From: Pedro Alves <palves@redhat.com>
To: markus.t.metzger@intel.com
Cc: gdb-patches@sourceware.org, markus.t.metzger@gmail.com,
jan.kratochvil@redhat.com, palves@redhat.com,
tromey@redhat.com, kettenis@gnu.org
Subject: Re: [patch v4 09/13] gdbserver, btrace: add generic btrace support
Date: Wed, 28 Nov 2012 20:32:00 -0000 [thread overview]
Message-ID: <50B674CC.6050009@redhat.com> (raw)
In-Reply-To: <1354013351-14791-10-git-send-email-markus.t.metzger@intel.com>
On 11/27/2012 10:49 AM, markus.t.metzger@intel.com wrote:
> From: Markus Metzger <markus.t.metzger@intel.com>
>
> Add support to gdbserver to understand branch trace related packages.
>
> 2012-11-27 Markus Metzger <markus.t.metzger@intel.com>
>
> gdbserver/
> * target.h (struct target_ops): Add btrace ops.
> (target_supports_btrace): New macro.
> (target_btrace_has_changed): New macro.
> (target_enable_btrace): New macro.
> (target_disable_btrace): New macro.
> (target_read_btrace): New macro.
> * gdbthread.h (struct thread_info): Add btrace field.
> * server.c (handle_btrace_general_set): New function.
> (handle_btrace_enable): New function.
> (handle_btrace_disable): New function.
> (handle_general_set): Call handle_btrace_general_set.
> (handle_qxfer_btrace): New function.
> (struct qxfer qxfer_packets[]): Add btrace entry.
> (handle_btrace_query): New function.
> (handle_query): Add btrace to supported query, call handle_btrace_query.
> * inferiors.c (remove_thread): Disable btrace.
>
>
> ---
> gdb/gdbserver/gdbthread.h | 5 ++
> gdb/gdbserver/inferiors.c | 3 +
> gdb/gdbserver/server.c | 167 +++++++++++++++++++++++++++++++++++++++++++++
> gdb/gdbserver/target.h | 33 +++++++++
> 4 files changed, 208 insertions(+), 0 deletions(-)
>
> diff --git a/gdb/gdbserver/gdbthread.h b/gdb/gdbserver/gdbthread.h
> index adc23da..6238ed2 100644
> --- a/gdb/gdbserver/gdbthread.h
> +++ b/gdb/gdbserver/gdbthread.h
> @@ -22,6 +22,8 @@
>
> #include "server.h"
>
> +struct btrace_target_info;
> +
> struct thread_info
> {
> struct inferior_list_entry entry;
> @@ -58,6 +60,9 @@ struct thread_info
> Each item in the list holds the current step of the while-stepping
> action. */
> struct wstep_state *while_stepping;
> +
> + /* Branch trace target information for this thread. */
> + struct btrace_target_info *btrace;
> };
>
> extern struct inferior_list all_threads;
> diff --git a/gdb/gdbserver/inferiors.c b/gdb/gdbserver/inferiors.c
> index 76abaf5..76f5731 100644
> --- a/gdb/gdbserver/inferiors.c
> +++ b/gdb/gdbserver/inferiors.c
> @@ -161,6 +161,9 @@ free_one_thread (struct inferior_list_entry *inf)
> void
> remove_thread (struct thread_info *thread)
> {
> + if (thread->btrace)
> + target_disable_btrace (thread->btrace);
> +
> remove_inferior (&all_threads, (struct inferior_list_entry *) thread);
> free_one_thread (&thread->entry);
> }
> diff --git a/gdb/gdbserver/server.c b/gdb/gdbserver/server.c
> index fae9199..2758d40 100644
> --- a/gdb/gdbserver/server.c
> +++ b/gdb/gdbserver/server.c
> @@ -442,6 +442,82 @@ write_qxfer_response (char *buf, const void *data, int len, int is_more)
>
> /* Handle all of the extended 'Q' packets. */
This comment above should be kept close to the
corresponding function.
>
> +/* Handle btrace enabling. */
> +static int
> +handle_btrace_enable (char *ptid_str)
> +{
> + ptid_t ptid = read_ptid (ptid_str, NULL);
> + struct thread_info *thread = find_thread_ptid (ptid);
> + int err = 1;
> +
> + if (!thread)
> + fprintf (stderr, "No such thread: %s.\n", ptid_str);
> + else if (thread->btrace)
> + fprintf (stderr, "Btrace already enabled for %s.\n", ptid_str);
> + else if (!the_target->enable_btrace)
> + fprintf (stderr, "Target does not support branch tracing.");
> + else
> + {
> + thread->btrace = target_enable_btrace (thread->entry.id);
> + if (!thread->btrace)
> + fprintf (stderr, "Could not enable btrace for %s.\n", ptid_str);
> + else
> + err = 0;
> + }
> +
> + return err;
Instead of printing those errors to stderr, you can pass them
down to GDB, by sending "E.ERRTXT" instead of E01. (See
packet_check_result in remote.c).
> +}
> +
> +/* Handle btrace disabling. */
> +static int
> +handle_btrace_disable (char *ptid_str)
> +{
> + ptid_t ptid = read_ptid (ptid_str, NULL);
> + struct thread_info *thread = find_thread_ptid (ptid);
> + int err = 1;
> +
> + if (!thread)
> + fprintf (stderr, "No such thread: %s.\n", ptid_str);
> + else if (!thread->btrace)
> + fprintf (stderr, "Branch tracing not enabled for %s.\n", ptid_str);
> + else if (!the_target->disable_btrace)
> + fprintf (stderr, "Target does not support branch tracing.");
> + else if (target_disable_btrace (thread->btrace) != 0)
> + fprintf (stderr, "Could not disable branch tracing for %s.\n", ptid_str);
> + else
> + {
> + thread->btrace = NULL;
> + err = 0;
> + }
> +
> + return err;
> +}
Same comment about handle_btrace_enable/handle_btrace_disable
being almost identical.
> +
> +/* Handle the "Qbtrace" packet. */
> +static int
> +handle_btrace_general_set (char *own_buf)
> +{
> + char *operation = own_buf + strlen ("Qbtrace:");
> + int err = 1;
> +
> + if (strncmp ("Qbtrace:", own_buf, strlen ("Qbtrace:")) != 0)
> + return 0;
> +
> + if (strncmp ("on:", operation, strlen ("on:")) == 0)
> + err = handle_btrace_enable (operation + strlen ("on:"));
> + else if (strncmp ("off:", operation, strlen ("off:")) == 0)
> + err = handle_btrace_disable (operation + strlen ("off:"));
> + else
> + fprintf (stderr, "Unknown btrace operation: %s. Use on or off.\n", own_buf);
> +
> + if (err)
> + write_enn (own_buf);
> + else
> + write_ok (own_buf);
> +
> + return 1;
> +}
> +
> static void
> handle_general_set (char *own_buf)
> {
> @@ -596,6 +672,9 @@ handle_general_set (char *own_buf)
> return;
> }
>
> + if (handle_btrace_general_set (own_buf))
> + return;
> +
> /* Otherwise we didn't know what packet it was. Say we didn't
> understand it. */
> own_buf[0] = 0;
> @@ -1295,6 +1374,55 @@ handle_qxfer_fdpic (const char *annex, gdb_byte *readbuf,
> return (*the_target->read_loadmap) (annex, offset, readbuf, len);
> }
>
> +/* Handle branch trace data transfer. */
> +static int
> +handle_qxfer_btrace (const char *annex,
> + gdb_byte *readbuf, const gdb_byte *writebuf,
> + ULONGEST offset, LONGEST len)
> +{
> + static struct buffer cache;
> + struct thread_info *thread;
> + ptid_t ptid;
> +
> + if (!the_target->read_btrace || writebuf)
> + return -2;
> +
> + if (!target_running () || !annex)
> + return -1;
> +
> + ptid = read_ptid ((char *) annex, NULL);
> + thread = find_thread_ptid (ptid);
> + if (!thread)
> + {
> + fprintf (stderr, "No such thread: %s\n", annex);
> + return -1;
> + }
> + else if (!thread->btrace)
> + {
> + fprintf (stderr, "Btrace not enabled for %s\n", annex);
> + return -1;
> + }
> +
> + if (!offset)
> + {
> + buffer_free (&cache);
> +
> + target_read_btrace (thread->btrace, &cache);
> + }
> + else if (offset >= cache.used_size)
> + {
> + buffer_free (&cache);
> + return 0;
I think this should be -1.
> + }
> +
> + if (len > (cache.used_size - offset))
> + len = cache.used_size - offset;
Unnecessary parens.
> +
> + memcpy (readbuf, cache.buffer + offset, len);
> +
> + return len;
> +}
> +
> static const struct qxfer qxfer_packets[] =
> {
> { "auxv", handle_qxfer_auxv },
> @@ -1308,6 +1436,7 @@ static const struct qxfer qxfer_packets[] =
> { "statictrace", handle_qxfer_statictrace },
> { "threads", handle_qxfer_threads },
> { "traceframe-info", handle_qxfer_traceframe_info },
> + { "btrace", handle_qxfer_btrace },
> };
Please keep the list sorted.
--
Pedro Alves
next prev parent reply other threads:[~2012-11-28 20:32 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-27 10:50 [patch v4 00/13] branch tracing support for Atom markus.t.metzger
2012-11-27 10:49 ` [patch v4 01/13] disas: add precise instructions flag markus.t.metzger
2012-11-27 16:49 ` Pedro Alves
2012-11-27 16:52 ` Jan Kratochvil
2012-11-27 16:56 ` Pedro Alves
2012-11-27 17:26 ` Metzger, Markus T
2012-11-27 17:33 ` Pedro Alves
2012-11-28 14:44 ` Metzger, Markus T
2012-11-27 10:49 ` [patch v4 03/13] cli, btrace: add btrace cli markus.t.metzger
2012-11-27 21:53 ` Tom Tromey
2012-11-30 15:09 ` Metzger, Markus T
2012-11-30 15:16 ` Jan Kratochvil
2012-11-30 15:23 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 02/13] thread, btrace: add generic branch trace support markus.t.metzger
2012-11-27 18:32 ` Pedro Alves
2012-11-27 18:38 ` Pedro Alves
2012-11-28 0:41 ` Pedro Alves
2012-11-29 16:39 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 10/13] gdbserver, linux, btrace: add btrace support for linux-low markus.t.metzger
2012-11-28 20:44 ` Pedro Alves
2012-12-05 9:27 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 11/13] test, btrace: add branch tracing tests markus.t.metzger
2012-11-27 10:50 ` [patch v4 09/13] gdbserver, btrace: add generic btrace support markus.t.metzger
2012-11-28 20:32 ` Pedro Alves [this message]
2012-12-04 14:50 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 05/13] linux, btrace: perf_event based branch tracing markus.t.metzger
2012-11-28 17:31 ` Pedro Alves
2012-12-03 14:38 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 13/13] btrace, x86: restrict to Atom markus.t.metzger
2012-11-27 11:19 ` Mark Kettenis
2012-11-27 11:49 ` Metzger, Markus T
2012-11-27 14:42 ` Mark Kettenis
2012-11-27 15:14 ` Metzger, Markus T
2012-11-27 15:32 ` Pedro Alves
2012-11-27 13:05 ` Jan Kratochvil
2012-11-27 14:04 ` Metzger, Markus T
2012-11-27 14:29 ` Jan Kratochvil
2012-11-27 15:14 ` Metzger, Markus T
2012-11-27 15:50 ` Pedro Alves
2012-11-27 15:54 ` Metzger, Markus T
2012-12-06 10:15 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 08/13] remote, btrace: add branch trace remote ops markus.t.metzger
2012-11-28 19:23 ` Pedro Alves
2012-12-04 12:47 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 06/13] linux, i386, amd64: enable btrace for 32bit and 64bit linux native markus.t.metzger
2012-11-28 18:40 ` Pedro Alves
2012-12-03 16:24 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 04/13] configure: add check for perf_event header markus.t.metzger
2012-11-28 10:11 ` Pedro Alves
2012-11-28 14:52 ` Metzger, Markus T
2012-11-28 14:55 ` Pedro Alves
2012-11-27 10:50 ` [patch v4 07/13] xml, btrace: define btrace xml document style markus.t.metzger
2012-11-28 18:53 ` Pedro Alves
2012-12-04 10:35 ` Metzger, Markus T
2012-11-27 10:50 ` [patch v4 12/13] test, btrace: more branch tracing tests markus.t.metzger
2012-11-27 13:11 ` [patch v4 00/13] branch tracing support for Atom Jan Kratochvil
2012-11-27 14:26 ` Metzger, Markus T
2012-11-27 14:32 ` Jan Kratochvil
2012-11-27 14:40 ` Metzger, Markus T
2012-11-27 15:36 ` Jan Kratochvil
2012-11-27 16:17 ` Metzger, Markus T
2012-11-27 16:28 ` Jan Kratochvil
2012-11-27 17:30 ` Metzger, Markus T
2012-11-27 18:31 ` Jan Kratochvil
2012-11-27 18:56 ` Markus Metzger
2012-11-28 19:01 ` Jan Kratochvil
2012-11-29 9:13 ` Metzger, Markus T
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=50B674CC.6050009@redhat.com \
--to=palves@redhat.com \
--cc=gdb-patches@sourceware.org \
--cc=jan.kratochvil@redhat.com \
--cc=kettenis@gnu.org \
--cc=markus.t.metzger@gmail.com \
--cc=markus.t.metzger@intel.com \
--cc=tromey@redhat.com \
/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