staging: lustre: o2iblnd: Fix crash in kiblnd_handle_early_rxs()
authorDoug Oucharek <dougso@me.com>
Thu, 10 May 2018 18:07:01 +0000 (11:07 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 14 May 2018 11:55:05 +0000 (13:55 +0200)
commita8da8e528cb0a7f5f7ad9880f13c3359cfb31181
tree0b5c7b57217736f5429ccc523b235fa23869ec71
parentfbae47076e9307154fac38d57b260e8dd7c79b1d
staging: lustre: o2iblnd: Fix crash in kiblnd_handle_early_rxs()

Under upstream staging commit 5a2ca43fa54f561c252c2, the list handling
code in kiblnd_handle_early_rxs() got changed to list_for_each_safe().
That protects against the current thread from deleting the current entry
it is looking at. It does not protect against another thread from deleting
the next item in the list (which the tmp variable points to). The way this
routine holds then releases a lock opens the door to other threads doing
just that.

This patch reverts this commit on this routine.

Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9886
Fixes: 5a2ca43fa54f ("Staging: lustre: Iterate list using list_for_each_entry")
Signed-off-by: Doug Oucharek <dougso@me.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c