Re: [PATCH v3 1/2] security: add fault injection capability

From: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
To: Aleksandr Nogikh <aleksandrnogikh@gmail.com>,
	jmorris@namei.org, serge@hallyn.com, akinobu.mita@gmail.com
Cc: andreyknvl@google.com, dvyukov@google.com, elver@google.com,
	glider@google.com, keescook@google.com, casey@schaufler-ca.com,
	linux-kernel@vger.kernel.org,
	linux-security-module@vger.kernel.org,
	Aleksandr Nogikh <nogikh@google.com>
Subject: Re: [PATCH v3 1/2] security: add fault injection capability
Date: Tue, 10 Nov 2020 13:43:06 +0900	[thread overview]
Message-ID: <04d8c32a-06cd-d71a-43d9-47b1de6c7684@i-love.sakura.ne.jp> (raw)
In-Reply-To: <20201029183526.2131776-2-aleksandrnogikh@gmail.com>

On 2020/10/30 3:35, Aleksandr Nogikh wrote:
> +#ifdef CONFIG_FAIL_LSM_HOOKS
> +
> +static struct {
> +	struct fault_attr attr;
> +	int retval;
> +} fail_lsm_hooks = {
> +	.attr = FAULT_ATTR_INITIALIZER,
> +	.retval = -EACCES
> +};
> +
> +static int __init setup_fail_lsm_hooks(char *str)
> +{
> +	return setup_fault_attr(&fail_lsm_hooks.attr, str);
> +}
> +__setup("fail_lsm_hooks=", setup_fail_lsm_hooks);
> +
> +static int lsm_hooks_inject_fail(void)
> +{
> +	return should_fail(&fail_lsm_hooks.attr, 1) ? fail_lsm_hooks.retval : 0;
> +}
> +
> +#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS
> +
> +static int __init fail_lsm_hooks_debugfs(void)
> +{
> +	umode_t mode = S_IFREG | 0600;
> +	struct dentry *dir;
> +
> +	dir = fault_create_debugfs_attr("fail_lsm_hooks", NULL,
> +					&fail_lsm_hooks.attr);
> +	debugfs_create_u32("retval", mode, dir, &fail_lsm_hooks.retval);

Since production kernels will use CONFIG_FAIL_LSM_HOOKS=n, we won't need to worry about userspace ABI.

Reviewed-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>

By the way, fail_lsm_hooks.retval is "signed int" but debugfs_create_u32() handles "unsigned int".
Do we want to allow lsm_hooks_inject_fail() to inject arbitrary !IS_ERR_VALUE() values?

> +	return 0;
> +}