Quoting r. Roland Dreier <[EMAIL PROTECTED]>:
> OK, this looks fine. Applied (with slight renames of function names).
>
> Thanks,
> Roland
>
Now that ib_umem_release is always called in the context of a system
call, we dont have to get_task_mm, since current->mm is not going anywere.
Signed-off-by: Michael S. Tsirkin <[EMAIL PROTECTED]>
Index: infiniband/core/uverbs_mem.c
===================================================================
--- infiniband/core/uverbs_mem.c (revision 2740)
+++ infiniband/core/uverbs_mem.c (working copy)
@@ -174,14 +174,10 @@ void ib_umem_release(struct ib_device *d
__ib_umem_release(dev, umem, 1);
- mm = get_task_mm(current);
- if (!mm)
- return;
-
+ mm = current->mm;
down_write(&mm->mmap_sem);
mm->locked_vm -= PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT;
up_write(&mm->mmap_sem);
- mmput(mm);
}
static void ib_umem_account(void *work_ptr)
--
MST
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
