From: Alan Modra <amodra@gmail.com>
Date: Sun, 30 Aug 2020 10:57:27 +0000 (+0930)
Subject: readelf: ubsan: shift exponent 32 is too large
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=54b8331d0b6b58d6f7256a22e7a9457d64891a43;p=binutils-gdb.git

readelf: ubsan: shift exponent 32 is too large

When compiled on a 32-bit host, a temp var is too small for possible
64-bit values to be calculated.

	* readelf.c (print_gnu_build_attribute_name): Make "bytes"
	unsigned long long.
---

diff --git a/binutils/ChangeLog b/binutils/ChangeLog
index ca9028688f9..9872504871b 100644
--- a/binutils/ChangeLog
+++ b/binutils/ChangeLog
@@ -1,3 +1,8 @@
+2020-08-30  Alan Modra  <amodra@gmail.com>
+
+	* readelf.c (print_gnu_build_attribute_name): Use unsigned long
+	long for "bytes".
+
 2020-08-28  Nick Clifton  <nickc@redhat.com>
 
 	PR 26548
diff --git a/binutils/readelf.c b/binutils/readelf.c
index f02848e4681..94aa876b5ae 100644
--- a/binutils/readelf.c
+++ b/binutils/readelf.c
@@ -19662,7 +19662,7 @@ print_gnu_build_attribute_name (Elf_Internal_Note * pnote)
 
 	while (bytes --)
 	  {
-	    unsigned long byte = (* name ++) & 0xff;
+	    unsigned long long byte = *name++ & 0xff;
 
 	    val |= byte << shift;
 	    shift += 8;