They are allocated dynamically, and thus need sysfs_attr_init()
treatment.
Signed-off-by: David S. Miller <[email protected]>
diff --git a/drivers/scsi/scsi_transport_sas.c b/drivers/scsi/scsi_transport_sas.c
index 927e99c..f9c61b9 100644
--- a/drivers/scsi/scsi_transport_sas.c
+++ b/drivers/scsi/scsi_transport_sas.c
@@ -1679,15 +1679,17 @@ static int sas_user_scan(struct Scsi_Host *shost, uint channel,
*/
#define SETUP_TEMPLATE(attrb, field, perm, test) \
- i->private_##attrb[count] = dev_attr_##field; \
+ i->private_##attrb[count] = dev_attr_##field; \
i->private_##attrb[count].attr.mode = perm; \
+ sysfs_attr_init(&i->private_##attrb[count].attr); \
i->attrb[count] = &i->private_##attrb[count]; \
if (test) \
count++
#define SETUP_TEMPLATE_RW(attrb, field, perm, test, ro_test, ro_perm) \
- i->private_##attrb[count] = dev_attr_##field; \
+ i->private_##attrb[count] = dev_attr_##field; \
i->private_##attrb[count].attr.mode = perm; \
+ sysfs_attr_init(&i->private_##attrb[count].attr); \
if (ro_test) { \
i->private_##attrb[count].attr.mode = ro_perm; \
i->private_##attrb[count].store = NULL; \
On Sat, 2010-04-10 at 21:40 -0700, David Miller wrote:
> They are allocated dynamically, and thus need sysfs_attr_init()
> treatment.
This should all be taken care of by
commit ebd09ec93c90c8ec571d7e166832fb1fc705bf5e
Author: James Bottomley <[email protected]>
Date: Sat Mar 20 12:44:12 2010 -0500
[SCSI] attirbute_container: Initialize sysfs attributes with
sysfs_attr_init
I can't check the FC class, but I have checked the SAS one, and I don't
see any more lockdep warn ons ... unless there's one I'm not seeing
because of the options I've compiled with?
James
From: James Bottomley <[email protected]>
Date: Sun, 11 Apr 2010 07:35:12 -0500
> On Sat, 2010-04-10 at 21:40 -0700, David Miller wrote:
>> They are allocated dynamically, and thus need sysfs_attr_init()
>> treatment.
>
> This should all be taken care of by
>
> commit ebd09ec93c90c8ec571d7e166832fb1fc705bf5e
> Author: James Bottomley <[email protected]>
> Date: Sat Mar 20 12:44:12 2010 -0500
Thanks, that should take care of it, my sparc-2.6 tree hadn't
sync'd with Linus's tree in a while. :-)