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;
Message-ID: <68507426-54bb-3902-a8b6-e11d25327d65@infradead.org>
Date:   Wed, 22 Jun 2022 07:39:41 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.10.0
Subject: Re: [PATCH 4.9 13/20] x86/speculation/mmio: Add mitigation for
 Processor MMIO Stale Data
Content-Language: en-US
To:     Pavel Machek <pavel@denx.de>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc:     linux-kernel@vger.kernel.org, stable@vger.kernel.org,
        Pawan Gupta <pawan.kumar.gupta@linux.intel.com>,
        Borislav Petkov <bp@suse.de>,
        Thomas Gleixner <tglx@linutronix.de>,
        Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
References: <20220614183722.061550591@linuxfoundation.org>
 <20220614183725.181834522@linuxfoundation.org>
 <20220622114844.GA6854@duo.ucw.cz>
From:   Randy Dunlap <rdunlap@infradead.org>
In-Reply-To: <20220622114844.GA6854@duo.ucw.cz>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 6/22/22 04:48, Pavel Machek wrote:
> Hi!
> 
> 
>> +static int __init mmio_stale_data_parse_cmdline(char *str)
>> +{
>> +	if (!boot_cpu_has_bug(X86_BUG_MMIO_STALE_DATA))
>> +		return 0;
>> +
>> +	if (!str)
>> +		return -EINVAL;
>> +
>> +	if (!strcmp(str, "off")) {
>> +		mmio_mitigation = MMIO_MITIGATION_OFF;
>> +	} else if (!strcmp(str, "full")) {
>> +		mmio_mitigation = MMIO_MITIGATION_VERW;
>> +	} else if (!strcmp(str, "full,nosmt")) {
>> +		mmio_mitigation = MMIO_MITIGATION_VERW;
>> +		mmio_nosmt = true;
>> +	}
>> +
>> +	return 0;
>> +}
> 
> This is wrong, AFAICT. Returning 0 will pollute init's environment;
> Randy was cleaning those lately and we are even seeing them in
> -stable. See for example b793a01000122d2bd133ba451a76cc135b5e162c.
> 
> The early return 0 should disappear, too; we should validate the
> option even on non-buggy machines.

It's good to be on the lookout for such problems, but __setup()
functions (like I was cleaning) are the opposite (sad:(  of
early_param() functions, which this one is.

early_param() does return 0 on success and non-zero
on error, so this looks OK to me.

-- 
~Randy