Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 6 Aug 2020 17:19:12 -0700
From:   Guenter Roeck <linux@roeck-us.net>
To:     Qian Cai <cai@lca.pw>
Cc:     torvalds@linux-foundation.org, daniel.diaz@linaro.org,
        tytso@mit.edu, catalin.marinas@arm.com, will@kernel.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        sfr@canb.auug.org.au, linux-next@vger.kernel.org
Subject: Re: [PATCH] random: fix a compilation error on arm64
Message-ID: <20200807001912.GA29367@roeck-us.net>
References: <20200806135836.31736-1-cai@lca.pw>
 <4edcf552-ef82-cdd2-3fda-38d42ebfecc7@roeck-us.net>
 <20200806224313.GA36490@lca.pw>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200806224313.GA36490@lca.pw>
User-Agent: Mutt/1.9.4 (2018-02-28)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Aug 06, 2020 at 06:43:14PM -0400, Qian Cai wrote:
> On Thu, Aug 06, 2020 at 07:50:23AM -0700, Guenter Roeck wrote:
> > On 8/6/20 6:58 AM, Qian Cai wrote:
> > > linux-next failed to compile using this .config,
> > > https://gitlab.com/cailca/linux-mm/-/blob/master/arm64.config
> > > 
> > > arch/arm64/kernel/kaslr.c: In function 'kaslr_early_init':
> > > arch/arm64/kernel/kaslr.c:128:6: error: implicit declaration of function
> > > '__early_cpu_has_rndr'; did you mean '__early_pfn_to_nid'?
> > > [-Werror=implicit-function-declaration]
> > >   if (__early_cpu_has_rndr()) {
> > >       ^~~~~~~~~~~~~~~~~~~~
> > >       __early_pfn_to_nid
> > > arch/arm64/kernel/kaslr.c:131:7: error: implicit declaration of function
> > > '__arm64_rndr' [-Werror=implicit-function-declaration]
> > >    if (__arm64_rndr(&raw))
> > >        ^~~~~~~~~~~~
> > > cc1: some warnings being treated as errors
> > > 
> > 
> > This happens because ARCH_RANDOM is disabled. May be I am wrong, but I
> > would argue that __arm64_rndr() should not be called directly in the
> > first place. arch_get_random_seed_long_early() should be called instead.
> 
> I am not sure about if that would cure possible ARCH_RANDOM=n compiling
> failures on all arches. Since my patch would not compile on mips etc, an
> alternative is to revert the buggy commit 585524081ecd for now. Any thought?
> 
It is quite unlikely that any other architecture would attempt to compile
any arm64 code (which is the only code touched by commit 585524081ecd).
Also, commit 585524081ecd fixes a different problem caused by the
recursive inclusion of asm/archrandom.h. On top of that, I think using
arch_get_random_seed_long_early() would be much cleaner than the current
code since it doesn't override/bypass ARCH_RANDOM.

Let's do it this way: I'll submit an alternate patch using
arch_get_random_seed_long_early(). This way Linus can decide how to
proceed, and we don't have to go back and forth.

Thanks,
Guenter

> > 
> > Thanks,
> > Guenter
> > 
> > > Fixes: 585524081ecd ("random: random.h should include archrandom.h, not the other way around")
> > > Signed-off-by: Qian Cai <cai@lca.pw>
> > > ---
> > >  include/linux/random.h | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/include/linux/random.h b/include/linux/random.h
> > > index f45b8be3e3c4..da782c16c20c 100644
> > > --- a/include/linux/random.h
> > > +++ b/include/linux/random.h
> > > @@ -12,6 +12,8 @@
> > >  #include <linux/list.h>
> > >  #include <linux/once.h>
> > >  
> > > +#include <asm/archrandom.h>
> > > +
> > >  #include <uapi/linux/random.h>
> > >  
> > >  struct random_ready_callback {
> > > 
> >