linux-trace-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: linux-trace-devel@vger.kernel.org
Cc: "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com>
Subject: [PATCH 2/8] trace-cmd: Create API tracecmd_read_pre_headers()
Date: Mon, 01 Mar 2021 09:37:26 -0500	[thread overview]
Message-ID: <20210301143856.948525416@goodmis.org> (raw)
In-Reply-To: 20210301143724.540985351@goodmis.org

From: "Steven Rostedt (VMware)" <rostedt@goodmis.org>

The trace-cmd restore operation can create a partial header to read the
trace event formats and kallsyms and other data into a stand alone header
before it has access to the cpu data. Then it will also read this header to
put together a broken trace, and it reads the header that does not have the
cpu data attached to it.

In order to handle this case, it needs a way to read the headers but stop
short of reading the CPU information. That requires breaking up
tracecmd_read_headers() with just stopping short of adding the cpu data.

A new API is added called tracecmd_read_pre_headers() that does exactly that.

Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
---
 .../include/private/trace-cmd-private.h       |  1 +
 lib/trace-cmd/trace-input.c                   | 25 ++++++++++++++++---
 tracecmd/trace-restore.c                      |  2 +-
 3 files changed, 24 insertions(+), 4 deletions(-)

diff --git a/lib/trace-cmd/include/private/trace-cmd-private.h b/lib/trace-cmd/include/private/trace-cmd-private.h
index fc968cc9efe1..c7ef3af7c8f7 100644
--- a/lib/trace-cmd/include/private/trace-cmd-private.h
+++ b/lib/trace-cmd/include/private/trace-cmd-private.h
@@ -153,6 +153,7 @@ typedef void (*tracecmd_handle_init_func)(struct tracecmd_input *handle,
 struct tracecmd_input *tracecmd_alloc(const char *file, int flags);
 struct tracecmd_input *tracecmd_alloc_fd(int fd, int flags);
 void tracecmd_ref(struct tracecmd_input *handle);
+int tracecmd_read_pre_headers(struct tracecmd_input *handle);
 int tracecmd_read_headers(struct tracecmd_input *handle);
 int tracecmd_get_parsing_failures(struct tracecmd_input *handle);
 int tracecmd_long_size(struct tracecmd_input *handle);
diff --git a/lib/trace-cmd/trace-input.c b/lib/trace-cmd/trace-input.c
index 9ef7b9f16951..9e1a44540201 100644
--- a/lib/trace-cmd/trace-input.c
+++ b/lib/trace-cmd/trace-input.c
@@ -772,14 +772,17 @@ static int read_cpus(struct tracecmd_input *handle)
 }
 
 /**
- * tracecmd_read_headers - read the header information from trace.dat
+ * tracecmd_read_pre_headers - read the header information from trace.dat
  * @handle: input handle for the trace.dat file
  *
  * This reads the trace.dat file for various information. Like the
  * format of the ring buffer, event formats, ftrace formats, kallsyms
- * and printk.
+ * and printk, but stops before reading cpu and options.
+ *
+ * This is needed by the restore operation where the header does not
+ * have the CPU information yet.
  */
-int tracecmd_read_headers(struct tracecmd_input *handle)
+int tracecmd_read_pre_headers(struct tracecmd_input *handle)
 {
 	int ret;
 
@@ -815,6 +818,22 @@ int tracecmd_read_headers(struct tracecmd_input *handle)
 		return -1;
 	handle->file_state = TRACECMD_FILE_CMD_LINES;
 
+	return 0;
+}
+
+/**
+ * tracecmd_read_headers - read the header information from trace.dat
+ * @handle: input handle for the trace.dat file
+ *
+ * This reads the trace.dat file for various information. Like the
+ * format of the ring buffer, event formats, ftrace formats, kallsyms
+ * and printk.
+ */
+int tracecmd_read_headers(struct tracecmd_input *handle)
+{
+	if (tracecmd_read_pre_headers(handle))
+		return -1;
+
 	if (read_cpus(handle) < 0)
 		return -1;
 	handle->file_state = TRACECMD_FILE_CPU_COUNT;
diff --git a/tracecmd/trace-restore.c b/tracecmd/trace-restore.c
index 13f803053582..bf6940991178 100644
--- a/tracecmd/trace-restore.c
+++ b/tracecmd/trace-restore.c
@@ -122,7 +122,7 @@ void trace_restore (int argc, char **argv)
 		if (!ihandle)
 			die("error reading file %s", input);
 		/* make sure headers are ok */
-		if (tracecmd_read_headers(ihandle) < 0)
+		if (tracecmd_read_pre_headers(ihandle) < 0)
 			die("error reading file %s headers", input);
 
 		handle = tracecmd_copy(ihandle, output);
-- 
2.30.0



  parent reply	other threads:[~2021-03-01 14:39 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-01 14:37 [PATCH 0/8] trace-cmd: Fixes for trace-cmd restore Steven Rostedt
2021-03-01 14:37 ` [PATCH 1/8] trace-cmd restore: Fix to add saved cmdlines after calling tracecmd_create_init_file_override() Steven Rostedt
2021-03-01 14:37 ` Steven Rostedt [this message]
2021-03-02  4:49   ` [PATCH 2/8] trace-cmd: Create API tracecmd_read_pre_headers() Tzvetomir Stoyanov
2021-03-02 14:13     ` Steven Rostedt
2021-03-01 14:37 ` [PATCH 3/8] trace-cmd: Move tracecmd_write_cmdlines() out of tracecmd_append_cpu_data() Steven Rostedt
2021-03-01 14:37 ` [PATCH 4/8] trace-cmd: Move the output state updates into the functions that change the state Steven Rostedt
2021-03-01 14:37 ` [PATCH 5/8] trace-cmd: Move the input " Steven Rostedt
2021-03-01 14:37 ` [PATCH 6/8] trace-cmd output: Set file_state of output handle after copy of headers Steven Rostedt
2021-03-02  8:10   ` Tzvetomir Stoyanov
2021-03-02 14:19     ` Steven Rostedt
2021-03-02 14:51       ` Tzvetomir Stoyanov
2021-03-02 15:48         ` Steven Rostedt
2021-03-02 17:35           ` Tzvetomir Stoyanov
2021-03-02 19:59             ` Steven Rostedt
2021-03-02 14:22     ` Steven Rostedt
2021-03-01 14:37 ` [PATCH 7/8] trace-cmd input: Validate the input handle when copying from it Steven Rostedt
2021-03-01 14:37 ` [PATCH 8/8] trace-cmd input: Add validation updates to the copy of a handle Steven Rostedt

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=20210301143856.948525416@goodmis.org \
    --to=rostedt@goodmis.org \
    --cc=linux-trace-devel@vger.kernel.org \
    --cc=tz.stoyanov@gmail.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;
as well as URLs for NNTP newsgroup(s).