Re: [PATCH v3 06/28] driver core: fw_devlink: Introduce fw_devlink_set_device()

[Herve Codina]

Hi Saravana,

On Mon, 16 Jun 2025 09:04:06 +0200
Herve Codina <herve.codina@xxxxxxxxxxx> wrote:

> Hi Saravana,
> 
> On Fri, 13 Jun 2025 14:13:49 -0700
> Saravana Kannan <saravanak@xxxxxxxxxx> wrote:
> 
> > On Fri, Jun 13, 2025 at 6:49 AM Herve Codina <herve.codina@xxxxxxxxxxx> wrote:  
> > >
> > > Setting fwnode->dev is specific to fw_devlink.
> > >
> > > In order to avoid having a direct 'fwnode->dev = dev;' in several
> > > place in the kernel, introduce fw_devlink_set_device() helper to perform
> > > this operation.
> > >    
> > 
> > This should not be set anywhere outside the driver core files. I'll
> > get to reviewing the series, but until then, NACK to this.
> > 
> > Is there a specific patch that explain why we need to set this outside
> > driver core?  
> 
> We need to set it in case of creating device-tree node for PCI.
> 
> Usually, fwnode are created (based on DT or ACPI) and then, dev are
> created.
> 
> In the PCI DT node creation case, device are already created and then, based
> on information already computed by the kernel, DT node are created.
> 
> You can see that on patch 11 (dev setting was already upstream and it is
> replace by a call to the helper for PCI host bridge) and on patch 13 (PCI
> device).
> 
> Other patches (8, 9 and 10) replace the existing direct setting of the dev
> member by a call to the helper.
> 

Have you got time to look at the series, patches I pointed out and the reply
from Andy?

Are modifications still nacked on your side?
If so, what kind of modification would you like to see in order to move
forward?

Best regards,
Hervé