Re: [PATCH 2/4] devcoredump: Add dev_coredump_timeout_set()

[Souza, Jose]

On Wed, 2024-02-28 at 13:56 +0530, Mukesh Ojha wrote:
> 
> On 2/28/2024 2:30 AM, José Roberto de Souza wrote:
> > Add function to set a custom coredump timeout.
> > 
> > Current 5-minute timeout may be too short for users to search and
> > understand what needs to be done to capture coredump to report bugs.
> > 
> > Cc: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
> > Cc: Mukesh Ojha <quic_mojha@xxxxxxxxxxx>
> > Cc: Johannes Berg <johannes@xxxxxxxxxxxxxxxx>
> > Cc: Jonathan Cavitt <jonathan.cavitt@xxxxxxxxx>
> > Signed-off-by: José Roberto de Souza <jose.souza@xxxxxxxxx>
> > ---
> >   drivers/base/devcoredump.c  | 23 +++++++++++++++++++++++
> >   include/linux/devcoredump.h |  6 ++++++
> >   2 files changed, 29 insertions(+)
> > 
> > diff --git a/drivers/base/devcoredump.c b/drivers/base/devcoredump.c
> > index e96427411b87c..2857ceb3eb3aa 100644
> > --- a/drivers/base/devcoredump.c
> > +++ b/drivers/base/devcoredump.c
> > @@ -326,6 +326,29 @@ void dev_coredump_put(struct device *dev)
> >   }
> >   EXPORT_SYMBOL_GPL(dev_coredump_put);
> >   
> > +/**
> > + * dev_coredump_timeout_set - set coredump timeout
> > + * @dev: the struct device for the crashed device
> > + * @secs_timeout: new timeout in seconds
> > + *
> > + * If giving device has a coredump this sets a new timeout for coredump.
> > + */
> > +void dev_coredump_timeout_set(struct device *dev, unsigned long secs_timeout)
> > +{
> > +	struct device *existing;
> > +	struct devcd_entry *devcd;
> > +
> > +	existing = class_find_device(&devcd_class, NULL, dev,
> > +				     devcd_match_failing);
> > +	if (!existing)
> > +		return;
> > +
> > +	devcd = dev_to_devcd(existing);
> > +	mod_delayed_work(system_wq, &devcd->del_wk, HZ * secs_timeout);
> > +	put_device(existing);
> > +}
> > +EXPORT_SYMBOL_GPL(dev_coredump_timeout_set);
> 
> Why don't we provide a new exported function like [1] as calling
> dev_coredumpm() and just another function to configure the timeout
> does not look good.

sounds good too, tried to be less intrusive.

will wait feedback in the first patch to send a new version.

thank you

> 
> [1]
> 
> dev_coredump_timeout(..., timeout);
>   __dev_coredump_timeout(...);
> 
> dev_coredumpm(..)
>   __dev_coredump_timeout(..., DEVCD_TIMEOUT);
> 
> dev_coredumpv()
>   __dev_coredump_timeout(..., DEVCD_TIMEOUT);
> 
> dev_coredumpsg()
>    __dev_coredump_timeout(..., DEVCD_TIMEOUT);
> 
> -Mukesh
> 
> > +
> >   /**
> >    * dev_coredumpm - create device coredump with read/free methods
> >    * @dev: the struct device for the crashed device
> > diff --git a/include/linux/devcoredump.h b/include/linux/devcoredump.h
> > index c8f7eb6cc1915..70fe72a5c6d36 100644
> > --- a/include/linux/devcoredump.h
> > +++ b/include/linux/devcoredump.h
> > @@ -65,6 +65,7 @@ void dev_coredumpsg(struct device *dev, struct scatterlist *table,
> >   		    size_t datalen, gfp_t gfp);
> >   
> >   void dev_coredump_put(struct device *dev);
> > +void dev_coredump_timeout_set(struct device *dev, unsigned long secs_timeout);
> >   #else
> >   static inline void dev_coredumpv(struct device *dev, void *data,
> >   				 size_t datalen, gfp_t gfp)
> > @@ -90,6 +91,11 @@ static inline void dev_coredumpsg(struct device *dev, struct scatterlist *table,
> >   static inline void dev_coredump_put(struct device *dev)
> >   {
> >   }
> > +
> > +static inline void dev_coredump_timeout_set(struct device *dev,
> > +					    unsigned long secs_timeout)
> > +{
> > +}
> >   #endif /* CONFIG_DEV_COREDUMP */
> >   
> >   #endif /* __DEVCOREDUMP_H */