[PATCH 01/59] drm: Add devm_drm_dev_alloc macro

Daniel Vetter daniel at ffwll.ch
Tue Apr 28 13:06:28 UTC 2020


On Tue, Apr 21, 2020 at 10:32:45PM +0200, Sam Ravnborg wrote:
> Hi
> 
> > > Hm, I see the point of this (and the dev_field below, although I'd go
> > > with dev_member there for some consistency with other macros using
> > > offset_of or container_of), but I'm not sure about the dev_ prefix.
> > > Drivers use that sometimes for the struct device *, and usage for
> > > struct drm_device * is also very inconsistent. I've seen ddev, drm,
> > > dev and base (that one only for embedded structs ofc). So not sure
> > > which prefix to pick, aside from dev_ seems the most confusing. Got
> > > ideas?
> > 
> > We have pdev for the PCI device, dev for the abstract device, and things
> > like mdev for struct mga_device in mgag200. So I'd go with ddev. I don't
> > like drm, because it could be anything in DRM. I guess struct drm_driver
> > is more 'drm' than struct drm_device.
> > 
> > But all of this is bikeshedding. It's probably best to keep the patch
> > as-is, and maybe rename variables later if we ever find consent on the
> > naming.
> 
> bikeshedding - I know.
> But reading code is is quite natural for me that drm equals the central
> drm_device data structure. Maybe thats because this was is in the code
> I started looking at.
> 
> So as an example:
> 
> 	drm_err(drm, "bla bla\n");
> 
> This parses nicely and is easy to type and get right.
> And matches nicely that drm_device => drm.
> But bikeshedding  - I will go to bed...
> (Whatever is the conclusion we should not hold back the patch in
> questions).

Ok, since we can't agree on dev vs ddev vs drm vs something else I just
left it as-is. We can always repaint this later on.

Thanks everyone for comments and review.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch


More information about the dri-devel mailing list