public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Alan Modra <amodra@gmail.com>
To: binutils@sourceware.org
Subject: NS32K arg_bufs uninitialised
Date: Sun, 22 Mar 2020 23:23:53 +1030	[thread overview]
Message-ID: <20200322125353.GO4583@bubble.grove.modra.org> (raw)

git commit d1e304bc27 was aimed at stopping uninitialised memory
access to the index_offset array.  Unfortunately that patch resulted
in a different array being uninitialised for all instructions with
more than two arguments.

	* ns32k-dis.c (print_insn_arg): Update comment.
	(print_insn_ns32k): Reduce size of index_offset array, and
	initialize, passing -1 to print_insn_arg for args that are not
	an index.  Don't exit arg loop early.  Abort on bad arg number.

diff --git a/opcodes/ns32k-dis.c b/opcodes/ns32k-dis.c
index d505edd774..12df182d0a 100644
--- a/opcodes/ns32k-dis.c
+++ b/opcodes/ns32k-dis.c
@@ -446,7 +446,7 @@ invalid_float (float_type_u *p, int len)
    bit position of the addressing extension.  BUFFER contains the
    instruction.  ADDR is where BUFFER was read from.  Put the disassembled
    version of the operand in RESULT.  INDEX_OFFSET is the bit position
-   of the index byte (it contains garbage if this operand is not a
+   of the index byte (it contains -1 if this operand is not a
    general operand using scaled indexed addressing mode).  */
 
 static int
@@ -790,10 +790,8 @@ print_insn_ns32k (bfd_vma memaddr, disassemble_info *info)
   if (*d)
     {
       /* Offset in bits of the first thing beyond each index byte.
-	 Element 0 is for operand A and element 1 is for operand B.
-	 The rest are irrelevant, but we put them here so we don't
-	 index outside the array.  */
-      int index_offset[MAX_ARGS];
+	 Element 0 is for operand A and element 1 is for operand B.  */
+      int index_offset[2];
 
       /* 0 for operand A, 1 for operand B, greater for other args.  */
       int whicharg = 0;
@@ -806,6 +804,8 @@ print_insn_ns32k (bfd_vma memaddr, disassemble_info *info)
 	 if we are using scaled indexed addressing mode, since the index
 	 bytes occur right after the basic instruction, not as part
 	 of the addressing extension.  */
+      index_offset[0] = -1;
+      index_offset[1] = -1;
       if (Is_gen (d[1]))
 	{
 	  int bitoff = d[1] == 'f' ? 10 : 5;
@@ -832,15 +832,16 @@ print_insn_ns32k (bfd_vma memaddr, disassemble_info *info)
       while (*d)
 	{
 	  argnum = *d - '1';
+	  if (argnum >= MAX_ARGS)
+	    abort ();
 	  d++;
-	  if (argnum > maxarg && argnum < MAX_ARGS)
+	  if (argnum > maxarg)
 	    maxarg = argnum;
 	  ioffset = print_insn_arg (*d, ioffset, &aoffset, buffer,
 				    memaddr, arg_bufs[argnum],
-				    index_offset[whicharg]);
+				    whicharg > 1 ? -1 : index_offset[whicharg]);
 	  d++;
-	  if (whicharg++ >= 1)
-	    break;
+	  whicharg++;
 	}
 
       for (argnum = 0; argnum <= maxarg; argnum++)

-- 
Alan Modra
Australia Development Lab, IBM

                 reply	other threads:[~2020-03-22 12:54 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20200322125353.GO4583@bubble.grove.modra.org \
    --to=amodra@gmail.com \
    --cc=binutils@sourceware.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 read-only IMAP folder(s) and NNTP newsgroup(s).