[PATCH 2/3][-mm] reclassify the sg_sysfs_class mutex

[PATCH 2/3][-mm] reclassify the sg_sysfs_class mutex

[Dave Young]

[Please first apply the patch 1/3 before this]

To please lockdep here we use class_reclassify to change
the lock class of sg_sysfs_class

Signed-off-by: Dave Young <hidave.darkstar@gmail.com>

---
 drivers/scsi/sg.c |    2 ++
 1 file changed, 2 insertions(+)

--- linux/drivers/scsi/sg.c	2008-05-19 12:41:05.000000000 +0800
+++ linux.new/drivers/scsi/sg.c	2008-05-19 14:08:19.000000000 +0800
@@ -1579,6 +1579,8 @@ init_sg(void)
 		rc = PTR_ERR(sg_sysfs_class);
 		goto err_out;
         }
+
+	class_reclassify(sg_sysfs_class);
 	sg_sysfs_valid = 1;
 	rc = scsi_register_interface(&sg_interface);
 	if (0 == rc) {

Re: [PATCH 2/3][-mm] reclassify the sg_sysfs_class mutex

[Greg KH]

On Tue, May 20, 2008 at 05:58:51PM +0800, Dave Young wrote:
> [Please first apply the patch 1/3 before this]
> 
> To please lockdep here we use class_reclassify to change
> the lock class of sg_sysfs_class

Are you suggesting we do this for every struct class in the kernel?  If
so, why not just do it in the class core, instead of having to modify
every single caller?

I don't think the overall idea of this patch set is acceptable anyway :(

thanks,

greg k-h

Re: [PATCH 2/3][-mm] reclassify the sg_sysfs_class mutex

[Dave Young]

On Wed, May 21, 2008 at 1:22 AM, Greg KH <greg@kroah.com> wrote:
> On Tue, May 20, 2008 at 05:58:51PM +0800, Dave Young wrote:
>> [Please first apply the patch 1/3 before this]
>>
>> To please lockdep here we use class_reclassify to change
>> the lock class of sg_sysfs_class
>
> Are you suggesting we do this for every struct class in the kernel?  If
> so, why not just do it in the class core, instead of having to modify
> every single caller?

I don't think it necessary to do this for every class, because only
class_interface_* calls could cause the issue, and the
class_interface_* are only used by pcmcia & scsi, do you think there
will be other users of them?

>
> I don't think the overall idea of this patch set is acceptable anyway :(
>
> thanks,
>
> greg k-h
>

Re: [PATCH 2/3][-mm] reclassify the sg_sysfs_class mutex

[Greg KH]

On Wed, May 21, 2008 at 09:25:58AM +0800, Dave Young wrote:
> On Wed, May 21, 2008 at 1:22 AM, Greg KH <greg@kroah.com> wrote:
> > On Tue, May 20, 2008 at 05:58:51PM +0800, Dave Young wrote:
> >> [Please first apply the patch 1/3 before this]
> >>
> >> To please lockdep here we use class_reclassify to change
> >> the lock class of sg_sysfs_class
> >
> > Are you suggesting we do this for every struct class in the kernel?  If
> > so, why not just do it in the class core, instead of having to modify
> > every single caller?
> 
> I don't think it necessary to do this for every class, because only
> class_interface_* calls could cause the issue, and the
> class_interface_* are only used by pcmcia & scsi, do you think there
> will be other users of them?

I do not know, there might be in the future.

thanks,

greg k-h