Bug#1031561: v4l2loopback-dkms: Fails to build on cloud kernel image

[IOhannes m zmoelnig]

On Sun, 19 Feb 2023 21:36:56 +0100 Andreas Beckmann <a...@debian.org> wrote:

> i would therefore ask the dkms maintainers to consider making > 'BUILD_EXCLUSE_KERNEL' a soft-failure (that is ignored in the automatic > case).

Not trivial.

ah, my naive assumption was that i could do something like this

```
diff --git a/etc/kernel/postinst.d/dkms b/etc/kernel/postinst.d/dkms
index c0b9cda..b1fc2ff 100755
--- a/etc/kernel/postinst.d/dkms
+++ b/etc/kernel/postinst.d/dkms
@@ -35,7 +35,17 @@ case "${uname_s}" in
 esac


+    ret=$?
+    case $ret in
+           9)
+                   # ignore failures due to BUILD_EXCLUDE_...
+                   ret=0
+                   ;;
+           *)
+                   ;;
+    esac
+    exit $ret
 fi

 if ! _check_kernel_dir "$inst_kern" ; then
```

(or something simpler).
which looks trivial enough.
i haven't tested it though.



"9" is used for:
- BUILD_EXCLUSIVE
- missing write permissions on tarball
- unable to install ... using rpm
- missing dkms.conf file



sigh.



gdmasr
IOhannes

if [ -x /usr/lib/etc/kernel/postinst.d/dkms/etc/kernel/postinst.d/dkms_autoinstaller ]; then - exec /usr/lib/etc/kernel/postinst.d/dkms/etc/kernel/postinst.d/dkms_autoinstaller start "$inst_kern" + /usr/lib/etc/kernel/postinst.d/dkms/etc/kernel/postinst.d/dkms_autoinstaller start "$inst_kern" and on second thought i see that the error-codes are used inconsistently in dkms, and what's worse they are re-used. of course, according to errno.h, '9' is EBADF, aka "Bad file number", which makes no sense in any of the cases. so i guess a good start would be to convince upstream to issue unique (and hopefully meaningful) status-codes for the various errors...

OpenPGP_signature
Description: OpenPGP digital signature