linux-kernel - Re: [RFC PATCH 1/3] watchdog: Add Meson GXBB Watchdog Driver

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

Date:	Fri, 27 May 2016 06:48:53 -0700
From:	Guenter Roeck <linux@...ck-us.net>
To:	Neil Armstrong <narmstrong@...libre.com>,
	Wim Van Sebroeck <wim@...ana.be>
Cc:	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-amlogic@...ts.infradead.org, linux-watchdog@...r.kernel.org
Subject: Re: [RFC PATCH 1/3] watchdog: Add Meson GXBB Watchdog Driver

On 05/27/2016 01:25 AM, Neil Armstrong wrote:
[ ... ]
>>> +    data->wdt_dev.max_hw_heartbeat_ms = GXBB_WDT_TCNT_SETUP_MASK;
>>> +    data->wdt_dev.min_hw_heartbeat_ms = 1;
>>
>> Does the device require a minimum time between heartbeats ?
>> Just asking, because you violate it yourself below.
>>
>> If you want to set the minimum timeout, that would be min_timeout.
>
> Ok, these values are not very clear actually.
>
Hmmm .. yes, reading the description again, it doesn't really describe well
what it is doing. Essentially, min_hw_heartbeat_ms is enforced by the watchdog
core, and should be used if a watchdog hardware can not tolerate short intervals
between heartbeats. min_timeout is the minimum timeout value configurable from
user space.

>>> +    data->wdt_dev.timeout = DEFAULT_TIMEOUT;
>>> +    watchdog_set_drvdata(&data->wdt_dev, data);
>>> +
>>> +    watchdog_init_timeout(&data->wdt_dev, DEFAULT_TIMEOUT, &pdev->dev);
>>> +
>>
>> DEFAULT_TIMEOUT is unnecessary here. Since the default timeout is already set,
>> it makes the call useless.
>>
>>> +    ret = watchdog_register_device(&data->wdt_dev);
>>> +    if (ret)
>>> +        return ret;
>>> +
>>> +    /* Setup with 1ms timebase */
>>> +    writel(((clk_get_rate(data->clk) / 1000) & GXBB_WDT_CTRL_DIV_MASK) |
>>> +        GXBB_WDT_CTRL_EE_RESET |
>>> +        GXBB_WDT_CTRL_CLK_EN |
>>> +        GXBB_WDT_CTRL_CLKDIV_EN,
>>> +        data->reg_base + GXBB_WDT_CTRL_REG);
>>> +    meson_gxbb_wdt_set_timeout(&data->wdt_dev, data->wdt_dev.timeout);
>>
>> set_timeout already calls the ping function. Also, the functions should be called
>> prior to watchdog registration to avoid race conditions.
>>
>>> +    meson_gxbb_wdt_ping(&data->wdt_dev);
>>
>> This is unusual. If the watchdog can be already running, it might make sense
>> to tell the core about it (set WDOG_HW_RUNNING in the status field), so it
>> can send heartbeats until user space opens the device.
>
> Yes, since meson_gxbb_wdt_set_timeout() already ping, this is useless.
>

Not only that - if the watchdog _is_ already running at boot time, you should
really set WDOG_HW_RUNNING to let the watchdog core know. You status function
would come handy there.

	if (meson_gxbb_wdt_status(data))	// note the changed parameter
		set_bit(WDOG_HW_RUNNING, &data->wdt_dev.status);

Thanks,
Guenter