Re: [PATCH v4 3.0-rc2-tip 7/22] 7: uprobes: mmap and fork hooks.

From: Peter Zijlstra
Date: Fri Jun 24 2011 - 03:43:38 EST


On Fri, 2011-06-24 at 07:36 +0530, Srikar Dronamraju wrote:
> >
> > so I am thinking of a solution that includes most of your ideas along
> > with using i_mmap_mutex in mmap_uprobe path.
> >
>
> Addressing Peter's comments given on irc wrt i_mmap_mutex.
>

> void _unregister_uprobe(...)
> {
> if (!del_consumer(...)) { // includes tree removal on last consumer
> return;
> }
> if (uprobe->consumers)
> return;
>
> mutex_lock(&mapping->i_mmap_mutex); //sync with mmap.
> vma_prio_tree_foreach() {
> // create list
> }
>
> mutex_unlock(&mapping->i_mmap_mutex);
>
> list_for_each_entry_safe() {
> // remove from list
> down_read(&mm->mmap_sem);
> remove_breakpoint(); // unconditional, if it wasn't there
> up_read(&mm->mmap_sem);
> }
>
> mutex_lock(&mapping->i_mmap_mutex);
> delete_uprobe(uprobe);
> mutex_unlock(&mapping->i_mmap_mutex);
>
> inode->uprobes_count--;
> mutex_unlock(&inode->i_mutex);

Right, so this lonesome unlock got me puzzled for a while, I always find
it best not to do asymmetric locking like this, keep the lock and unlock
in the same function.

> }
>
> int register_uprobe(...)
> {
> uprobe = alloc_uprobe(...); // find or insert in tree
>
> mutex_lock(&inode->i_mutex); // sync with register/unregister
> if (uprobe->consumers) {
> add_consumer();
> goto put_unlock;
> }
> add_consumer();
> inode->uprobes_count++;
> mutex_lock(&mapping->i_mmap_mutex); //sync with mmap.
> vma_prio_tree_foreach(..) {
> // get mm ref, add to list blah blah
> }
>
> mutex_unlock(&mapping->i_mmap_mutex);
> list_for_each_entry_safe() {
> if (ret) {
> // del from list etc..
> //
> continue;
> }
> down_read(mm->mmap_sem);
> ret = install_breakpoint();
> up_read(..);
> // del from list etc..
> //
> if (ret && (ret == -ESRCH || ret == -EEXIST))
> ret = 0;
> }
>
> if (ret)
> _unregister_uprobe();
>
> put_unlock:
> mutex_unlock(&inode->i_mutex);

You see, now this is a double unlock

> put_uprobe(uprobe);
> return ret;
> }
>
> void unregister_uprobe(...)
> {
> mutex_lock(&inode->i_mutex); // sync with register/unregister
> uprobe = find_uprobe(); // ref++
> _unregister_uprobe();
> mutex_unlock(&inode->i_mutex);

idem

> put_uprobe(uprobe);
> }
>
> int mmap_uprobe(struct vm_area_struct *vma)
> {
> struct list_head tmp_list;
> struct uprobe *uprobe, *u;
> struct mm_struct *mm;
> struct inode *inode;
> int ret = 0;
>
> if (!valid_vma(vma))
> return ret; /* Bail-out */
>
> mm = vma->vm_mm;
> inode = vma->vm_file->f_mapping->host;
> if (inode->uprobes_count)
> return ret;
> __iget(inode);
>
> INIT_LIST_HEAD(&tmp_list);
>
> mutex_lock(&mapping->i_mmap_mutex);
> add_to_temp_list(vma, inode, &tmp_list);
> list_for_each_entry_safe(uprobe, u, &tmp_list, pending_list) {
> loff_t vaddr;
>
> list_del(&uprobe->pending_list);
> if (ret)
> continue;
>
> vaddr = vma->vm_start + uprobe->offset;
> vaddr -= vma->vm_pgoff << PAGE_SHIFT;
> ret = install_breakpoint(mm, uprobe, vaddr);

Right, so this is the problem, you cannot do allocations under
i_mmap_mutex, however I think you can under i_mutex.

> if (ret && (ret == -ESRCH || ret == -EEXIST))
> ret = 0;
> }
>
> mutex_unlock(&mapping->i_mmap_mutex);
> iput(inode);
> return ret;
> }
>
> int munmap_uprobe(struct vm_area_struct *vma)
> {
> struct list_head tmp_list;
> struct uprobe *uprobe, *u;
> struct mm_struct *mm;
> struct inode *inode;
> int ret = 0;
>
> if (!valid_vma(vma))
> return ret; /* Bail-out */
>
> mm = vma->vm_mm;
> inode = vma->vm_file->f_mapping->host;
> if (inode->uprobes_count)
> return ret;

Should that be !->uprobes_count?

> // walk thro RB tree and decrement mm->uprobes_count
> walk_rbtree_and_dec_uprobes_count(); //hold treelock.
>
> return ret;
> }

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/