CI: do not build gpu for musl targets

Do not build the GPU device for musl targets. To compile it properly would
require setting up a build enivorment with all the native dependencies
compiled with musl as well.

Signed-off-by: Matej Hrica <mhrica@redhat.com>
This commit is contained in:
Matej Hrica 2025-01-21 13:59:54 +01:00 committed by Stefano Garzarella
parent a2f0f91857
commit 7488f44220
2 changed files with 7 additions and 2 deletions

View File

@ -11,7 +11,7 @@
},
{
"test_name": "staging: build-musl",
"command": "cd staging && RUSTFLAGS=\"-D warnings\" cargo build --release --target {target_platform}-unknown-linux-musl",
"command": "cd staging && RUSTFLAGS=\"-D warnings\" cargo build --release --workspace --exclude vhost-device-gpu --target {target_platform}-unknown-linux-musl",
"soft_fail": "true",
"platform": [
"x86_64",
@ -33,7 +33,7 @@
},
{
"test_name": "staging: unittests-musl",
"command": "cd staging && cargo test --all-features --workspace --target {target_platform}-unknown-linux-musl",
"command": "cd staging && cargo test --all-features --workspace --exclude vhost-device-gpu --target {target_platform}-unknown-linux-musl",
"soft_fail": "true",
"platform": [
"x86_64",

View File

@ -24,6 +24,11 @@ crate has been built with the `gfxstream` feature, which is the default.
## Limitations
This device links native libaries (because of the usage of Rutabaga) compiled
with GNU libc, so the CI is setup to not build this device for musl targets.
It might be possible to build those libraries using musl and then build the gpu
device, but this is not tested.
We are currently only supporting sharing the display output to QEMU through a
socket using the transfer_read operation triggered by
`VIRTIO_GPU_CMD_TRANSFER_FROM_HOST_3D` to transfer data from and to virtio-gpu 3D