Skip to contentSkip to conversation

watchdog: rti_wdt: Allow timeout config with timeout-sec

PATCH In review Judith Mendez 2 participants 3 messages Aug 21, 2024 – Aug 22, 2024 Watchdog device drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 watchdog: rti_wdt: Allow timeout config with timeout-sec

Currently rti_wdt does not allow timeout to be configured via DT property timeout-sec, so fix watchdog_init_timeout to be able to use timeout-sec.

---
 drivers/watchdog/rti_wdt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/watchdog/rti_wdt.c +2 -1
index 8e1be7ba01039..7260c67e60a25 100644
@@ -332,7 +332,8 @@ static int rti_wdt_probe(struct platform_device *pdev)
 		memunmap(vaddr);
 	}
 
-	watchdog_init_timeout(wdd, heartbeat, dev);
+	wdd->timeout = heartbeat;
+	watchdog_init_timeout(wdd, 0, dev);
 
 	ret = watchdog_register_device(wdd);
 	if (ret) {

base-commit: 860bbe8e618fd62446309e286ab4a83d38201c0a

Signature · 1 line
2.46.0
maintainer
#1 lore raw
Quoted 14 lines · Judith Mendez

Currently rti_wdt does not allow timeout to be configured via DT property timeout-sec, so fix watchdog_init_timeout to be able to use timeout-sec.

---
 drivers/watchdog/rti_wdt.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/watchdog/rti_wdt.c +0 -0
index 8e1be7ba01039..7260c67e60a25 100644

drivers/watchdog/rti_wdt.c static int rti_wdt_probe(struct platform_device *pdev) · line 335

Quoted 5 lines
drivers/watchdog/rti_wdt.c +1 -1
@@ -332,7 +332,8 @@ static int rti_wdt_probe(struct platform_device *pdev)
 		memunmap(vaddr);
 	}
 
-	watchdog_init_timeout(wdd, heartbeat, dev);
+	wdd->timeout = heartbeat;

The proper fix would be to initialize the 'heartbeat' variable with 0, set wdt->timeout to DEFAULT_HEARTBEAT, and to keep passing 'heartbeat' to watchdog_init_timeout(). That used to be done but was explicitly changed in an earlier commit, presumably on purpose. I am not inclined to accept a (partial) revert unless the author of commit 5527483f confirms that this is acceptable and desirable. Otherwise we'll just end up in an edit war, which I really don't want to get into.

Guenter
`

Quoted 9 lines
drivers/watchdog/rti_wdt.c +1 -0
+	watchdog_init_timeout(wdd, 0, dev);
 
 	ret = watchdog_register_device(wdd);
 	if (ret) {

base-commit: 860bbe8e618fd62446309e286ab4a83d38201c0a

Signature · 1 line
2.46.0
#2 lore raw

Hi Guenter,

Quoted 18 lines · Guenter Roeck
Quoted 17 lines · Judith Mendez

Currently rti_wdt does not allow timeout to be configured via DT property timeout-sec, so fix watchdog_init_timeout to be able to use timeout-sec.

---
  drivers/watchdog/rti_wdt.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/watchdog/rti_wdt.c +0 -0
index 8e1be7ba01039..7260c67e60a25 100644
@@ -332,7 +332,8 @@ static int rti_wdt_probe(struct platform_device *pdev)
  		memunmap(vaddr);
  	}
 
Quoted 10 lines

drivers/watchdog/rti_wdt.c static int rti_wdt_probe(struct platform_device *pdev) · line 335

Quoted 2 lines
drivers/watchdog/rti_wdt.c +1 -1
-	watchdog_init_timeout(wdd, heartbeat, dev);
+	wdd->timeout = heartbeat;

The proper fix would be to initialize the 'heartbeat' variable with 0, set wdt->timeout to DEFAULT_HEARTBEAT, and to keep passing 'heartbeat' to watchdog_init_timeout(). That used to be done but was explicitly changed in an earlier commit, presumably on purpose. I am not inclined to accept a (partial) revert unless the author of commit 5527483f confirms that this is acceptable and desirable. Otherwise we'll just end up in an edit war, which I really don't want to get into.

You are right, I completely missed this change in this commit due to the subject line not being related. I will try to align with Tero Kristo to see if this is something we can add back. Thanks.

~ Judith

Quoted 13 lines

Guenter
`

Quoted 9 lines
drivers/watchdog/rti_wdt.c +1 -0
+	watchdog_init_timeout(wdd, 0, dev);
 
  	ret = watchdog_register_device(wdd);
  	if (ret) {

base-commit: 860bbe8e618fd62446309e286ab4a83d38201c0a

Signature · 1 line
2.46.0