This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
RE: [PATCH] new MI command for pattern filling of memory regions
- From: Giuseppe MONTALTO <giuseppe dot montalto at st dot com>
- To: Tom Tromey <tromey at redhat dot com>, "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Cc: Giuseppe MONTALTO <giuseppe dot montalto at st dot com>
- Date: Tue, 12 Jun 2012 11:07:16 +0200
- Subject: RE: [PATCH] new MI command for pattern filling of memory regions
- References: <76FE3225DF13124EA2D05B290B624C95E65F2D466E@SAFEX1MAIL1.st.com> <87sjf9rzr0.fsf@fleche.redhat.com> <76FE3225DF13124EA2D05B290B624C95E65FECE356@SAFEX1MAIL1.st.com> <871umsqgu8.fsf@fleche.redhat.com> <76FE3225DF13124EA2D05B290B624C95E65FECE48F@SAFEX1MAIL1.st.com> <87397624b9.fsf@fleche.redhat.com>
Hi!
Sorry to bother you, but I received no feedback on the patch after three weeks; is there any chance
for it to be reviewed on time for 7.5?
> -----Original Message-----
> From: Giuseppe MONTALTO
> Sent: Friday, May 25, 2012 4:35 PM
> To: 'Tom Tromey'
> Cc: 'gdb-patches@sourceware.org'
> Subject: 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