Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 27 Jul 2023 12:49:51 +0200
From:   Christian Brauner <brauner@kernel.org>
To:     Randy Dunlap <rdunlap@infradead.org>
Cc:     Loic Poulain <loic.poulain@linaro.org>, viro@zeniv.linux.org.uk,
        corbet@lwn.net, linux-kernel@vger.kernel.org,
        linux-doc@vger.kernel.org, linux-fsdevel@vger.kernel.org,
        hch@infradead.org
Subject: Re: [PATCH v2] init: Add support for rootwait timeout parameter
Message-ID: <20230727-speerwerfen-tiefpunkt-c4cde40994af@brauner>
References: <20230726152232.932288-1-loic.poulain@linaro.org>
 <1d363177-2629-1ab3-7a4b-bc67d94bb87a@infradead.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <1d363177-2629-1ab3-7a4b-bc67d94bb87a@infradead.org>
Precedence: bulk

On Wed, Jul 26, 2023 at 11:08:35PM -0700, Randy Dunlap wrote:
> Hi--
> 
> On 7/26/23 08:22, Loic Poulain wrote:
> > Add an optional timeout arg to 'rootwait' as the maximum time in
> > seconds to wait for the root device to show up before attempting
> > forced mount of the root filesystem.
> > 
> > Use case:
> > In case of device mapper usage for the rootfs (e.g. root=/dev/dm-0),
> > if the mapper is not able to create the virtual block for any reason
> > (wrong arguments, bad dm-verity signature, etc), the `rootwait` param
> > causes the kernel to wait forever. It may however be desirable to only
> > wait for a given time and then panic (force mount) to cause device reset.
> > This gives the bootloader a chance to detect the problem and to take some
> > measures, such as marking the booted partition as bad (for A/B case) or
> > entering a recovery mode.
> > 
> > In success case, mounting happens as soon as the root device is ready,
> > unlike the existing 'rootdelay' parameter which performs an unconditional
> > pause.
> > 
> > Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
> > ---
> >  v2: rebase + reword: add use case example
> > 
> >  .../admin-guide/kernel-parameters.txt         |  4 ++++
> >  init/do_mounts.c                              | 19 +++++++++++++++++--
> >  2 files changed, 21 insertions(+), 2 deletions(-)
> > 
> 
> > diff --git a/init/do_mounts.c b/init/do_mounts.c
> > index 1aa015883519..118f2bbe7b38 100644
> > --- a/init/do_mounts.c
> > +++ b/init/do_mounts.c
> > @@ -18,6 +18,7 @@
> >  #include <linux/slab.h>
> >  #include <linux/ramfs.h>
> >  #include <linux/shmem_fs.h>
> > +#include <linux/ktime.h>
> >  
> >  #include <linux/nfs_fs.h>
> >  #include <linux/nfs_fs_sb.h>
> > @@ -71,12 +72,20 @@ static int __init rootwait_setup(char *str)
> >  {
> >  	if (*str)
> >  		return 0;
> > -	root_wait = 1;
> > +	root_wait = -1;
> >  	return 1;
> >  }
> >  
> >  __setup("rootwait", rootwait_setup);
> >  
> > +static int __init rootwait_timeout_setup(char *str)
> > +{
> > +	root_wait = simple_strtoul(str, NULL, 0);
> 
> Better to use kstrtoul().  simple_strtoul() says:
> 
>  * This function has caveats. Please use kstrtoul instead.
> 
> and kstrtoul() says:
> 
>  * Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error.
>  * Preferred over simple_strtoul(). Return code must be checked.

Yes, this should check and at least log an error that rootwait is
ignored and fall back to either indefinite waiting or no waiting.