From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.3 required=3.0 tests=DKIM_INVALID,DKIM_SIGNED, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_PASS,URIBL_BLOCKED,USER_AGENT_MUTT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C85D3C43218 for ; Fri, 26 Apr 2019 14:16:57 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 8108F206C1 for ; Fri, 26 Apr 2019 14:16:57 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="LKMS42x4" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726255AbfDZOQ4 (ORCPT ); Fri, 26 Apr 2019 10:16:56 -0400 Received: from mail-pg1-f195.google.com ([209.85.215.195]:40120 "EHLO mail-pg1-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726060AbfDZOQz (ORCPT ); Fri, 26 Apr 2019 10:16:55 -0400 Received: by mail-pg1-f195.google.com with SMTP id d31so1698736pgl.7; Fri, 26 Apr 2019 07:16:54 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=sender:date:from:to:cc:subject:message-id:references:mime-version :content-disposition:in-reply-to:user-agent; bh=Tir7SnBnfr/XF97ydPG18V+eRaue4+SKDKSBbqDKBhA=; b=LKMS42x4Zc1pjS1qCA8TsF8d/46IKZulZRw2rTtuHKXEGktLPcqFbsgdYPfyQjudHQ 6/PstUIl4AgHtI9qugdQJK8JnwvohQ104h/K69PFU8qF0+7GWoBLT669lG4ySS68iCs4 g5G5lNN/mdKNBswBJ2PLmGTY9y9qh+5bLsZ/kttYvahI7TIRs0Wyd3ykTOdAkjpAzH6u aey6N9NuJTTMzXEIMWjbexs5NuUtvNab095El9SPS4MaCSEhkJh01lQjZS9yO5uGS/dK qCEdMKw+yT4QsNWbsaii+JMpDHZ2IwKMq5O1L+jkOFa+cwYxrbVBb8L1uklNXN51esTJ MvcA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:sender:date:from:to:cc:subject:message-id :references:mime-version:content-disposition:in-reply-to:user-agent; bh=Tir7SnBnfr/XF97ydPG18V+eRaue4+SKDKSBbqDKBhA=; b=Kb2+U+RWOQ8T58afWfXuyRnkE4oxAZd6lLXhZvggwu+1uWXKTMrXDodIOuovJoZEZm tsgsN2aiQS04HUXdiqRwt2lAtRzvNOmW1+TrHagAncW03rFP7w8fokKr+XWsJtivPUJ7 igAbz6OZrmLG7kKaCNz6pG3sghpplCO+X3v9nu62zxrvoYxdT482IBFTn4G0S3kWO0ZC BzlRRi34vwQZUbfZX/6ohnwQdrvv/Lpi6ceM3Wa647sQQTJE4HJ/sywUcW7/7PTb9fO0 ewBYnTPFOJ0LkjOe0FtXFJZ/jw/7tIhqse/g+bH+M5jRAeiBPMATneiXF9kNg/QUUUqS ypLw== X-Gm-Message-State: APjAAAUmyj2qqKwGQWAf/SAak8X7sivoVUBIQ/D5lj/hmxTQBboTc9hB +be/U5hUxXUjNs7J5vMFIfw= X-Google-Smtp-Source: APXvYqyq9MZTGy6XHIHbb3+8/8JmCmI4WFpmfpsOH5lDIYKmMWlqDE6H1r3xvBkwtLGIS2igB4SN6w== X-Received: by 2002:a65:5682:: with SMTP id v2mr14825221pgs.100.1556288214107; Fri, 26 Apr 2019 07:16:54 -0700 (PDT) Received: from localhost ([2600:1700:e321:62f0:329c:23ff:fee3:9d7c]) by smtp.gmail.com with ESMTPSA id r5sm10285213pgv.52.2019.04.26.07.16.52 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 26 Apr 2019 07:16:52 -0700 (PDT) Date: Fri, 26 Apr 2019 07:16:51 -0700 From: Guenter Roeck To: Ludovic Barre Cc: Wim Van Sebroeck , Rob Herring , Maxime Coquelin , Alexandre Torgue , linux-watchdog@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, devicetree@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com Subject: Re: [PATCH] watchdog: stm32: add dynamic prescaler support Message-ID: <20190426141651.GA1715@roeck-us.net> References: <1556286075-29496-1-git-send-email-ludovic.Barre@st.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1556286075-29496-1-git-send-email-ludovic.Barre@st.com> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Apr 26, 2019 at 03:41:15PM +0200, Ludovic Barre wrote: > From: Ludovic Barre > > This patch allows to define the max prescaler by compatible. > To set a large range of timeout, the prescaler should be set > dynamically (from the timeout request) to improve the resolution > in order to have a timeout close to the expected value. > > Signed-off-by: Ludovic Barre > --- > drivers/watchdog/stm32_iwdg.c | 70 +++++++++++++++++++++++++------------------ > 1 file changed, 41 insertions(+), 29 deletions(-) > > diff --git a/drivers/watchdog/stm32_iwdg.c b/drivers/watchdog/stm32_iwdg.c > index e00e3b3..91d0a89 100644 > --- a/drivers/watchdog/stm32_iwdg.c > +++ b/drivers/watchdog/stm32_iwdg.c > @@ -34,18 +34,10 @@ > #define KR_KEY_EWA 0x5555 /* write access enable */ > #define KR_KEY_DWA 0x0000 /* write access disable */ > > -/* IWDG_PR register bit values */ > -#define PR_4 0x00 /* prescaler set to 4 */ > -#define PR_8 0x01 /* prescaler set to 8 */ > -#define PR_16 0x02 /* prescaler set to 16 */ > -#define PR_32 0x03 /* prescaler set to 32 */ > -#define PR_64 0x04 /* prescaler set to 64 */ > -#define PR_128 0x05 /* prescaler set to 128 */ > -#define PR_256 0x06 /* prescaler set to 256 */ > +#define PR_SHIFT 2 > > /* IWDG_RLR register values */ > -#define RLR_MIN 0x07C /* min value supported by reload register */ > -#define RLR_MAX 0xFFF /* max value supported by reload register */ > +#define RLR_MAX GENMASK(11, 0) /* max value of reload register */ > > /* IWDG_SR register bit mask */ > #define FLAG_PVU BIT(0) /* Watchdog prescaler value update */ > @@ -55,15 +47,28 @@ > #define TIMEOUT_US 100000 > #define SLEEP_US 1000 > > -#define HAS_PCLK true > +struct stm32_iwdg_data { > + bool has_pclk; > + u32 max_prescaler; > +}; > + > +static const struct stm32_iwdg_data stm32_iwdg_data = { > + .has_pclk = false, > + .max_prescaler = 256, > +}; > + > +static const struct stm32_iwdg_data stm32mp1_iwdg_data = { > + .has_pclk = true, > + .max_prescaler = 1024, > +}; > > struct stm32_iwdg { > struct watchdog_device wdd; > + const struct stm32_iwdg_data *data; > void __iomem *regs; > struct clk *clk_lsi; > struct clk *clk_pclk; > unsigned int rate; > - bool has_pclk; > }; > > static inline u32 reg_read(void __iomem *base, u32 reg) > @@ -80,21 +85,30 @@ static int stm32_iwdg_start(struct watchdog_device *wdd) > { > struct stm32_iwdg *wdt = watchdog_get_drvdata(wdd); > u32 val = FLAG_PVU | FLAG_RVU; > - u32 reload; > + u32 timeout, presc, iwdg_rlr, iwdg_pr; > int ret; > > dev_dbg(wdd->parent, "%s\n", __func__); > > - /* prescaler fixed to 256 */ > - reload = clamp_t(unsigned int, ((wdd->timeout * wdt->rate) / 256) - 1, > - RLR_MIN, RLR_MAX); > + timeout = clamp_t(unsigned int, wdd->timeout, > + wdd->min_timeout, wdd->max_hw_heartbeat_ms / 1000); > + > + if (timeout != wdd->timeout) > + dev_warn(wdd->parent, "timeout skrinked to %d\n", timeout); > + Valid values for timeout should be set in the set_timeout function, not here. As such, there is no need for this warning. More specifically, if the selected timeout is between min_timeout and max_hw_heartbeat_ms, and the hardware can not meet the exact requested value, the set_timeout function should adjust wdd->timeout value accordingly. Example: The requested timeout is 55 seconds, but the hardware can only support either 32 or 64 seconds. The set_timeout function should then set wdd->timeout to either 32 or 64. Furthermore, this is a valid condition. For example, the timeout could be set for one hour (or day), and the maximum heartbeat could be one minute. In that situation, the watchdog core handles heartbeat/ping requests. Again, this does not warrant a warning. On top of all that, if the hardware can not set a minimum timeout of 1 second, min_timeout should not be set to 1 second. It should be set to the actual minimum supportable value if that value is larger than 1 second. > + presc = DIV_ROUND_UP(timeout * wdt->rate, RLR_MAX + 1); > + > + /* The prescaler is align on power of 2 and start at 2 ^ PR_SHIFT. */ > + presc = roundup_pow_of_two(presc); > + iwdg_pr = presc <= 1 << PR_SHIFT ? 0 : ilog2(presc) - PR_SHIFT; > + iwdg_rlr = ((timeout * wdt->rate) / presc) - 1; > > /* enable write access */ > reg_write(wdt->regs, IWDG_KR, KR_KEY_EWA); > > /* set prescaler & reload registers */ > - reg_write(wdt->regs, IWDG_PR, PR_256); /* prescaler fix to 256 */ > - reg_write(wdt->regs, IWDG_RLR, reload); > + reg_write(wdt->regs, IWDG_PR, iwdg_pr); > + reg_write(wdt->regs, IWDG_RLR, iwdg_rlr); > reg_write(wdt->regs, IWDG_KR, KR_KEY_ENABLE); > > /* wait for the registers to be updated (max 100ms) */ > @@ -150,7 +164,7 @@ static int stm32_iwdg_clk_init(struct platform_device *pdev, > } > > /* optional peripheral clock */ > - if (wdt->has_pclk) { > + if (wdt->data->has_pclk) { > wdt->clk_pclk = devm_clk_get(&pdev->dev, "pclk"); > if (IS_ERR(wdt->clk_pclk)) { > dev_err(&pdev->dev, "Unable to get pclk clock\n"); > @@ -191,8 +205,8 @@ static const struct watchdog_ops stm32_iwdg_ops = { > }; > > static const struct of_device_id stm32_iwdg_of_match[] = { > - { .compatible = "st,stm32-iwdg", .data = (void *)!HAS_PCLK }, > - { .compatible = "st,stm32mp1-iwdg", .data = (void *)HAS_PCLK }, > + { .compatible = "st,stm32-iwdg", .data = &stm32_iwdg_data }, > + { .compatible = "st,stm32mp1-iwdg", .data = &stm32mp1_iwdg_data }, > { /* end node */ } > }; > MODULE_DEVICE_TABLE(of, stm32_iwdg_of_match); > @@ -200,20 +214,17 @@ MODULE_DEVICE_TABLE(of, stm32_iwdg_of_match); > static int stm32_iwdg_probe(struct platform_device *pdev) > { > struct watchdog_device *wdd; > - const struct of_device_id *match; > struct stm32_iwdg *wdt; > struct resource *res; > int ret; > > - match = of_match_device(stm32_iwdg_of_match, &pdev->dev); > - if (!match) > - return -ENODEV; > - > wdt = devm_kzalloc(&pdev->dev, sizeof(*wdt), GFP_KERNEL); > if (!wdt) > return -ENOMEM; > > - wdt->has_pclk = match->data; > + wdt->data = of_device_get_match_data(&pdev->dev); > + if (!wdt->data) > + return -ENODEV; > > /* This is the timer base. */ > res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > @@ -231,8 +242,9 @@ static int stm32_iwdg_probe(struct platform_device *pdev) > wdd = &wdt->wdd; > wdd->info = &stm32_iwdg_info; > wdd->ops = &stm32_iwdg_ops; > - wdd->min_timeout = ((RLR_MIN + 1) * 256) / wdt->rate; > - wdd->max_hw_heartbeat_ms = ((RLR_MAX + 1) * 256 * 1000) / wdt->rate; > + wdd->min_timeout = 1; Is this guaranteed to be no larger than 1 under all circumstances ? [ Not that the old code was necessarily correct - it might have set min_timeout to 0 if wdt->rate is high ] > + wdd->max_hw_heartbeat_ms = ((RLR_MAX + 1) * wdt->data->max_prescaler * > + 1000) / wdt->rate; > wdd->parent = &pdev->dev; > > watchdog_set_drvdata(wdd, wdt); > -- > 2.7.4 >