2019-01-10 21:12:29

by Alistair Francis

[permalink] [raw]
Subject: [PATCH v2] RISC-V: defconfig: Enable Generic PCIE by default

Enable generic PCIe by default in the RISC-V defconfig, this allows us
to use QEMU's PCIe support out of the box.

Signed-off-by: Alistair Francis <[email protected]>
---
arch/riscv/configs/defconfig | 3 +++
1 file changed, 3 insertions(+)

diff --git a/arch/riscv/configs/defconfig b/arch/riscv/configs/defconfig
index f399659d3b8d..ccfd3f05223c 100644
--- a/arch/riscv/configs/defconfig
+++ b/arch/riscv/configs/defconfig
@@ -14,6 +14,8 @@ CONFIG_EXPERT=y
CONFIG_BPF_SYSCALL=y
CONFIG_SMP=y
CONFIG_PCI=y
+CONFIG_PCIEPORTBUS=y
+CONFIG_PCI_HOST_GENERIC=y
CONFIG_PCIE_XILINX=y
CONFIG_MODULES=y
CONFIG_MODULE_UNLOAD=y
@@ -77,5 +79,6 @@ CONFIG_NFS_V4_1=y
CONFIG_NFS_V4_2=y
CONFIG_ROOT_NFS=y
CONFIG_CRYPTO_USER_API_HASH=y
+CONFIG_CRYPTO_DEV_VIRTIO=y
CONFIG_PRINTK_TIME=y
# CONFIG_RCU_TRACE is not set
--
2.19.1



2019-01-11 08:23:53

by Paul Walmsley

[permalink] [raw]
Subject: Re: [PATCH v2] RISC-V: defconfig: Enable Generic PCIE by default

Hi Alistair,

Looks like you're still sending patches to "palmpalmer" ?

On Thu, 10 Jan 2019, Alistair Francis wrote:

> Enable generic PCIe by default in the RISC-V defconfig, this allows us
> to use QEMU's PCIe support out of the box.
>
> Signed-off-by: Alistair Francis <[email protected]>
> ---
> arch/riscv/configs/defconfig | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/arch/riscv/configs/defconfig b/arch/riscv/configs/defconfig
> index f399659d3b8d..ccfd3f05223c 100644
> --- a/arch/riscv/configs/defconfig
> +++ b/arch/riscv/configs/defconfig
> @@ -14,6 +14,8 @@ CONFIG_EXPERT=y
> CONFIG_BPF_SYSCALL=y
> CONFIG_SMP=y
> CONFIG_PCI=y
> +CONFIG_PCIEPORTBUS=y
> +CONFIG_PCI_HOST_GENERIC=y
> CONFIG_PCIE_XILINX=y
> CONFIG_MODULES=y
> CONFIG_MODULE_UNLOAD=y
> @@ -77,5 +79,6 @@ CONFIG_NFS_V4_1=y
> CONFIG_NFS_V4_2=y
> CONFIG_ROOT_NFS=y
> CONFIG_CRYPTO_USER_API_HASH=y
> +CONFIG_CRYPTO_DEV_VIRTIO=y
> CONFIG_PRINTK_TIME=y
> # CONFIG_RCU_TRACE is not set

Not sure what commit this patch is against - I picked v5.0-rc1. Before
applying your patch, I generated a seedconfig with "make defconfig
savedefconfig; mv defconfig defconfig.1". Then applied your patch and did
a "make defconfig savedefconfig" and then ran "diff -u defconfig.1
defconfig".

That should have shown me just the patch above. But it also included a
line removing CONFIG_RAS=y. I am sorry to ask you to fix such a nitpicky
item, but it is good practice for these defconfig patches. Otherwise the
defconfig deltas gradually get out of control.

When you repost it, if you verify that the ./defconfig delta from running
"make savedefconfig" before and after your patch applied is simply your
patch, then feel free to add a Reviewed-by: line from my E-mail address
above.


- Paul

2019-01-11 20:55:20

by Palmer Dabbelt

[permalink] [raw]
Subject: Re: [PATCH v2] RISC-V: defconfig: Enable Generic PCIE by default

On Thu, 10 Jan 2019 11:35:01 PST (-0800), Alistair Francis wrote:
> Enable generic PCIe by default in the RISC-V defconfig, this allows us
> to use QEMU's PCIe support out of the box.
>
> Signed-off-by: Alistair Francis <[email protected]>
> ---
> arch/riscv/configs/defconfig | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/arch/riscv/configs/defconfig b/arch/riscv/configs/defconfig
> index f399659d3b8d..ccfd3f05223c 100644
> --- a/arch/riscv/configs/defconfig
> +++ b/arch/riscv/configs/defconfig
> @@ -14,6 +14,8 @@ CONFIG_EXPERT=y
> CONFIG_BPF_SYSCALL=y
> CONFIG_SMP=y
> CONFIG_PCI=y
> +CONFIG_PCIEPORTBUS=y
> +CONFIG_PCI_HOST_GENERIC=y
> CONFIG_PCIE_XILINX=y
> CONFIG_MODULES=y
> CONFIG_MODULE_UNLOAD=y
> @@ -77,5 +79,6 @@ CONFIG_NFS_V4_1=y
> CONFIG_NFS_V4_2=y
> CONFIG_ROOT_NFS=y
> CONFIG_CRYPTO_USER_API_HASH=y
> +CONFIG_CRYPTO_DEV_VIRTIO=y
> CONFIG_PRINTK_TIME=y
> # CONFIG_RCU_TRACE is not set
> --
> 2.19.1

Odd -- when your QEMU patches came through I tried doing this, which resulted
in a non-booting kernel. For some reason this patch is fine, though. I'm
going to just assume I screwed this up the first time, so it looks like this
and the QEMU patch are good to go.

Thanks!