Re: [patch 5/7] genhd: send async notification on media change

[James Bottomley <James.Bottomley@SteelEye.com>]

On Wed, 2007-05-23 at 11:26 -0700, Kristen Carlson Accardi wrote:
> On Wed, 23 May 2007 12:03:55 -0500
> James Bottomley <James.Bottomley@SteelEye.com> wrote:
> 
> > On Wed, 2007-05-23 at 09:31 -0700, Kristen Carlson Accardi wrote:
> > > On Tue, 22 May 2007 14:12:11 -0700
> > > Andrew Morton <akpm@linux-foundation.org> wrote:
> > > 
> > > > On Wed, 9 May 2007 16:38:35 -0700
> > > > Kristen Carlson Accardi <kristen.c.accardi@intel.com> wrote:
> > > > 
> > > > > Send an uevent to user space to indicate that a media change event has occurred.
> > > > > 
> > > > > Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
> > > > > 
> > > > > Index: 2.6-git/block/genhd.c
> > > > > ===================================================================
> > > > > --- 2.6-git.orig/block/genhd.c
> > > > > +++ 2.6-git/block/genhd.c
> > > > > @@ -643,6 +643,27 @@ struct seq_operations diskstats_op = {
> > > > >  	.show	= diskstats_show
> > > > >  };
> > > > >  
> > > > > +static void media_change_notify_thread(struct work_struct *work)
> > > > > +{
> > > > > +	struct gendisk *gd = container_of(work, struct gendisk, async_notify);
> > > > > +	char event[] = "MEDIA_CHANGE=1";
> > > > > +	char *envp[] = { event, NULL };
> > > > > +
> > > > > +	/*
> > > > > +	 * set enviroment vars to indicate which event this is for
> > > > > +	 * so that user space will know to go check the media status.
> > > > > +	 */
> > > > > +	kobject_uevent_env(&gd->kobj, KOBJ_CHANGE, envp);
> > > > > +	put_device(gd->driverfs_dev);
> > > > > +}
> > > > > +
> > > > > +void genhd_media_change_notify(struct gendisk *disk)
> > > > > +{
> > > > > +	get_device(disk->driverfs_dev);
> > > > > +	schedule_work(&disk->async_notify);
> > > > > +}
> > > > > +EXPORT_SYMBOL_GPL(genhd_media_change_notify);
> > > > > +
> > > > >  struct gendisk *alloc_disk(int minors)
> > > > >  {
> > > > >  	return alloc_disk_node(minors, -1);
> > > > > @@ -672,6 +693,8 @@ struct gendisk *alloc_disk_node(int mino
> > > > >  		kobj_set_kset_s(disk,block_subsys);
> > > > >  		kobject_init(&disk->kobj);
> > > > >  		rand_initialize_disk(disk);
> > > > > +		INIT_WORK(&disk->async_notify,
> > > > > +			media_change_notify_thread);
> > > > >  	}
> > > > >  	return disk;
> > > > 
> > > > Why does this do a schedule_work() rather than calling kobject_uevent_env()
> > > > directly?
> > > > 
> > > 
> > > Because it is called at Interrupt time.
> > 
> > It better not be ... there's two GFP_KERNEL allocations just above this
> > line in the file.
> > 
> > James
> > 
> 
> Where?  We are talking about the call to genhd_media_change_notify().
> the calling path is this:
> ahci_host_intr()->ata_scsi_media_change_notify()->genhd_media_change_notify().
> 
> I don't see the allocations - are they hidden somewhere?

Sorry, I thought you were saying alloc_disk_node() could be called from
interrupt context.

<gets back on ball>

If you just want to invoke guaranteed user context from a place in the
code which *may* be called from interrupt, then
execute_in_process_context() might be a better way of doing it ... at
least it avoids setting up a workqueue where none is needed.

James