[PATCH] bcachefs: fix NULL pointer dereference in try_alloc_bucket

[PATCH] bcachefs: fix NULL pointer dereference in try_alloc_bucket

[Mikulas Patocka]

On Mon, 29 May 2023, Mikulas Patocka wrote:

> The oops happens in set_btree_iter_dontneed and it is caused by the fact 
> that iter->path is NULL. The code in try_alloc_bucket is buggy because it 
> sets "struct btree_iter iter = { NULL };" and then jumps to the "err" 
> label that tries to dereference values in "iter".

Here I'm sending a patch for it.



From: Mikulas Patocka <mpatocka@redhat.com>

The function try_alloc_bucket sets the variable "iter" to NULL and then
(on various error conditions) jumps to the label "err". On the "err"
label, it calls "set_btree_iter_dontneed" that tries to dereference
"iter->trans" and "iter->path".

So, we get an oops on error condition.

This patch fixes the crash by testing that iter.trans and iter.path is
non-zero before calling set_btree_iter_dontneed.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>

---
 fs/bcachefs/alloc_foreground.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Index: bcachefs/fs/bcachefs/alloc_foreground.c
===================================================================
--- bcachefs.orig/fs/bcachefs/alloc_foreground.c
+++ bcachefs/fs/bcachefs/alloc_foreground.c
@@ -371,7 +371,8 @@ static struct open_bucket *try_alloc_buc
 	if (!ob)
 		iter.path->preserve = false;
 err:
-	set_btree_iter_dontneed(&iter);
+	if (iter.trans && iter.path)
+		set_btree_iter_dontneed(&iter);
 	bch2_trans_iter_exit(trans, &iter);
 	printbuf_exit(&buf);
 	return ob;

Re: [PATCH] bcachefs: fix NULL pointer dereference in try_alloc_bucket

[Kent Overstreet]

On Tue, May 30, 2023 at 08:15:41AM -0400, Mikulas Patocka wrote:
> 
> 
> On Mon, 29 May 2023, Mikulas Patocka wrote:
> 
> > The oops happens in set_btree_iter_dontneed and it is caused by the fact 
> > that iter->path is NULL. The code in try_alloc_bucket is buggy because it 
> > sets "struct btree_iter iter = { NULL };" and then jumps to the "err" 
> > label that tries to dereference values in "iter".
> 
> Here I'm sending a patch for it.
> 
> 
> 
> +		set_btree_iter_dontneed(&iter);
>  	bch2_trans_iter_exit(trans, &iter);

I need to look at this code a bit more, perhaps we'll want to move the
check into set_btree_iter_dontneed(), since iter_exit() is safe to call
on an uninitialized-but-zeroed iterator