Message ID | 20240607-rockchip-pcie-ep-v1-v5-0-0a042d6b0049@kernel.org (mailing list archive) |
---|---|
Headers | show |
Series | PCI: dw-rockchip: Add endpoint mode support | expand |
On Fri, Jun 07, 2024 at 01:14:20PM +0200, Niklas Cassel wrote: > Hello all, > > This series adds PCIe endpoint mode support for the rockchip rk3588 and > rk3568 SoCs. > > This series is based on: pci/next > (git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git) > > This series can also be found in git: > https://github.com/floatious/linux/commits/rockchip-pcie-ep-v5 > > Testing done: > This series has been tested with two rock5b:s, one running in RC mode and > one running in EP mode. This series has also been tested with an Intel x86 > host and rock5b running in EP mode. (snip) Hello PCI maintainers, If there is anything more I can do to get this picked up, please tell me. Kind regards, Niklas
Hello,
[...]
> If there is anything more I can do to get this picked up, please tell me.
Looks good! As such...
Applied to dt-bindings, thank you!
[01/06] dt-bindings: PCI: snps,dw-pcie-ep: Add vendor specific reg-name
https://git.kernel.org/pci/pci/c/3b287269ab60
[02/06] dt-bindings: PCI: snps,dw-pcie-ep: Add vendor specific interrupt-names
https://git.kernel.org/pci/pci/c/b96353773d24
[03/06] dt-bindings: PCI: snps,dw-pcie-ep: Add tx_int{a,b,c,d} legacy IRQs
https://git.kernel.org/pci/pci/c/6f308c017c27
[04/06] dt-bindings: PCI: rockchip-dw-pcie: Prepare for Endpoint mode support
https://git.kernel.org/pci/pci/c/9b0b9b588c00
[05/06] dt-bindings: PCI: rockchip-dw-pcie: Fix description of legacy IRQ
https://git.kernel.org/pci/pci/c/5f262f67cbc5
[06/06] dt-bindings: rockchip: Add DesignWare based PCIe Endpoint controller
https://git.kernel.org/pci/pci/c/ff36edde817e
Applied to controller/rockchip, thank you!
[01/04] PCI: dw-rockchip: Fix weird indentation
https://git.kernel.org/pci/pci/c/e7e8872191af
[02/04] PCI: dw-rockchip: Add rockchip_pcie_get_ltssm() helper
https://git.kernel.org/pci/pci/c/cbb2d4ae3fdc
[03/04] PCI: dw-rockchip: Add endpoint mode support
https://git.kernel.org/pci/pci/c/67fe449bcd85
[04/04] PCI: dw-rockchip: Refactor the driver to prepare for EP mode
https://git.kernel.org/pci/pci/c/ecdc98a3a912
Applied to endpoint, thank you!
[1/1] misc: pci_endpoint_test: Add support for Rockchip rk3588
https://git.kernel.org/pci/pci/c/657463e393d1
Krzysztof
Krzysztof, thank you very much for applying! Heiko, now when the DT-binding and driver has been queued for 6.11, is there any chance of getting patches 12/13 and 13/13 applied to the linux-rockchip tree? Kind regards, Niklas
On Sat, Jun 22, 2024 at 04:39:37AM +0900, Krzysztof Wilczyński wrote: > Hello, > > [...] > > If there is anything more I can do to get this picked up, please tell me. > > Looks good! As such... > > Applied to controller/rockchip, thank you! > > [01/04] PCI: dw-rockchip: Fix weird indentation > https://git.kernel.org/pci/pci/c/e7e8872191af > > [02/04] PCI: dw-rockchip: Add rockchip_pcie_get_ltssm() helper > https://git.kernel.org/pci/pci/c/cbb2d4ae3fdc > > [03/04] PCI: dw-rockchip: Add endpoint mode support > https://git.kernel.org/pci/pci/c/67fe449bcd85 > > [04/04] PCI: dw-rockchip: Refactor the driver to prepare for EP mode > https://git.kernel.org/pci/pci/c/ecdc98a3a912 Krzysztof, unfortunately, the controller/rockchip branch currently doesn't build: drivers/pci/controller/dwc/pcie-dw-rockchip.c: In function ‘rockchip_pcie_ep_sys_irq_thread’: drivers/pci/controller/dwc/pcie-dw-rockchip.c:407:17: error: implicit declaration of function ‘dw_pcie_ep_linkdown’; did you mean ‘dw_pcie_ep_linkup’? [-Wimplicit-function-declaration] 407 | dw_pcie_ep_linkdown(&pci->ep); | ^~~~~~~~~~~~~~~~~~~ | dw_pcie_ep_linkup Could you possibly include the commit: 3d2e425263e2 ("PCI: dwc: ep: Add a generic dw_pcie_ep_linkdown() API to handle Link Down event") from the controller/dwc branch in the controller/rockchip as well, or rebase the controller/rockchip branch on top of the controller/dwc branch, or merge the controller/dwc branch to the controller/rockchip branch? Additionally, since you picked up Mani's series which removes dw_pcie_ep_init_notify() on the controller/dwc branch: 9eba2f70362f ("PCI: dwc: ep: Remove dw_pcie_ep_init_notify() wrapper") You will need to pick up this patch as well: https://lore.kernel.org/linux-pci/20240622132024.2927799-2-cassel@kernel.org/T/#u Otherwise there will be a build error when merging the controller/dwc and the controller/rockchip branch to for-next. The patch that I sent out can be picked up to the controller/rockchip right now (since the API that Mani is switching to already exists in Linus's tree). May I ask why all the branches for the different DWC glue drivers are not based on the controller/dwc branch? They are obviously going to be tightly related. Kind regards, Niklas
Hello, [...] > Could you possibly include the commit: > 3d2e425263e2 ("PCI: dwc: ep: Add a generic dw_pcie_ep_linkdown() API to handle Link Down event") > from the controller/dwc branch in the controller/rockchip as well, > or rebase the controller/rockchip branch on top of the controller/dwc branch, > or merge the controller/dwc branch to the controller/rockchip branch? > Done. > Additionally, since you picked up Mani's series which removes > dw_pcie_ep_init_notify() on the controller/dwc branch: > 9eba2f70362f ("PCI: dwc: ep: Remove dw_pcie_ep_init_notify() wrapper") > > You will need to pick up this patch as well: > https://lore.kernel.org/linux-pci/20240622132024.2927799-2-cassel@kernel.org/T/#u > Otherwise there will be a build error when merging the controller/dwc > and the controller/rockchip branch to for-next. > The patch that I sent out can be picked up to the controller/rockchip right > now (since the API that Mani is switching to already exists in Linus's tree). Done. Hopefully, this settles things for a bit. > May I ask why all the branches for the different DWC glue drivers are not > based on the controller/dwc branch? No worries. > They are obviously going to be tightly related. Normally, we prefer to apply things to specific topic branches, but I will revisit this approach going forward, since changes between Endpoint, DWC and specific controller drivers are often tightly coupled, as you noted, which can make things a bit of a mess, unnecessarily. We are going to do some clean up once Bjorn sends his Pull Request. Krzysztof
On Sun, Jun 23, 2024 at 12:43:24AM +0900, Krzysztof Wilczyński wrote: > Hello, > > [...] > > Could you possibly include the commit: > > 3d2e425263e2 ("PCI: dwc: ep: Add a generic dw_pcie_ep_linkdown() API to handle Link Down event") > > from the controller/dwc branch in the controller/rockchip as well, > > or rebase the controller/rockchip branch on top of the controller/dwc branch, > > or merge the controller/dwc branch to the controller/rockchip branch? > > > > Done. > > > Additionally, since you picked up Mani's series which removes > > dw_pcie_ep_init_notify() on the controller/dwc branch: > > 9eba2f70362f ("PCI: dwc: ep: Remove dw_pcie_ep_init_notify() wrapper") > > > > You will need to pick up this patch as well: > > https://lore.kernel.org/linux-pci/20240622132024.2927799-2-cassel@kernel.org/T/#u > > Otherwise there will be a build error when merging the controller/dwc > > and the controller/rockchip branch to for-next. > > The patch that I sent out can be picked up to the controller/rockchip right > > now (since the API that Mani is switching to already exists in Linus's tree). > > Done. > > Hopefully, this settles things for a bit. Everything looks good! :) I'm glad that we got this sorted quickly, thank you Krzysztof! Kind regards, Niklas
On Fri, 07 Jun 2024 13:14:20 +0200, Niklas Cassel wrote: > This series adds PCIe endpoint mode support for the rockchip rk3588 and > rk3568 SoCs. > > This series is based on: pci/next > (git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git) > > This series can also be found in git: > https://github.com/floatious/linux/commits/rockchip-pcie-ep-v5 > > [...] Applied, thanks! [12/13] arm64: dts: rockchip: Add PCIe endpoint mode support commit: 2fe9fe4e54f5763b8b681478dda9ac61fd42ecaf [13/13] arm64: dts: rockchip: Add rock5b overlays for PCIe endpoint mode commit: 41367db58cbf51ecb89ca017b7473688345caa7b I've dropped the overlay-symbol-enablement for now. As this creates massive size increases there have actually been concerns of things like TF-A getting overwhelmed by the size if I remember correctly. In any case, right now we don't have an established way on how to handle overlay symbold for Rockchip boards. For example broadcom enables symbols for all DTs, Nvidia and TI do it for select boards only, while for example Mediatek and Freescale do not handle symbols at all right now. So I'll just postpone that decision for a bit. Best regards,
On Wed, Jun 26, 2024 at 05:32:49PM +0200, Heiko Stuebner wrote: > On Fri, 07 Jun 2024 13:14:20 +0200, Niklas Cassel wrote: > > This series adds PCIe endpoint mode support for the rockchip rk3588 and > > rk3568 SoCs. > > > > This series is based on: pci/next > > (git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git) > > > > This series can also be found in git: > > https://github.com/floatious/linux/commits/rockchip-pcie-ep-v5 > > > > [...] > > Applied, thanks! > > [12/13] arm64: dts: rockchip: Add PCIe endpoint mode support > commit: 2fe9fe4e54f5763b8b681478dda9ac61fd42ecaf > [13/13] arm64: dts: rockchip: Add rock5b overlays for PCIe endpoint mode > commit: 41367db58cbf51ecb89ca017b7473688345caa7b > > I've dropped the overlay-symbol-enablement for now. > As this creates massive size increases there have actually > been concerns of things like TF-A getting overwhelmed by > the size if I remember correctly. > > In any case, right now we don't have an established way on > how to handle overlay symbold for Rockchip boards. > > For example broadcom enables symbols for all DTs, Nvidia and TI do > it for select boards only, while for example Mediatek and Freescale > do not handle symbols at all right now. > > So I'll just postpone that decision for a bit. Okay, I see your argument. Thank you for applying, I just realized that rk3588.dtsi has been renamed to rk3588-extra.dtsi, so I was about to rebase and resend these two patches. The conflict was trivial, and it looks correct in your tree, so thanks a lot for fixing this up! Kind regards, Niklas