diff options
author | Nicky Sielicki <nlsielicki@wisc.edu> | 2015-06-06 07:51:54 -0500 |
---|---|---|
committer | Patrick Georgi <pgeorgi@google.com> | 2015-06-11 20:18:49 +0200 |
commit | 1f38d7986828fb58ced4a30bb17accb03943e798 (patch) | |
tree | f8890b807ef3024684acda822ddfbfd7cc630dc5 /util/lint | |
parent | 538ce57d47b9e017cea6f167c50f96a7ac702a45 (diff) |
Clearly define printk log level use cases.
The proper log level for any given printk statement is up to the
interpretation of the developer. This results in console output with
somewhat inconsistent levels of verbosity. This patch clearly defines each
log level and its use case, hopefully resulting in less ambiguity for
developers.
The concern with this patch might be that it leaves a lot of preexisting
printk statements using a log level that is inconsistent with the
description. I think that *most* statements map to these extended
definitions very nicely. The most discrepancies are between debug and
spew, but I'm willing to say that 95% of statements with a level lower than
debug are correct by these definitions.
There was some discussion dating back to 2010 on the mailing list about
renaming these constants to lose the 'BIOS_' prefix and to consolidate
some of them into a single constant. I disagree that it is necessary
to merge any of them, I think they all have unique use cases. But I do
think that if you all agree with these definitions, it might be useful to
rename them to reflect their use cases.
I also will add that I believe removing BIOS_NEVER is a good idea. I do
not see the use case, and it's used in only 4 files.
Change-Id: I8aefdd9dee4cb4ad2fc78ee7133a93f8ddf0720b
Signed-off-by: Nicky Sielicki <nlsielicki@wisc.edu>
Reviewed-on: http://review.coreboot.org/10444
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
Diffstat (limited to 'util/lint')
0 files changed, 0 insertions, 0 deletions