[Bug gdb/22388] New: assertion "xfered_len>0" in target.c for remote connection

["pcarroll at codesourcery dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=22388

            Bug ID: 22388
           Summary: assertion "xfered_len>0" in target.c for remote
                    connection
           Product: gdb
           Version: HEAD
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: gdb
          Assignee: unassigned at sourceware dot org
          Reporter: pcarroll at codesourcery dot com
  Target Milestone: ---

We have a customer who is using a Corelis gdb server to connect to gdb.
Occasionally, the gdb server will send a 0-byte block of memory for a read.
When this happens, gdb gives an assertion from target.c:

internal-error: target_xfer_partial: Assertion `*xfered_len > 0' failed.

This problem is almost identical to that fixed in
https://sourceware.org/ml/gdb-patches/2014-02/msg00636.html

In this case, remote.c needs to be modified to return TARGET_XFER_EOF instead
of TARGET_XFER_OK or TARGET_XFER_UNAVAILABLE when 0 bytes are to be
transferred. 

The proposed fix would be:

--- fsf/remote.c        2017-11-02 15:21:31.358857300 -0500
+++ fix/remote.c        2017-11-02 15:24:35.536391700 -0500
@@ -8264,7 +8264,7 @@ remote_write_bytes_aux (const char *head
   /* Return UNITS_WRITTEN, not TODO_UNITS, in case escape chars caused us to
      send fewer units than we'd planned.  */
   *xfered_len_units = (ULONGEST) units_written;
-  return TARGET_XFER_OK;
+  return (*xfered_len_units != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF;
 }

 /* Write memory data directly to the remote machine.
@@ -8358,7 +8358,7 @@ remote_read_bytes_1 (CORE_ADDR memaddr,
   decoded_bytes = hex2bin (p, myaddr, todo_units * unit_size);
   /* Return what we have.  Let higher layers handle partial reads.  */
   *xfered_len_units = (ULONGEST) (decoded_bytes / unit_size);
-  return TARGET_XFER_OK;
+  return (*xfered_len_units != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF;
 }

 /* Using the set of read-only target sections of remote, read live
@@ -8455,13 +8455,14 @@ remote_read_bytes (struct target_ops *op
              res = remote_xfer_live_readonly_partial (ops, myaddr, memaddr,
                                                       len, unit_size,
xfered_len);
              if (res == TARGET_XFER_OK)
-               return TARGET_XFER_OK;
+               return (*xfered_len != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF;
              else
                {
                  /* No use trying further, we know some memory starting
                     at MEMADDR isn't available.  */
                  *xfered_len = len;
-                 return TARGET_XFER_UNAVAILABLE;
+                 return (*xfered_len != 0) ?
+                   TARGET_XFER_UNAVAILABLE : TARGET_XFER_EOF;
                }
            }

@@ -10386,7 +10387,7 @@ remote_write_qxfer (struct target_ops *o
   unpack_varlen_hex (rs->buf, &n);

   *xfered_len = n;
-  return TARGET_XFER_OK;
+  return (*xfered_len != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF;
 }

 /* Read OBJECT_NAME/ANNEX from the remote target using a qXfer packet.
@@ -10687,7 +10688,7 @@ remote_xfer_partial (struct target_ops *
   strcpy ((char *) readbuf, rs->buf);

   *xfered_len = strlen ((char *) readbuf);
-  return TARGET_XFER_OK;
+  return (*xfered_len != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF;
 }

 /* Implementation of to_get_memory_xfer_limit.  */

-- 
You are receiving this mail because:
You are on the CC list for the bug.