linux-kernel - Re: [PATCH] watchdog: sama5d4: fix WDD value to be always set to max

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:   Thu, 9 May 2019 06:04:51 +0000
From:   <Eugen.Hristev@...rochip.com>
To:     <linux@...ck-us.net>
CC:     <Nicolas.Ferre@...rochip.com>, <linux-kernel@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>, <wim@...ux-watchdog.org>,
        <linux-watchdog@...r.kernel.org>
Subject: Re: [PATCH] watchdog: sama5d4: fix WDD value to be always set to max



On 08.05.2019 19:34, Guenter Roeck wrote:
> On Wed, May 08, 2019 at 02:15:03PM +0000, Eugen.Hristev@...rochip.com wrote:
>> From: Eugen Hristev <eugen.hristev@...rochip.com>
>>
>> WDD value must be always set to max (0xFFF) otherwise the hardware
>> block will reset the board on the first ping of the watchdog.
>>
> Not sure why setting WDD to the same value as WDV would do that,
> but on the other side it looks like setting WDD to anything but
> the maximum doesn't add any value either, so

It's a hardware issue.
Normally, setting a WDD lower would allow the system for a little time 
to print out something, before the inevitable reset happens...
In our case, setting WDD lower than max would not allow a single 
watchdog feed, as the first attempt to write the CR would trigger the reset.
This bug was not obvious because the driver was always setting both WDD 
and WDV to max... until timeout-sec property came up in DT which changes 
this.

Eugen
> 
> Reviewed-by: Guenter Roeck <linux@...ck-us.net>
> 
> Guenter
> 
>> Signed-off-by: Eugen Hristev <eugen.hristev@...rochip.com>
>> ---
>>   drivers/watchdog/sama5d4_wdt.c | 4 +---
>>   1 file changed, 1 insertion(+), 3 deletions(-)
>>
>> diff --git a/drivers/watchdog/sama5d4_wdt.c b/drivers/watchdog/sama5d4_wdt.c
>> index 1e93c1b..d495336 100644
>> --- a/drivers/watchdog/sama5d4_wdt.c
>> +++ b/drivers/watchdog/sama5d4_wdt.c
>> @@ -111,9 +111,7 @@ static int sama5d4_wdt_set_timeout(struct watchdog_device *wdd,
>>   	u32 value = WDT_SEC2TICKS(timeout);
>>   
>>   	wdt->mr &= ~AT91_WDT_WDV;
>> -	wdt->mr &= ~AT91_WDT_WDD;
>>   	wdt->mr |= AT91_WDT_SET_WDV(value);
>> -	wdt->mr |= AT91_WDT_SET_WDD(value);
>>   
>>   	/*
>>   	 * WDDIS has to be 0 when updating WDD/WDV. The datasheet states: When
>> @@ -251,7 +249,7 @@ static int sama5d4_wdt_probe(struct platform_device *pdev)
>>   
>>   	timeout = WDT_SEC2TICKS(wdd->timeout);
>>   
>> -	wdt->mr |= AT91_WDT_SET_WDD(timeout);
>> +	wdt->mr |= AT91_WDT_SET_WDD(WDT_SEC2TICKS(MAX_WDT_TIMEOUT));
>>   	wdt->mr |= AT91_WDT_SET_WDV(timeout);
>>   
>>   	ret = sama5d4_wdt_init(wdt);
>> -- 
>> 2.7.4
>>
> 
>