Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

1 files changed, 353 insertions, 0 deletions

diff --git a/Documentation/fb/framebuffer.rst b/Documentation/fb/framebuffer.rst
new file mode 100644
index 000000000..7fe087310
--- /dev/null
+++ b/Documentation/fb/framebuffer.rst
@@ -0,0 +1,353 @@
+=======================
+The Frame Buffer Device
+=======================
+
+Last revised: May 10, 2001
+
+
+0. Introduction
+---------------
+
+The frame buffer device provides an abstraction for the graphics hardware. It
+represents the frame buffer of some video hardware and allows application
+software to access the graphics hardware through a well-defined interface, so
+the software doesn't need to know anything about the low-level (hardware
+register) stuff.
+
+The device is accessed through special device nodes, usually located in the
+/dev directory, i.e. /dev/fb*.
+
+
+1. User's View of /dev/fb*
+--------------------------
+
+From the user's point of view, the frame buffer device looks just like any
+other device in /dev. It's a character device using major 29; the minor
+specifies the frame buffer number.
+
+By convention, the following device nodes are used (numbers indicate the device
+minor numbers)::
+
+      0 = /dev/fb0	First frame buffer
+      1 = /dev/fb1	Second frame buffer
+	  ...
+     31 = /dev/fb31	32nd frame buffer
+
+For backwards compatibility, you may want to create the following symbolic
+links::
+
+    /dev/fb0current -> fb0
+    /dev/fb1current -> fb1
+
+and so on...
+
+The frame buffer devices are also `normal` memory devices, this means, you can
+read and write their contents. You can, for example, make a screen snapshot by::
+
+  cp /dev/fb0 myfile
+
+There also can be more than one frame buffer at a time, e.g. if you have a
+graphics card in addition to the built-in hardware. The corresponding frame
+buffer devices (/dev/fb0 and /dev/fb1 etc.) work independently.
+
+Application software that uses the frame buffer device (e.g. the X server) will
+use /dev/fb0 by default (older software uses /dev/fb0current). You can specify
+an alternative frame buffer device by setting the environment variable
+$FRAMEBUFFER to the path name of a frame buffer device, e.g. (for sh/bash
+users)::
+
+    export FRAMEBUFFER=/dev/fb1
+
+or (for csh users)::
+
+    setenv FRAMEBUFFER /dev/fb1
+
+After this the X server will use the second frame buffer.
+
+
+2. Programmer's View of /dev/fb*
+--------------------------------
+
+As you already know, a frame buffer device is a memory device like /dev/mem and
+it has the same features. You can read it, write it, seek to some location in
+it and mmap() it (the main usage). The difference is just that the memory that
+appears in the special file is not the whole memory, but the frame buffer of
+some video hardware.
+
+/dev/fb* also allows several ioctls on it, by which lots of information about
+the hardware can be queried and set. The color map handling works via ioctls,
+too. Look into <linux/fb.h> for more information on what ioctls exist and on
+which data structures they work. Here's just a brief overview:
+
+  - You can request unchangeable information about the hardware, like name,
+    organization of the screen memory (planes, packed pixels, ...) and address
+    and length of the screen memory.
+
+  - You can request and change variable information about the hardware, like
+    visible and virtual geometry, depth, color map format, timing, and so on.
+    If you try to change that information, the driver maybe will round up some
+    values to meet the hardware's capabilities (or return EINVAL if that isn't
+    possible).
+
+  - You can get and set parts of the color map. Communication is done with 16
+    bits per color part (red, green, blue, transparency) to support all
+    existing hardware. The driver does all the computations needed to apply
+    it to the hardware (round it down to less bits, maybe throw away
+    transparency).
+
+All this hardware abstraction makes the implementation of application programs
+easier and more portable. E.g. the X server works completely on /dev/fb* and
+thus doesn't need to know, for example, how the color registers of the concrete
+hardware are organized. XF68_FBDev is a general X server for bitmapped,
+unaccelerated video hardware. The only thing that has to be built into
+application programs is the screen organization (bitplanes or chunky pixels
+etc.), because it works on the frame buffer image data directly.
+
+For the future it is planned that frame buffer drivers for graphics cards and
+the like can be implemented as kernel modules that are loaded at runtime. Such
+a driver just has to call register_framebuffer() and supply some functions.
+Writing and distributing such drivers independently from the kernel will save
+much trouble...
+
+
+3. Frame Buffer Resolution Maintenance
+--------------------------------------
+
+Frame buffer resolutions are maintained using the utility `fbset`. It can
+change the video mode properties of a frame buffer device. Its main usage is
+to change the current video mode, e.g. during boot up in one of your `/etc/rc.*`
+or `/etc/init.d/*` files.
+
+Fbset uses a video mode database stored in a configuration file, so you can
+easily add your own modes and refer to them with a simple identifier.
+
+
+4. The X Server
+---------------
+
+The X server (XF68_FBDev) is the most notable application program for the frame
+buffer device. Starting with XFree86 release 3.2, the X server is part of
+XFree86 and has 2 modes:
+
+  - If the `Display` subsection for the `fbdev` driver in the /etc/XF86Config
+    file contains a::
+
+	Modes "default"
+
+    line, the X server will use the scheme discussed above, i.e. it will start
+    up in the resolution determined by /dev/fb0 (or $FRAMEBUFFER, if set). You
+    still have to specify the color depth (using the Depth keyword) and virtual
+    resolution (using the Virtual keyword) though. This is the default for the
+    configuration file supplied with XFree86. It's the most simple
+    configuration, but it has some limitations.
+
+  - Therefore it's also possible to specify resolutions in the /etc/XF86Config
+    file. This allows for on-the-fly resolution switching while retaining the
+    same virtual desktop size. The frame buffer device that's used is still
+    /dev/fb0current (or $FRAMEBUFFER), but the available resolutions are
+    defined by /etc/XF86Config now. The disadvantage is that you have to
+    specify the timings in a different format (but `fbset -x` may help).
+
+To tune a video mode, you can use fbset or xvidtune. Note that xvidtune doesn't
+work 100% with XF68_FBDev: the reported clock values are always incorrect.
+
+
+5. Video Mode Timings
+---------------------
+
+A monitor draws an image on the screen by using an electron beam (3 electron
+beams for color models, 1 electron beam for monochrome monitors). The front of
+the screen is covered by a pattern of colored phosphors (pixels). If a phosphor
+is hit by an electron, it emits a photon and thus becomes visible.
+
+The electron beam draws horizontal lines (scanlines) from left to right, and
+from the top to the bottom of the screen. By modifying the intensity of the
+electron beam, pixels with various colors and intensities can be shown.
+
+After each scanline the electron beam has to move back to the left side of the
+screen and to the next line: this is called the horizontal retrace. After the
+whole screen (frame) was painted, the beam moves back to the upper left corner:
+this is called the vertical retrace. During both the horizontal and vertical
+retrace, the electron beam is turned off (blanked).
+
+The speed at which the electron beam paints the pixels is determined by the
+dotclock in the graphics board. For a dotclock of e.g. 28.37516 MHz (millions
+of cycles per second), each pixel is 35242 ps (picoseconds) long::
+
+    1/(28.37516E6 Hz) = 35.242E-9 s
+
+If the screen resolution is 640x480, it will take::
+
+    640*35.242E-9 s = 22.555E-6 s
+
+to paint the 640 (xres) pixels on one scanline. But the horizontal retrace
+also takes time (e.g. 272 `pixels`), so a full scanline takes::
+
+    (640+272)*35.242E-9 s = 32.141E-6 s
+
+We'll say that the horizontal scanrate is about 31 kHz::
+
+    1/(32.141E-6 s) = 31.113E3 Hz
+
+A full screen counts 480 (yres) lines, but we have to consider the vertical
+retrace too (e.g. 49 `lines`). So a full screen will take::
+
+    (480+49)*32.141E-6 s = 17.002E-3 s
+
+The vertical scanrate is about 59 Hz::
+
+    1/(17.002E-3 s) = 58.815 Hz
+
+This means the screen data is refreshed about 59 times per second. To have a
+stable picture without visible flicker, VESA recommends a vertical scanrate of
+at least 72 Hz. But the perceived flicker is very human dependent: some people
+can use 50 Hz without any trouble, while I'll notice if it's less than 80 Hz.
+
+Since the monitor doesn't know when a new scanline starts, the graphics board
+will supply a synchronization pulse (horizontal sync or hsync) for each
+scanline.  Similarly it supplies a synchronization pulse (vertical sync or
+vsync) for each new frame. The position of the image on the screen is
+influenced by the moments at which the synchronization pulses occur.
+
+The following picture summarizes all timings. The horizontal retrace time is
+the sum of the left margin, the right margin and the hsync length, while the
+vertical retrace time is the sum of the upper margin, the lower margin and the
+vsync length::
+
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |upper_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------###############################################----------+-------+
+  |          #                ↑                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |   left   #                |                            #  right   | hsync |
+  |  margin  #                |       xres                 #  margin  |  len  |
+  |<-------->#<---------------+--------------------------->#<-------->|<----->|
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |yres                        #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                |                            #          |       |
+  |          #                ↓                            #          |       |
+  +----------###############################################----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |lower_margin                |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+  |          |                ↑                            |          |       |
+  |          |                |vsync_len                   |          |       |
+  |          |                ↓                            |          |       |
+  +----------+---------------------------------------------+----------+-------+
+
+The frame buffer device expects all horizontal timings in number of dotclocks
+(in picoseconds, 1E-12 s), and vertical timings in number of scanlines.
+
+
+6. Converting XFree86 timing values info frame buffer device timings
+--------------------------------------------------------------------
+
+An XFree86 mode line consists of the following fields::
+
+ "800x600"     50      800  856  976 1040    600  637  643  666
+ < name >     DCF       HR  SH1  SH2  HFL     VR  SV1  SV2  VFL
+
+The frame buffer device uses the following fields:
+
+  - pixclock: pixel clock in ps (pico seconds)
+  - left_margin: time from sync to picture
+  - right_margin: time from picture to sync
+  - upper_margin: time from sync to picture
+  - lower_margin: time from picture to sync
+  - hsync_len: length of horizontal sync
+  - vsync_len: length of vertical sync
+
+1) Pixelclock:
+
+   xfree: in MHz
+
+   fb: in picoseconds (ps)
+
+   pixclock = 1000000 / DCF
+
+2) horizontal timings:
+
+   left_margin = HFL - SH2
+
+   right_margin = SH1 - HR
+
+   hsync_len = SH2 - SH1
+
+3) vertical timings:
+
+   upper_margin = VFL - SV2
+
+   lower_margin = SV1 - VR
+
+   vsync_len = SV2 - SV1
+
+Good examples for VESA timings can be found in the XFree86 source tree,
+under "xc/programs/Xserver/hw/xfree86/doc/modeDB.txt".
+
+
+7. References
+-------------
+
+For more specific information about the frame buffer device and its
+applications, please refer to the Linux-fbdev website:
+
+    http://linux-fbdev.sourceforge.net/
+
+and to the following documentation:
+
+  - The manual pages for fbset: fbset(8), fb.modes(5)
+  - The manual pages for XFree86: XF68_FBDev(1), XF86Config(4/5)
+  - The mighty kernel sources:
+
+      - linux/drivers/video/
+      - linux/include/linux/fb.h
+      - linux/include/video/
+
+
+
+8. Mailing list
+---------------
+
+There is a frame buffer device related mailing list at kernel.org:
+linux-fbdev@vger.kernel.org.
+
+Point your web browser to http://sourceforge.net/projects/linux-fbdev/ for
+subscription information and archive browsing.
+
+
+9. Downloading
+--------------
+
+All necessary files can be found at
+
+    ftp://ftp.uni-erlangen.de/pub/Linux/LOCAL/680x0/
+
+and on its mirrors.
+
+The latest version of fbset can be found at
+
+    http://www.linux-fbdev.org/
+
+
+10. Credits
+-----------
+
+This readme was written by Geert Uytterhoeven, partly based on the original
+`X-framebuffer.README` by Roman Hodek and Martin Schaller. Section 6 was
+provided by Frank Neumann.
+
+The frame buffer device abstraction was designed by Martin Schaller.