Re: [PATCH] RFC: platform driver registering via initcall tables

[Greg KH]

On Tue, Dec 17, 2019 at 03:43:56PM +0100, Enrico Weigelt, metux IT consult wrote:
> On 17.12.19 15:06, Greg KH wrote:
> 
> > That's not needed, and you are going to break the implicit ordering we
> > already have with link order.  
> 
> Ups, 10 points for you - I didn't consider that.
> 
> > You are going to have to figure out what
> > bus type the driver is, to determine what segment it was in, to figure
> > out what was loaded before what.
> 
> hmm, if it's just the ordering by bus type (but not within one bus
> type), then it shouldn't be the big deal to fix, as I'll need one table
> and register-loop per bus-type anyways.
> 
> By the way: how is there init order ensured with dynamically loaded
> modules ? (for cases where there aren't explicit symbol dependencies)

See the recent work in the driver core for DT fixes for that very issue.

greg k-h