All of lore.kernel.org
 help / color / mirror / Atom feed
From: Lukasz Majewski <lukma@denx.de>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH v3 14/14] log: Add documentation
Date: Tue, 21 Nov 2017 11:19:03 +0100	[thread overview]
Message-ID: <20171121111903.3bc4d64e@jawa> (raw)
In-Reply-To: <20171120223335.45852-15-sjg@chromium.org>

On Mon, 20 Nov 2017 15:33:35 -0700
Simon Glass <sjg@chromium.org> wrote:

> Add documentation for the log system.
> 
> Signed-off-by: Simon Glass <sjg@chromium.org>
> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
> ---
> 
> Changes in v3:
> - Rebase to master
> 
> Changes in v2:
> - Drop the special log() functions from the README
> 
>  doc/README.log | 214
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file
> changed, 214 insertions(+) create mode 100644 doc/README.log
> 
> diff --git a/doc/README.log b/doc/README.log
> new file mode 100644
> index 00000000000..f653fe7d797
> --- /dev/null
> +++ b/doc/README.log
> @@ -0,0 +1,214 @@
> +Logging in U-Boot
> +=================
> +
> +Introduction
> +------------
> +
> +U-Boot's internal operation involves many different steps and
> actions. From +setting up the board to displaying a start-up screen
> to loading an Operating +System, there are many component parts each
> with many actions. +
> +Most of the time this internal detail is not useful. Displaying it
> on the +console would delay booting (U-Boot's primary purpose) and
> confuse users. +
> +But for digging into what is happening in a particular area, or for
> debugging +a problem it is often useful to see what U-Boot is doing
> in more detail than +is visible from the basic console output.
> +
> +U-Boot's logging feature aims to satisfy this goal for both users and
> +developers.
> +
> +
> +Logging levels
> +--------------
> +
> +There are a number logging levels available, in increasing order of
> verbosity: +
> +   LOGL_EMERG	- Printed before U-Boot halts
> +   LOGL_ALERT	- Indicates action must be taken immediate or
> U-Boot will crash
> +   LOGL_CRIT	- Indicates a critical error that will cause
> boot failure
> +   LOGL_ERR	- Indicates an error that may cause boot failure
> +   LOGL_WARNING	- Warning about an unexpected condition
> +   LOGL_NOTE	- Important information about progress
> +   LOGL_INFO	- Information about normal boot progress
> +   LOGL_DEBUG	- Debug information (useful for debugging a
> driver or subsystem)
> +   LOGL_DEBUG_CONTENT	- Debug message showing full message
> content
> +   LOGL_DEBUG_IO	- Debug message showing hardware I/O access
> +
> +
> +Logging category
> +----------------
> +
> +Logging can come from a wide variety of places within U-Boot. Each
> log message +has a category which is intended to allow messages to be
> filtered according to +their source.
> +
> +The following main categories are defined:
> +
> +   LOGC_NONE	- Unknown category (e.g. a debug() statement)
> +   UCLASS_...	- Related to a particular uclass (e.g.
> UCLASS_USB)
> +   LOGC_ARCH	- Related to architecture-specific code
> +   LOGC_BOARD	- Related to board-specific code
> +   LOGC_CORE	- Related to core driver-model support
> +   LOGC_DT	- Related to device tree control
> +
> +
> +Enabling logging
> +----------------
> +
> +The following options are used to enable logging at compile time:
> +
> +   CONFIG_LOG		- Enables the logging system
> +   CONFIG_MAX_LOG_LEVEL - Max log level to build (anything higher is
> compiled
> +				out)
> +   CONFIG_LOG_CONSOLE	- Enable writing log records to the
> console +
> +If CONFIG_LOG is not set, then no logging will be available.
> +
> +The above have SPL versions also, e.g. CONFIG_SPL_MAX_LOG_LEVEL.
> +
> +
> +Using DEBUG
> +-----------
> +
> +U-Boot has traditionally used a #define called DEBUG to enable
> debugging on a +file-by-file basis. The debug() macro compiles to a
> printf() statement if +DEBUG is enabled, and an empty statement if
> not. +
> +With logging enabled, debug() statements are interpreted as logging
> output +with a level of LOGL_DEBUG and a category of LOGC_NONE.
> +
> +The logging facilities are intended to replace DEBUG, but if DEBUG
> is defined +at the top of a file, then it takes precedence. This
> means that debug() +statements will result in output to the console
> and this output will not be +logged.
> +
> +
> +Logging destinations
> +--------------------
> +
> +If logging information goes nowhere then it serves no purpose.
> U-Boot provides +several possible determinations for logging
> information, all of which can be +enabled or disabled independently:
> +
> +   console - goes to stdout
> +
> +
> +Filters
> +-------
> +
> +Filters are attached to log drivers to control what those drivers
> emit. Only +records that pass through the filter make it to the
> driver. +
> +Filters can be based on several criteria:
> +
> +   - maximum log level
> +   - in a set of categories
> +   - in a set of files
> +
> +If no filters are attached to a driver then a default filter is
> used, which +limits output to records with a level less than
> CONFIG_LOG_MAX_LEVEL. +
> +
> +Logging statements
> +------------------
> +
> +The main logging function is:
> +
> +   log(category, level, format_string, ...)
> +
> +Also debug() and error() will generate log records  - these use
> LOG_CATEGORY +as the category, so you should #define this right at
> the top of the source +file to ensure the category is correct.
> +
> +
> +Code size
> +---------
> +
> +Code size impact depends largely on what is enabled. The following
> numbers are +for snow, which is a Thumb-2 board:
> +
> +This series: adds bss +20.0 data +4.0 rodata +4.0 text +44.0

	It would be more informative to add units (KiB, B).

> +CONFIG_LOG: bss -52.0 data +92.0 rodata -635.0 text +1048.0
> +CONFIG_LOG_MAX_LEVEL=7: bss +188.0 data +4.0 rodata +49183.0 text
> +98124.0 +
> +The last option turns every debug() statement into a logging call,
> which +bloats the code hugely. The advantage is that it is then
> possible to enable +all logging within U-Boot.
> +
> +
> +To Do
> +-----
> +
> +There are lots of useful additions that could be made. None of the
> below is +implemented! If you do one, please add a test in
> test/py/tests/test_log.py +
> +Convenience functions to support setting the category:
> +
> +   log_arch(level, format_string, ...) - category LOGC_ARCH
> +   log_board(level, format_string, ...) - category LOGC_BOARD
> +   log_core(level, format_string, ...) - category LOGC_CORE
> +   log_dt(level, format_string, ...) - category LOGC_DT
> +
> +Convenience functions to support a category defined for a single
> file, for +example:
> +
> +   #define LOG_CATEGORY   UCLASS_USB
> +
> +all of these can use LOG_CATEGORY as the category, and a log level
> +corresponding to the function name:
> +
> +   logc(level, format_string, ...)
> +
> +More logging destinations:
> +
> +   device - goes to a device (e.g. serial)
> +   buffer - recorded in a memory buffer
> +
> +Convert debug() statements in the code to log() statements
> +
> +Support making printf() emit log statements a L_INFO level
> +
> +Convert error() statements in the code to log() statements
> +
> +Figure out what to do with BUG(), BUG_ON() and warn_non_spl()
> +
> +Figure out what to do with assert()
> +
> +Add a way to browse log records
> +
> +Add a way to record log records for browsing using an external tool
> +
> +Add commands to add and remove filters
> +
> +Add commands to add and remove log devices
> +
> +Allow sharing of printf format strings in log records to reduce
> storage size +for large numbers of log records
> +
> +Add a command-line option to sandbox to set the default logging level
> +
> +Convert core driver model code to use logging
> +
> +Convert uclasses to use logging with the correct category
> +
> +Consider making log() calls emit an automatic newline, perhaps with
> a logn()
> +   function to avoid that
> +
> +Passing log records through to linux (e.g. via device tree /chosen)
> +
> +Provide a command to access the number of log records generated, and
> the +number dropped due to them being generated before the log system
> was ready. +
> +Add a printf() format string pragma so that log statements are
> checked properly +
> +Enhance the log console driver to show level / category / file / line
> +information
> +
> +Add a command to add new log records and delete existing records.
> +
> +Provide additional log() functions - e.g. logc() to specify the
> category +
> +--
> +Simon Glass <sjg@chromium.org>
> +15-Sep-17

As I said before - this is a great piece of SW.

Reviewed-by: Lukasz Majewski <lukma@denx.de>

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20171121/c7ed448c/attachment.sig>

  reply	other threads:[~2017-11-21 10:19 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-20 22:33 [U-Boot] [PATCH v3 00/14] log: Add a new logging feature Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 01/14] Revert "sandbox: remove os_putc() and os_puts()" Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 02/14] sandbox: Adjust pre-console address to avoid conflict Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 03/14] Revert "sandbox: Drop special case console code for sandbox" Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 04/14] Move debug and logging support to a separate header Simon Glass
2017-11-21  9:41   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 05/14] mtdparts: Correct use of debug() Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 06/14] Drop the log buffer Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 07/14] log: Add an implemention of logging Simon Glass
2017-11-21  9:55   ` Lukasz Majewski
2017-11-24  1:49     ` Simon Glass
2017-11-20 22:33 ` [U-Boot] [PATCH v3 08/14] log: Add a console driver Simon Glass
2017-11-21  9:57   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 09/14] log: Add a 'log level' command Simon Glass
2017-11-21  9:58   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 10/14] log: Add a test command Simon Glass
2017-11-21 10:00   ` Lukasz Majewski
2017-11-30  3:35   ` [U-Boot] [U-Boot,v3,10/14] " Tom Rini
2017-11-30 16:27     ` Simon Glass
2017-11-30 16:38       ` Tom Rini
2017-11-20 22:33 ` [U-Boot] [PATCH v3 11/14] log: Plumb logging into the init sequence Simon Glass
2017-11-21 10:01   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 12/14] log: sandbox: Enable logging Simon Glass
2017-11-21 10:01   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 13/14] log: test: Add a pytest for logging Simon Glass
2017-11-21 10:03   ` Lukasz Majewski
2017-11-20 22:33 ` [U-Boot] [PATCH v3 14/14] log: Add documentation Simon Glass
2017-11-21 10:19   ` Lukasz Majewski [this message]
2018-04-02  8:43     ` Simon Glass

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=20171121111903.3bc4d64e@jawa \
    --to=lukma@denx.de \
    --cc=u-boot@lists.denx.de \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.