[PATCH] drm: change logs to print connectors in the form CONNECTOR:id:name
jim.cromie at gmail.com
jim.cromie at gmail.com
Thu Nov 18 03:41:11 UTC 2021
On Mon, Nov 15, 2021 at 12:40 PM Claudio Suarez <cssk at net-c.es> wrote:
>
> On Mon, Nov 15, 2021 at 12:24:26PM +0200, Jani Nikula wrote:
> > On Sun, 14 Nov 2021, Claudio Suarez <cssk at net-c.es> wrote:
> > > On Sat, Nov 13, 2021 at 09:39:46PM +0100, Sam Ravnborg wrote:
> > >> Hi Claudio,
> > >>
> > >> On Sat, Nov 13, 2021 at 08:27:30PM +0100, Claudio Suarez wrote:
> > >> > The prefered way to log connectors is [CONNECTOR:id:name]. Change it in
> > >> > drm core programs.
> > >> >
> > >> > Suggested-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > >> > Signed-off-by: Claudio Suarez <cssk at net-c.es>
> > >>
> > >> While touching all these logging calls, could you convernt them to the
> > >> newer drm_dbg_kms variants?
> > >> DRM_DEBUG_* are all deprecated.
> > >>
> > >
> > > Yes, I can, but it is recommended to do it in a different patch:
> > >
> > > https://www.kernel.org/doc/html/latest/process/submitting-patches.html#separate-your-changes
> > >
> > > C&P:
> > > "Separate your changes
> > > Separate each logical change into a separate patch.
> > > For example, if your changes include both bug fixes and performance enhancements..."
> > >
> > >
> > > I will study and send a new separate patch with your suggestion.
> >
> > I feel these logging changes are the types of changes where I'd err on
> > the side of fewer patches than strictly following the above guidelines.
>
> To size the problem:
> - there are about 3434 references to DRM_DEBUG_* in all the drm code, all drivers.
> - there are 413 references to DRM_DEBUG_* in the drm core code, only 66 of
> them are related to connectors.
> - there are 62 references to DRM_ERR* and 7 references to DRM_INFO in
> the drm core programs.
>
> I meant I can make two patches:
> 1 - this one with the change to CONNECTOR:id:name (29 changes)
Is there a "committee" that makes or coordinates these log-facing changes ?
The reason I ask is that Ive proposed that the DRM_UT_<Categories>
be replaced by a set of prefix strings; "drm:core:" "drm:kms:" etc.
If the DRM_DEBUG_* etc api were converted to use pr_debug,
then dyndbg could optimize away the drm_debug_enabled() overheads.
dyndbg would enable those classes of drm-debug callsites using
#> echo module drm format drm:kms: > /proc/dyndbg/control
the patchset includes a macro to declare a bit-field to implement drm.debug
https://patchwork.freedesktop.org/series/96328/
how would one (hope to) coordinate changes like this ?
More information about the dri-devel
mailing list