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] [thread-next>] [day] [month] [year] [list]
Message-ID: <00d859b0-f766-4322-fe58-095d4f84e954@roeck-us.net>
Date:   Fri, 21 Oct 2022 21:00:22 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Eddie James <eajames@...ux.ibm.com>
Cc:     linux-watchdog@...r.kernel.org, linux-aspeed@...ts.ozlabs.org,
        linux-kernel@...r.kernel.org, wim@...ux-watchdog.org,
        andrew@...id.au, joel@....id.au, robh+dt@...nel.org,
        krzysztof.kozlowski+dt@...aro.org, devicetree@...r.kernel.org
Subject: Re: [PATCH 1/2] watchdog: aspeed: Add pre-timeout interrupt support

On 10/21/22 12:39, Eddie James wrote:
> 
> On 10/21/22 11:56, Guenter Roeck wrote:
>> On Fri, Oct 21, 2022 at 10:15:58AM -0500, Eddie James wrote:
>>> Enable the pre-timeout interrupt if requested by device property.
>>>
>> I am not inclined to accept this patch without detailed explanation.
>> Why would it make sense and/or be desirable to completely bypass the
>> watchdog core with this pretimeout support ?
> 
> 
> Sorry, I should add more detail.
> 
> It doesn't necessarily bypass the watchdog core. It can, if you specify reset-type="none". But if not, the watchdog will still fire at the appropriate time.
> 
> The purpose is to get a stack dump from a kernel panic rather than a hard reset from the watchdog. The interrupt will fire a certain number of microseconds (configurable by dts property) before the watchdog does. The interrupt handler then panics, and all the CPU stacks are dumped, so hopefully you can catch where another processor was stuck.
> 
> 
> I can submit v2 with this information in the commit message and/or comments.
> 

You did not answer the question why you do not use the pretimeout functionality
supported by the watchdog core.

Guenter

> Thanks,
> 
> Eddie
> 
> 
>>
>> Thanks,
>> Guenter
>>
>>> Signed-off-by: Eddie James <eajames@...ux.ibm.com>
>>> ---
>>>   drivers/watchdog/aspeed_wdt.c | 53 +++++++++++++++++++++++++++++++++--
>>>   1 file changed, 51 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/watchdog/aspeed_wdt.c b/drivers/watchdog/aspeed_wdt.c
>>> index 0cff2adfbfc9..8e12181a827e 100644
>>> --- a/drivers/watchdog/aspeed_wdt.c
>>> +++ b/drivers/watchdog/aspeed_wdt.c
>>> @@ -5,11 +5,14 @@
>>>    * Joel Stanley <joel@....id.au>
>>>    */
>>> +#include <linux/bits.h>
>>>   #include <linux/delay.h>
>>> +#include <linux/interrupt.h>
>>>   #include <linux/io.h>
>>>   #include <linux/kernel.h>
>>>   #include <linux/module.h>
>>>   #include <linux/of.h>
>>> +#include <linux/of_irq.h>
>>>   #include <linux/platform_device.h>
>>>   #include <linux/watchdog.h>
>>> @@ -26,20 +29,32 @@ struct aspeed_wdt {
>>>   struct aspeed_wdt_config {
>>>       u32 ext_pulse_width_mask;
>>> +    u32 irq_shift;
>>> +    u32 irq_mask;
>>>   };
>>>   static const struct aspeed_wdt_config ast2400_config = {
>>>       .ext_pulse_width_mask = 0xff,
>>> +    .irq_shift = 0,
>>> +    .irq_mask = 0,
>>>   };
>>>   static const struct aspeed_wdt_config ast2500_config = {
>>>       .ext_pulse_width_mask = 0xfffff,
>>> +    .irq_shift = 12,
>>> +    .irq_mask = GENMASK(31, 12),
>>> +};
>>> +
>>> +static const struct aspeed_wdt_config ast2600_config = {
>>> +    .ext_pulse_width_mask = 0xfffff,
>>> +    .irq_shift = 0,
>>> +    .irq_mask = GENMASK(31, 10),
>>>   };
>>>   static const struct of_device_id aspeed_wdt_of_table[] = {
>>>       { .compatible = "aspeed,ast2400-wdt", .data = &ast2400_config },
>>>       { .compatible = "aspeed,ast2500-wdt", .data = &ast2500_config },
>>> -    { .compatible = "aspeed,ast2600-wdt", .data = &ast2500_config },
>>> +    { .compatible = "aspeed,ast2600-wdt", .data = &ast2600_config },
>>>       { },
>>>   };
>>>   MODULE_DEVICE_TABLE(of, aspeed_wdt_of_table);
>>> @@ -58,6 +73,7 @@ MODULE_DEVICE_TABLE(of, aspeed_wdt_of_table);
>>>   #define   WDT_CTRL_RESET_SYSTEM        BIT(1)
>>>   #define   WDT_CTRL_ENABLE        BIT(0)
>>>   #define WDT_TIMEOUT_STATUS    0x10
>>> +#define   WDT_TIMEOUT_STATUS_IRQ        BIT(2)
>>>   #define   WDT_TIMEOUT_STATUS_BOOT_SECONDARY    BIT(1)
>>>   #define WDT_CLEAR_TIMEOUT_STATUS    0x14
>>>   #define   WDT_CLEAR_TIMEOUT_AND_BOOT_CODE_SELECTION    BIT(0)
>>> @@ -243,6 +259,17 @@ static const struct watchdog_info aspeed_wdt_info = {
>>>       .identity    = KBUILD_MODNAME,
>>>   };
>>> +static irqreturn_t aspeed_wdt_irq(int irq, void *arg)
>>> +{
>>> +    struct aspeed_wdt *wdt = arg;
>>> +    u32 status = readl(wdt->base + WDT_TIMEOUT_STATUS);
>>> +
>>> +    if (status & WDT_TIMEOUT_STATUS_IRQ)
>>> +        panic("Watchdog pre-timeout IRQ");
>>> +
>>> +    return IRQ_NONE;
>>> +}
>>> +
>>>   static int aspeed_wdt_probe(struct platform_device *pdev)
>>>   {
>>>       struct device *dev = &pdev->dev;
>>> @@ -253,6 +280,7 @@ static int aspeed_wdt_probe(struct platform_device *pdev)
>>>       const char *reset_type;
>>>       u32 duration;
>>>       u32 status;
>>> +    u32 timeout = 0;
>>>       int ret;
>>>       wdt = devm_kzalloc(dev, sizeof(*wdt), GFP_KERNEL);
>>> @@ -291,6 +319,27 @@ static int aspeed_wdt_probe(struct platform_device *pdev)
>>>       if (of_device_is_compatible(np, "aspeed,ast2400-wdt"))
>>>           wdt->ctrl = WDT_CTRL_1MHZ_CLK;
>>> +    if (config->irq_mask) {
>>> +        if (!of_property_read_u32(np, "aspeed,pre-timeout-irq-us", &timeout) && timeout) {
>>> +            int irq =  platform_get_irq(pdev, 0);
>>> +
>>> +            if (irq < 0) {
>>> +                dev_warn(dev, "Couldn't find IRQ: %d\n", irq);
>>> +                timeout = 0;
>>> +            } else {
>>> +                ret = devm_request_irq(dev, irq, aspeed_wdt_irq, IRQF_SHARED,
>>> +                               dev_name(dev), wdt);
>>> +                if (ret) {
>>> +                    dev_warn(dev, "Couldn't request IRQ:%d\n", ret);
>>> +                    timeout = 0;
>>> +                } else {
>>> +                    wdt->ctrl |= ((timeout << config->irq_shift) &
>>> +                              config->irq_mask) | WDT_CTRL_WDT_INTR;
>>> +                }
>>> +            }
>>> +        }
>>> +    }
>>> +
>>>       /*
>>>        * Control reset on a per-device basis to ensure the
>>>        * host is not affected by a BMC reboot
>>> @@ -308,7 +357,7 @@ static int aspeed_wdt_probe(struct platform_device *pdev)
>>>           else if (!strcmp(reset_type, "system"))
>>>               wdt->ctrl |= WDT_CTRL_RESET_MODE_FULL_CHIP |
>>>                        WDT_CTRL_RESET_SYSTEM;
>>> -        else if (strcmp(reset_type, "none"))
>>> +        else if (strcmp(reset_type, "none") && !timeout)
>>>               return -EINVAL;
>>>       }
>>>       if (of_property_read_bool(np, "aspeed,external-signal"))
>>> -- 
>>> 2.31.1
>>>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ