| Date | Thu, 16 Jun 2011 11:10:22 +0530 | From | Srikar Dronamraju <> | Subject | Re: [PATCH v4 3.0-rc2-tip 4/22] 4: Uprobes: register/unregister probes. |
| |
* Peter Zijlstra <peterz@infradead.org> [2011-06-10 01:03:24]:
> On Tue, 2011-06-07 at 18:29 +0530, Srikar Dronamraju wrote: > > + vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, 0, 0) { > > + loff_t vaddr; > > + struct task_struct *tsk; > > + > > + if (!atomic_inc_not_zero(&vma->vm_mm->mm_users)) > > + continue; > > + > > + mm = vma->vm_mm; > > + if (!valid_vma(vma)) { > > + mmput(mm); > > + continue; > > + } > > + > > + vaddr = vma->vm_start + offset; > > + vaddr -= vma->vm_pgoff << PAGE_SHIFT; > > + if (vaddr < vma->vm_start || vaddr > vma->vm_end) { > > + /* Not in this vma */ > > + mmput(mm); > > + continue; > > + } > > + tsk = get_mm_owner(mm); > > + if (tsk && vaddr > TASK_SIZE_OF(tsk)) { > > + /* > > + * We cannot have a virtual address that is > > + * greater than TASK_SIZE_OF(tsk) > > + */ > > + put_task_struct(tsk); > > + mmput(mm); > > + continue; > > + } > > + put_task_struct(tsk); > > + mm->uprobes_vaddr = (unsigned long) vaddr; > > + list_add(&mm->uprobes_list, &try_list); > > + } > > This still falls flat on its face when there's multiple maps of the same > text in one mm. >
To address this we will use a uprobe_info structure.
struct uprobe_info { unsigned long uprobes_vaddr; struct mm_struct *mm; struct list_head uprobes_list; };
and remove the uprobes_list and uprobes_vaddr entries from mm structure.
the uprobe_info structures will be created in the vma_prio_tree_foreach loop as and when required. Since we now have i_mmap_mutex, allocating this uprobe_info structure as when required should be okay.
Agree?
-- Thanks and Regards Srikar
|