On 11/11/21 11:27 AM, Brijesh Singh wrote:

Hi Randy,

On 11/10/21 4:27 PM, Randy Dunlap wrote:

quoted

Hi,

On 11/10/21 2:07 PM, Brijesh Singh wrote:

quoted

diff --git a/drivers/virt/coco/sevguest/Kconfig b/drivers/virt/coco/sevguest/Kconfig
new file mode 100644
index 000000000000..96190919cca8
--- /dev/null
+++ b/drivers/virt/coco/sevguest/Kconfig

@@ -0,0 +1,9 @@
+config SEV_GUEST
+    tristate "AMD SEV Guest driver"
+    default y

For this to remain as "default y", you need to justify it.
E.g., if a board cannot boot with an interrupt controller,
the driver for the interrupt controller can be "default y".

So why is this default y?
No other drivers in drivers/virt/ are default y.

I choose the default "y" for two reasons:

1.  The driver is built if the user enables the AMD memory encryption support. If the user has selected the AMD memory encryption support, they will be querying an attestation report to verify that the guest is running on AMD memory encryption enabled hardware.

OK, I see. I'm OK with this.

2. Typically, an attestation report is retrieved from an initial ramdisk (before mounting the disk). IIUC, the standard initramfs build tools may not include the driver by default and requires the user to go through hoops.

However, I have no strong reason to keep it to "y" if other prefers "m".

"m" is no better than "y" in this case.

thanks.
-- 
~Randy