mbox series

[00/11] ARM: dts: mvebu: Add definitions for PCIe legacy INTx interrupts

Message ID 20220706183114.30783-1-pali@kernel.org (mailing list archive)
Headers show
Series ARM: dts: mvebu: Add definitions for PCIe legacy INTx interrupts | expand

Message

Pali Rohár July 6, 2022, 6:31 p.m. UTC
This patch series add definitions for PCIe legacy INTx interrupts into
every DTS file used by the pci-mvebu.c controller driver.

It was tested on 88F6820 (A385) and 88F6281 (Kirkwood) SoCs.

Pali Rohár (11):
  ARM: dts: kirkwood: Add definitions for PCIe legacy INTx interrupts
  ARM: dts: dove: Add definitions for PCIe legacy INTx interrupts
  ARM: dts: armada-370.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-xp-98dx3236.dtsi: Add definitions for PCIe legacy
    INTx interrupts
  ARM: dts: armada-xp-mv78230.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-xp-mv78260.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-xp-mv78460.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-375.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-380.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-385.dtsi: Add definitions for PCIe legacy INTx
    interrupts
  ARM: dts: armada-39x.dtsi: Add definitions for PCIe legacy INTx
    interrupts

 arch/arm/boot/dts/armada-370.dtsi         |  28 ++++-
 arch/arm/boot/dts/armada-375.dtsi         |  28 ++++-
 arch/arm/boot/dts/armada-380.dtsi         |  42 ++++++-
 arch/arm/boot/dts/armada-385.dtsi         |  52 ++++++--
 arch/arm/boot/dts/armada-39x.dtsi         |  56 +++++++--
 arch/arm/boot/dts/armada-xp-98dx3236.dtsi |  14 ++-
 arch/arm/boot/dts/armada-xp-mv78230.dtsi  |  70 +++++++++--
 arch/arm/boot/dts/armada-xp-mv78260.dtsi  | 126 ++++++++++++++++---
 arch/arm/boot/dts/armada-xp-mv78460.dtsi  | 140 ++++++++++++++++++----
 arch/arm/boot/dts/dove.dtsi               |  28 ++++-
 arch/arm/boot/dts/kirkwood-6192.dtsi      |  14 ++-
 arch/arm/boot/dts/kirkwood-6281.dtsi      |  14 ++-
 arch/arm/boot/dts/kirkwood-6282.dtsi      |  28 ++++-
 arch/arm/boot/dts/kirkwood-98dx4122.dtsi  |  14 ++-
 14 files changed, 560 insertions(+), 94 deletions(-)

Comments

Andrew Lunn July 6, 2022, 8:53 p.m. UTC | #1
On Wed, Jul 06, 2022 at 08:31:03PM +0200, Pali Rohár wrote:
> This patch series add definitions for PCIe legacy INTx interrupts into
> every DTS file used by the pci-mvebu.c controller driver.

What you fail to explain in the commit message is Why?

     Andrew
Marek Behún July 7, 2022, 2:38 p.m. UTC | #2
On Wed, 6 Jul 2022 22:53:46 +0200
Andrew Lunn <andrew@lunn.ch> wrote:

> On Wed, Jul 06, 2022 at 08:31:03PM +0200, Pali Rohár wrote:
> > This patch series add definitions for PCIe legacy INTx interrupts into
> > every DTS file used by the pci-mvebu.c controller driver.  
> 
> What you fail to explain in the commit message is Why?
> 
>      Andrew

Hi Andrew,

this is required for example in a scenario where a driver requests only
one of the legacy interrupts (INTA). Without this change, the driver
would be notified on events on all 4 (INTA, INTB, INTC, INTD), even if
it requested only one of them.

Marek
Andrew Lunn July 7, 2022, 4:50 p.m. UTC | #3
On Thu, Jul 07, 2022 at 04:38:18PM +0200, Marek Behún wrote:
> On Wed, 6 Jul 2022 22:53:46 +0200
> Andrew Lunn <andrew@lunn.ch> wrote:
> 
> > On Wed, Jul 06, 2022 at 08:31:03PM +0200, Pali Rohár wrote:
> > > This patch series add definitions for PCIe legacy INTx interrupts into
> > > every DTS file used by the pci-mvebu.c controller driver.  
> > 
> > What you fail to explain in the commit message is Why?
> > 
> >      Andrew
> 
> Hi Andrew,
> 
> this is required for example in a scenario where a driver requests only
> one of the legacy interrupts (INTA). Without this change, the driver
> would be notified on events on all 4 (INTA, INTB, INTC, INTD), even if
> it requested only one of them.

O.K, so please add that to the commit message.

Is this purely theoretical, or are there known broken platforms out
there?  Should a subset of these patches be added to stable?

	Andrew