[PATCH 0/2] omap4: panda: convert to device model

[Tom Rini]

On Mon, Jun 15, 2020 at 06:01:00PM +0300, Tero Kristo wrote:
> On 15/06/2020 17:57, Tom Rini wrote:
> > On Mon, Jun 15, 2020 at 05:53:39PM +0300, Tero Kristo wrote:
> > > On 15/06/2020 17:48, Vignesh Raghavendra wrote:
> > > > 
> > > > 
> > > > On 15/06/20 11:49 am, Tero Kristo wrote:
> > > > > On 14/06/2020 16:49, Lokesh Vutla wrote:
> > > > > > Hi Tero,
> > > > > > 
> > > > > > On 02/06/20 4:49 pm, Tero Kristo wrote:
> > > > > > > Hi,
> > > > > > > 
> > > > > > > As there is looming death to OMAP4 Panda board u-boot support, I decided
> > > > > > > to take a shot and convert it to device model myself. With these patches
> > > > > > > it boots up fine, and there are no DM_SPL conversion complaints during
> > > > > > > compile time anymore. I think USB ethernet does not work anymore with
> > > > > > > this, but its better than dropping the support for the board completely.
> > > > > > > USB itself appears working, so it should be relatively easy for someone
> > > > > > > to fix the networking support if they need it.
> > > > > > 
> > > > > > I see the below warning with these patches:
> > > > > > 
> > > > > > board/ti/panda/panda.c:329:35: warning: ‘struct ehci_hcor’ declared
> > > > > > inside parameter list will not be visible outside of this definition
> > > > > > or declaration
> > > > > >      329 |   struct ehci_hccr **hccr, struct ehci_hcor **hcor)
> > > > > >          |                                   ^~~~~~~~~
> > > > > > board/ti/panda/panda.c:329:10: warning: ‘struct ehci_hccr’ declared
> > > > > > inside parameter list will not be visible outside of this definition
> > > > > > or declaration
> > > > > >      329 |   struct ehci_hccr **hccr, struct ehci_hcor **hcor)
> > > > > >          |          ^~~~~~~~~
> > > > > > board/ti/panda/panda.c: In function ‘ehci_hcd_init’:
> > > > > > board/ti/panda/panda.c:339:8: warning: implicit declaration of
> > > > > > function ‘omap_ehci_hcd_init’; did you mean ‘ehci_hcd_init’?
> > > > > > [-Wimplicit-function-declaration]
> > > > > >      339 |  ret = omap_ehci_hcd_init(index, &usbhs_bdata, hccr, hcor);
> > > > > >          |        ^~~~~~~~~~~~~~~~~~
> > > > > >          |        ehci_hcd_init
> > > > > > board/ti/panda/panda.c: In function ‘ehci_hcd_stop’:
> > > > > > board/ti/panda/panda.c:348:9: warning: implicit declaration of
> > > > > > function ‘omap_ehci_hcd_stop’; did you mean ‘ehci_hcd_stop’?
> > > > > > [-Wimplicit-function-declaration]
> > > > > >      348 |  return omap_ehci_hcd_stop();
> > > > > >          |         ^~~~~~~~~~~~~~~~~~
> > > > > >          |         ehci_hcd_stop
> > > > > 
> > > > > Yeah, I actually see these also. I think we could maybe drop all the
> > > > > offending code as I did later with the omap5-uevm. On omap5, USB seemed
> > > > > to work even without these.
> > > > > 
> > > > > Maybe some USB expert can comment on these (Vignesh)?
> > > > > 
> > > > 
> > > > 
> > > > Offending code won't compile with DM_USB enabled. I suggest to drop them
> > > > for now in order to avoid entire board support from being dropped. USB
> > > > can be added back later if needed...
> > > > 
> > > > Or you would have to disable DM_USB give this a try..
> > > 
> > > Yeah, dropping DM_USB USB stops working completely I believe. And, USB
> > > appears working (at least partially: usb start, usb info provides some sane
> > > data) with these two patches, just the compiler warnings are there.
> > > 
> > > However, I eliminated the offending code in v2 I posted today, and USB still
> > > appears functional as far as I can see. Trying to compile mass storage
> > > support in to test USB mass storage devices causes a mayhem so I left that
> > > for someone more experienced with USB.
> > 
> > Please note that DM_USB migration also has a deadline of almost a year
> > ago so this is another area to fix or remove-now-readd-later.  Thanks!
> > 
> 
> Yeah, so with these patches, building u-boot for omap4 panda does not
> complain about any DM_xyz conversion deadlines anymore, does it mean we are
> clean? :)

Not quite as you're not dealing with DM_SPI but that's calling make's
warning and not echo'ing like everything else so it's easier to miss.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20200615/db8e545e/attachment.sig>