> AFAICT, there are no kernels where this file should have
> ever existed. And it definitely doesn't exist today. Is this
> just dead code, that always fails the open()?

At the time that the code was created, the existing kernels did
provide the file.

I wouldn't be surprised that newer kernels didn't maintain that,
and I kind of think that most of the hardware that had the kernel
support is either EOU, or has had a kernel upgrade and lost the
support.

Dropping ownership:  I'm not responsible for this code anymore.

> At the time that the code was created, the existing kernels did
> provide the file.

Ohh....I've never looked at a chromeos-2.x.y branch before, but apparently we had this back in 2.6.38:

https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/4892

It didn't survive beyond that kernel (and was never upstreamed).

Methinks this has been dead long enough that it can't possibly be important. I'll see about extricating this part.

> Dropping ownership:  I'm not responsible for this code anymore.

https://chromium-review.googlesource.com/c/chromiumos/platform2/+/1101481

https://chromium-review.googlesource.com/c/chromiumos/platform2/+/1101704

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform2/+/659b8223ec5dbab70cd46773bb02a6b9dc90f2e9

commit 659b8223ec5dbab70cd46773bb02a6b9dc90f2e9
Author: Brian Norris <briannorris@chromium.org>
Date: Sat Jun 16 01:51:08 2018

bootstat: stop writing to non-existent debugfs file

/sys/kernel/debug/bootstage/mark was introduced way back in
chromeos-2.6.38:

https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/4892

This never was upstreamed, and it didn't get ported to newer kernels.
Thus, it has been dead for a very long time.

Additionally, I can't find anyone who reads the companion file
"/sys/kernel/debug/bootstage/report", so I don't think this code would
really do anything anyway. Users like the platform_BootPerf autotest
only look at the files we stash in /tmp/ anyway (see append_logdata()),
which mostly duplicate this info, except for the few in-kernel
stages that we were trying to track -- but if we wanted to track this
stuff, we'd probably do it differently today anyway.

Remove the dead code, so we don't continue trying to drag it along
(e.g.,  crbug.com/804941 ).

BUG= chromium:852869 
TEST=precq; manual tests

Change-Id: I129b167f6065ba1b1fd937d8425cf42a3a80fedb
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1101704
Reviewed-by: Richard Barnette <jrbarnette@chromium.org>
Reviewed-by: Chung-yih Wang <cywang@chromium.org>
Reviewed-by: Mike Frysinger <vapier@chromium.org>

[modify] https://crrev.com/659b8223ec5dbab70cd46773bb02a6b9dc90f2e9/bootstat/bootstat_log.c