[PATCH] RFCv2: omapdrm DRM/KMS driver for TI OMAP platforms

Rob Clark rob.clark at linaro.org
Fri Oct 14 07:03:21 PDT 2011

On Mon, Sep 26, 2011 at 8:55 AM, Konrad Rzeszutek Wilk
<konrad.wilk at oracle.com> wrote:
>> I'm a bit undecided on some of this error handling at startup..  I
>> guess ENOMEM is clear enough.  But some of the other parts, like
>> connector initialization, could fail just because some hw is not
>> present/populated.  Like missing some LCD.  I guess that it is best to
>> try and limp along as best as possible and hope to get some pixels on
>> some screen that the user can see.  If you give up and all the
>> screens/monitors/etc stay dark, it might be a bit inconvenient to
>> debug..
> Perhaps then just do WARN_ON, like:
> if (create_encode(..))
>    WARN_ON(1,"Danger Danger! Limping along\n");

fwiw, in cases where create_{crtc,encoder,connector}() return !=0,
there is a dev_err() in the create_*() fxn.  I think this should be
sufficient, no need to double up the messages in the log..

I've cleaned up the rest of the things that you had pointed out, and
will send a v3 shortly.  Thanks!


More information about the dri-devel mailing list