Re: virtio-pci in qemu-system-arm is broken in 8.2

[Michael Tokarev]

08.01.2024 20:07, Peter Maydell wrote:
..
> I'm trying to figure out what kind of kernel config causes
> this. If I do an upstream build of v6.6.9 with this config:
> https://people.linaro.org/~peter.maydell/arm-6.6.9-config
>
> then it successfully probes the PCI device even with -cpu max:

(with -cpu max it worked before but broke in 8.2 (after that commit).
Without -cpu max it doesn't work for me neither before nor in 8.2.
But ok.)

Speaking of config. This is standard debian config, I'm attaching one
to this email.  It can be found in the package, eg
http://deb.debian.org/debian/pool/main/l/linux/linux-image-6.6.9-armmp_6.6.9-1_armhf.deb
in /boot/config-$(uname -r).

I haven't tried very hard to compare the two, but maybe this change
is worth checking already?

@@ -601,3 +420,2 @@
-CONFIG_VMSPLIT_3G=y
-# CONFIG_VMSPLIT_3G_OPT is not set
-# CONFIG_VMSPLIT_2G is not set
+# CONFIG_VMSPLIT_3G is not set
+CONFIG_VMSPLIT_2G=y

I tried to bisect kernel, when it started failing with this (without
-cpu max), - as a base I used this config from debian 6.6.9 package,
and went a few releases back, being able to reproduce the failure
easily (building just bzImage, with a few virtio modules built-in).
But going back it failed to build with current gcc, so stopped there.

/mjt

config-6.6.9-armmp
Description: Text document