This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: [PATCH] new MI command for pattern filling of memory regions


I reworked the patch, including the test script;
Accordingly with your feedback, the sscanf cycle is now performed only once.

Here's the new patch, please let me know if there is anything more to change.

Regards,
	Giuseppe

gdb/ChangeLog                           |    5 ++
 gdb/mi/mi-main.c                        |   50 +++++++++++++++++----
 gdb/testsuite/ChangeLog                 |    4 ++
 gdb/testsuite/gdb.mi/mi-fill-memory.exp |   72 +++++++++++++++++++++++++++++++
 4 files changed, 122 insertions(+), 9 deletions(-)

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 309473e..9ec8882 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,8 @@
+2012-05-09  Giuseppe Montalto  <giuseppe.montalto@st.com>
+
+	* mi/mi-main.c  (mi_cmd_data_write_memory): additional
+	parameter for pattern filling of memory regions
+
 2012-05-09  Pedro Alves  <palves@redhat.com>
 
 	* target.c (set_maintenance_target_async_permitted): Rename to ...
diff --git a/gdb/mi/mi-main.c b/gdb/mi/mi-main.c
index 90fb624..7315848 100644
--- a/gdb/mi/mi-main.c
+++ b/gdb/mi/mi-main.c
@@ -1658,7 +1658,8 @@ mi_cmd_data_write_memory (char *command, char **argv, int argc)
 /* Implementation of the -data-write-memory-bytes command.
 
    ADDR: start address
-   DATA: string of bytes to write at that address.  */
+   DATA: string of bytes to write at that address
+   COUNT: number of bytes to be filled (decimal integer).  */
 
 void
 mi_cmd_data_write_memory_bytes (char *command, char **argv, int argc)
@@ -1666,27 +1667,58 @@ mi_cmd_data_write_memory_bytes (char *command, char **argv, int argc)
   CORE_ADDR addr;
   char *cdata;
   gdb_byte *data;
-  int len, r, i;
+  gdb_byte *databuf;
+  int len, r, i, steps, reminder;
+  long int count, j;
   struct cleanup *back_to;
 
-  if (argc != 2)
-    error (_("Usage: ADDR DATA."));
+  if (argc != 2 && argc != 3)
+    error (_("Usage: ADDR DATA [COUNT]."));
 
   addr = parse_and_eval_address (argv[0]);
   cdata = argv[1];
-  len = strlen (cdata)/2;
+  len = strlen (cdata) / 2;
+  if (argc == 3)
+    count = strtol(argv[2], NULL, 10);
+  else
+    count = len;
 
-  data = xmalloc (len);
-  back_to = make_cleanup (xfree, data);
+  databuf = xmalloc (len * sizeof(gdb_byte));
 
   for (i = 0; i < len; ++i)
     {
       int x;
       sscanf (cdata + i * 2, "%02x", &x);
-      data[i] = (gdb_byte) x;
+      databuf[i] = (gdb_byte)x;
     }
 
-  r = target_write_memory (addr, data, len);
+  if (len < count)
+    {
+      /* pattern is made of less bytes than count: 
+         repeat pattern to fill memory.  */
+      data = xmalloc (count);
+      back_to = make_cleanup (xfree, data);
+    
+      steps = count / len;
+      reminder = count % len; /* there may be some spare bytes.  */
+      for (j = 0; j < steps; j++)
+        memcpy (data + j * len, databuf, (size_t) len);
+
+      if (reminder > 0) /* copy spare bytes too.  */
+        memcpy (data + steps * len, databuf, (size_t) reminder);
+    }
+  else
+    {
+      /* pattern is longer than (or equal to) count: 
+         just copy len bytes.  */
+      data = xmalloc (len);
+      back_to = make_cleanup (xfree, data);
+      memcpy (data, databuf, (size_t) len);
+    }
+
+  xfree(databuf);
+
+  r = target_write_memory (addr, data, count);
   if (r != 0)
     error (_("Could not write memory"));
 
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 43549bf..a49274f 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,7 @@
+2012-05-09  Giuseppe Montalto  <giuseppe.montalto@st.com>
+
+	* gdb.mi/mi-fill-memory.exp: New test.
+
 2012-05-08  Maciej W. Rozycki  <macro@codesourcery.com>
 
 	* gdb.mi/mi-var-display.exp: Check for the existence of $fp
diff --git a/gdb/testsuite/gdb.mi/mi-fill-memory.exp b/gdb/testsuite/gdb.mi/mi-fill-memory.exp
new file mode 100644
index 0000000..b6961b1
--- /dev/null
+++ b/gdb/testsuite/gdb.mi/mi-fill-memory.exp
@@ -0,0 +1,72 @@
+# Copyright (C) 2012 Free Software Foundation, Inc.
+# Copyright (C) 2012 STMicroelectronics
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+#
+# test basic Machine interface (MI) operations
+#
+# Verify that, using the MI, we can load a program and do
+# other basic things that are used by all test files through  mi_gdb_exit,
+# mi_gdb_start, mi_delete_breakpoints, mi_gdb_reinitialize_dir and
+# mi_gdb_load, so we can safely use those.
+#
+# The goal is not to test gdb functionality, which is done by other tests,
+# but the command syntax and correct output response to MI operations.
+# 
+# added for testing the -data-write-memory-bytes MI command enhancements
+#
+
+load_lib mi-support.exp
+set MIFLAGS "-i=mi"
+
+gdb_exit
+if [mi_gdb_start] {
+    continue
+}
+
+set testfile "mi-read-memory"
+set srcfile ${testfile}.c
+set binfile ${objdir}/${subdir}/${testfile}
+if  { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable {debug additional_flags=-DFAKEARGV}] != "" } {
+     untested mi-read-memory.exp
+     return -1
+}
+
+
+mi_run_to_main
+mi_next_to "main" "" "mi-read-memory.c" "20" "next at main"
+
+mi_gdb_test "1-data-write-memory-bytes"\
+	"1\\\^error,msg=\"Usage: ADDR DATA \\\[COUNT\\\]\.\""\
+	"no arguments"
+
+mi_gdb_test "2-data-write-memory-bytes 8"\
+	"2\\\^error,msg=\"Usage: ADDR DATA \\\[COUNT\\\]\.\""\
+	"one argument missing"
+
+mi_gdb_test "3-data-write-memory-bytes \$pc ab"\
+	"3\\\^done" \
+	"memory successfully written"
+
+mi_gdb_test "4-data-write-memory-bytes \$pc ab 8"\
+	"4\\\^done" \
+	"memory successfully filled (8 bytes)"
+
+mi_gdb_test "5-interpreter-exec console \"x \$pc\"" \
+    ".*0xabababab.*" \
+    "pattern correctly read from memory"
+
+mi_gdb_exit
+return 0



> -----Original Message-----
> From: Giuseppe MONTALTO
> Sent: Friday, May 11, 2012 6:06 PM
> To: 'Tom Tromey'
> Cc: gdb-patches@sourceware.org
> Subject: RE: [PATCH] new MI command for pattern filling of memory
> regions
> 
> Thanks,
> 
> my answers are interleaved below.
> 
> > -----Original Message-----
> > From: Tom Tromey [mailto:tromey@redhat.com]
> > Sent: Friday, May 11, 2012 4:14 PM
> > To: Giuseppe MONTALTO
> > Cc: gdb-patches@sourceware.org
> > Subject: Re: [PATCH] new MI command for pattern filling of memory
> > regions
> >
> > The indentation is wrong here and elsewhere in the patch.
> > There are missing spaces as well.
> > See the GNU Coding Standards for details.
> 
> I'll try to better format the code.
> 
> >
> > I don't think the cast here, or other casts in the patch, are
> > necessary.
> 
> Consider them gone.
> 
> >
> > Giuseppe> +	  /* pattern is made of less bytes than count:
> > Giuseppe> +	     repeat pattern to fill memory.  */
> > Giuseppe> +	  data = xmalloc (count/len);
> >
> > This allocates count/len bytes...
> >
> > Giuseppe> +	  steps = count/len;
> > Giuseppe> +	  for (j = 0; j < steps; j++)
> > Giuseppe> +		{
> > Giuseppe> +		  for (i = 0; i < len; ++i)
> > Giuseppe> +			{
> > Giuseppe> +			  int x;
> > Giuseppe> +			  sscanf (cdata + i * 2, "%02x", &x);
> > Giuseppe> +			  data[i + j * len] = (gdb_byte)x;
> >
> > ... but writes 'count' bytes.
> 
> you're right; I was supposed to allocate 'count' bytes.
> 
> Incidentally, there's a worse error there:
> is that I'm counting the cycles basing on the result of an integer
> division, hence not taking into account the spare bytes!
> 
> I'll fix that too...
> 
> 	Giuseppe


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]