netdev - Re: [PATCH v2] netconsole: Append kernel version to message

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date: Fri, 14 Jul 2023 06:49:37 -0700
From: Breno Leitao <leitao@...ian.org>
To: Petr Mladek <pmladek@...e.com>
Cc: "David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Jonathan Corbet <corbet@....net>, sergey.senozhatsky@...il.com,
	tj@...nel.org, stephen@...workplumber.org,
	Dave Jones <davej@...emonkey.org.uk>,
	"open list:NETWORKING [GENERAL]" <netdev@...r.kernel.org>,
	"open list:DOCUMENTATION" <linux-doc@...r.kernel.org>,
	open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] netconsole: Append kernel version to message

On Fri, Jul 14, 2023 at 01:41:55PM +0200, Petr Mladek wrote:
> On Fri 2023-07-07 06:29:11, Breno Leitao wrote:
> > @@ -254,6 +267,11 @@ static ssize_t extended_show(struct config_item *item, char *buf)
> >  	return snprintf(buf, PAGE_SIZE, "%d\n", to_target(item)->extended);
> >  }
> >  
> > +static ssize_t release_show(struct config_item *item, char *buf)
> > +{
> > +	return snprintf(buf, PAGE_SIZE, "%d\n", to_target(item)->release);
> 
> I have learned recently that sysfs_emit() was preferred over snprintf() in the
> _show() callbacks.

I didn't know either, I just read about it in the thread. Thanks for the
heads up. We probably want to change it for the other _show() structs.

> > +}
> > +
> >  static ssize_t dev_name_show(struct config_item *item, char *buf)
> >  {
> >  	return snprintf(buf, PAGE_SIZE, "%s\n", to_target(item)->np.dev_name);
> > @@ -366,6 +389,38 @@ static ssize_t enabled_store(struct config_item *item,
> >  	return err;
> >  }
> >  
> > +static ssize_t release_store(struct config_item *item, const char *buf,
> > +			     size_t count)
> > +{
> > +	struct netconsole_target *nt = to_target(item);
> > +	int release;
> > +	int err;
> > +
> > +	mutex_lock(&dynamic_netconsole_mutex);
> > +	if (nt->enabled) {
> > +		pr_err("target (%s) is enabled, disable to update parameters\n",
> > +		       config_item_name(&nt->item));
> > +		err = -EINVAL;
> > +		goto out_unlock;
> > +	}
> > +
> > +	err = kstrtoint(buf, 10, &release);
> > +	if (err < 0)
> > +		goto out_unlock;
> > +	if (release < 0 || release > 1) {
> > +		err = -EINVAL;
> > +		goto out_unlock;
> > +	}
> 
> You might consider using:
> 
> 	bool enabled;
> 
> 	err = kstrtobool(buf, &enabled);
> 	if (err)
> 		goto unlock;
> 
> 
> It accepts more input values, for example, 1/0, y/n, Y/N, ...
> 
> Well, I see that kstrtoint() is used also in enabled_store().
> It might be confusing when "/enabled" supports only "1/0"
> and "/release" supports more variants.

Right. we probably want to move a few _stores to kstrtobool(). Here is
what I have in mind:
	* enabled_store()
	* release_store()
	* extended_store()

That said, there are two ways moving forward:

1) I forward fix it. I've send v3 earlier today[1], I can send a patch
on top of it.
2) I fix this in a v4 patch. Probably a patchset of 3 patches:
	a) Move the current snprintf to emit_sysfs()
	b) Move kstrtoint() to kstrtobool()
	c) This new feature using emit_sysfs() and kstrtobool().

What is the best way moving forward?


Thanks for the review!
[1] Link: https://lore.kernel.org/all/20230714111330.3069605-1-leitao@debian.org/