linux-kernel - Re: [PATCH V3] watchdog: mtk: support pre-timeout when the bark irq is available

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]

Message-ID: <7555c705-a12d-5d51-3881-68feb5d078c1@roeck-us.net>
Date:   Thu, 15 Apr 2021 06:29:09 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     王擎 <wangqing@...o.com>
Cc:     Wim Van Sebroeck <wim@...ux-watchdog.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        linux-watchdog@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH V3] watchdog: mtk: support pre-timeout when the bark irq
 is available

On 4/15/21 2:40 AM, 王擎 wrote:

> 
> Yes, as mentioned before, the behavior of WDT_MODE_IRQ_EN is use irq instead of
> reset, so we must use WDT_MODE_IRQ_EN | WDT_MODE_DUAL_EN if like you said
> "the first time generating an interrupt and the second time resetting the system" . 
> 
> The Dual mode is mentioned in the MTK datasheet:
> In this mode, the watchdog will be AUTO-RESTART after interrupt is triggered. 
> AP need to clear WDT_STA after receiving interrupt from TOPRGU, or system reset
> will be triggered after watchdog timer expires.
> Instructions for use：
> Set wdt_en = 1'b1.
> Set dual_mode = 1'b1.
> Set wdt_irq = 1'b1.
> 
> We can use Dual mode to achieve pretimeout behavior, only in this way can we
> get more information during pretimeout processing, instead of directly resetting.
> 

If so, the pretimeout would always be half of the full timeout, and the timeout
to configure would be timeout / 2. The code would have to adjust the pretimeout
time accordingly. Also, to avoid unexpected and changed behavior, the prefimeout
should by default be disabled.

Guenter