aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
...
* | kernelshark: Move capture display output to main capture windowSteven Rostedt2010-06-22
| | | | | | | | | | | | The capture output belongs in the main window. Move it there. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Restructure capture dialogSteven Rostedt2010-06-22
| | | | | | | | | | | | | | | | Change the capture dialog around to include the events window. Also use table widget instead of vpaned to layout the widgets nicer. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Have trace_create_combo_box() be able to just return combo boxSteven Rostedt2010-06-22
| | | | | | | | | | | | | | | | | | | | | | Instead of requiring trace_create_combo_box() to work with a hbox, change it so passing in a NULL hbox will just produce a combo box and ignore the text (for the label). This will allow us to still use this helper function when we need to create the label and combo on something other than a hbox. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Make ~/.trace-cmd/filters store filtersSteven Rostedt2010-06-22
| | | | | | | | | | | | | | Have the filters stored locally in the users .trace-cmd directory. This will allow the filters to be read and saved from there. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Add extension to files being savedSteven Rostedt2010-06-21
| | | | | | | | | | | | | | | | Change the file dialog to when a "waring" is set (which is when we plan on saving a file) to automatically add the extension if the extension does not already exist. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernlshark: Add TRACE_DIALOG_FILTER_FILTERSteven Rostedt2010-06-21
| | | | | | | | | | | | | | Add TRACE_DIALOG_FILTER_FILTER and have the KernelShark filter files use the *.ksf extension. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Set default file filtersSteven Rostedt2010-06-21
| | | | | | | | | | | | | | | | | | | | Set the default file filters for the specified commands. A load of trace data should filter on *.dat. A load of a settings should filter on *.kss. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Add trace_get_file_dialog_filter() to set default filterSteven Rostedt2010-06-21
| | | | | | | | | | | | | | Add trace_get_file_dialog_filter() to add an interface to set the default filter of files for a given operation. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Add filters to dialog menuSteven Rostedt2010-06-21
| | | | | | | | | | | | | | Add filters to the file chooser dialog menu. Now the user can pick from *.dat or *.kss files. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Move capture execute to dialog buttonSteven Rostedt2010-06-21
| | | | | | | | | | | | | | Move the execute button down with the cancel button. Reported-by: Darren Hart <dvhltc@us.ibm.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Make capture window close button be called "Close"Steven Rostedt2010-06-21
| | | | | | | | | | | | | | | | "Done" is confusing, the close of the capture window should be called "close". Reported-by: Darren Hart <dvhltc@us.ibm.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Make a pull down of capture settings and save in home dirSteven Rostedt2010-06-21
| | | | | | | | | | | | | | | | Have the settings saved in a ~/.trace-cmd/settings under *.kss. Then have the user able to load and save them there with simple combo and dialog box respectively. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Move the event tree into the capture dialog boxSteven Rostedt2010-06-18
| | | | | | | | | | | | | | | | Since the events are a major selection feature of the capture, move the event tree into the main dialog box. Requested-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Add trace_update_event_view()Steven Rostedt2010-06-18
| | | | | | | | | | | | | | Add trace_update_event_view() to be able to update the events set in an already existing event_view. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | trace-plot: Fix filtering of event boxes in graphSteven Rostedt2010-06-18
| | | | | | | | | | | | | | | | | | The graph was using the "orig_pid" to determine if it should display the box (process running) or not. But this only worked when we had sched events available. When schedule events were enabled, this would fail the tests. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | trace-cmd: Fix split record size with large eventsSteven Rostedt2010-06-18
| | | | | | | | | | | | | | Large events store the size in the array[0] portion of the header of the record. This size is the actual size and not a division. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | trace-cmd: Convert length field endian for large events in splitSteven Rostedt2010-06-18
| | | | | | | | | | | | | | | | trace-cmd split processing a different endian file than the host fails to do the conversion of the length field. Reported-by: Will Schmidt <will_schmidt@vnet.ibm.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Remove open coded event list addition with generic functionSteven Rostedt2010-06-18
| | | | | | | | | | | | | | Remove the open coded event list from the trace-capture.c with the generic tracecmd_add_id() routine. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Remove the add_event() functions and use genericSteven Rostedt2010-06-18
| | | | | | | | | | | | | | Remove the custom add_event() and add_event_int() from trace-filter.c and use the tracecmd_add_id() generic routine. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | trace-cmd: Add helper function tracecmd_add_id()Steven Rostedt2010-06-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Several places are creating list of event ids as an array with the ids ending in -1. To use this, simply: int *events = NULL; intt len = 0; /* Fictious for statement */ for (event = get_event(); event; event = next_event(event)) events = tracecmd_add_id(events, event->id, len++); /* ... */ free(events); Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Remove the add_systems() functions and use genericSteven Rostedt2010-06-18
| | | | | | | | | | | | | | Remove the custom add_system() and add_system_str() from trace-filter.c and use the tracecmd_add_list() generic routine. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kernelshark: Add trace_extract_event_list_view() to extract eventsSteven Rostedt2010-06-18
| | | | | | | | | | | | | | | | | | Add function trace_extract_event_list_view() to extract events from the event list view created with: trace_create_event_list_view(). Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | kerenlshark: Make the event tree widge global to allSteven Rostedt2010-06-18
|/ | | | | | | | Make the create_event_list_view() into a global function called trace_create_event_list_view() that exports the creation of a list view of events for other parts of kernelshark to use. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Add -k option to keep record from resetting buffersSteven Rostedt2010-06-17
| | | | | | | | Add -k option to trace-cmd record that will keep it from resetting the buffers when it is done. This option brings back the old behavior of trace-cmd record. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Disable buffers after printing statsSteven Rostedt2010-06-17
| | | | | | | | | | If we disable the buffers before we print the stats, then the stats will be zero. That is because disabling the buffers also zeros out the stats. Print out the stats _before_ we disable the buffers. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* parse-events: Add helper functions pevent_get_field_val()Steven Rostedt2010-06-17
| | | | | | | | Several plugins were implementing their own helper function to set a value of the data from a field. Move the helper function into the core code. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* parse-events: Add helper function pevent_print_num_field()Steven Rostedt2010-06-17
| | | | | | | | | | | The _print_field() and print_field() functions were being used by multiple plugins. Each implementing their own helper function to get a field and print its value. Instead of having multiple functions to do the same job in different plugins, just add it to the core code. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* parse-events: Concatinate two strings that are adjacentSteven Rostedt2010-06-17
| | | | | | | When two or more strings are adjacent ("string1" "string2") concatenate them into one string token. ("string1string2") Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* parse-events: Use all the printf fields of a %sSteven Rostedt2010-06-17
| | | | | | | If a %s has options like %8.2s, we still want to process it. Pass the format off to the string processing to do so. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* parse-events: Handle '*' in TP_printk formatSteven Rostedt2010-06-17
| | | | | | | | | When the "*" (next arg is length) is used in the TP_printk format, it is currently not handled. This will cause a shift in the arguments that are passed in causing all arguments after to be corrupted. Reported-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Use correct size for rest of recordSteven Rostedt2010-06-17
| | | | | | | | | | | | | | | When the field does not have a lenth, use rest of the data item. But account for the offset of the field. Field arrays with [0] denote to use the rest of the record. But the size set was the size of the record and not the size starting from the field to the end of the record. This was causing SEGFAULTS when the record was at the end of a page, and the copy was going beyond the page. Reported-by: Will Schmidt <will_schmidt@vnet.ibm.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Disable everything at end of recordSteven Rostedt2010-06-17
| | | | | | | Disable all tracing at the end of a record, and set the tracing_on back to the value it was before the record started. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Always enabled tracing_enabledSteven Rostedt2010-06-16
| | | | | | | | | | | | | | | The tracing_enabled file really needs to be deprecated. It should not be used. But if the tracing_enabled file gets set to zero somehow (by either the user or tool - latencytop?) then it will disable output. Since trace-cmd does not touch the file, it will confuse users when they see no data recorded. The simple answer is to always set tracing_enabled to one if the file exists. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Show only one pop up warning message after loading fileSteven Rostedt2010-06-16
| | | | | | | | Instead of showing every little error that happened while loading a trace file, show only one message at the end. Have the warnings go to the display status and the user can reference them there. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Have gui warning reset errnoSteven Rostedt2010-06-16
| | | | | | After printing a gui warning, reset errno. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Add newline between perror and message in warningSteven Rostedt2010-06-16
| | | | | | | In showing a warning dialog, if errno exists, then add a new line between the message and the errno message. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-graph: Account for events not found in plotSteven Rostedt2010-06-16
| | | | | | | | | | | | | In the plot code, it is possible that an event exists in the data file that was not stored in the events log. This happened with the 2.6.35 kernel where some events had formats that were "TOO BIG" and that event format was not saved. But the events still existed in the trace.dat file. The plot code assumed that events were found and when they were not it crashed. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Fix bad loading of capture settingsSteven Rostedt2010-06-15
| | | | | | Add a "return" if the file is not found in load capture settings. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Manually look for trace-cmd in PATHSteven Rostedt2010-06-15
| | | | | | | | When executing a record, walk the PATH variable to search for trace-cmd. If it is not found we can nicely give a warning instead of trying to execute it. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: append to CFLAGS instead of being overridenChase Douglas2010-06-15
| | | | | | | | | | Most package builders apply their own CFLAGS, often set during the make invocation. The trace-cmd internal CFLAGS is overriden in this case. Make sure the important flags are appended. Signed-off-by: Chase Douglas <chase.douglas@canonical.com> LKML-Reference: <1276449108-21328-1-git-send-email-chase.douglas@canonical.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* trace-cmd: Prevent print_graph_duration buffer overflowChase Douglas2010-06-15
| | | | | | | | | | | | | Passing n > sizeof(string) to snprintf can cause a glibc buffer overflow condition. We know the exact size of nsecs_str, so use it along with the the math to determine the longest string size we want. Note that an overflow isn't really possible given the format of the string. However, glibc would abort due to a runtime check. Signed-off-by: Chase Douglas <chase.douglas@canonical.com> LKML-Reference: <1276613224-19247-1-git-send-email-chase.douglas@canonical.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* python: Ignore vpr_statSteven Rostedt2010-06-15
| | | | | | | Python can not handle va_list, so ignore vpr_stat from the swig code. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Add load and store of capture settingsSteven Rostedt2010-06-15
| | | | | | | | Add the load and store buttons to the capture dialog that lets the user store and load at another time their capture settings. This lets them store events and plugins and retrieve it later. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Keep capture state persistentSteven Rostedt2010-06-15
| | | | | | | | Make the changes to the capture dialog persistent. That is if the user makes updates to the dialog then closes it, the next time they open it, it will still contain their changes. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Automatically load file after executionSteven Rostedt2010-06-15
| | | | | | After the user records a trace, automatically load the file. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Warn on start of capture if file existsSteven Rostedt2010-06-14
| | | | | | | | Instead of warning that a file exists at the file pop up, warn about it when we execute. Otherwise, we may miss a warning since the user can simply type the file name in. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernel-shark: Only warn in file dialog if file existsSteven Rostedt2010-06-14
| | | | | | | The warning of the file dialog will happen everytime even if the file does not exist. Fix that. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Have stop dialog close when command is finishedSteven Rostedt2010-06-14
| | | | | | | | | | When the trace-cmd record command finishes, close the stop dialog. Luckily we know when it finishes because the command will close the pipe which will cause the monitor_pipes thread to detect it. Then it will close the dialog box by emitting the delete-event signal to it. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Allow to record no commandSteven Rostedt2010-06-14
| | | | | | | When no command is selected, the recording happens until the stop button is hit. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* kernelshark: Have file dialog ask if we want to overwriteSteven Rostedt2010-06-14
| | | | | | | | If someone selects a file that already exists and the command will be used to overwrite the file. Ask the user if they want to overwrite it instead of just doing so. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
generated by cgit v1.2.2 (git 2.25.0) at 2025-04-19 20:07:42 -0400