Re: [PATCH] firmware class: remove from pending list on load failure

[Sasha Levin <sasha.levin@oracle.com>]

On 01/06/2015 11:52 PM, Ming Lei wrote:
> On Mon, Jan 5, 2015 at 11:41 PM, Sasha Levin <sasha.levin@oracle.com> wrote:
>> > If we failed loading the firmware we have to make sure it leaves the pending
>> > list if abort wasn't executed for it.
>> >
>> > Otherwise we'd free an object still on the pending list and corrupt it.
>> >
>> > Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
>> > ---
>> >  drivers/base/firmware_class.c | 11 ++++++++++-
>> >  1 file changed, 10 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
>> > index 58470c3..8ccf6cf4 100644
>> > --- a/drivers/base/firmware_class.c
>> > +++ b/drivers/base/firmware_class.c
>> > @@ -929,9 +929,18 @@ static int _request_firmware_load(struct firmware_priv *fw_priv,
>> >         cancel_delayed_work_sync(&fw_priv->timeout_work);
>> >         if (is_fw_load_aborted(buf))
>> >                 retval = -EAGAIN;
>> > -       else if (!buf->data)
>> > +       else if (!buf->data) {
>> >                 retval = -ENOMEM;
>> >
>> > +               /*
>> > +                * We failed loading, but abort was never done so we
>> > +                * need to remove it from the pending list ourselves.
>> > +                */
>> > +               mutex_lock(&fw_lock);
>> > +               list_del_init(&buf->pending_list);
>> > +               mutex_unlock(&fw_lock);
> The buf is always removed before the complete_all(), isn't it? Or did
> you observe the issue?

Not in the case where userspacehelper fails. Yes, this is not a theoretical
thing and can be easily reproduced.


Thanks,
Sasha