W dniu sob, 10.03.2018 o godzinie 01∶05 +0000, użytkownik Robin H.
Johnson napisał:
> On Thu, Mar 08, 2018 at 06:05:43PM +0100, Michał Górny wrote:
> > Make linux-info_get_any_version die if it can't determine any version
> > of the Linux kernel. This indicates a problem with the eclass code
> > (as it should not happen on Linux) and the missing KV_* variables
> > are going to cause random misbehavior and failures.
> 
> This change worries me a bit. get_running_version calls get_version in
> some cases, and there are corner cases there which could trip this up.
> 
> linux-info_get_any_version -> get_running_version -> get_version
> with get_version returning non-zero.
> 
> get_version has ALREADY returned non-zero once in that path, and we're
> potentially tweaking KERNEL_DIR/KBUILD_OUTPUT before calling it again.
> 
> Most of the breakage cases IIRC are where
> /lib/modules/${KV_FULL}/*/Makefile still exist, but point to kernel
> sources in weird places that are broken.

Yes, and this needs to be fixed one way or another. Otherwise you don't
get KV_* exported and a lot of other logic trips on unset variables.

-- 
Best regards,
Michał Górny


Reply via email to