[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240904161133.oyjuwjrm4rjdfivx@treble>
Date: Wed, 4 Sep 2024 09:11:33 -0700
From: Josh Poimboeuf <jpoimboe@...nel.org>
To: Peter Zijlstra <peterz@...radead.org>
Cc: live-patching@...r.kernel.org, linux-kernel@...r.kernel.org,
x86@...nel.org, Miroslav Benes <mbenes@...e.cz>,
Petr Mladek <pmladek@...e.com>,
Joe Lawrence <joe.lawrence@...hat.com>,
Jiri Kosina <jikos@...nel.org>,
Marcos Paulo de Souza <mpdesouza@...e.com>,
Song Liu <song@...nel.org>
Subject: Re: [RFC 29/31] objtool: Calculate function checksums
On Wed, Sep 04, 2024 at 09:54:33AM +0200, Peter Zijlstra wrote:
> > +LIBXXHASH_FLAGS := $(shell $(HOSTPKG_CONFIG) libxxhash --cflags 2>/dev/null)
> > +LIBXXHASH_LIBS := $(shell $(HOSTPKG_CONFIG) libxxhash --libs 2>/dev/null || echo -lxxhash)
>
> This was not installed on my system and I got a nasty build fail. Should
> we make all this depend on CONFIG_LIVEPATCH or force world+dog to
> install this as yet another kernel build dependency?
Hm, that would be nice, but as of now the objtool build is independent
of the CONFIG stuff and I'm not sure how to integrate them cleanly.
Maybe it could only link to the library if it exists, and then complain
later at runtime when the feature gets specified on the cmdline?
Or we could try dlopen?
--
Josh
Powered by blists - more mailing lists