| Message ID | 20230606105122.520421-1-richard.purdie@linuxfoundation.org |
|---|---|
| State | Accepted, archived |
| Commit | 37ccd11cb0b89416b8e23160445186269b6c0c8a |
| Headers | show |
| Series | v86d: Improve kernel dependency | expand |
diff --git a/meta/recipes-bsp/v86d/v86d_0.1.10.bb b/meta/recipes-bsp/v86d/v86d_0.1.10.bb index 5f342b1120e..b4fe362f8e3 100644 --- a/meta/recipes-bsp/v86d/v86d_0.1.10.bb +++ b/meta/recipes-bsp/v86d/v86d_0.1.10.bb @@ -6,7 +6,6 @@ DESCRIPTION = "v86d provides a backend for kernel drivers that need to execute x LICENSE = "GPL-2.0-only" LIC_FILES_CHKSUM = "file://README;md5=94ac1971e4f2309dc322d598e7b1f7dd" -DEPENDS = "virtual/kernel" RRECOMMENDS:${PN} = "kernel-module-uvesafb" PR = "r2"
Working with enabling SPDX, an issue was observerd where v86d wasn't rebuilding when the kernel was changed from linux-yocto to linux-yocto-rt. This is due to the code in sstatesig.py which was seeing the RRECOMMENDS on a kernel module and ignoring the DEPENDS. The v86d is technically a kernel module since it uses kernel header files. There are two ways to address this, we could inherit the module-base class and the dependency code does the correct thing. It appears the code doesn't look into STAGING_KERNEL_DIR though and doesn't use the kernel sources. We can therefore drop the DEPENDS and the code will the do the correct thing. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> --- meta/recipes-bsp/v86d/v86d_0.1.10.bb | 1 - 1 file changed, 1 deletion(-)