linux-trace-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com>
Cc: linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH v2 28/87] trace-cmd library: Move CPU flyrecord trace metadata into the buffer option, for trace file version 7
Date: Tue, 17 Aug 2021 11:40:38 -0400	[thread overview]
Message-ID: <20210817114038.2a163dc2@oasis.local.home> (raw)
In-Reply-To: <20210729050959.12263-29-tz.stoyanov@gmail.com>

On Thu, 29 Jul 2021 08:09:00 +0300
"Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com> wrote:

> Extended the BUFFER trace option in trace file version 7 with CPU
> flyrecord trace metadata. In the version 6 of the trace file, this metadata
> is located at several places in the file. Part of the metadata is only for
> the top trace instance, thus limiting per-instance configuration. Moving
> all CPU trace related metadata in the BUFFER option simplifies the parsing
> and makes per-instance configuration more flexible. In the new file
> structure, the top instance is treated as any other instances.
> The format of the extended BUFFER option is:
>  - offset of the buffer in the trace file
>  - name of the buffer
>  - trace clock, used in this buffer for events timestamps
>  - count of CPUs with trace data
>  - array, describing each CPU with trace data:
>    - CPU id
>    - offset of CPU trace data in the trace file
>    - size of the recorded CPU trace data
> 
> Signed-off-by: Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com>
> ---
>  lib/trace-cmd/trace-output.c | 171 ++++++++++++++++++++++++++---------
>  1 file changed, 127 insertions(+), 44 deletions(-)
> 
> diff --git a/lib/trace-cmd/trace-output.c b/lib/trace-cmd/trace-output.c
> index d60bd456..21555d6d 100644
> --- a/lib/trace-cmd/trace-output.c
> +++ b/lib/trace-cmd/trace-output.c
> @@ -1826,7 +1826,7 @@ int tracecmd_append_options(struct tracecmd_output *handle)
>  }
>  
>  static struct tracecmd_option *
> -add_buffer_option(struct tracecmd_output *handle, const char *name, int cpus)
> +add_buffer_option_v6(struct tracecmd_output *handle, const char *name, int cpus)
>  {
>  	struct tracecmd_option *option;
>  	char *buf;
> @@ -1876,8 +1876,11 @@ int tracecmd_write_buffer_info(struct tracecmd_output *handle)
>  	struct tracecmd_option *option;
>  	struct tracecmd_buffer *buf;
>  
> +	if (handle->file_version >= 7)
> +		return 0;
> +
>  	list_for_each_entry(buf, &handle->buffers, list) {
> -		option = add_buffer_option(handle, buf->name, buf->cpus);
> +		option = add_buffer_option_v6(handle, buf->name, buf->cpus);
>  		if (!option)
>  			return -1;
>  		buf->option = option;
> @@ -1938,6 +1941,98 @@ int tracecmd_write_cmdlines(struct tracecmd_output *handle)
>  	return 0;
>  }
>  
> +static char *get_clock(struct tracecmd_output *handle)
> +{
> +	struct tracefs_instance *inst;
> +
> +	if (handle->trace_clock)
> +		return handle->trace_clock;
> +
> +	/*
> +	 * If no clock is set on this handle, get the trace clock of
> +	 * the top instance in the handle's tracing dir
> +	 */
> +	inst = tracefs_instance_alloc(handle->tracing_dir, NULL);
> +	if (!inst)
> +		return NULL;
> +	handle->trace_clock = tracefs_get_clock(inst);
> +	tracefs_instance_free(inst);
> +	return handle->trace_clock;
> +}
> +
> +__hidden struct tracecmd_option *
> +out_add_buffer_option_v7(struct tracecmd_output *handle, const char *name,
> +			 unsigned short id, unsigned long long data_offset,
> +			 int cpus, struct data_file_write *cpu_data)


Why is this hidden, and not static?

I noticed that you are using "out_" as a normal prefix. Is that really
needed?


> +{
> +	struct tracecmd_option *option;
> +	int i, j = 0, k = 0;
> +	int *cpu_ids = NULL;
> +	struct iovec *vect;
> +	char *clock;
> +
> +	if (handle->file_version < 7)
> +		return NULL;
> +
> +	clock = get_clock(handle);
> +
> +	/* Buffer flyrecord option, v7:

Nit, for consistency, use:

	/*
	 * Buffer flyrecord option, v7:

> +	 *  - trace data offset in the file

Is the above the "meta data". If so, perhaps we should call it that.

> +	 *  - buffer name
> +	 *  - buffer clock
> +	 *  - CPU count
> +	 *  - for each CPU:
> +	 *    - CPU id
> +	 *    - CPU trace data offset in the file
> +	 *    - CPU trace data size
> +	 */
> +
> +	/* Buffer latency option, v7:
> +	 *  - trace data offset in the file

Is this the actual latency?

> +	 *  - buffer name
> +	 *  - buffer clock
> +	 */
> +
> +	vect = calloc(5 + (cpus * 3), sizeof(struct iovec));

Comment what these magic numbers are: 5 and 3.

-- Steve

> +	if (!vect)
> +		return NULL;
> +	if (cpus) {
> +		cpu_ids = calloc(cpus, sizeof(int));
> +		if (!cpu_ids) {
> +			free(vect);
> +			return NULL;
> +		}
> +	}
> +	vect[j].iov_base = (void *) &data_offset;
> +	vect[j++].iov_len = 8;
> +	vect[j].iov_base = (void *) name;
> +	vect[j++].iov_len = strlen(name) + 1;
> +	vect[j].iov_base = (void *) clock;
> +	vect[j++].iov_len = strlen(clock) + 1;
> +	if (id == TRACECMD_OPTION_BUFFER) {
> +		vect[j].iov_base = (void *) &k;
> +		vect[j++].iov_len = 4;
> +		for (i = 0; i < cpus; i++) {
> +			if (!cpu_data[i].file_size)
> +				continue;
> +			cpu_ids[i] = i;
> +			vect[j].iov_base = &cpu_ids[i];
> +			vect[j++].iov_len = 4;
> +			vect[j].iov_base = &cpu_data[i].data_offset;
> +			vect[j++].iov_len = 8;
> +			vect[j].iov_base = &cpu_data[i].write_size;
> +			vect[j++].iov_len = 8;
> +			k++;
> +		}
> +	}
> +
> +	option = tracecmd_add_option_v(handle, id, vect, j);
> +	free(vect);
> +	free(cpu_ids);
> +
> +	return option;
> +}
> +

  reply	other threads:[~2021-08-17 15:41 UTC|newest]

Thread overview: 118+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-29  5:08 [PATCH v2 00/87] Trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 01/87] trace-cmd library: Read option id with correct endian Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 02/87] trace-cmd report: Fix typos in error messages Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 03/87] tarce-cmd library: Fix version string memory leak Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 04/87] trace-cmd library: Fixed a memory leak on input handler close Tzvetomir Stoyanov (VMware)
2021-07-29 19:36   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 05/87] trace-cmd library: Fix possible memory corruption on processing a trace buffer Tzvetomir Stoyanov (VMware)
2021-07-29 19:39   ` Steven Rostedt
2021-07-29 19:52   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 06/87] trace-cmd library: Add constructor and destructor Tzvetomir Stoyanov (VMware)
2021-07-29 20:06   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 07/87] trace-cmd library: Add cache functionality to network message handler Tzvetomir Stoyanov (VMware)
2021-07-29 20:33   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 08/87] trace-cmd library: Add support for compression algorithms Tzvetomir Stoyanov (VMware)
2021-07-29 21:02   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 09/87] trace-cmd list: Show supported " Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 10/87] trace-cmd library: Internal helpers for compressing data Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 11/87] trace-cmd library: Internal helpers for uncompressing data Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 12/87] trace-cmd library: Define trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 13/87] trace-cmd library: Refactor APIs for creating output handler Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 14/87] trace-cmd library: Reuse within the library the function that checks file state Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 15/87] trace-cmd library: New API to get the version of output handler Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 16/87] trace-cmd library: Inherit compression algorithm from input file Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 17/87] trace-cmd library: New API to configure compression on an output handler Tzvetomir Stoyanov (VMware)
2021-08-05 21:15   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 18/87] trace-cmd record: Add compression to the trace context Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 19/87] trace-cmd library: Write compression header in the trace file Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 20/87] trace-cmd library: Compress part of " Tzvetomir Stoyanov (VMware)
2021-08-05 21:27   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 21/87] trace-cmd library: Add internal helper functon for writing headers before file sections Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 22/87] trace-cmd library: Write header " Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 23/87] trace-cmd library: Refactor the logic for writing trace data in the file Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 24/87] trace-cmd library: Add local helper function for data compression Tzvetomir Stoyanov (VMware)
2021-08-17 14:53   ` Steven Rostedt
2021-07-29  5:08 ` [PATCH v2 25/87] trace-cmd library: Compress the trace data Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 26/87] tarce-cmd library: Add multiple options sections in trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:08 ` [PATCH v2 27/87] trace-cmd library: Do not write CPU count section in trace files " Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 28/87] trace-cmd library: Move CPU flyrecord trace metadata into the buffer option, for trace file " Tzvetomir Stoyanov (VMware)
2021-08-17 15:40   ` Steven Rostedt [this message]
2021-09-02 13:20     ` Tzvetomir Stoyanov
2021-07-29  5:09 ` [PATCH v2 29/87] trace-cmd record: Append trace options after the trace data are written Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 30/87] trace-cmd library: Add section header before flyrecord trace data Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 31/87] trace-cmd library: Fit CPU latency trace data in the new trace file version 7 format Tzvetomir Stoyanov (VMware)
2021-08-17 15:44   ` Steven Rostedt
2021-09-02 12:48     ` Tzvetomir Stoyanov
2021-08-19 19:10   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 32/87] trace-cmd library: Do not write CPUs with empty trace data Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 33/87] trace-cmd library: Add macro to check file state on reading Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 34/87] trace-cmd library: Introduce sections in trace file reading logic Tzvetomir Stoyanov (VMware)
2021-08-19 17:53   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 35/87] trace-cmd library: Initialize internal sections database on file read Tzvetomir Stoyanov (VMware)
2021-08-19 17:57   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 36/87] trace-cmd library: Use sections database when reading parts of the trace file Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 37/87] trace-cmd library: Set log size to the input tep handler when it is read from the file Tzvetomir Stoyanov (VMware)
2021-08-19 18:01   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 38/87] trace-cmd library: Fix possible memory leak in read_ftrace_files() Tzvetomir Stoyanov (VMware)
2021-08-19 18:07   ` Steven Rostedt
2021-08-19 18:08   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 39/87] trace-cmd library: Fix possible memory leak in read_event_files() Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 40/87] trace-cmd library: Fix possible memory leak in read_proc_kallsyms() Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 41/87] trace-cmd library: Fix possible memory leak in read_ftrace_printk() Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 42/87] trace-cmd library: Fix possible memory leak in read_and_parse_cmdlines() Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 43/87] trace-cmd library: Track maximum CPUs count in input handler Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 44/87] trace-cmd library: Set input handler default values in allocation function Tzvetomir Stoyanov (VMware)
2021-08-19 18:11   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 45/87] trace-cmd library: Read headers from trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 46/87] tarce-cmd library: Do not use local variables when reading CPU stat option Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 47/87] trace-cmd library: Read handle header and compression of the option section Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 48/87] trace-cmd library: Read extended BUFFER option Tzvetomir Stoyanov (VMware)
2021-08-19 18:54   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 49/87] trace-cmd library: Handle the extended DONE option Tzvetomir Stoyanov (VMware)
2021-08-19 19:13   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 50/87] trace-cmd library: Read compression header Tzvetomir Stoyanov (VMware)
2021-08-19 19:15   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 51/87] trace-cmd library: Extend the input handler with trace data decompression context Tzvetomir Stoyanov (VMware)
2021-08-19 19:18   ` Steven Rostedt
2021-09-02 12:46     ` Tzvetomir Stoyanov
2021-07-29  5:09 ` [PATCH v2 52/87] trace-cmd library: Initialize CPU data decompression logic Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 53/87] trace-cmd library: Initialize CPU data for reading from version 7 trace files Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 54/87] trace-cmd library: Add logic for in-memory decompression Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 55/87] trace-cmd library: Handle latency trace in version 7 files Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 56/87] trace-cmd library: Handle buffer trace data init for " Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 57/87] trace-cmd report: Use the new latency API to read data Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 58/87] trace-cmd report: Close input file handlers on exit Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 59/87] trace-cmd report: Do not print empty buffer name Tzvetomir Stoyanov (VMware)
2021-08-19 19:21   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 60/87] trace-cmd report: Init the top trace instance earlier Tzvetomir Stoyanov (VMware)
2021-08-19 19:22   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 61/87] trace-cmd: Call additional APIs when creating trace file Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 62/87] trace-cmd dump: Add helpers for processing trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 63/87] trace-cmd dump: Print compression header Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 64/87] trace-cmd dump: Add helpers for processing trace file sections Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 65/87] trace-cmd dump: Read recursively all options sections Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 66/87] trace-cmd dump: Read extended BUFFER option Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 67/87] trace-cmd dump: Dump sections Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 68/87] trace-cmd dump: Dump trace file version 7 Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 69/87] trace-cmd dump: Dump sections content Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 70/87] trace-cmd dump: Add new argument --sections Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 71/87] trace-cmd dump: Align better the output of flyrecord dump Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 72/87] trace-cmd library: Add zlib compression algorithm Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 73/87] trace-cmd library: Reuse local function that writes to output handler Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 74/87] trace-cmd library: Use output handler when copying data from input file Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 75/87] trace-cmd library: Handle version 7 files when copying headers between files Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 76/87] tarce-cmd library: Copy CPU count between trace files Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 77/87] tarce-cmd library: New API to copy buffer description " Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 78/87] tarce-cmd library: New API to copy options " Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 79/87] tarce-cmd library: New API to copy trace data " Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 80/87] trace-cmd library: Extend tracecmd_copy() API Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 81/87] trace-cmd library: Set correct CPU to the record, retrieved with tracecmd_peek_data Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 82/87] trace-cmd: Add new subcommand "convert" Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 83/87] trace-cmd report: Add new parameters for version 7 trace files Tzvetomir Stoyanov (VMware)
2021-08-19 19:26   ` Steven Rostedt
2021-07-29  5:09 ` [PATCH v2 84/87] trace-cmd: Update bash completion Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 85/87] tarce-cmd: Man page for "trace-cmd convert" Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 86/87] tarce-cmd: Update record man page Tzvetomir Stoyanov (VMware)
2021-07-29  5:09 ` [PATCH v2 87/87] trace-cmd: Document trace file version 7 Tzvetomir Stoyanov (VMware)
2021-08-19 19:33   ` Steven Rostedt
2021-09-02 13:07     ` Tzvetomir Stoyanov

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=20210817114038.2a163dc2@oasis.local.home \
    --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).