[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2024091428-unison-relocate-4c2b@gregkh>
Date: Sat, 14 Sep 2024 09:09:51 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: Thorsten Leemhuis <regressions@...mhuis.info>
Cc: "stable@...r.kernel.org" <stable@...r.kernel.org>,
Linux kernel regressions list <regressions@...ts.linux.dev>,
Sasha Levin <sashal@...nel.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: Bug 219269 - v6.10.10 kernel fails to build when CONFIG_MODULES
is disabled
On Sat, Sep 14, 2024 at 07:02:26AM +0200, Thorsten Leemhuis wrote:
> Hi Greg! I noticed a bug report in bz:
> https://bugzilla.kernel.org/show_bug.cgi?id=219269
>
> > Fair enough, you get a compiler warning:
> >
> > kernel/trace/trace_kprobe.c: In function ‘validate_probe_symbol’:
> > kernel/trace/trace_kprobe.c:810:23: error: implicit declaration of function ‘find_module’; did you mean init_module’? [-Wimplicit-function-declaration]
> > 810 | mod = find_module(modname);
> > | ^~~~~~~~~~~
> > | init_module
> > kernel/trace/trace_kprobe.c:810:21: error: assignment to ‘struct module *’ from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
> > 810 | mod = find_module(modname);
> > |
> >
> > but there is no find_module symbol when CONFIG_MODULES is disabled.
>
> I *very briefly* looked into this. I might be wrong, but looked a bit
> like "tracing/kprobes: Add symbol counting check when module loads"
> caused this and backporting b10545b6b86b7a ("tracing/kprobes: Fix build
> error when find_module() is not available") [v6.11-rc1] would fix this
> (which applies cleanly).
>
> Shall I ask the reporter to confirm or is that already enough for you?
This is enough, now queued up, thanks.
greg k-h
Powered by blists - more mailing lists