drm/i915/display: Do not use ids from enum pipe in TP_printk()

drm/i915/display: Do not use ids from enum pipe in TP_printk()

Because much of kernel tracepoints is implemented at the C preprocessor
level, C identifiers used in TP_printk() are saved verbatim in the event
format, even when they represent compile-time constant values.

As an example, we can look at the format for the intel_pipe_enable
event:

    # cat /sys/kernel/debug/tracing/events/i915/intel_pipe_enable/format | grep '^print fmt'
    print fmt: "dev %s, pipe %c enable, pipe A: frame=%u, scanline=%u, pipe B: frame=%u, scanline=%u, pipe C: frame=%u, scanline=%u", __get_str(dev), REC->pipe_name, REC->frame[PIPE_A], REC->scanline[PIPE_A], REC->frame[PIPE_B], REC->scanline[PIPE_B], REC->frame[PIPE_C], REC->scanline[PIPE_C]

We see that PIPE_A, PIPE_B and PIPE_C are pasted directly in the format.
Because tools that interact with kernel tracepoints don't know about
those ids, they'll endup failing to parse the format or produce
corrupted output.

For example, we can see below that trace-cmd repeats PIPE_A's
frame/scanline counts for all pipes (probably because it evaluates
unknown ids as zero):

   $ trace-cmd report -F intel_pipe_enable | tail -n5
        testdisplay-8616  [000] 22048.276758: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=861, scanline=480, pipe B: frame=861, scanline=480, pipe C: frame=861, scanline=480
        testdisplay-8616  [001] 22048.490287: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=867, scanline=480, pipe B: frame=867, scanline=480, pipe C: frame=867, scanline=480
        testdisplay-8616  [003] 22048.700181: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=872, scanline=400, pipe B: frame=872, scanline=400, pipe C: frame=872, scanline=400
        testdisplay-8616  [002] 22049.054220: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=881, scanline=2170, pipe B: frame=881, scanline=2170, pipe C: frame=881, scanline=2170
        testdisplay-8616  [002] 22049.166851: intel_pipe_enable:    dev 0000:00:02.0, pipe B enable, pipe A: frame=887, scanline=1632, pipe B: frame=887, scanline=1632, pipe C: frame=887, scanline=1632

, while in fact we have different values for each pipe, which can be
confirmed with the raw view of the events:

    $ trace-cmd report -R -F intel_pipe_enable | tail -n5
         testdisplay-8616  [000] 22048.276758: intel_pipe_enable:     dev=0000:00:02.0 frame=ARRAY[5d, 03, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] scanline=ARRAY[e0, 01, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] pipe_name=A
         testdisplay-8616  [001] 22048.490287: intel_pipe_enable:     dev=0000:00:02.0 frame=ARRAY[63, 03, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] scanline=ARRAY[e0, 01, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] pipe_name=A
         testdisplay-8616  [003] 22048.700181: intel_pipe_enable:     dev=0000:00:02.0 frame=ARRAY[68, 03, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] scanline=ARRAY[90, 01, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] pipe_name=A
         testdisplay-8616  [002] 22049.054220: intel_pipe_enable:     dev=0000:00:02.0 frame=ARRAY[71, 03, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] scanline=ARRAY[7a, 08, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] pipe_name=A
         testdisplay-8616  [002] 22049.166851: intel_pipe_enable:     dev=0000:00:02.0 frame=ARRAY[77, 03, 00, 00, 01, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] scanline=ARRAY[60, 06, 00, 00, 39, 04, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00] pipe_name=B

To fix that, we need a fix that looks more like a hack: use macros that
result to integer constants instead of enum pipe values. This fixes the
issue, but could break if, for whatever unlikely reason, the underlying
values in the enum are changed.

In the future, we should find a better way to handle this, but for now,
the hack took care of the job:

    $ trace-cmd report -F intel_pipe_enable | tail -n5
         testdisplay-9224  [003] 24324.455375: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=1103, scanline=480, pipe B: frame=0, scanline=0, pipe C: frame=0, scanline=0
         testdisplay-9224  [002] 24324.669845: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=1109, scanline=480, pipe B: frame=0, scanline=0, pipe C: frame=0, scanline=0
         testdisplay-9224  [003] 24324.900105: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=1115, scanline=31, pipe B: frame=0, scanline=0, pipe C: frame=0, scanline=0
         testdisplay-9224  [002] 24325.256408: intel_pipe_enable:    dev 0000:00:02.0, pipe A enable, pipe A: frame=1124, scanline=2171, pipe B: frame=0, scanline=0, pipe C: frame=0, scanline=0
         testdisplay-9224  [003] 24325.380789: intel_pipe_enable:    dev 0000:00:02.0, pipe B enable, pipe A: frame=1131, scanline=979, pipe B: frame=1, scanline=1082, pipe C: frame=0, scanline=0

v2:
  - Statically assert that PIPE_A == _TRACE_PIPE_A. (MattR)

Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241016135300.21428-5-gustavo.sousa@intel.com