTIONS, NULL };
 
@@ -948,6 +952,7 @@ static const char *data_sections[] = { D
 static const char *head_sections[] = { ".head.text*", NULL };
 static const char *linker_symbols[] =
 	{ "__init_begin", "_sinittext", "_einittext", NULL };
+static const char *const optim_symbols[] = { "*.constprop.*", NULL };
 
 enum mismatch {
 	TEXT_TO_ANY_INIT,
@@ -1105,6 +1110,17 @@ static const struct sectioncheck *sectio
  *   This pattern is identified by
  *   refsymname = __init_begin, _sinittext, _einittext
  *
+ * Pattern 5:
+ *   GCC may optimize static inlines when fed constant arg(s) resulting
+ *   in functions like cpumask_empty() -- generating an associated symbol
+ *   cpumask_empty.constprop.3 that appears in the audit.  If the const that
+ *   is passed in comes from __init, like say nmi_ipi_mask, we get a
+ *   meaningless section warning.  May need to add isra symbols too...
+ *   This pattern is identified by
+ *   tosec   = init section
+ *   fromsec = text section
+ *   refsymname = *.constprop.*
+ *
  **/
 static int secref_whitelist(const struct sectioncheck *mismatch,
 			    const char *fromsec, const char *fromsym,
@@ -1137,6 +1153,12 @@ static int secref_whitelist(const struct
 	if (match(tosym, linker_symbols))
 		return 0;
 
+	/* Check for pattern 5 */
+	if (match(fromsec, text_sections) &&
+	    match(tosec, init_sections) &&
+	    match(fromsym, optim_symbols))
+		return 0;
+
 	return 1;
 }