[PATCH] drm/vmwgfx: respect 'nomodeset'

[Rob Clark]

On Wed, Oct 15, 2014 at 4:44 PM, Thomas Hellstrom <thellstrom at vmware.com> wrote:
> On 10/15/2014 10:38 PM, Rob Clark wrote:
>> On Wed, Oct 15, 2014 at 4:35 PM, Rob Clark <robdclark at gmail.com> wrote:
>>> On Wed, Oct 15, 2014 at 4:17 PM, Thomas Hellstrom <thellstrom at vmware.com> wrote:
>>>> On 10/15/2014 09:46 PM, Rob Clark wrote:
>>>>> On Wed, Oct 15, 2014 at 3:24 PM, Thomas Hellstrom <thellstrom at vmware.com> wrote:
>>>>>> On 10/15/2014 09:00 PM, Rob Clark wrote:
>>>>>>> Signed-off-by: Rob Clark <robdclark at gmail.com>
>>>>>>> ---
>>>>>>>  drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 7 +++++++
>>>>>>>  1 file changed, 7 insertions(+)
>>>>>>>
>>>>>>> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
>>>>>>> index 18b54ac..f0267b8 100644
>>>>>>> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
>>>>>>> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
>>>>>>> @@ -25,6 +25,7 @@
>>>>>>>   *
>>>>>>>   **************************************************************************/
>>>>>>>  #include <linux/module.h>
>>>>>>> +#include <linux/console.h>
>>>>>>>
>>>>>>>  #include <drm/drmP.h>
>>>>>>>  #include "vmwgfx_drv.h"
>>>>>>> @@ -1453,6 +1454,12 @@ static int vmw_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>>>>>>>  static int __init vmwgfx_init(void)
>>>>>>>  {
>>>>>>>       int ret;
>>>>>>> +
>>>>>>> +#ifdef CONFIG_VGA_CONSOLE
>>>>>>> +     if (vgacon_text_force())
>>>>>>> +             return -EINVAL;
>>>>>>> +#endif
>>>>>>> +
>>>>>> Hmm,
>>>>>>
>>>>>> From the function name vgacon_text_force() it sounds like this should
>>>>>> just stop the driver from initializing fbcon? Not refuse to load?
>>>>> yeah, the function is badly named.. it perhaps should be
>>>>> vgacon_is_text_forced() or something like that.  But basically it
>>>>> returns whether we are forced to text mode.
>>>>>
>>>>> BR,
>>>>> -R
>>>> So then I guess it would be more correct to use the output of that
>>>> function when determining the value of dev_priv->enable_fb (vmwgfx can
>>>> enable the user-space modesetting API without turning on vmwgfx fbcon).
>>> well, other drivers, 'nomodeset' forces the driver not to load (to
>>> work around buggyness, etc)..  I suppose for vmwgfx perhaps oddball
>>> "hardware" is less of a concern.  Although it seems like it would be
>>> nice if vmwgfx behaved consistently with the other drivers.
>>>
>>> Most/all for the drivers have an additional module param that lets you
>>> override this and load the driver for UMS in case of 'nomodeset'..
>>> which would give you the behaviour you describe.  But I think in the
>>> absence of an additional module param specified, the default
>>> 'nomodeset' behaviour should be that the driver does not load.
>>>
>>> But I can add such a module param if you think it is useful..
>> oh, wait.. you already have an 'enable_fbdev'..  although I guess that
>> is actually meaning "enable_kms"?
>
> It should actually have been "enable_fbcon_and_fbdev" instead of
> "enable_fbdev". KMS is always enabled with the vmwgfx driver, we have no
> UMS driver using DRM functionality.
>
> I think what confuses me is how "nomodeset" translates to
> "use_text_console", (sounds orthogonal to me) but if that's the way that
> parameter is presented to the drivers, I guess the correct way after all
> is to stop the driver from loading....

the:

  __setup("nomodeset", text_mode);

bit causes text_mode() to be called which sets vgacon_text_mode_force

(assuming that was the question)

At least users seem to expect that 'nomodeset' means don't load the
modesetting driver, since I got a bug about it ;-)

BR,
-R

>
> /Thomas
>