LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* Is read barrier missed in kfifo?
@ 2018-05-11 7:25 Xiao Guangrong
2018-05-11 7:33 ` Stefani Seibold
2018-05-11 8:32 ` Peter Zijlstra
0 siblings, 2 replies; 6+ messages in thread
From: Xiao Guangrong @ 2018-05-11 7:25 UTC (permalink / raw)
To: paulmck; +Cc: peterz, rostedt, Lai Jiangshan, stefani, linux-kernel
Hi,
Currently, there is no read barrier between reading the index
(kfifo.in) and fetching the real data from the fifo.
I am afraid that will cause the vfifo is observed as not empty
however the data is not actually ready for read. Right?
Thanks!
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Is read barrier missed in kfifo?
2018-05-11 7:25 Is read barrier missed in kfifo? Xiao Guangrong
@ 2018-05-11 7:33 ` Stefani Seibold
2018-05-11 8:32 ` Peter Zijlstra
1 sibling, 0 replies; 6+ messages in thread
From: Stefani Seibold @ 2018-05-11 7:33 UTC (permalink / raw)
To: Xiao Guangrong, paulmck; +Cc: peterz, rostedt, Lai Jiangshan, linux-kernel
My guts thinks you are right. Feel free to send a patch...
Am Freitag, den 11.05.2018, 15:25 +0800 schrieb Xiao Guangrong:
> Hi,
>
> Currently, there is no read barrier between reading the index
> (kfifo.in) and fetching the real data from the fifo.
>
> I am afraid that will cause the vfifo is observed as not empty
> however the data is not actually ready for read. Right?
>
> Thanks!
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Is read barrier missed in kfifo?
2018-05-11 7:25 Is read barrier missed in kfifo? Xiao Guangrong
2018-05-11 7:33 ` Stefani Seibold
@ 2018-05-11 8:32 ` Peter Zijlstra
2018-05-11 16:20 ` Paul E. McKenney
1 sibling, 1 reply; 6+ messages in thread
From: Peter Zijlstra @ 2018-05-11 8:32 UTC (permalink / raw)
To: Xiao Guangrong; +Cc: paulmck, rostedt, Lai Jiangshan, stefani, linux-kernel
On Fri, May 11, 2018 at 03:25:18PM +0800, Xiao Guangrong wrote:
>
> Hi,
>
> Currently, there is no read barrier between reading the index
> (kfifo.in) and fetching the real data from the fifo.
>
> I am afraid that will cause the vfifo is observed as not empty
> however the data is not actually ready for read. Right?
That code is decidedly dodgy indeed. I can only see smp_wmb() but no
matching barriers at all -- therefore the code is almost certainly as
good as not having any barriers at all.
I would suggest you try and convert the code to smp_store_release() and
smp_load_acquire() while you're at it.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Is read barrier missed in kfifo?
2018-05-11 8:32 ` Peter Zijlstra
@ 2018-05-11 16:20 ` Paul E. McKenney
2018-05-14 6:57 ` Peter Zijlstra
0 siblings, 1 reply; 6+ messages in thread
From: Paul E. McKenney @ 2018-05-11 16:20 UTC (permalink / raw)
To: Peter Zijlstra
Cc: Xiao Guangrong, rostedt, Lai Jiangshan, stefani, linux-kernel
On Fri, May 11, 2018 at 10:32:42AM +0200, Peter Zijlstra wrote:
> On Fri, May 11, 2018 at 03:25:18PM +0800, Xiao Guangrong wrote:
> >
> > Hi,
> >
> > Currently, there is no read barrier between reading the index
> > (kfifo.in) and fetching the real data from the fifo.
> >
> > I am afraid that will cause the vfifo is observed as not empty
> > however the data is not actually ready for read. Right?
>
> That code is decidedly dodgy indeed. I can only see smp_wmb() but no
> matching barriers at all -- therefore the code is almost certainly as
> good as not having any barriers at all.
>
> I would suggest you try and convert the code to smp_store_release() and
> smp_load_acquire() while you're at it.
Isn't this one of the places where we rely on control dependencies?
Thanx, Paul
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Is read barrier missed in kfifo?
2018-05-11 16:20 ` Paul E. McKenney
@ 2018-05-14 6:57 ` Peter Zijlstra
2018-05-14 13:25 ` Paul E. McKenney
0 siblings, 1 reply; 6+ messages in thread
From: Peter Zijlstra @ 2018-05-14 6:57 UTC (permalink / raw)
To: Paul E. McKenney
Cc: Xiao Guangrong, rostedt, Lai Jiangshan, stefani, linux-kernel
On Fri, May 11, 2018 at 09:20:53AM -0700, Paul E. McKenney wrote:
> On Fri, May 11, 2018 at 10:32:42AM +0200, Peter Zijlstra wrote:
> > On Fri, May 11, 2018 at 03:25:18PM +0800, Xiao Guangrong wrote:
> > >
> > > Hi,
> > >
> > > Currently, there is no read barrier between reading the index
> > > (kfifo.in) and fetching the real data from the fifo.
> > >
> > > I am afraid that will cause the vfifo is observed as not empty
> > > however the data is not actually ready for read. Right?
> >
> > That code is decidedly dodgy indeed. I can only see smp_wmb() but no
> > matching barriers at all -- therefore the code is almost certainly as
> > good as not having any barriers at all.
> >
> > I would suggest you try and convert the code to smp_store_release() and
> > smp_load_acquire() while you're at it.
>
> Isn't this one of the places where we rely on control dependencies?
Then it bloody well should have a comment. But at least one side of the
fifo needs a read barrier I think. We can rely on a ctrl-dep on the
write side, where we read the head/tail values, compute space and then
conditionally allow writes to happen.
But on the read side it's all reads and ctrl-dep doesn't help anything.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Is read barrier missed in kfifo?
2018-05-14 6:57 ` Peter Zijlstra
@ 2018-05-14 13:25 ` Paul E. McKenney
0 siblings, 0 replies; 6+ messages in thread
From: Paul E. McKenney @ 2018-05-14 13:25 UTC (permalink / raw)
To: Peter Zijlstra
Cc: Xiao Guangrong, rostedt, Lai Jiangshan, stefani, linux-kernel
On Mon, May 14, 2018 at 08:57:55AM +0200, Peter Zijlstra wrote:
> On Fri, May 11, 2018 at 09:20:53AM -0700, Paul E. McKenney wrote:
> > On Fri, May 11, 2018 at 10:32:42AM +0200, Peter Zijlstra wrote:
> > > On Fri, May 11, 2018 at 03:25:18PM +0800, Xiao Guangrong wrote:
> > > >
> > > > Hi,
> > > >
> > > > Currently, there is no read barrier between reading the index
> > > > (kfifo.in) and fetching the real data from the fifo.
> > > >
> > > > I am afraid that will cause the vfifo is observed as not empty
> > > > however the data is not actually ready for read. Right?
> > >
> > > That code is decidedly dodgy indeed. I can only see smp_wmb() but no
> > > matching barriers at all -- therefore the code is almost certainly as
> > > good as not having any barriers at all.
> > >
> > > I would suggest you try and convert the code to smp_store_release() and
> > > smp_load_acquire() while you're at it.
> >
> > Isn't this one of the places where we rely on control dependencies?
>
> Then it bloody well should have a comment. But at least one side of the
> fifo needs a read barrier I think. We can rely on a ctrl-dep on the
> write side, where we read the head/tail values, compute space and then
> conditionally allow writes to happen.
>
> But on the read side it's all reads and ctrl-dep doesn't help anything.
Agreed, for a control depdendency to help, there does need to be a
control dependency. ;-)
Thanx, Paul
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2018-05-14 13:24 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-11 7:25 Is read barrier missed in kfifo? Xiao Guangrong
2018-05-11 7:33 ` Stefani Seibold
2018-05-11 8:32 ` Peter Zijlstra
2018-05-11 16:20 ` Paul E. McKenney
2018-05-14 6:57 ` Peter Zijlstra
2018-05-14 13:25 ` Paul E. McKenney
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).