Re: [bug report] RDMA/rxe: Rewrite rxe_task.c

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Mar 29, 2023 at 02:09:17PM -0500, Bob Pearson wrote:
> On 3/29/23 01:48, Dan Carpenter wrote:
> > On Wed, Mar 29, 2023 at 09:27:26AM +0300, Dan Carpenter wrote:
> >> Hello Bob Pearson,
> >>
> >> The patch d94671632572: "RDMA/rxe: Rewrite rxe_task.c" from Mar 4,
> >> 2023, leads to the following Smatch static checker warning:
> >>
> >> 	drivers/infiniband/sw/rxe/rxe_task.c:24 __reserve_if_idle()
> >> 	warn: bitwise AND condition is false here
> >>
> >> drivers/infiniband/sw/rxe/rxe_task.c
> >>     20 static bool __reserve_if_idle(struct rxe_task *task)
> >>     21 {
> >>     22         WARN_ON(rxe_read(task->qp) <= 0);
> >>     23 
> >> --> 24         if (task->tasklet.state & TASKLET_STATE_SCHED)
> >>                                          ^^^^^^^^^^^^^^^^^^^
> >> This is zero.  Should the check be == TASKLET_STATE_SCHED?
> >>
> > 
> > The next function as well.
> > 
> > drivers/infiniband/sw/rxe/rxe_task.c:49 __is_done() warn: bitwise AND condition is false here
> > 
> > regards,
> > dan carpenter
> > 
> 
> Good catch. I was trying to open code the test in tasklet_schedule which was
> test_and_set_bit(TASKLET_STATE_SCHED, &t->state). I should have typed
> 
> 	if (task->tasklet.state & (1 << TASKLET_STATE_BIT)) or similar.

Thanks.  I'm sorry, I didn't look at this carefully. I just looked at
the next lines which are talking about a different state.  (My bug
reports are the much lazier option than my patches because it's faster
to be lazy).

Also I'm supposed to have a different check for shifter vs mask bugs
which should have warned about this.  I'll look into that.

regards,
dan carpenter




[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Photo]     [Yosemite News]     [Yosemite Photos]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux