On 9/21/19 11:20 PM, DJ Lucas via lfs-dev wrote:
On 9/21/2019 11:13 PM, Bruce Dubbs via lfs-dev wrote:
Seems odd to depend on the overall setting. If loadable modules
support is set, but you don't build any modules, why should the SW care?
I agree that the above would be a little unusual, but the test should
not be whether the kernel is wired for modules, but if the module
files actually exist.
I don't have an answer for that. I'm not sure that it should. I guess
I'll dig through kernel archives to see if this is something legacy and
forgotten, or a deliberate choice. Short answer is that we should
probably be doing this, even if we've built no modules. I just don't
know how to handle the case where "# CONFIG_MODULES is not set" -
obviously this can't be a supported configuration for firewalld as it is
currently written, and I don't know if there is a more elegant solution.
I wonder if the firewalld authors have considered the case of a kernel
without modules. I don't know of any of the major distros that do not
use modules. User configured kernels are a relatively rare situation
(not for us of course). You can make a case for saying the not building
any modules but enabling module support is a misconfiguration.
What we have is:
Install the modules, if the kernel configuration uses them:
make modules_install
Perhaps a simple rewording to:
Install the modules, unless module support has been disabled in the
kernel configuration:
Or reversing:
Unless module support has been disabled in the kernel configuration,
install the modules with:
-- Bruce
--
http://lists.linuxfromscratch.org/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
