Re: [PATCH v2 20/23] staging: qlge: Fix CHECK: usleep_range is preferred over udelay

[Dan Carpenter]

On Thu, Dec 12, 2019 at 05:00:57AM -0600, Scott Schafer wrote:
> On Thu, Dec 12, 2019 at 01:45:57PM +0300, Sergei Shtylyov wrote:
> > Hello!
> > 
> > On 11.12.2019 21:12, Scott Schafer wrote:
> > 
> > > chage udelay() to usleep_range()
> > 
> >    Change?
> > 
> > > Signed-off-by: Scott Schafer <schaferjscott@xxxxxxxxx>
> > > ---
> > >   drivers/staging/qlge/qlge_main.c | 2 +-
> > >   1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/staging/qlge/qlge_main.c b/drivers/staging/qlge/qlge_main.c
> > > index e18aa335c899..9427386e4a1e 100644
> > > --- a/drivers/staging/qlge/qlge_main.c
> > > +++ b/drivers/staging/qlge/qlge_main.c
> > > @@ -147,7 +147,7 @@ int ql_sem_spinlock(struct ql_adapter *qdev, u32 sem_mask)
> > >   	do {
> > >   		if (!ql_sem_trylock(qdev, sem_mask))
> > >   			return 0;
> > > -		udelay(100);
> > > +		usleep_range(100, 200);
> > 
> >    I hope you're not in atomic context...
> > 
> > >   	} while (--wait_count);
> > >   	return -ETIMEDOUT;
> > >   }
> > 
> > MBR, Sergei
> 
> Im not quite what you mean by "I hope you're not in atomic context",
> could you please explain why you said this? 

You can't sleep from certain IRQs or when you are holding certain locks
(spin_locks and rwlocks).  The we have preempt_disable() then you can't
sleep.

regards,
dan carpenter