Re: Memory barrier -- fix sample code

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Tue, Aug 02, 2016 at 10:32:26AM +1000, Balbir Singh wrote:
> From: Balbir singh <bsingharora@gmail.com>
> Date: Tue, 2 Aug 2016 10:26:34 +1000
> Subject: [PATCH] Correct a sample in memorybarriers.tex
> 
> We probably don't want to show GFP_KERNEL allocation
> under spin_lock, GFP_ATOMIC seems much better.
> 
> Signed-off-by: Balbir singh <bsingharora@gmail.com>

Good catch, applied, thank you!

However, I had to apply this by hand due to some recent mass changes to
figures containing code.  Could you please move to current perfbook.git
for future submissions?

							Thanx, Paul

> ---
>  advsync/memorybarriers.tex | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/advsync/memorybarriers.tex b/advsync/memorybarriers.tex
> index 9eaa0ea..efbd4a2 100644
> --- a/advsync/memorybarriers.tex
> +++ b/advsync/memorybarriers.tex
> @@ -936,7 +936,7 @@ Then the following code might leak memory!
>  \begin{verbatim}
>  spin_lock(&mylock);
>  if (p == NULL)
> -  p = kmalloc(sizeof(*p), GFP_KERNEL);
> +  p = kmalloc(sizeof(*p), GFP_ATOMIC);
>  spin_unlock(&mylock);
>  \end{verbatim}
>  \end{minipage}
> -- 
> 2.7.4
>