From: Andrew Jeffery <andrew@codeconstruct.com.au>
To: Chin-Ting Kuo <chin-ting_kuo@aspeedtech.com>,
patrick@stwcx.xyz, robh@kernel.org, krzk+dt@kernel.org,
conor+dt@kernel.org, joel@jms.id.au, wim@linux-watchdog.org,
linux@roeck-us.net, linux-arm-kernel@lists.infradead.org,
linux-aspeed@lists.ozlabs.org, linux-kernel@vger.kernel.org,
linux-watchdog@vger.kernel.org
Cc: Peter.Yin@quantatw.com, Patrick_NC_Lin@wiwynn.com,
Bonnie_Lo@wiwynn.com, DELPHINE_CHIU@wiwynn.com,
bmc-sw@aspeedtech.com, chnguyen@amperecomputing.com
Subject: Re: [PATCH v3 2/2] watchdog: aspeed: Add support for SW restart
Date: Thu, 31 Oct 2024 10:33:40 +1030 [thread overview]
Message-ID: <f7148ed6d4ef835d60fd1aa7141b5f0a9cb69eff.camel@codeconstruct.com.au> (raw)
In-Reply-To: <20241030104717.168324-3-chin-ting_kuo@aspeedtech.com>
On Wed, 2024-10-30 at 18:47 +0800, Chin-Ting Kuo wrote:
> Since AST2600, except for HW WDT counter timeout, HW WDT
> reset can also be triggered by just cinfiguring some
> HW registers by SW directly. We named it "SW restart".
> Although it is "SW" restart, its mechanism is implemented
> by HW.
>
> Originally, system can only know it is reset by WDT
> through a reset flag. However, since AST2600, SW can
> trigger the reset event consciously and directly without
> wait for WDT timeout. WDT counter is not enabled when
> SW restart is adopted. After that, an independent reset
> event flag will be set after systemis reset by SW.
>
> Signed-off-by: Chin-Ting Kuo <chin-ting_kuo@aspeedtech.com>
> ---
> drivers/watchdog/aspeed_wdt.c | 40
> +++++++++++++++++++++++++++++++++++
> 1 file changed, 40 insertions(+)
>
> diff --git a/drivers/watchdog/aspeed_wdt.c
> b/drivers/watchdog/aspeed_wdt.c
> index add76be3ee42..1e9808d42023 100644
> --- a/drivers/watchdog/aspeed_wdt.c
> +++ b/drivers/watchdog/aspeed_wdt.c
> @@ -42,6 +42,9 @@ MODULE_PARM_DESC(nowayout, "Watchdog cannot be
> stopped once started (default="
>
> #define WDT_REG_OFFSET_MASK 0x00000fff
>
> +/* WDT behavior control flag */
> +#define WDT_RESTART_SYSTEM_SW_SUPPORT 0x00000001
> +
> struct aspeed_wdt_scu {
> const char *compatible;
> u32 reset_status_reg;
> @@ -55,6 +58,7 @@ struct aspeed_wdt_config {
> u32 irq_shift;
> u32 irq_mask;
> u32 reg_size;
> + u32 flags;
Why add the flags member rather than change the restart callback for
the 2600? The latter seems more direct to me.
> struct aspeed_wdt_scu scu;
> };
>
> @@ -71,6 +75,7 @@ static const struct aspeed_wdt_config
> ast2400_config = {
> .irq_shift = 0,
> .irq_mask = 0,
> .reg_size = 0x20,
> + .flags = 0,
> .scu = {
> .compatible = "aspeed,ast2400-scu",
> .reset_status_reg = AST2400_SCU_SYS_RESET_STATUS,
> @@ -85,6 +90,7 @@ static const struct aspeed_wdt_config
> ast2500_config = {
> .irq_shift = 12,
> .irq_mask = GENMASK(31, 12),
> .reg_size = 0x20,
> + .flags = 0,
> .scu = {
> .compatible = "aspeed,ast2500-scu",
> .reset_status_reg = AST2400_SCU_SYS_RESET_STATUS,
> @@ -99,6 +105,7 @@ static const struct aspeed_wdt_config
> ast2600_config = {
> .irq_shift = 0,
> .irq_mask = GENMASK(31, 10),
> .reg_size = 0x40,
> + .flags = WDT_RESTART_SYSTEM_SW_SUPPORT,
> .scu = {
> .compatible = "aspeed,ast2600-scu",
> .reset_status_reg = AST2600_SCU_SYS_RESET_STATUS,
> @@ -136,6 +143,11 @@ MODULE_DEVICE_TABLE(of, aspeed_wdt_of_table);
> #define WDT_CLEAR_TIMEOUT_AND_BOOT_CODE_SELECTION BIT(0)
> #define WDT_RESET_MASK1 0x1c
> #define WDT_RESET_MASK2 0x20
> +#define WDT_SW_RESET_CTRL 0x24
> +#define WDT_SW_RESET_COUNT_CLEAR 0xDEADDEAD
> +#define WDT_SW_RESET_ENABLE 0xAEEDF123
> +#define WDT_SW_RESET_MASK1 0x28
> +#define WDT_SW_RESET_MASK2 0x2c
>
> /*
> * WDT_RESET_WIDTH controls the characteristics of the external
> pulse (if
> @@ -255,10 +267,31 @@ static int aspeed_wdt_set_pretimeout(struct
> watchdog_device *wdd,
> return 0;
> }
>
> +static void aspeed_wdt_sw_reset(struct watchdog_device *wdd)
> +{
> + struct aspeed_wdt *wdt = to_aspeed_wdt(wdd);
> + u32 ctrl = WDT_CTRL_RESET_MODE_SOC |
> + WDT_CTRL_RESET_SYSTEM;
> +
> + writel(ctrl, wdt->base + WDT_CTRL);
> + writel(WDT_SW_RESET_COUNT_CLEAR,
> + wdt->base + WDT_SW_RESET_CTRL);
> + writel(WDT_SW_RESET_ENABLE, wdt->base + WDT_SW_RESET_CTRL);
> +
> + /* system must be reset immediately */
> + mdelay(1000);
> +}
> +
> static int aspeed_wdt_restart(struct watchdog_device *wdd,
> unsigned long action, void *data)
> {
> struct aspeed_wdt *wdt = to_aspeed_wdt(wdd);
> + const struct aspeed_wdt_config *cfg = wdt->cfg;
> +
> + if (cfg->flags & WDT_RESTART_SYSTEM_SW_SUPPORT) {
> + aspeed_wdt_sw_reset(wdd);
> + return 0;
> + }
>
> wdt->ctrl &= ~WDT_CTRL_BOOT_SECONDARY;
> aspeed_wdt_enable(wdt, 128 * WDT_RATE_1MHZ / 1000);
> @@ -529,6 +562,13 @@ static int aspeed_wdt_probe(struct
> platform_device *pdev)
> if (nrstmask > 1)
> writel(reset_mask[1], wdt->base +
> WDT_RESET_MASK2);
> }
> +
> + if (wdt->cfg->flags & WDT_RESTART_SYSTEM_SW_SUPPORT)
This condition could be a match against the compatible if you drop the
flags member.
Or, assuming the software reset masks are not adjusted elsewhere, move
the copies below into the ast2600-specific restart callback
implementation?
Andrew
> {
> + reg = readl(wdt->base + WDT_RESET_MASK1);
> + writel(reg, wdt->base + WDT_SW_RESET_MASK1);
> + reg = readl(wdt->base + WDT_RESET_MASK2);
> + writel(reg, wdt->base + WDT_SW_RESET_MASK2);
> + }
> }
>
> if (!of_property_read_u32(np, "aspeed,ext-pulse-duration",
> &duration)) {
next prev parent reply other threads:[~2024-10-31 0:03 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-30 10:47 [PATCH v3 0/2] Update ASPEED WDT bootstatus Chin-Ting Kuo
2024-10-30 10:47 ` [PATCH v3 1/2] watchdog: aspeed: Update bootstatus handling Chin-Ting Kuo
2024-10-30 23:53 ` Andrew Jeffery
2024-10-31 6:24 ` Chin-Ting Kuo
2024-10-30 10:47 ` [PATCH v3 2/2] watchdog: aspeed: Add support for SW restart Chin-Ting Kuo
2024-10-31 0:03 ` Andrew Jeffery [this message]
2024-10-31 6:25 ` Chin-Ting Kuo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=f7148ed6d4ef835d60fd1aa7141b5f0a9cb69eff.camel@codeconstruct.com.au \
--to=andrew@codeconstruct.com.au \
--cc=Bonnie_Lo@wiwynn.com \
--cc=DELPHINE_CHIU@wiwynn.com \
--cc=Patrick_NC_Lin@wiwynn.com \
--cc=Peter.Yin@quantatw.com \
--cc=bmc-sw@aspeedtech.com \
--cc=chin-ting_kuo@aspeedtech.com \
--cc=chnguyen@amperecomputing.com \
--cc=conor+dt@kernel.org \
--cc=joel@jms.id.au \
--cc=krzk+dt@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-aspeed@lists.ozlabs.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-watchdog@vger.kernel.org \
--cc=linux@roeck-us.net \
--cc=patrick@stwcx.xyz \
--cc=robh@kernel.org \
--cc=wim@linux-watchdog.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®