Reduce the log overhead when the traces are filtered,
by moving the vsnprintf call from _log() to log_dispatch().
This patch avoids the printf treatment when LOG features is
activated, but trace is filtered, for example when
MAX_LOG_LEVEL=8 and LOG_DEFAULT_LEVEL=6.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Use the console puts functions to output the traces before
the log initialization (when CONFIG_LOG is not activated).
This patch allows to display the first U-Boot traces
(with macro debug) when CONFIG_DEBUG_UART is activated
and not only drop them.
For example for traces in board_f.c requested by the macro debug,
when LOG_DEBUG is defined and CONFIG_LOG is activated.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Update _log function to drop any traces when log is yet initialized:
vsnprintf is no more executed in this case.
This patch allows to reduce the cost for the dropped early debug trace.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Add logging for each part of the boot process, using a new
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Igor Opaniuk <igor.opaniuk@toradex.com>
This is the complement of the existing behavior to match only messages with
a log level less than a threshold. This is primarily useful in conjunction
with LOGFF_DENY.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Without this flag, log filters can only explicitly accept messages.
Allowing denial makes it easier to filter certain subsystems. Unlike
allow-ing filters, deny-ing filters are added to the beginning of the
filter list. This should do the Right Thing most of the time, but it's
less-universal than allowing filters to be inserted anywhere. If this
becomes a problem, then perhaps log_filter_add* should take a filter number
to insert before/after.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This function exposes a way to specify flags when creating a filter.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
These functions are required by "cmd: log: Add commands to manipulate
filters" and "test: Add a test for log filter-*".
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Both these arrays and their members are const. Fixes checkpatch complaint.
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Errors returned should be negative.
Fixes: 45fac9fc18 ("log: Correct missing free() on error in log_add_filter()")
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The log command has led to NULL dereferences if an unknown category name
name was used due to missing entries in the list of category names.
Add compile time checks for the array sizes of log_cat_name and
log_lvl_name to avoid future mishaps.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Some drivers use macro pr_cont() for continuing a message sent via printk.
Hence if we want to convert printk messaging to using the logging system,
we must support continuation of log messages too.
As pr_cont() does not provide a message level we need a means of
remembering the last log level.
With the patch a pseudo log level LOGL_CONT as well as a pseudo log
category LOGC_CONT are introduced. Using these results in the application
of the same log level and category as in the previous log message.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Replace the static variable processing_msg by a field in the global data.
Make the field bool at it can only be true or false.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present all log devices are enabled by default. Add a function to allow
devices to be disabled or enabled at runtime.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present there is no way to disable a log driver. But the syslog driver
causes (attempted) network traffic in sandbox every time a log message
is printed, which is often.
Add a flag to enable a log driver. Adjust struct log_device to use a short
for next_filter_num so that no more memory is used for devices. Also fix
a missing line in the struct log_driver comment while here.
To maintain compatibility, enable it for all drivers for now.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present if CONFIG_LOG enabled, putting LOG_DEBUG at the top of a file
(before log.h inclusion) causes _log() to be executed for every log()
call, regardless of the build- or run-time logging level.
However there is no guarantee that the log record will actually be
displayed. If the current log level is lower than LOGL_DEBUG then it will
not be.
Add a way to signal that the log record should always be displayed and
update log_passes_filters() to handle this.
With the new behaviour, log_debug() will always log if LOG_DEBUG is
enabled.
Move log_test_syslog_nodebug() into its own file since it cannot be made
to work where it is, with LOG_DEBUG defined.
Signed-off-by: Simon Glass <sjg@chromium.org>
When a message is written by a log driver (e.g. via the network stack) this
may result in the generation of further messages. We cannot allow these
additional messages to be emitted as this might result in an infinite
recursion.
Up to now only the syslog driver was safeguarded. We should safeguard all
log drivers instead.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
The name of the function emitting a log message may be of interest for a
developer but is distracting for normal users. See the example below:
try_load_entry() Booting: Debian
Make the default format for log messages customizable. By default show
only the message text.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
If CONFIG_SPL_DM=n and CONFIG_SPL_LOG=y a build error occurs:
ld.bfd: common/built-in.o: in function `log_get_cat_name':
common/log.c:48: undefined reference to `uclass_get_name'
make[1]: *** [scripts/Makefile.spl:422: spl/u-boot-spl] Error 1
Call uclass_get_name() only if DM is enabled.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
There is no need to cast from (void *) before assigning to a pointer.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present the default log level is set to LOGL_INFO on start-up. Allow
this to be controlled from Kconfig.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
It is useful to be able to set the default log level from the command line
when running sandbox. Add a new -L command-line flag for this. The log
level is set using the enum log_level_t in log.h. At present a number must
be specified, e.g. -L7 for debug.
Signed-off-by: Simon Glass <sjg@chromium.org>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
If there is a problem with the parameters to log_add_filter(), the memory
allocated is currently not freed. Fix this.
Reported-by: Coverity (CID: 171962)
Signed-off-by: Simon Glass <sjg@chromium.org>
The EFI implementation does not fit into any of the existing categories.
Provide LOGC_EFI so that EFI related message can be filtered.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
It is useful to be able to control the output format of log records on the
console. As a starting point, add definitions for controlling which
elements of the log record are displayed. Use function and message as the
default, since these are the most useful fields.
Signed-off-by: Simon Glass <sjg@chromium.org>
Category and level both use an enum for their ID values. Add functions to
convert these IDs to strings and vice versa. This will allow the log to
output the strings instead of the (inscrutable) values.
At the same time, add a new 'driver-model' category, to cover core
driver-model functions and fix an incorrect value for LOGL_MAX.
Tests will be added with the new 'log' subcommands.
Signed-off-by: Simon Glass <sjg@chromium.org>
(Updated to correct clang warnings)
