From: Christoph Lameter <cl@linux-foundation.org>
To: Tejun Heo <tj@kernel.org>
Cc: akpm@linux-foundation.org, linux-mm@kvack.org, mingo@elte.hu,
rusty@rustcorp.com.au, davem@davemloft.net
Subject: Re: [this_cpu_xx V2 10/19] this_cpu: X86 optimized this_cpu operations
Date: Thu, 18 Jun 2009 10:07:38 -0400 (EDT) [thread overview]
Message-ID: <alpine.DEB.1.10.0906181001420.15556@gentwo.org> (raw)
In-Reply-To: <4A39ADBF.1000505@kernel.org>
On Thu, 18 Jun 2009, Tejun Heo wrote:
> > +#define __this_cpu_read(pcp) percpu_from_op("mov", pcp)
> ^^^^
> missing parentheses
> and maybe adding () around val is a good idea too?
() around pcp right.
> Also, I'm not quite sure these macros would operate on the correct
> address. Checking... yeap, the following function,
>
> DEFINE_PER_CPU(int, my_pcpu_cnt);
> void my_func(void)
> {
> int *ptr = &per_cpu__my_pcpu_cnt;
>
> *(int *)this_cpu_ptr(ptr) = 0;
> this_cpu_add(ptr, 1);
Needs to be this_cpu_add(*ptr, 1). this_cpu_add does not take a pointer
to an int but a lvalue. The typical use case is with a struct. I.e.
struct {
int x;
} * ptr = &per_cpu_var(m_cpu_pnt);
then do
this_cpu_add(ptr->x, 1)
> percpu_add(my_pcpu_cnt, 1);
> }
>
> So, this_cpu_add(ptr, 1) ends up accessing the wrong address. Also,
> please note the use of 'addq' instead of 'addl' as the pointer
> variable is being modified.
You incremented the pointer instead of the value pointed to. Look at the
patches that use this_cpu_add(). You pass the object to be incremented not
a pointer. If the convention would be different then the address would
have to be taken of these objects everywhere.
> > +#define irqsafe_cpu_add(pcp, val) percpu_to_op("add", (pcp), val)
> > +#define irqsafe_cpu_sub(pcp, val) percpu_to_op("sub", (pcp), val)
> > +#define irqsafe_cpu_and(pcp, val) percpu_to_op("and", (pcp), val)
> > +#define irqsafe_cpu_or(pcp, val) percpu_to_op("or", (pcp), val)
> > +#define irqsafe_cpu_xor(pcp, val) percpu_to_op("xor", (pcp), val)
>
> Wouldn't it be clearer / easier to define preempt and irqsafe versions
> as aliases of __ prefixed ones?
Ok will do that.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2009-06-18 14:07 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-06-17 20:33 [this_cpu_xx V2 00/19] Introduce this_cpu_xx operations cl
2009-06-17 20:33 ` [this_cpu_xx V2 01/19] Fix handling of pagesets for downed cpus cl
2009-06-17 20:33 ` [this_cpu_xx V2 02/19] Introduce this_cpu_ptr() and generic this_cpu_* operations cl
2009-06-18 1:50 ` Tejun Heo
2009-06-18 2:29 ` Tejun Heo
2009-06-18 13:54 ` Christoph Lameter
2009-06-18 14:49 ` Tejun Heo
2009-06-17 20:33 ` [this_cpu_xx V2 03/19] Use this_cpu operations for SNMP statistics cl
2009-06-18 1:55 ` Tejun Heo
2009-06-17 20:33 ` [this_cpu_xx V2 04/19] Use this_cpu operations for NFS statistics cl
2009-06-18 2:03 ` Tejun Heo
2009-06-17 20:33 ` [this_cpu_xx V2 05/19] use this_cpu ops for network statistics cl
2009-06-17 20:33 ` [this_cpu_xx V2 06/19] this_cpu_ptr: Straight transformations cl
2009-06-17 20:33 ` [this_cpu_xx V2 07/19] this_cpu_ptr: Elimninate get/put_cpu cl
2009-06-17 20:33 ` [this_cpu_xx V2 08/19] this_cpu_ptr: xfs_icsb_modify_counters does not need "cpu" variable cl
2009-06-17 20:33 ` [this_cpu_xx V2 09/19] Use this_cpu_ptr in crypto subsystem cl
2009-06-17 20:33 ` [this_cpu_xx V2 10/19] this_cpu: X86 optimized this_cpu operations cl
2009-06-18 3:00 ` Tejun Heo
2009-06-18 14:07 ` Christoph Lameter [this message]
2009-06-18 14:48 ` Tejun Heo
2009-06-18 15:39 ` Christoph Lameter
2009-06-18 16:06 ` Tejun Heo
2009-06-18 16:15 ` Tejun Heo
2009-06-18 17:05 ` Christoph Lameter
2009-06-19 5:41 ` Rusty Russell
2009-06-23 18:00 ` Christoph Lameter
2009-06-17 20:33 ` [this_cpu_xx V2 11/19] Use this_cpu ops for VM statistics cl
2009-06-18 3:05 ` Tejun Heo
2009-06-17 20:33 ` [this_cpu_xx V2 12/19] RCU: Use this_cpu operations cl
2009-06-17 20:33 ` [this_cpu_xx V2 13/19] Use this_cpu operations in slub cl
2009-06-18 6:20 ` Pekka Enberg
2009-06-18 6:25 ` Pekka Enberg
2009-06-18 13:59 ` Christoph Lameter
2009-06-25 7:12 ` Pekka Enberg
2009-06-18 6:49 ` Tejun Heo
2009-06-18 7:35 ` Pekka Enberg
2009-06-18 13:59 ` Christoph Lameter
2009-06-25 7:11 ` Pekka Enberg
2009-06-17 20:33 ` [this_cpu_xx V2 14/19] this_cpu: Remove slub kmem_cache fields cl
2009-06-17 20:33 ` [this_cpu_xx V2 15/19] Make slub statistics use this_cpu_inc cl
2009-06-17 20:33 ` [this_cpu_xx V2 16/19] this_cpu: slub aggressive use of this_cpu operations in the hotpaths cl
2009-06-18 6:33 ` Pekka Enberg
2009-06-18 11:59 ` Mathieu Desnoyers
2009-06-18 14:00 ` Christoph Lameter
2009-06-17 20:33 ` [this_cpu_xx V2 17/19] Move early initialization of pagesets out of zone_wait_table_init() cl
2009-06-18 3:13 ` Tejun Heo
2009-06-17 20:33 ` [this_cpu_xx V2 18/19] this_cpu_ops: page allocator conversion cl
2009-06-17 20:33 ` [this_cpu_xx V2 19/19] this_cpu ops: Remove pageset_notifier cl
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.DEB.1.10.0906181001420.15556@gentwo.org \
--to=cl@linux-foundation.org \
--cc=akpm@linux-foundation.org \
--cc=davem@davemloft.net \
--cc=linux-mm@kvack.org \
--cc=mingo@elte.hu \
--cc=rusty@rustcorp.com.au \
--cc=tj@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox