[PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()

[Jeff Liu]

From: Jie Liu <jeff...@oracle.com>

Simplify the code in current_tail() via list_last_entry_or_null().

Signed-off-by: Jie Liu <jeff...@oracle.com>
---
fs/gfs2/log.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
index 610613f..555f767 100644
--- a/fs/gfs2/log.c
+++ b/fs/gfs2/log.c
@@ -441,13 +441,9 @@ static unsigned int current_tail(struct gfs2_sbd *sdp)

spin_lock(&sdp->sd_ail_lock);

- if (list_empty(&sdp->sd_ail1_list)) {
- tail = sdp->sd_log_head;
- } else {
- tr = list_entry(sdp->sd_ail1_list.prev, struct gfs2_trans,
- tr_list);
- tail = tr->tr_first;
- }
+ tr = list_last_entry_or_null(&sdp->sd_ail1_list, struct gfs2_trans,
+ tr_list);
+ tail = tr ? tr->tr_first : sdp->sd_log_head;

spin_unlock(&sdp->sd_ail_lock);

--
1.8.3.2
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

[Oleg Nesterov]

On 11/15, Jeff Liu wrote:
>
> @@ -441,13 +441,9 @@ static unsigned int current_tail(struct gfs2_sbd *sdp)
>
> spin_lock(&sdp->sd_ail_lock);
>
> - if (list_empty(&sdp->sd_ail1_list)) {
> - tail = sdp->sd_log_head;
> - } else {
> - tr = list_entry(sdp->sd_ail1_list.prev, struct gfs2_trans,
> - tr_list);
> - tail = tr->tr_first;
> - }
> + tr = list_last_entry_or_null(&sdp->sd_ail1_list, struct gfs2_trans,
> + tr_list);
> + tail = tr ? tr->tr_first : sdp->sd_log_head;
>

Personally I agree with Steven. At least in this case
list_last_entry_or_null() doesn't really help to simplify the code.

But probably list_last_entry() makes sense in the "else" branch,
athough this is minor.


Off-topic. Not sure this really makes sense, but I was thinking about

list_get_first(pos, head, member) \
((pos) = list_first_entry(head, typeof(*pos), member))

and list_get_first() last of course. The obvious advantage is that
compared to

tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);

above you do not need to type "struct gfs2_trans",

list_get_last(tr, sdp->sd_ail1_list, tr_list);

looks a bit better.

Oleg.

[Oleg Nesterov]

On 11/15, Oleg Nesterov wrote:
>
> But probably list_last_entry() makes sense in the "else" branch,
> athough this is minor.
>
>
> Off-topic. Not sure this really makes sense, but I was thinking about
>
> list_get_first(pos, head, member) \
> ((pos) = list_first_entry(head, typeof(*pos), member))
>
> and list_get_first() last of course. The obvious advantage is that
> compared to
>
> tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);
>
> above you do not need to type "struct gfs2_trans",
>
> list_get_last(tr, sdp->sd_ail1_list, tr_list);
>
> looks a bit better.

And list.h can use it too.

So how about the patch below? I can't decide whether it cleanups or
uglfies the code, I do not trust my taste ;) But to me it looks a
bit better this way.

Oleg.
--

diff --git a/include/linux/list.h b/include/linux/list.h
index ef95941..f52aba8 100644
--- a/include/linux/list.h
+++ b/include/linux/list.h
@@ -372,6 +372,12 @@ static inline void list_splice_tail_init(struct list_head *list,
#define list_last_entry(ptr, type, member) \
list_entry((ptr)->prev, type, member)

+#define list_get_first(pos, head, member) \
+ ((pos) = list_first_entry(head, typeof(*(pos)), member))
+
+#define list_get_last(pos, head, member) \
+ ((pos) = list_last_entry(head, typeof(*(pos)), member))
+
/**
* list_first_entry_or_null - get the first element from a list
* @ptr: the list head to take the element from.
@@ -443,7 +449,7 @@ static inline void list_splice_tail_init(struct list_head *list,
* @member: the name of the list_struct within the struct.
*/
#define list_for_each_entry(pos, head, member) \
- for (pos = list_first_entry(head, typeof(*pos), member); \
+ for (list_get_first(pos, head, member); \
&pos->member != (head); \
pos = list_next_entry(pos, member))

@@ -454,7 +460,7 @@ static inline void list_splice_tail_init(struct list_head *list,
* @member: the name of the list_struct within the struct.
*/
#define list_for_each_entry_reverse(pos, head, member) \
- for (pos = list_last_entry(head, typeof(*pos), member); \
+ for (list_get_last(pos, head, member); \
&pos->member != (head); \
pos = list_prev_entry(pos, member))

@@ -517,7 +523,7 @@ static inline void list_splice_tail_init(struct list_head *list,
* @member: the name of the list_struct within the struct.
*/
#define list_for_each_entry_safe(pos, n, head, member) \
- for (pos = list_first_entry(head, typeof(*pos), member), \
+ for (list_get_first(pos, head, member), \
n = list_next_entry(pos, member); \
&pos->member != (head); \
pos = n, n = list_next_entry(n, member))
@@ -564,7 +570,7 @@ static inline void list_splice_tail_init(struct list_head *list,
* of list entry.
*/
#define list_for_each_entry_safe_reverse(pos, n, head, member) \
- for (pos = list_last_entry(head, typeof(*pos), member), \
+ for (list_get_last(pos, head, member), \
n = list_prev_entry(pos, member); \
&pos->member != (head); \
pos = n, n = list_prev_entry(n, member))

[Jeff Liu]

Hi Oleg,

Sorry for my late response!

On 11/16/2013 01:12 AM, Oleg Nesterov wrote:

> On 11/15, Oleg Nesterov wrote:
>>
>> But probably list_last_entry() makes sense in the "else" branch,
>> athough this is minor.
>>
>>
>> Off-topic. Not sure this really makes sense, but I was thinking about
>>
>> list_get_first(pos, head, member) \
>> ((pos) = list_first_entry(head, typeof(*pos), member))
>>
>> and list_get_first() last of course. The obvious advantage is that
>> compared to
>>
>> tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);
>>
>> above you do not need to type "struct gfs2_trans",
>>
>> list_get_last(tr, sdp->sd_ail1_list, tr_list);
>>
>> looks a bit better.
>
> And list.h can use it too.
>
> So how about the patch below? I can't decide whether it cleanups or
> uglfies the code, I do not trust my taste ;) But to me it looks a
> bit better this way.

I personally think this is better for GFS2 case if I understood correctly,
and it can make the existing list helper a bit neat per my taste :)

The current list_first_entry() works fine consider a common pattern below:
while (!list_empty(ptr)) {
struct dummy *p = list_first_entry(ptr, struct dummy, member);

....
}

But I'm not sure if we can find out more external use cases can get
benefits with list_get_first/last...

Thanks,
-Jeff

> _______________________________________________
> xfs mailing list
> x...@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs