On Fri, Jun 05, 2020 at 04:29:22AM -0700, Matthew Wilcox wrote:
> On Thu, Jun 04, 2020 at 08:55:55PM -0700, Eric Biggers wrote:
> > Possibly a bug in lib/radix-tree.c? this_cpu_ptr() in
> > radix_tree_node_alloc()
> > can be reached without a prior preempt_disable(). Or is the caller of
> > idr_
On Thu, Jun 04, 2020 at 08:55:55PM -0700, Eric Biggers wrote:
> Possibly a bug in lib/radix-tree.c? this_cpu_ptr() in radix_tree_node_alloc()
> can be reached without a prior preempt_disable(). Or is the caller of
> idr_alloc() doing something wrong?
Yes, the idr_alloc() call is plainly wrong:
syzbot has bisected this bug to:
commit e42671084361302141a09284fde9bbc14fdd16bf
Author: Manivannan Sadhasivam
Date: Thu May 7 12:53:06 2020 +
net: qrtr: Do not depend on ARCH_QCOM
bisection log: https://syzkaller.appspot.com/x/bisect.txt?x=17e2221210
start commit: acf25aa6 Mer
[+Cc Matthew Wilcox]
Possibly a bug in lib/radix-tree.c? this_cpu_ptr() in radix_tree_node_alloc()
can be reached without a prior preempt_disable(). Or is the caller of
idr_alloc() doing something wrong?
On Thu, Jun 04, 2020 at 07:02:18PM -0700, syzbot wrote:
> Hello,
>
> syzbot found the foll
