Re: lockup in rb_get_reader_page

[Jiaying Zhang <jiayingz@google.com>]

On Wed, Mar 31, 2010 at 5:28 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
> On Tue, 2010-03-30 at 21:01 -0700, Jiaying Zhang wrote:
>
>> I am going to patch rb_get_reader_page() to print out some debugging
>> message when it enters into an infinitely loop. I will keep you updated
>> if I find any interesting info.
>
> I was just thinking about this some more. One restriction to the
> ring_buffer is that a reader can not preempt a writer, which means that
> a read can not happen in an interrupt or NMI. If a read happens there
> and preempts a writer as the writer was moving the head page, it will go
> into an infinite loop waiting for the writer to finish its move.
>
> You can only do reads from interrupt context if the writer disables
> interrupts before reserving and committing the data.
>
> Are you doing reads from interrupt context?
>
No. We are using a user-space reader that reads ring buffer via syscall
or ioctl. But even though we don't read from interrupt context, as I
understand, we can still have a reader on another cpuB reads cpuA's
buffer when the writer on cpuA moves the head page. Will we have
any problem in that case?

Jiaying

> -- Steve
>
>
>