RE: [PATCH V10] powerpc/fsl-pci: Unify pci/pcie initialization code

Hi Kumar,=0A=
=0A=
Do you or Ben has any comment on this?=0A=
Please let me know.=0A=
=0A=
Thanks.=0A=
=0A=
________________________________________=0A=
From: Jia Hongtao-B38951=0A=
Sent: Friday, August 31, 2012 10:25 AM=0A=
To: Kumar Gala=0A=
Cc: linuxppc-dev@lists.ozlabs.org; Li Yang-R58472; Wood Scott-B07421=0A=
Subject: RE: [PATCH V10] powerpc/fsl-pci: Unify pci/pcie initialization cod=
e=0A=
=0A=

-----Original Message-----=0A=
From: Kumar Gala [mailto:galak@kernel.crashing.org]=0A=
Sent: Friday, August 31, 2012 12:54 AM=0A=
To: Jia Hongtao-B38951=0A=
Cc: linuxppc-dev@lists.ozlabs.org; Li Yang-R58472; Wood Scott-B07421=0A=
Subject: Re: [PATCH V10] powerpc/fsl-pci: Unify pci/pcie initialization=

=0A=

code=0A=
=0A=
=0A=
On Aug 28, 2012, at 2:44 AM, Jia Hongtao wrote:=0A=
=0A=

quoted

We unified the Freescale pci/pcie initialization by changing the=0A=
fsl_pci to a platform driver. In previous PCI code architecture the=0A=
initialization routine is called at board_setup_arch stage. Now the=0A=
initialization is done in probe function which is architectural=0A=
better. Also It's convenient for adding PM support for PCI controller=

=0A=

in later patch.=0A=

quoted

=0A=
Now we registered pci controllers as platform devices. So we combine=0A=
two initialization code as one platform driver.=0A=
=0A=
Signed-off-by: Jia Hongtao <redacted>=0A=
Signed-off-by: Li Yang <redacted>=0A=
Signed-off-by: Chunhe Lan <redacted>=0A=
---=0A=
Changes for V10:=0A=
* Just update comments for mpc85xx_cds_pci_assign_primary=0A=

=0A=
What do we have to say w/regards to Ben's comment about breaking of=0A=
pci_fixup_final by moving to a platform driver?=0A=
=0A=
- k=0A=

=0A=
If Ben mean the fixup order I have already gave some explanations as follow=
:=0A=
=0A=
I have already done some investigations and the sequence of fixup=0A=
(including early, header, final) will not be changed in platform driver.=0A=
=0A=
We register and init PCI controllers as platform devices at arch_initcall=
=0A=
stage and PCI scanning (pcibios_init) is at subsys_initcall stage in which=
=0A=
early and header fixup will be done in right sequence. The final fixup will=
=0A=
be start at rootfs_initcall stage which is later than early and header fixu=
p.=0A=
=0A=
- Hongtao.=0A=