* [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer
@ 2014-11-27 4:54 Naidu Tellapati
[not found] ` <1417064045-13274-1-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
0 siblings, 1 reply; 6+ messages in thread
From: Naidu Tellapati @ 2014-11-27 4:54 UTC (permalink / raw)
To: wim-IQzOog9fTRqzQB+pC5nmwQ, linux-0h96xk9xTtrk1uMJSBkQmQ,
abrestic-F7+t8E8rja9g9hUCZPvPmw,
ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA,
james.hartley-1AXoQHu6uovQT0dZR+AlfA,
james.hogan-1AXoQHu6uovQT0dZR+AlfA
Cc: naidu.tellapati-1AXoQHu6uovQT0dZR+AlfA,
arul.ramasamy-1AXoQHu6uovQT0dZR+AlfA,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA, Naidu Tellapati
This patchset provides support for the PowerDown Controller (PDC) Watchdog Timer
found on Pistachio SOC from Imagination Technologies (ImgTec).
The series is based on 3.18-rc5.
I am re-sending the series with correct Signed-off-by headers.
Changes from v5:
* Removed unused struct pdc_wdt_dev variables.
* Removed out_wdt label in probe function.
Changes from v4:
* Added warning message if initial timeout value exceeds maximum timeout.
Changes from v3:
* Removed PDC_WD_MAX_DELAY.
* Removed spinlock.
* Removed timeout check inside pdc_wdt_set_timeout function.
* Changed identity to IMG PDC Watchdog.
* Removed pdc_wdt_set_timeout call from the probe function.
* Removed pdc_wdt_shutdown function call from pdc_wdt_remove function
and added pdc_wdt_stop function call instead.
Changes from v2:
* Removed interrupt.h header file.
* Removed clock rate-change notifier callback.
Changes from v1:
* Arranged all include files in alphabetical order.
* Arranged register field masks/shifts defines next to their
corresponding register
* Registered clock rate-change notifier callback.
* Removed all references to 32 kHz fixed clock rate.
Now the clock rate is read dynamically using clk_get_rate().
* Defined a macro PDC_WD_CONFIG_ENABLE.
* Added code for getting system clock.
* Moved module_param stuffs next to the corresponding variable defines.
* Added code in the function pdc_wdt_remove for stopping watchdog.
* Removed MODULE_ALIAS.
* Removed few unwanted spaces and unwanted blank lines.
Naidu Tellapati (2):
watchdog: ImgTec PDC Watchdog Timer Driver
DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation
.../devicetree/bindings/watchdog/imgpdc-wdt.txt | 18 ++
drivers/watchdog/Kconfig | 11 +
drivers/watchdog/Makefile | 1 +
drivers/watchdog/imgpdc_wdt.c | 298 ++++++++++++++++++++
4 files changed, 328 insertions(+), 0 deletions(-)
create mode 100644 Documentation/devicetree/bindings/watchdog/imgpdc-wdt.txt
create mode 100644 drivers/watchdog/imgpdc_wdt.c
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver
[not found] ` <1417064045-13274-1-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
@ 2014-11-27 4:54 ` Naidu Tellapati
[not found] ` <1417064045-13274-2-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2014-11-27 4:54 ` [PATCH RESEND v6 2/2] DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation Naidu Tellapati
2014-12-05 11:54 ` [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer Ezequiel Garcia
2 siblings, 1 reply; 6+ messages in thread
From: Naidu Tellapati @ 2014-11-27 4:54 UTC (permalink / raw)
To: wim-IQzOog9fTRqzQB+pC5nmwQ, linux-0h96xk9xTtrk1uMJSBkQmQ,
abrestic-F7+t8E8rja9g9hUCZPvPmw,
ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA,
james.hartley-1AXoQHu6uovQT0dZR+AlfA,
james.hogan-1AXoQHu6uovQT0dZR+AlfA
Cc: naidu.tellapati-1AXoQHu6uovQT0dZR+AlfA,
arul.ramasamy-1AXoQHu6uovQT0dZR+AlfA,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA, Naidu Tellapati,
Naidu Tellapati, Jude.Abraham
This commit adds support for ImgTec PowerDown Controller Watchdog Timer.
Signed-off-by: Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
Signed-off-by: Jude.Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
Reviewed-by: Andrew Bresticker <abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
Reviewed-by: Ezequiel Garcia <ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
---
drivers/watchdog/Kconfig | 11 ++
drivers/watchdog/Makefile | 1 +
drivers/watchdog/imgpdc_wdt.c | 298 +++++++++++++++++++++++++++++++++++++++++
3 files changed, 310 insertions(+), 0 deletions(-)
create mode 100644 drivers/watchdog/imgpdc_wdt.c
diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
index d0107d4..a184f23 100644
--- a/drivers/watchdog/Kconfig
+++ b/drivers/watchdog/Kconfig
@@ -1235,6 +1235,17 @@ config BCM_KONA_WDT_DEBUG
If in doubt, say 'N'.
+config IMGPDC_WDT
+ tristate "Imagination Technologies PDC Watchdog Timer"
+ depends on HAS_IOMEM
+ depends on METAG || MIPS || COMPILE_TEST
+ help
+ Driver for Imagination Technologies PowerDown Controller
+ Watchdog Timer.
+
+ To compile this driver as a loadable module, choose M here.
+ The module will be called imgpdc_wdt.
+
config LANTIQ_WDT
tristate "Lantiq SoC watchdog"
depends on LANTIQ
diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
index c569ec8..d4dfbb4 100644
--- a/drivers/watchdog/Makefile
+++ b/drivers/watchdog/Makefile
@@ -142,6 +142,7 @@ obj-$(CONFIG_OCTEON_WDT) += octeon-wdt.o
octeon-wdt-y := octeon-wdt-main.o octeon-wdt-nmi.o
obj-$(CONFIG_LANTIQ_WDT) += lantiq_wdt.o
obj-$(CONFIG_RALINK_WDT) += rt2880_wdt.o
+obj-$(CONFIG_IMGPDC_WDT) += imgpdc_wdt.o
# PARISC Architecture
diff --git a/drivers/watchdog/imgpdc_wdt.c b/drivers/watchdog/imgpdc_wdt.c
new file mode 100644
index 0000000..efb00b5
--- /dev/null
+++ b/drivers/watchdog/imgpdc_wdt.c
@@ -0,0 +1,298 @@
+/*
+ * Imagination Technologies PowerDown Controller Watchdog Timer.
+ *
+ * Copyright (c) 2014 Imagination Technologies Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * Based on drivers/watchdog/sunxi_wdt.c Copyright (c) 2013 Carlo Caione
+ * 2012 Henrik Nordstrom
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/log2.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/watchdog.h>
+
+/* registers */
+#define PDC_WD_SW_RESET 0x000
+
+#define PDC_WD_CONFIG 0x004
+#define PDC_WD_CONFIG_ENABLE BIT(31)
+#define PDC_WD_CONFIG_DELAY_MASK 0x0000001f
+#define PDC_WD_CONFIG_DELAY_SHIFT 0
+
+#define PDC_WD_TICKLE1 0x008
+#define PDC_WD_TICKLE1_MAGIC 0xabcd1234
+
+#define PDC_WD_TICKLE2 0x00c
+#define PDC_WD_TICKLE2_MAGIC 0x4321dcba
+
+#define PDC_WD_TICKLE_STATUS_MASK 0x00000007
+#define PDC_WD_TICKLE_STATUS_SHIFT 0
+#define PDC_WD_TICKLE_STATUS_HRESET 0x0 /* Hard reset */
+#define PDC_WD_TICKLE_STATUS_TIMEOUT 0x1 /* Timeout */
+#define PDC_WD_TICKLE_STATUS_TICKLE 0x2 /* Tickled incorrectly */
+#define PDC_WD_TICKLE_STATUS_SRESET 0x3 /* Soft reset */
+#define PDC_WD_TICKLE_STATUS_USER 0x4 /* User reset */
+
+/* timeout in seconds */
+#define PDC_WD_MIN_TIMEOUT 1
+#define PDC_WD_DEFAULT_TIMEOUT 64
+
+static int timeout = PDC_WD_DEFAULT_TIMEOUT;
+module_param(timeout, int, 0);
+MODULE_PARM_DESC(timeout, "PDC watchdog delay in seconds (default 64s)");
+
+static bool nowayout = WATCHDOG_NOWAYOUT;
+module_param(nowayout, bool, 0);
+MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started");
+
+struct pdc_wdt_dev {
+ struct watchdog_device wdt_dev;
+ struct clk *wdt_clk;
+ struct clk *sys_clk;
+ int min_delay;
+ void __iomem *base;
+};
+
+static int pdc_wdt_keepalive(struct watchdog_device *wdt_dev)
+{
+ struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
+
+ writel(PDC_WD_TICKLE1_MAGIC, wdt->base + PDC_WD_TICKLE1);
+ writel(PDC_WD_TICKLE2_MAGIC, wdt->base + PDC_WD_TICKLE2);
+
+ return 0;
+}
+
+static int pdc_wdt_stop(struct watchdog_device *wdt_dev)
+{
+ unsigned int val;
+ struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
+
+ val = readl(wdt->base + PDC_WD_CONFIG);
+ val &= ~PDC_WD_CONFIG_ENABLE;
+ writel(val, wdt->base + PDC_WD_CONFIG);
+
+ /* Must tickle to finish the stop */
+ pdc_wdt_keepalive(wdt_dev);
+
+ return 0;
+}
+
+static int pdc_wdt_set_timeout(struct watchdog_device *wdt_dev,
+ unsigned int new_timeout)
+{
+ unsigned int val;
+ struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
+
+ wdt->wdt_dev.timeout = new_timeout;
+ /* round up to the next power of 2 */
+ new_timeout = order_base_2(new_timeout);
+ val = readl(wdt->base + PDC_WD_CONFIG);
+ val &= ~PDC_WD_CONFIG_DELAY_MASK;
+ val |= (new_timeout + wdt->min_delay) << PDC_WD_CONFIG_DELAY_SHIFT;
+ writel(val, wdt->base + PDC_WD_CONFIG);
+
+ return 0;
+}
+
+/* Start the watchdog timer (delay should already be set) */
+static int pdc_wdt_start(struct watchdog_device *wdt_dev)
+{
+ unsigned int val;
+ struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
+
+ val = readl(wdt->base + PDC_WD_CONFIG);
+ val |= PDC_WD_CONFIG_ENABLE;
+ writel(val, wdt->base + PDC_WD_CONFIG);
+
+ return 0;
+}
+
+static struct watchdog_info pdc_wdt_info = {
+ .identity = "IMG PDC Watchdog",
+ .options = WDIOF_SETTIMEOUT |
+ WDIOF_KEEPALIVEPING |
+ WDIOF_MAGICCLOSE,
+};
+
+/* Kernel interface */
+static const struct watchdog_ops pdc_wdt_ops = {
+ .owner = THIS_MODULE,
+ .start = pdc_wdt_start,
+ .stop = pdc_wdt_stop,
+ .ping = pdc_wdt_keepalive,
+ .set_timeout = pdc_wdt_set_timeout,
+};
+
+static int pdc_wdt_probe(struct platform_device *pdev)
+{
+ int ret, val;
+ unsigned long clk_rate;
+ struct resource *res;
+ struct pdc_wdt_dev *pdc_wdt;
+
+ pdc_wdt = devm_kzalloc(&pdev->dev, sizeof(*pdc_wdt), GFP_KERNEL);
+ if (!pdc_wdt)
+ return -ENOMEM;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ pdc_wdt->base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(pdc_wdt->base))
+ return PTR_ERR(pdc_wdt->base);
+
+ pdc_wdt->sys_clk = devm_clk_get(&pdev->dev, "sys");
+ if (IS_ERR(pdc_wdt->sys_clk)) {
+ dev_err(&pdev->dev, "failed to get the sys clock.\n");
+ ret = PTR_ERR(pdc_wdt->wdt_clk);
+ return ret;
+ }
+
+ pdc_wdt->wdt_clk = devm_clk_get(&pdev->dev, "wdt");
+ if (IS_ERR(pdc_wdt->wdt_clk)) {
+ dev_err(&pdev->dev, "failed to get the wdt clock.\n");
+ ret = PTR_ERR(pdc_wdt->wdt_clk);
+ return ret;
+ }
+
+ ret = clk_prepare_enable(pdc_wdt->sys_clk);
+ if (ret) {
+ dev_err(&pdev->dev, "could not prepare or enable sys clock.\n");
+ return ret;
+ }
+
+ ret = clk_prepare_enable(pdc_wdt->wdt_clk);
+ if (ret) {
+ dev_err(&pdev->dev, "could not prepare or enable wdt clock.\n");
+ goto disable_sys_clk;
+ }
+
+ /* We use the clock rate to calculate the max timeout */
+ clk_rate = clk_get_rate(pdc_wdt->wdt_clk);
+ if (clk_rate == 0) {
+ dev_err(&pdev->dev, "failed to get clock rate\n");
+ ret = -EINVAL;
+ goto disable_wdt_clk;
+ }
+
+ if (order_base_2(clk_rate) > PDC_WD_CONFIG_DELAY_MASK + 1) {
+ dev_err(&pdev->dev, "invalid clock rate\n");
+ ret = -EINVAL;
+ goto disable_wdt_clk;
+ }
+
+ if (order_base_2(clk_rate) == 0)
+ pdc_wdt->wdt_dev.min_timeout = PDC_WD_MIN_TIMEOUT + 1;
+ else
+ pdc_wdt->wdt_dev.min_timeout = PDC_WD_MIN_TIMEOUT;
+
+ pdc_wdt->min_delay = order_base_2(clk_rate) - 1;
+
+ pdc_wdt->wdt_dev.info = &pdc_wdt_info;
+ pdc_wdt->wdt_dev.ops = &pdc_wdt_ops;
+ pdc_wdt->wdt_dev.max_timeout =
+ (1 << (PDC_WD_CONFIG_DELAY_MASK - pdc_wdt->min_delay));
+ pdc_wdt->wdt_dev.parent = &pdev->dev;
+
+ ret = watchdog_init_timeout(&pdc_wdt->wdt_dev, timeout, &pdev->dev);
+ if (ret < 0) {
+ pdc_wdt->wdt_dev.timeout = pdc_wdt->wdt_dev.max_timeout;
+ dev_warn(&pdev->dev,
+ "Initial timeout out of range! setting max timeout\n");
+ }
+
+ pdc_wdt_stop(&pdc_wdt->wdt_dev);
+
+ /* Find what caused the last reset */
+ val = readl(pdc_wdt->base + PDC_WD_TICKLE1);
+ val = (val & PDC_WD_TICKLE_STATUS_MASK) >> PDC_WD_TICKLE_STATUS_SHIFT;
+ switch (val) {
+ case PDC_WD_TICKLE_STATUS_TICKLE:
+ case PDC_WD_TICKLE_STATUS_TIMEOUT:
+ pdc_wdt->wdt_dev.bootstatus |= WDIOF_CARDRESET;
+ dev_info(&pdev->dev,
+ "watchdog module last reset due to timeout\n");
+ break;
+ case PDC_WD_TICKLE_STATUS_HRESET:
+ dev_info(&pdev->dev,
+ "watchdog module last reset due to hard reset\n");
+ break;
+ case PDC_WD_TICKLE_STATUS_SRESET:
+ dev_info(&pdev->dev,
+ "watchdog module last reset due to soft reset\n");
+ break;
+ case PDC_WD_TICKLE_STATUS_USER:
+ dev_info(&pdev->dev,
+ "watchdog module last reset due to user reset\n");
+ break;
+ default:
+ dev_info(&pdev->dev,
+ "contains an illegal status code (%08x)\n", val);
+ break;
+ }
+
+ watchdog_set_nowayout(&pdc_wdt->wdt_dev, nowayout);
+
+ platform_set_drvdata(pdev, pdc_wdt);
+ watchdog_set_drvdata(&pdc_wdt->wdt_dev, pdc_wdt);
+
+ ret = watchdog_register_device(&pdc_wdt->wdt_dev);
+ if (ret)
+ goto disable_wdt_clk;
+
+ return 0;
+
+disable_wdt_clk:
+ clk_disable_unprepare(pdc_wdt->wdt_clk);
+disable_sys_clk:
+ clk_disable_unprepare(pdc_wdt->sys_clk);
+ return ret;
+}
+
+static void pdc_wdt_shutdown(struct platform_device *pdev)
+{
+ struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
+
+ pdc_wdt_stop(&pdc_wdt->wdt_dev);
+}
+
+static int pdc_wdt_remove(struct platform_device *pdev)
+{
+ struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
+
+ pdc_wdt_stop(&pdc_wdt->wdt_dev);
+ watchdog_unregister_device(&pdc_wdt->wdt_dev);
+ clk_disable_unprepare(pdc_wdt->wdt_clk);
+ clk_disable_unprepare(pdc_wdt->sys_clk);
+
+ return 0;
+}
+
+static const struct of_device_id pdc_wdt_match[] = {
+ { .compatible = "img,pdc-wdt" },
+ {}
+};
+MODULE_DEVICE_TABLE(of, pdc_wdt_match);
+
+static struct platform_driver pdc_wdt_driver = {
+ .driver = {
+ .name = "imgpdc-wdt",
+ .of_match_table = pdc_wdt_match,
+ },
+ .probe = pdc_wdt_probe,
+ .remove = pdc_wdt_remove,
+ .shutdown = pdc_wdt_shutdown,
+};
+module_platform_driver(pdc_wdt_driver);
+
+MODULE_AUTHOR("Jude Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
+MODULE_AUTHOR("Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
+MODULE_DESCRIPTION("Imagination Technologies PDC Watchdog Timer Driver");
+MODULE_LICENSE("GPL v2");
--
1.7.0.4
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH RESEND v6 2/2] DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation
[not found] ` <1417064045-13274-1-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2014-11-27 4:54 ` [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver Naidu Tellapati
@ 2014-11-27 4:54 ` Naidu Tellapati
2014-12-05 11:54 ` [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer Ezequiel Garcia
2 siblings, 0 replies; 6+ messages in thread
From: Naidu Tellapati @ 2014-11-27 4:54 UTC (permalink / raw)
To: wim-IQzOog9fTRqzQB+pC5nmwQ, linux-0h96xk9xTtrk1uMJSBkQmQ,
abrestic-F7+t8E8rja9g9hUCZPvPmw,
ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA,
james.hartley-1AXoQHu6uovQT0dZR+AlfA,
james.hogan-1AXoQHu6uovQT0dZR+AlfA
Cc: naidu.tellapati-1AXoQHu6uovQT0dZR+AlfA,
arul.ramasamy-1AXoQHu6uovQT0dZR+AlfA,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA, Naidu Tellapati,
Naidu Tellapati, Jude Abraham
Add the devicetree binding document for ImgTec PDC Watchdog Timer.
Signed-off-by: Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
Signed-off-by: Jude Abraham <Jude.Abraham-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Reviewed-by: Andrew Bresticker <abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
---
No changes from v5.
---
.../devicetree/bindings/watchdog/imgpdc-wdt.txt | 18 ++++++++++++++++++
1 files changed, 18 insertions(+), 0 deletions(-)
create mode 100644 Documentation/devicetree/bindings/watchdog/imgpdc-wdt.txt
diff --git a/Documentation/devicetree/bindings/watchdog/imgpdc-wdt.txt b/Documentation/devicetree/bindings/watchdog/imgpdc-wdt.txt
new file mode 100644
index 0000000..2f9ba77
--- /dev/null
+++ b/Documentation/devicetree/bindings/watchdog/imgpdc-wdt.txt
@@ -0,0 +1,18 @@
+*ImgTec PowerDown Controller (PDC) Watchdog Timer (WDT)
+
+Required properties:
+- compatible : Should be "img,pdc-wdt"
+- reg : Should contain WDT registers location and length
+- clocks: Must contain an entry for each entry in clock-names.
+- clock-names: Should contain "wdt" and "sys"
+- interrupts : Should contain WDT interrupt
+
+Examples:
+
+wdt@18102100 {
+ compatible = "img,pdc-wdt";
+ reg = <0x18102100 0x100>;
+ clocks = <&pdc_wdt_clk>, <&sys_clk>;
+ clock-names = "wdt", "sys";
+ interrupts = <0 52 IRQ_TYPE_LEVEL_HIGH>;
+};
--
1.7.0.4
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer
[not found] ` <1417064045-13274-1-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2014-11-27 4:54 ` [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver Naidu Tellapati
2014-11-27 4:54 ` [PATCH RESEND v6 2/2] DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation Naidu Tellapati
@ 2014-12-05 11:54 ` Ezequiel Garcia
2 siblings, 0 replies; 6+ messages in thread
From: Ezequiel Garcia @ 2014-12-05 11:54 UTC (permalink / raw)
To: Naidu Tellapati, wim-IQzOog9fTRqzQB+pC5nmwQ,
linux-0h96xk9xTtrk1uMJSBkQmQ, abrestic-F7+t8E8rja9g9hUCZPvPmw,
james.hartley-1AXoQHu6uovQT0dZR+AlfA,
james.hogan-1AXoQHu6uovQT0dZR+AlfA
Cc: naidu.tellapati-1AXoQHu6uovQT0dZR+AlfA,
arul.ramasamy-1AXoQHu6uovQT0dZR+AlfA,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA
Wim,
On 11/27/2014 01:54 AM, Naidu Tellapati wrote:
> This patchset provides support for the PowerDown Controller (PDC) Watchdog Timer
> found on Pistachio SOC from Imagination Technologies (ImgTec).
>
> The series is based on 3.18-rc5.
>
> I am re-sending the series with correct Signed-off-by headers.
>
> Changes from v5:
>
> * Removed unused struct pdc_wdt_dev variables.
>
> * Removed out_wdt label in probe function.
>
> Changes from v4:
>
> * Added warning message if initial timeout value exceeds maximum timeout.
>
> Changes from v3:
>
> * Removed PDC_WD_MAX_DELAY.
>
> * Removed spinlock.
>
> * Removed timeout check inside pdc_wdt_set_timeout function.
>
> * Changed identity to IMG PDC Watchdog.
>
> * Removed pdc_wdt_set_timeout call from the probe function.
>
> * Removed pdc_wdt_shutdown function call from pdc_wdt_remove function
> and added pdc_wdt_stop function call instead.
>
> Changes from v2:
>
> * Removed interrupt.h header file.
>
> * Removed clock rate-change notifier callback.
>
> Changes from v1:
>
> * Arranged all include files in alphabetical order.
>
> * Arranged register field masks/shifts defines next to their
> corresponding register
>
> * Registered clock rate-change notifier callback.
>
> * Removed all references to 32 kHz fixed clock rate.
> Now the clock rate is read dynamically using clk_get_rate().
>
> * Defined a macro PDC_WD_CONFIG_ENABLE.
>
> * Added code for getting system clock.
>
> * Moved module_param stuffs next to the corresponding variable defines.
>
> * Added code in the function pdc_wdt_remove for stopping watchdog.
>
> * Removed MODULE_ALIAS.
>
> * Removed few unwanted spaces and unwanted blank lines.
>
> Naidu Tellapati (2):
> watchdog: ImgTec PDC Watchdog Timer Driver
> DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation
>
Do you think it's possible to merge this series in time for v3.19?
Thanks!
--
Ezequiel
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver
[not found] ` <1417064045-13274-2-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
@ 2014-12-15 17:17 ` Guenter Roeck
[not found] ` <20141215171744.GA16516-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org>
0 siblings, 1 reply; 6+ messages in thread
From: Guenter Roeck @ 2014-12-15 17:17 UTC (permalink / raw)
To: Naidu Tellapati
Cc: wim-IQzOog9fTRqzQB+pC5nmwQ, abrestic-F7+t8E8rja9g9hUCZPvPmw,
ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA,
james.hartley-1AXoQHu6uovQT0dZR+AlfA,
james.hogan-1AXoQHu6uovQT0dZR+AlfA,
naidu.tellapati-1AXoQHu6uovQT0dZR+AlfA,
arul.ramasamy-1AXoQHu6uovQT0dZR+AlfA,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA, Jude.Abraham
On Thu, Nov 27, 2014 at 10:24:04AM +0530, Naidu Tellapati wrote:
> This commit adds support for ImgTec PowerDown Controller Watchdog Timer.
>
> Signed-off-by: Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Jude.Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> Reviewed-by: Andrew Bresticker <abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
> Reviewed-by: Ezequiel Garcia <ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> ---
> drivers/watchdog/Kconfig | 11 ++
> drivers/watchdog/Makefile | 1 +
> drivers/watchdog/imgpdc_wdt.c | 298 +++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 310 insertions(+), 0 deletions(-)
> create mode 100644 drivers/watchdog/imgpdc_wdt.c
>
> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> index d0107d4..a184f23 100644
> --- a/drivers/watchdog/Kconfig
> +++ b/drivers/watchdog/Kconfig
> @@ -1235,6 +1235,17 @@ config BCM_KONA_WDT_DEBUG
>
> If in doubt, say 'N'.
>
> +config IMGPDC_WDT
> + tristate "Imagination Technologies PDC Watchdog Timer"
> + depends on HAS_IOMEM
> + depends on METAG || MIPS || COMPILE_TEST
> + help
> + Driver for Imagination Technologies PowerDown Controller
> + Watchdog Timer.
> +
> + To compile this driver as a loadable module, choose M here.
> + The module will be called imgpdc_wdt.
> +
> config LANTIQ_WDT
> tristate "Lantiq SoC watchdog"
> depends on LANTIQ
> diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> index c569ec8..d4dfbb4 100644
> --- a/drivers/watchdog/Makefile
> +++ b/drivers/watchdog/Makefile
> @@ -142,6 +142,7 @@ obj-$(CONFIG_OCTEON_WDT) += octeon-wdt.o
> octeon-wdt-y := octeon-wdt-main.o octeon-wdt-nmi.o
> obj-$(CONFIG_LANTIQ_WDT) += lantiq_wdt.o
> obj-$(CONFIG_RALINK_WDT) += rt2880_wdt.o
> +obj-$(CONFIG_IMGPDC_WDT) += imgpdc_wdt.o
>
> # PARISC Architecture
>
> diff --git a/drivers/watchdog/imgpdc_wdt.c b/drivers/watchdog/imgpdc_wdt.c
> new file mode 100644
> index 0000000..efb00b5
> --- /dev/null
> +++ b/drivers/watchdog/imgpdc_wdt.c
> @@ -0,0 +1,298 @@
> +/*
> + * Imagination Technologies PowerDown Controller Watchdog Timer.
> + *
> + * Copyright (c) 2014 Imagination Technologies Ltd.
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms of the GNU General Public License version 2 as published by
> + * the Free Software Foundation.
> + *
> + * Based on drivers/watchdog/sunxi_wdt.c Copyright (c) 2013 Carlo Caione
> + * 2012 Henrik Nordstrom
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/io.h>
> +#include <linux/log2.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +#include <linux/watchdog.h>
> +
> +/* registers */
> +#define PDC_WD_SW_RESET 0x000
> +
> +#define PDC_WD_CONFIG 0x004
> +#define PDC_WD_CONFIG_ENABLE BIT(31)
> +#define PDC_WD_CONFIG_DELAY_MASK 0x0000001f
> +#define PDC_WD_CONFIG_DELAY_SHIFT 0
> +
> +#define PDC_WD_TICKLE1 0x008
> +#define PDC_WD_TICKLE1_MAGIC 0xabcd1234
> +
> +#define PDC_WD_TICKLE2 0x00c
> +#define PDC_WD_TICKLE2_MAGIC 0x4321dcba
> +
> +#define PDC_WD_TICKLE_STATUS_MASK 0x00000007
> +#define PDC_WD_TICKLE_STATUS_SHIFT 0
> +#define PDC_WD_TICKLE_STATUS_HRESET 0x0 /* Hard reset */
> +#define PDC_WD_TICKLE_STATUS_TIMEOUT 0x1 /* Timeout */
> +#define PDC_WD_TICKLE_STATUS_TICKLE 0x2 /* Tickled incorrectly */
> +#define PDC_WD_TICKLE_STATUS_SRESET 0x3 /* Soft reset */
> +#define PDC_WD_TICKLE_STATUS_USER 0x4 /* User reset */
> +
> +/* timeout in seconds */
> +#define PDC_WD_MIN_TIMEOUT 1
> +#define PDC_WD_DEFAULT_TIMEOUT 64
> +
> +static int timeout = PDC_WD_DEFAULT_TIMEOUT;
> +module_param(timeout, int, 0);
> +MODULE_PARM_DESC(timeout, "PDC watchdog delay in seconds (default 64s)");
It might be better to use __MODULE_STRING(PDC_WD_DEFAULT_TIMEOUT)
to encode the default timeout.
> +
> +static bool nowayout = WATCHDOG_NOWAYOUT;
> +module_param(nowayout, bool, 0);
> +MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started");
> +
> +struct pdc_wdt_dev {
> + struct watchdog_device wdt_dev;
> + struct clk *wdt_clk;
> + struct clk *sys_clk;
> + int min_delay;
> + void __iomem *base;
> +};
> +
> +static int pdc_wdt_keepalive(struct watchdog_device *wdt_dev)
> +{
> + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> +
> + writel(PDC_WD_TICKLE1_MAGIC, wdt->base + PDC_WD_TICKLE1);
> + writel(PDC_WD_TICKLE2_MAGIC, wdt->base + PDC_WD_TICKLE2);
> +
> + return 0;
> +}
> +
> +static int pdc_wdt_stop(struct watchdog_device *wdt_dev)
> +{
> + unsigned int val;
> + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> +
> + val = readl(wdt->base + PDC_WD_CONFIG);
> + val &= ~PDC_WD_CONFIG_ENABLE;
> + writel(val, wdt->base + PDC_WD_CONFIG);
> +
> + /* Must tickle to finish the stop */
> + pdc_wdt_keepalive(wdt_dev);
> +
> + return 0;
> +}
> +
> +static int pdc_wdt_set_timeout(struct watchdog_device *wdt_dev,
> + unsigned int new_timeout)
> +{
> + unsigned int val;
> + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> +
> + wdt->wdt_dev.timeout = new_timeout;
> + /* round up to the next power of 2 */
> + new_timeout = order_base_2(new_timeout);
> + val = readl(wdt->base + PDC_WD_CONFIG);
> + val &= ~PDC_WD_CONFIG_DELAY_MASK;
> + val |= (new_timeout + wdt->min_delay) << PDC_WD_CONFIG_DELAY_SHIFT;
> + writel(val, wdt->base + PDC_WD_CONFIG);
> +
> + return 0;
> +}
> +
> +/* Start the watchdog timer (delay should already be set) */
> +static int pdc_wdt_start(struct watchdog_device *wdt_dev)
> +{
> + unsigned int val;
> + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> +
> + val = readl(wdt->base + PDC_WD_CONFIG);
> + val |= PDC_WD_CONFIG_ENABLE;
> + writel(val, wdt->base + PDC_WD_CONFIG);
> +
> + return 0;
> +}
> +
> +static struct watchdog_info pdc_wdt_info = {
> + .identity = "IMG PDC Watchdog",
> + .options = WDIOF_SETTIMEOUT |
> + WDIOF_KEEPALIVEPING |
> + WDIOF_MAGICCLOSE,
> +};
> +
> +/* Kernel interface */
> +static const struct watchdog_ops pdc_wdt_ops = {
> + .owner = THIS_MODULE,
> + .start = pdc_wdt_start,
> + .stop = pdc_wdt_stop,
> + .ping = pdc_wdt_keepalive,
> + .set_timeout = pdc_wdt_set_timeout,
> +};
> +
> +static int pdc_wdt_probe(struct platform_device *pdev)
> +{
> + int ret, val;
> + unsigned long clk_rate;
> + struct resource *res;
> + struct pdc_wdt_dev *pdc_wdt;
> +
> + pdc_wdt = devm_kzalloc(&pdev->dev, sizeof(*pdc_wdt), GFP_KERNEL);
> + if (!pdc_wdt)
> + return -ENOMEM;
> +
> + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> + pdc_wdt->base = devm_ioremap_resource(&pdev->dev, res);
> + if (IS_ERR(pdc_wdt->base))
> + return PTR_ERR(pdc_wdt->base);
> +
> + pdc_wdt->sys_clk = devm_clk_get(&pdev->dev, "sys");
> + if (IS_ERR(pdc_wdt->sys_clk)) {
> + dev_err(&pdev->dev, "failed to get the sys clock.\n");
> + ret = PTR_ERR(pdc_wdt->wdt_clk);
sys_clk
> + return ret;
return PTR_ERR(pdc_wdt->sys_clk);
On a higher level, what is sys_clk needed and used for in the first place ?
The need for wdt_clk is obvious, but there is no explanation for the need
of sys_clk, and the bindings file doesn't give a hint either.
> + }
> +
> + pdc_wdt->wdt_clk = devm_clk_get(&pdev->dev, "wdt");
> + if (IS_ERR(pdc_wdt->wdt_clk)) {
> + dev_err(&pdev->dev, "failed to get the wdt clock.\n");
> + ret = PTR_ERR(pdc_wdt->wdt_clk);
> + return ret;
return PTR_ERR(pdc_wdt->wdt_clk);
> + }
> +
> + ret = clk_prepare_enable(pdc_wdt->sys_clk);
> + if (ret) {
> + dev_err(&pdev->dev, "could not prepare or enable sys clock.\n");
> + return ret;
> + }
> +
> + ret = clk_prepare_enable(pdc_wdt->wdt_clk);
> + if (ret) {
> + dev_err(&pdev->dev, "could not prepare or enable wdt clock.\n");
> + goto disable_sys_clk;
I would suggest to drop the '.' in those messages.
> + }
> +
> + /* We use the clock rate to calculate the max timeout */
> + clk_rate = clk_get_rate(pdc_wdt->wdt_clk);
> + if (clk_rate == 0) {
> + dev_err(&pdev->dev, "failed to get clock rate\n");
> + ret = -EINVAL;
> + goto disable_wdt_clk;
> + }
> +
> + if (order_base_2(clk_rate) > PDC_WD_CONFIG_DELAY_MASK + 1) {
> + dev_err(&pdev->dev, "invalid clock rate\n");
> + ret = -EINVAL;
> + goto disable_wdt_clk;
> + }
> +
> + if (order_base_2(clk_rate) == 0)
> + pdc_wdt->wdt_dev.min_timeout = PDC_WD_MIN_TIMEOUT + 1;
> + else
> + pdc_wdt->wdt_dev.min_timeout = PDC_WD_MIN_TIMEOUT;
> +
> + pdc_wdt->min_delay = order_base_2(clk_rate) - 1;
Ultimately this means that min_delay can be smaller than 0,
which makes it quite difficult to validate the rest of the code
to ensure that it handles this case correctly.
Give that a clock rate of 1 is somewhere between highly unlikely and
unreasonable, wouldn't it be much easier to consider it to be invalid
and not try to handle it ?
> +
> + pdc_wdt->wdt_dev.info = &pdc_wdt_info;
> + pdc_wdt->wdt_dev.ops = &pdc_wdt_ops;
> + pdc_wdt->wdt_dev.max_timeout =
> + (1 << (PDC_WD_CONFIG_DELAY_MASK - pdc_wdt->min_delay));
> + pdc_wdt->wdt_dev.parent = &pdev->dev;
> +
> + ret = watchdog_init_timeout(&pdc_wdt->wdt_dev, timeout, &pdev->dev);
> + if (ret < 0) {
> + pdc_wdt->wdt_dev.timeout = pdc_wdt->wdt_dev.max_timeout;
> + dev_warn(&pdev->dev,
> + "Initial timeout out of range! setting max timeout\n");
> + }
> +
> + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> +
> + /* Find what caused the last reset */
> + val = readl(pdc_wdt->base + PDC_WD_TICKLE1);
> + val = (val & PDC_WD_TICKLE_STATUS_MASK) >> PDC_WD_TICKLE_STATUS_SHIFT;
> + switch (val) {
> + case PDC_WD_TICKLE_STATUS_TICKLE:
> + case PDC_WD_TICKLE_STATUS_TIMEOUT:
> + pdc_wdt->wdt_dev.bootstatus |= WDIOF_CARDRESET;
> + dev_info(&pdev->dev,
> + "watchdog module last reset due to timeout\n");
> + break;
> + case PDC_WD_TICKLE_STATUS_HRESET:
> + dev_info(&pdev->dev,
> + "watchdog module last reset due to hard reset\n");
> + break;
> + case PDC_WD_TICKLE_STATUS_SRESET:
> + dev_info(&pdev->dev,
> + "watchdog module last reset due to soft reset\n");
> + break;
> + case PDC_WD_TICKLE_STATUS_USER:
> + dev_info(&pdev->dev,
> + "watchdog module last reset due to user reset\n");
> + break;
> + default:
> + dev_info(&pdev->dev,
> + "contains an illegal status code (%08x)\n", val);
> + break;
> + }
> +
> + watchdog_set_nowayout(&pdc_wdt->wdt_dev, nowayout);
> +
> + platform_set_drvdata(pdev, pdc_wdt);
> + watchdog_set_drvdata(&pdc_wdt->wdt_dev, pdc_wdt);
> +
> + ret = watchdog_register_device(&pdc_wdt->wdt_dev);
> + if (ret)
> + goto disable_wdt_clk;
> +
> + return 0;
> +
> +disable_wdt_clk:
> + clk_disable_unprepare(pdc_wdt->wdt_clk);
> +disable_sys_clk:
> + clk_disable_unprepare(pdc_wdt->sys_clk);
> + return ret;
> +}
> +
> +static void pdc_wdt_shutdown(struct platform_device *pdev)
> +{
> + struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
> +
> + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> +}
> +
> +static int pdc_wdt_remove(struct platform_device *pdev)
> +{
> + struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
> +
> + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> + watchdog_unregister_device(&pdc_wdt->wdt_dev);
> + clk_disable_unprepare(pdc_wdt->wdt_clk);
> + clk_disable_unprepare(pdc_wdt->sys_clk);
> +
> + return 0;
> +}
> +
> +static const struct of_device_id pdc_wdt_match[] = {
> + { .compatible = "img,pdc-wdt" },
Will "img" be added to the list of vendor prefixes with some other patch ?
> + {}
> +};
> +MODULE_DEVICE_TABLE(of, pdc_wdt_match);
> +
> +static struct platform_driver pdc_wdt_driver = {
> + .driver = {
> + .name = "imgpdc-wdt",
> + .of_match_table = pdc_wdt_match,
> + },
> + .probe = pdc_wdt_probe,
> + .remove = pdc_wdt_remove,
> + .shutdown = pdc_wdt_shutdown,
> +};
> +module_platform_driver(pdc_wdt_driver);
> +
> +MODULE_AUTHOR("Jude Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
> +MODULE_AUTHOR("Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
> +MODULE_DESCRIPTION("Imagination Technologies PDC Watchdog Timer Driver");
> +MODULE_LICENSE("GPL v2");
> --
> 1.7.0.4
>
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver
[not found] ` <20141215171744.GA16516-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org>
@ 2014-12-15 18:28 ` James Hartley
0 siblings, 0 replies; 6+ messages in thread
From: James Hartley @ 2014-12-15 18:28 UTC (permalink / raw)
To: Guenter Roeck, Naidu Tellapati
Cc: wim-IQzOog9fTRqzQB+pC5nmwQ@public.gmane.org,
abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org, Ezequiel Garcia,
James Hogan, Naidu Tellapati, Arul Ramasamy,
linux-watchdog-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Jude Abraham
Hi Guenter,
Thanks for your comments,
> -----Original Message-----
> From: Guenter Roeck [mailto:linux-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org]
> Sent: 15 December 2014 17:18
> To: Naidu Tellapati
> Cc: wim-IQzOog9fTRqzQB+pC5nmwQ@public.gmane.org; abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org; Ezequiel Garcia; James Hartley;
> James Hogan; Naidu Tellapati; Arul Ramasamy; linux-
> watchdog-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; Jude Abraham
> Subject: Re: [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer
> Driver
>
> On Thu, Nov 27, 2014 at 10:24:04AM +0530, Naidu Tellapati wrote:
> > This commit adds support for ImgTec PowerDown Controller Watchdog
> Timer.
> >
> > Signed-off-by: Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> > Signed-off-by: Jude.Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> > Reviewed-by: Andrew Bresticker <abrestic-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
> > Reviewed-by: Ezequiel Garcia <ezequiel.garcia-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>
> > ---
> > drivers/watchdog/Kconfig | 11 ++
> > drivers/watchdog/Makefile | 1 +
> > drivers/watchdog/imgpdc_wdt.c | 298
> > +++++++++++++++++++++++++++++++++++++++++
> > 3 files changed, 310 insertions(+), 0 deletions(-) create mode
> > 100644 drivers/watchdog/imgpdc_wdt.c
> >
> > diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig index
> > d0107d4..a184f23 100644
> > --- a/drivers/watchdog/Kconfig
> > +++ b/drivers/watchdog/Kconfig
> > @@ -1235,6 +1235,17 @@ config BCM_KONA_WDT_DEBUG
> >
> > If in doubt, say 'N'.
> >
> > +config IMGPDC_WDT
> > + tristate "Imagination Technologies PDC Watchdog Timer"
> > + depends on HAS_IOMEM
> > + depends on METAG || MIPS || COMPILE_TEST
> > + help
> > + Driver for Imagination Technologies PowerDown Controller
> > + Watchdog Timer.
> > +
> > + To compile this driver as a loadable module, choose M here.
> > + The module will be called imgpdc_wdt.
> > +
> > config LANTIQ_WDT
> > tristate "Lantiq SoC watchdog"
> > depends on LANTIQ
> > diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> > index c569ec8..d4dfbb4 100644
> > --- a/drivers/watchdog/Makefile
> > +++ b/drivers/watchdog/Makefile
> > @@ -142,6 +142,7 @@ obj-$(CONFIG_OCTEON_WDT) += octeon-wdt.o
> > octeon-wdt-y := octeon-wdt-main.o octeon-wdt-nmi.o
> > obj-$(CONFIG_LANTIQ_WDT) += lantiq_wdt.o
> > obj-$(CONFIG_RALINK_WDT) += rt2880_wdt.o
> > +obj-$(CONFIG_IMGPDC_WDT) += imgpdc_wdt.o
> >
> > # PARISC Architecture
> >
> > diff --git a/drivers/watchdog/imgpdc_wdt.c
> > b/drivers/watchdog/imgpdc_wdt.c new file mode 100644 index
> > 0000000..efb00b5
> > --- /dev/null
> > +++ b/drivers/watchdog/imgpdc_wdt.c
> > @@ -0,0 +1,298 @@
> > +/*
> > + * Imagination Technologies PowerDown Controller Watchdog Timer.
> > + *
> > + * Copyright (c) 2014 Imagination Technologies Ltd.
> > + *
> > + * This program is free software; you can redistribute it and/or
> > +modify it
> > + * under the terms of the GNU General Public License version 2 as
> > +published by
> > + * the Free Software Foundation.
> > + *
> > + * Based on drivers/watchdog/sunxi_wdt.c Copyright (c) 2013 Carlo
> Caione
> > + * 2012 Henrik Nordstrom
> > + */
> > +
> > +#include <linux/clk.h>
> > +#include <linux/io.h>
> > +#include <linux/log2.h>
> > +#include <linux/module.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/slab.h>
> > +#include <linux/watchdog.h>
> > +
> > +/* registers */
> > +#define PDC_WD_SW_RESET 0x000
> > +
> > +#define PDC_WD_CONFIG 0x004
> > +#define PDC_WD_CONFIG_ENABLE BIT(31)
> > +#define PDC_WD_CONFIG_DELAY_MASK 0x0000001f
> > +#define PDC_WD_CONFIG_DELAY_SHIFT 0
> > +
> > +#define PDC_WD_TICKLE1 0x008
> > +#define PDC_WD_TICKLE1_MAGIC 0xabcd1234
> > +
> > +#define PDC_WD_TICKLE2 0x00c
> > +#define PDC_WD_TICKLE2_MAGIC 0x4321dcba
> > +
> > +#define PDC_WD_TICKLE_STATUS_MASK 0x00000007
> > +#define PDC_WD_TICKLE_STATUS_SHIFT 0
> > +#define PDC_WD_TICKLE_STATUS_HRESET 0x0 /* Hard reset */
> > +#define PDC_WD_TICKLE_STATUS_TIMEOUT 0x1 /* Timeout */
> > +#define PDC_WD_TICKLE_STATUS_TICKLE 0x2 /* Tickled incorrectly */
> > +#define PDC_WD_TICKLE_STATUS_SRESET 0x3 /* Soft reset */
> > +#define PDC_WD_TICKLE_STATUS_USER 0x4 /* User reset */
> > +
> > +/* timeout in seconds */
> > +#define PDC_WD_MIN_TIMEOUT 1
> > +#define PDC_WD_DEFAULT_TIMEOUT 64
> > +
> > +static int timeout = PDC_WD_DEFAULT_TIMEOUT;
> module_param(timeout,
> > +int, 0); MODULE_PARM_DESC(timeout, "PDC watchdog delay in seconds
> > +(default 64s)");
>
> It might be better to use
> __MODULE_STRING(PDC_WD_DEFAULT_TIMEOUT)
> to encode the default timeout.
>
> > +
> > +static bool nowayout = WATCHDOG_NOWAYOUT;
> module_param(nowayout,
> > +bool, 0); MODULE_PARM_DESC(nowayout, "Watchdog cannot be
> stopped once
> > +started");
> > +
> > +struct pdc_wdt_dev {
> > + struct watchdog_device wdt_dev;
> > + struct clk *wdt_clk;
> > + struct clk *sys_clk;
> > + int min_delay;
> > + void __iomem *base;
> > +};
> > +
> > +static int pdc_wdt_keepalive(struct watchdog_device *wdt_dev) {
> > + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> > +
> > + writel(PDC_WD_TICKLE1_MAGIC, wdt->base + PDC_WD_TICKLE1);
> > + writel(PDC_WD_TICKLE2_MAGIC, wdt->base + PDC_WD_TICKLE2);
> > +
> > + return 0;
> > +}
> > +
> > +static int pdc_wdt_stop(struct watchdog_device *wdt_dev) {
> > + unsigned int val;
> > + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> > +
> > + val = readl(wdt->base + PDC_WD_CONFIG);
> > + val &= ~PDC_WD_CONFIG_ENABLE;
> > + writel(val, wdt->base + PDC_WD_CONFIG);
> > +
> > + /* Must tickle to finish the stop */
> > + pdc_wdt_keepalive(wdt_dev);
> > +
> > + return 0;
> > +}
> > +
> > +static int pdc_wdt_set_timeout(struct watchdog_device *wdt_dev,
> > + unsigned int new_timeout)
> > +{
> > + unsigned int val;
> > + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> > +
> > + wdt->wdt_dev.timeout = new_timeout;
> > + /* round up to the next power of 2 */
> > + new_timeout = order_base_2(new_timeout);
> > + val = readl(wdt->base + PDC_WD_CONFIG);
> > + val &= ~PDC_WD_CONFIG_DELAY_MASK;
> > + val |= (new_timeout + wdt->min_delay) <<
> PDC_WD_CONFIG_DELAY_SHIFT;
> > + writel(val, wdt->base + PDC_WD_CONFIG);
> > +
> > + return 0;
> > +}
> > +
> > +/* Start the watchdog timer (delay should already be set) */ static
> > +int pdc_wdt_start(struct watchdog_device *wdt_dev) {
> > + unsigned int val;
> > + struct pdc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
> > +
> > + val = readl(wdt->base + PDC_WD_CONFIG);
> > + val |= PDC_WD_CONFIG_ENABLE;
> > + writel(val, wdt->base + PDC_WD_CONFIG);
> > +
> > + return 0;
> > +}
> > +
> > +static struct watchdog_info pdc_wdt_info = {
> > + .identity = "IMG PDC Watchdog",
> > + .options = WDIOF_SETTIMEOUT |
> > + WDIOF_KEEPALIVEPING |
> > + WDIOF_MAGICCLOSE,
> > +};
> > +
> > +/* Kernel interface */
> > +static const struct watchdog_ops pdc_wdt_ops = {
> > + .owner = THIS_MODULE,
> > + .start = pdc_wdt_start,
> > + .stop = pdc_wdt_stop,
> > + .ping = pdc_wdt_keepalive,
> > + .set_timeout = pdc_wdt_set_timeout,
> > +};
> > +
> > +static int pdc_wdt_probe(struct platform_device *pdev) {
> > + int ret, val;
> > + unsigned long clk_rate;
> > + struct resource *res;
> > + struct pdc_wdt_dev *pdc_wdt;
> > +
> > + pdc_wdt = devm_kzalloc(&pdev->dev, sizeof(*pdc_wdt),
> GFP_KERNEL);
> > + if (!pdc_wdt)
> > + return -ENOMEM;
> > +
> > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > + pdc_wdt->base = devm_ioremap_resource(&pdev->dev, res);
> > + if (IS_ERR(pdc_wdt->base))
> > + return PTR_ERR(pdc_wdt->base);
> > +
> > + pdc_wdt->sys_clk = devm_clk_get(&pdev->dev, "sys");
> > + if (IS_ERR(pdc_wdt->sys_clk)) {
> > + dev_err(&pdev->dev, "failed to get the sys clock.\n");
> > + ret = PTR_ERR(pdc_wdt->wdt_clk);
>
> sys_clk
> > + return ret;
>
> return PTR_ERR(pdc_wdt->sys_clk);
>
> On a higher level, what is sys_clk needed and used for in the first place ?
> The need for wdt_clk is obvious, but there is no explanation for the need of
> sys_clk, and the bindings file doesn't give a hint either.
I believe that the watchdog registers are clocked from sys_clk, so without that they are not accessible.
>
> > + }
> > +
> > + pdc_wdt->wdt_clk = devm_clk_get(&pdev->dev, "wdt");
> > + if (IS_ERR(pdc_wdt->wdt_clk)) {
> > + dev_err(&pdev->dev, "failed to get the wdt clock.\n");
> > + ret = PTR_ERR(pdc_wdt->wdt_clk);
> > + return ret;
>
> return PTR_ERR(pdc_wdt->wdt_clk);
>
> > + }
> > +
> > + ret = clk_prepare_enable(pdc_wdt->sys_clk);
> > + if (ret) {
> > + dev_err(&pdev->dev, "could not prepare or enable sys
> clock.\n");
> > + return ret;
> > + }
> > +
> > + ret = clk_prepare_enable(pdc_wdt->wdt_clk);
> > + if (ret) {
> > + dev_err(&pdev->dev, "could not prepare or enable wdt
> clock.\n");
> > + goto disable_sys_clk;
>
> I would suggest to drop the '.' in those messages.
>
> > + }
> > +
> > + /* We use the clock rate to calculate the max timeout */
> > + clk_rate = clk_get_rate(pdc_wdt->wdt_clk);
> > + if (clk_rate == 0) {
> > + dev_err(&pdev->dev, "failed to get clock rate\n");
> > + ret = -EINVAL;
> > + goto disable_wdt_clk;
> > + }
> > +
> > + if (order_base_2(clk_rate) > PDC_WD_CONFIG_DELAY_MASK + 1) {
> > + dev_err(&pdev->dev, "invalid clock rate\n");
> > + ret = -EINVAL;
> > + goto disable_wdt_clk;
> > + }
> > +
> > + if (order_base_2(clk_rate) == 0)
> > + pdc_wdt->wdt_dev.min_timeout =
> PDC_WD_MIN_TIMEOUT + 1;
> > + else
> > + pdc_wdt->wdt_dev.min_timeout =
> PDC_WD_MIN_TIMEOUT;
> > +
> > + pdc_wdt->min_delay = order_base_2(clk_rate) - 1;
>
> Ultimately this means that min_delay can be smaller than 0, which makes it
> quite difficult to validate the rest of the code to ensure that it handles this
> case correctly.
>
> Give that a clock rate of 1 is somewhere between highly unlikely and
> unreasonable, wouldn't it be much easier to consider it to be invalid and not
> try to handle it ?
Sounds reasonable to me.
>
> > +
> > + pdc_wdt->wdt_dev.info = &pdc_wdt_info;
> > + pdc_wdt->wdt_dev.ops = &pdc_wdt_ops;
> > + pdc_wdt->wdt_dev.max_timeout =
> > + (1 << (PDC_WD_CONFIG_DELAY_MASK - pdc_wdt-
> >min_delay));
> > + pdc_wdt->wdt_dev.parent = &pdev->dev;
> > +
> > + ret = watchdog_init_timeout(&pdc_wdt->wdt_dev, timeout, &pdev-
> >dev);
> > + if (ret < 0) {
> > + pdc_wdt->wdt_dev.timeout = pdc_wdt-
> >wdt_dev.max_timeout;
> > + dev_warn(&pdev->dev,
> > + "Initial timeout out of range! setting max
> timeout\n");
> > + }
> > +
> > + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> > +
> > + /* Find what caused the last reset */
> > + val = readl(pdc_wdt->base + PDC_WD_TICKLE1);
> > + val = (val & PDC_WD_TICKLE_STATUS_MASK) >>
> PDC_WD_TICKLE_STATUS_SHIFT;
> > + switch (val) {
> > + case PDC_WD_TICKLE_STATUS_TICKLE:
> > + case PDC_WD_TICKLE_STATUS_TIMEOUT:
> > + pdc_wdt->wdt_dev.bootstatus |= WDIOF_CARDRESET;
> > + dev_info(&pdev->dev,
> > + "watchdog module last reset due to timeout\n");
> > + break;
> > + case PDC_WD_TICKLE_STATUS_HRESET:
> > + dev_info(&pdev->dev,
> > + "watchdog module last reset due to hard reset\n");
> > + break;
> > + case PDC_WD_TICKLE_STATUS_SRESET:
> > + dev_info(&pdev->dev,
> > + "watchdog module last reset due to soft reset\n");
> > + break;
> > + case PDC_WD_TICKLE_STATUS_USER:
> > + dev_info(&pdev->dev,
> > + "watchdog module last reset due to user reset\n");
> > + break;
> > + default:
> > + dev_info(&pdev->dev,
> > + "contains an illegal status code (%08x)\n", val);
> > + break;
> > + }
> > +
> > + watchdog_set_nowayout(&pdc_wdt->wdt_dev, nowayout);
> > +
> > + platform_set_drvdata(pdev, pdc_wdt);
> > + watchdog_set_drvdata(&pdc_wdt->wdt_dev, pdc_wdt);
> > +
> > + ret = watchdog_register_device(&pdc_wdt->wdt_dev);
> > + if (ret)
> > + goto disable_wdt_clk;
> > +
> > + return 0;
> > +
> > +disable_wdt_clk:
> > + clk_disable_unprepare(pdc_wdt->wdt_clk);
> > +disable_sys_clk:
> > + clk_disable_unprepare(pdc_wdt->sys_clk);
> > + return ret;
> > +}
> > +
> > +static void pdc_wdt_shutdown(struct platform_device *pdev) {
> > + struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
> > +
> > + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> > +}
> > +
> > +static int pdc_wdt_remove(struct platform_device *pdev) {
> > + struct pdc_wdt_dev *pdc_wdt = platform_get_drvdata(pdev);
> > +
> > + pdc_wdt_stop(&pdc_wdt->wdt_dev);
> > + watchdog_unregister_device(&pdc_wdt->wdt_dev);
> > + clk_disable_unprepare(pdc_wdt->wdt_clk);
> > + clk_disable_unprepare(pdc_wdt->sys_clk);
> > +
> > + return 0;
> > +}
> > +
> > +static const struct of_device_id pdc_wdt_match[] = {
> > + { .compatible = "img,pdc-wdt" },
>
> Will "img" be added to the list of vendor prefixes with some other patch ?
I think Img is already in the list of vendor prefixes.
>
> > + {}
> > +};
> > +MODULE_DEVICE_TABLE(of, pdc_wdt_match);
> > +
> > +static struct platform_driver pdc_wdt_driver = {
> > + .driver = {
> > + .name = "imgpdc-wdt",
> > + .of_match_table = pdc_wdt_match,
> > + },
> > + .probe = pdc_wdt_probe,
> > + .remove = pdc_wdt_remove,
> > + .shutdown = pdc_wdt_shutdown,
> > +};
> > +module_platform_driver(pdc_wdt_driver);
> > +
> > +MODULE_AUTHOR("Jude Abraham <Jude.Abraham-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
> > +MODULE_AUTHOR("Naidu Tellapati <Naidu.Tellapati-1AXoQHu6uovQT0dZR+AlfA@public.gmane.org>");
> > +MODULE_DESCRIPTION("Imagination Technologies PDC Watchdog Timer
> > +Driver"); MODULE_LICENSE("GPL v2");
> > --
> > 1.7.0.4
> >
James.
--
To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-12-15 18:28 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-11-27 4:54 [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer Naidu Tellapati
[not found] ` <1417064045-13274-1-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2014-11-27 4:54 ` [PATCH RESEND v6 1/2] watchdog: ImgTec PDC Watchdog Timer Driver Naidu Tellapati
[not found] ` <1417064045-13274-2-git-send-email-Naidu.Tellapati-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2014-12-15 17:17 ` Guenter Roeck
[not found] ` <20141215171744.GA16516-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org>
2014-12-15 18:28 ` James Hartley
2014-11-27 4:54 ` [PATCH RESEND v6 2/2] DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation Naidu Tellapati
2014-12-05 11:54 ` [PATCH RESEND v6 0/2] watchdog: Add support for ImgTec PowerDown Controller Watchdog Timer Ezequiel Garcia
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).