[PATCH] arm64/defconfig: Restore built-in 9P rootfs support

[PATCH] arm64/defconfig: Restore built-in 9P rootfs support

[Anshuman Khandual]

Commit d521c4ca039f ("arm64: defconfig: Switch Ethernet drivers to modules")
had changed 9P and virtio-9P support from built-in to modular instead. This
breaks simple kvmtool boots using a 9P root filesystem without an initramfs
, because the kernel must mount root filesystem before it can load the 9P
modules. Also there is no change in image size with these configs built-in.

Restore the relevant 9P options as built-in.

Cc: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Linus Walleij <linusw@kernel.org>
Cc: Alexandre Belloni <alexandre.belloni@bootlin.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
---
 arch/arm64/configs/defconfig | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 654a102cb5bc..be9715b1e005 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -200,8 +200,8 @@ CONFIG_MAC80211=m
 CONFIG_MAC80211_LEDS=y
 CONFIG_RFKILL=m
 CONFIG_RFKILL_GPIO=m
-CONFIG_NET_9P=m
-CONFIG_NET_9P_VIRTIO=m
+CONFIG_NET_9P=y
+CONFIG_NET_9P_VIRTIO=y
 CONFIG_NFC=m
 CONFIG_NFC_NCI=m
 CONFIG_NFC_NXP_NCI=m
@@ -1921,7 +1921,7 @@ CONFIG_NFS_V4=y
 CONFIG_NFS_V4_1=y
 CONFIG_NFS_V4_2=y
 CONFIG_ROOT_NFS=y
-CONFIG_9P_FS=m
+CONFIG_9P_FS=y
 CONFIG_NLS_CODEPAGE_437=y
 CONFIG_NLS_ISO8859_1=y
 CONFIG_SECURITY=y
-- 
2.30.2

Re: [PATCH] arm64/defconfig: Restore built-in 9P rootfs support

[Krzysztof Kozlowski]

On 16/07/2026 12:02, Anshuman Khandual wrote:
> Commit d521c4ca039f ("arm64: defconfig: Switch Ethernet drivers to modules")
> had changed 9P and virtio-9P support from built-in to modular instead. This
> breaks simple kvmtool boots using a 9P root filesystem without an initramfs
> , because the kernel must mount root filesystem before it can load the 9P


And that commit explained you must use initramfs. Nothing got broken,
because there was never a promise that a recent, virtualized environment
must work on defconfig without initramfs.

You fully control your software, so just build initramfs which is
absolutely trivial and we all moved on to that for arm64 long time ago.

> modules. Also there is no change in image size with these configs built-in.

How? How making a module a built in has no effect on size? How is this
even possible?




> 
> Restore the relevant 9P options as built-in.
> 
> Cc: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Eric Biggers <ebiggers@kernel.org>
> Cc: Linus Walleij <linusw@kernel.org>
> Cc: Alexandre Belloni <alexandre.belloni@bootlin.com>
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-kernel@vger.kernel.org
> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
> ---
>  arch/arm64/configs/defconfig | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> index 654a102cb5bc..be9715b1e005 100644
> --- a/arch/arm64/configs/defconfig
> +++ b/arch/arm64/configs/defconfig
> @@ -200,8 +200,8 @@ CONFIG_MAC80211=m
>  CONFIG_MAC80211_LEDS=y
>  CONFIG_RFKILL=m
>  CONFIG_RFKILL_GPIO=m
> -CONFIG_NET_9P=m
> -CONFIG_NET_9P_VIRTIO=m
> +CONFIG_NET_9P=y
> +CONFIG_NET_9P_VIRTIO=y

No, upgrade your build system from 90-ties to decent one to handle
initramfs.


Best regards,
Krzysztof

Re: [PATCH] arm64/defconfig: Restore built-in 9P rootfs support

[Arnd Bergmann]

On Thu, Jul 16, 2026, at 13:01, Krzysztof Kozlowski wrote:
> On 16/07/2026 12:02, Anshuman Khandual wrote:
>> Commit d521c4ca039f ("arm64: defconfig: Switch Ethernet drivers to modules")
>> had changed 9P and virtio-9P support from built-in to modular instead. This
>> breaks simple kvmtool boots using a 9P root filesystem without an initramfs
>> , because the kernel must mount root filesystem before it can load the 9P
>
>
> And that commit explained you must use initramfs. Nothing got broken,
> because there was never a promise that a recent, virtualized environment
> must work on defconfig without initramfs.
>
> You fully control your software, so just build initramfs which is
> absolutely trivial and we all moved on to that for arm64 long time ago.

I still regularly boot KVM guests on arm64 without an initramfs,
but I just use 'make kvm_guest.config', which turns on all the
commonly used virtio drivers as built-in. I wouldn't call
9pfs root trivial without initramfs, configuring 9p built-in
is already complicated enough.

So I agree we don't need this patch, but there is a trivial
workaround for Anshuman in case he wants to avoid doing this.

      Arnd

Re: [PATCH] arm64/defconfig: Restore built-in 9P rootfs support

[Anshuman Khandual]

On 16/07/26 5:01 PM, Arnd Bergmann wrote:
> On Thu, Jul 16, 2026, at 13:01, Krzysztof Kozlowski wrote:
>> On 16/07/2026 12:02, Anshuman Khandual wrote:
>>> Commit d521c4ca039f ("arm64: defconfig: Switch Ethernet drivers to modules")
>>> had changed 9P and virtio-9P support from built-in to modular instead. This
>>> breaks simple kvmtool boots using a 9P root filesystem without an initramfs
>>> , because the kernel must mount root filesystem before it can load the 9P
>>
>>
>> And that commit explained you must use initramfs. Nothing got broken,
>> because there was never a promise that a recent, virtualized environment
>> must work on defconfig without initramfs.
>>
>> You fully control your software, so just build initramfs which is
>> absolutely trivial and we all moved on to that for arm64 long time ago.
> 
> I still regularly boot KVM guests on arm64 without an initramfs,

Actually that is a very normal use case particularly with KVM tool.
> but I just use 'make kvm_guest.config', which turns on all the
> commonly used virtio drivers as built-in. I wouldn't call
> 9pfs root trivial without initramfs, configuring 9p built-in
> is already complicated enough.

Alright - fair enough.
> 
> So I agree we don't need this patch, but there is a trivial
> workaround for Anshuman in case he wants to avoid doing this.
The following sequence does indeed solve this problem for now.

make mrproper
make defconfig
make kvm_guest.config