2024-03-08 20:45:14

by Mina Almasry

[permalink] [raw]
Subject: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

The check is duplicated in 2 places, factor it out into a common helper.

Signed-off-by: Mina Almasry <[email protected]>
Reviewed-by: Yunsheng Lin <[email protected]>
---
net/core/page_pool.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/net/core/page_pool.c b/net/core/page_pool.c
index d706fe5548df..dd364d738c00 100644
--- a/net/core/page_pool.c
+++ b/net/core/page_pool.c
@@ -657,6 +657,11 @@ static bool page_pool_recycle_in_cache(struct page *page,
return true;
}

+static bool __page_pool_page_can_be_recycled(const struct page *page)
+{
+ return page_ref_count(page) == 1 && !page_is_pfmemalloc(page);
+}
+
/* If the page refcnt == 1, this will try to recycle the page.
* if PP_FLAG_DMA_SYNC_DEV is set, we'll try to sync the DMA area for
* the configured size min(dma_sync_size, pool->max_len).
@@ -678,7 +683,7 @@ __page_pool_put_page(struct page_pool *pool, struct page *page,
* page is NOT reusable when allocated when system is under
* some pressure. (page_is_pfmemalloc)
*/
- if (likely(page_ref_count(page) == 1 && !page_is_pfmemalloc(page))) {
+ if (likely(__page_pool_page_can_be_recycled(page))) {
/* Read barrier done in page_ref_count / READ_ONCE */

if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
@@ -793,7 +798,7 @@ static struct page *page_pool_drain_frag(struct page_pool *pool,
if (likely(page_pool_unref_page(page, drain_count)))
return NULL;

- if (page_ref_count(page) == 1 && !page_is_pfmemalloc(page)) {
+ if (__page_pool_page_can_be_recycled(page)) {
if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
page_pool_dma_sync_for_device(pool, page, -1);

--
2.44.0.278.ge034bb2e1d-goog



2024-03-09 00:04:42

by Mina Almasry

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

On Fri, Mar 8, 2024 at 3:50 PM David Wei <[email protected]> wrote:
>
> On 2024-03-08 12:44, Mina Almasry wrote:
> > The check is duplicated in 2 places, factor it out into a common helper.
> >
> > Signed-off-by: Mina Almasry <[email protected]>
> > Reviewed-by: Yunsheng Lin <[email protected]>
> > ---
> > net/core/page_pool.c | 9 +++++++--
> > 1 file changed, 7 insertions(+), 2 deletions(-)
> >
> > diff --git a/net/core/page_pool.c b/net/core/page_pool.c
> > index d706fe5548df..dd364d738c00 100644
> > --- a/net/core/page_pool.c
> > +++ b/net/core/page_pool.c
> > @@ -657,6 +657,11 @@ static bool page_pool_recycle_in_cache(struct page *page,
> > return true;
> > }
> >
> > +static bool __page_pool_page_can_be_recycled(const struct page *page)
>
> Could this be made inline?
>

Looking at the rest of the static functions in this file, they don't
specify inline, just static. I guess the compiler is smart enough to
inline static functions in .c files when it makes sense (and does not
when it doesn't)?

But this doesn't seem to be a kernel wide thing. net/core/dev.c does
have static inline functions in it, only page_pool.c doesn't do it. I
guess if there are no objections I can make it static inline to ask
the compiler to inline it. Likely after the merge window reopens if it
closes today.

2024-03-09 00:09:03

by Jakub Kicinski

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

On Fri, 8 Mar 2024 16:04:14 -0800 Mina Almasry wrote:
> > Could this be made inline?
> >
>
> Looking at the rest of the static functions in this file, they don't
> specify inline, just static. I guess the compiler is smart enough to
> inline static functions in .c files when it makes sense (and does not
> when it doesn't)?
>
> But this doesn't seem to be a kernel wide thing. net/core/dev.c does
> have static inline functions in it, only page_pool.c doesn't do it. I
> guess if there are no objections I can make it static inline to ask
> the compiler to inline it. Likely after the merge window reopens if it
> closes today.

It's all good. We have a policy in netdev of "no inline unless you can
prove it makes a difference". It will not make a difference here and it
will mute the "unused function" warning.

2024-03-11 08:32:49

by Ilias Apalodimas

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

On Fri, 8 Mar 2024 at 22:45, Mina Almasry <[email protected]> wrote:
>
> The check is duplicated in 2 places, factor it out into a common helper.
>
> Signed-off-by: Mina Almasry <[email protected]>
> Reviewed-by: Yunsheng Lin <[email protected]>
> ---
> net/core/page_pool.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/net/core/page_pool.c b/net/core/page_pool.c
> index d706fe5548df..dd364d738c00 100644
> --- a/net/core/page_pool.c
> +++ b/net/core/page_pool.c
> @@ -657,6 +657,11 @@ static bool page_pool_recycle_in_cache(struct page *page,
> return true;
> }
>
> +static bool __page_pool_page_can_be_recycled(const struct page *page)
> +{
> + return page_ref_count(page) == 1 && !page_is_pfmemalloc(page);
> +}
> +
> /* If the page refcnt == 1, this will try to recycle the page.
> * if PP_FLAG_DMA_SYNC_DEV is set, we'll try to sync the DMA area for
> * the configured size min(dma_sync_size, pool->max_len).
> @@ -678,7 +683,7 @@ __page_pool_put_page(struct page_pool *pool, struct page *page,
> * page is NOT reusable when allocated when system is under
> * some pressure. (page_is_pfmemalloc)
> */
> - if (likely(page_ref_count(page) == 1 && !page_is_pfmemalloc(page))) {
> + if (likely(__page_pool_page_can_be_recycled(page))) {
> /* Read barrier done in page_ref_count / READ_ONCE */
>
> if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
> @@ -793,7 +798,7 @@ static struct page *page_pool_drain_frag(struct page_pool *pool,
> if (likely(page_pool_unref_page(page, drain_count)))
> return NULL;
>
> - if (page_ref_count(page) == 1 && !page_is_pfmemalloc(page)) {
> + if (__page_pool_page_can_be_recycled(page)) {
> if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
> page_pool_dma_sync_for_device(pool, page, -1);
>
> --
> 2.44.0.278.ge034bb2e1d-goog
>
Reviewed-by: Ilias Apalodimas <[email protected]>

2024-03-11 20:30:42

by patchwork-bot+netdevbpf

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

Hello:

This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski <[email protected]>:

On Fri, 8 Mar 2024 12:44:58 -0800 you wrote:
> The check is duplicated in 2 places, factor it out into a common helper.
>
> Signed-off-by: Mina Almasry <[email protected]>
> Reviewed-by: Yunsheng Lin <[email protected]>
> ---
> net/core/page_pool.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)

Here is the summary with links:
- [net-next,v1] net: page_pool: factor out page_pool recycle check
https://git.kernel.org/netdev/net-next/c/46f40172b681

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html



2024-03-09 01:25:07

by David Wei

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

On 2024-03-08 16:04, Mina Almasry wrote:
> On Fri, Mar 8, 2024 at 3:50 PM David Wei <[email protected]> wrote:
>>
>> On 2024-03-08 12:44, Mina Almasry wrote:
>>> The check is duplicated in 2 places, factor it out into a common helper.
>>>
>>> Signed-off-by: Mina Almasry <[email protected]>
>>> Reviewed-by: Yunsheng Lin <[email protected]>
>>> ---
>>> net/core/page_pool.c | 9 +++++++--
>>> 1 file changed, 7 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/net/core/page_pool.c b/net/core/page_pool.c
>>> index d706fe5548df..dd364d738c00 100644
>>> --- a/net/core/page_pool.c
>>> +++ b/net/core/page_pool.c
>>> @@ -657,6 +657,11 @@ static bool page_pool_recycle_in_cache(struct page *page,
>>> return true;
>>> }
>>>
>>> +static bool __page_pool_page_can_be_recycled(const struct page *page)
>>
>> Could this be made inline?
>>
>
> Looking at the rest of the static functions in this file, they don't
> specify inline, just static. I guess the compiler is smart enough to
> inline static functions in .c files when it makes sense (and does not
> when it doesn't)?
>
> But this doesn't seem to be a kernel wide thing. net/core/dev.c does
> have static inline functions in it, only page_pool.c doesn't do it. I
> guess if there are no objections I can make it static inline to ask
> the compiler to inline it. Likely after the merge window reopens if it
> closes today.

Thanks for checking. Otherwise the change looks good to me, pulling out
the same check in two locations into a(n inline) function then calling
that function instead.

2024-03-08 23:50:09

by David Wei

[permalink] [raw]
Subject: Re: [PATCH net-next v1] net: page_pool: factor out page_pool recycle check

On 2024-03-08 12:44, Mina Almasry wrote:
> The check is duplicated in 2 places, factor it out into a common helper.
>
> Signed-off-by: Mina Almasry <[email protected]>
> Reviewed-by: Yunsheng Lin <[email protected]>
> ---
> net/core/page_pool.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/net/core/page_pool.c b/net/core/page_pool.c
> index d706fe5548df..dd364d738c00 100644
> --- a/net/core/page_pool.c
> +++ b/net/core/page_pool.c
> @@ -657,6 +657,11 @@ static bool page_pool_recycle_in_cache(struct page *page,
> return true;
> }
>
> +static bool __page_pool_page_can_be_recycled(const struct page *page)

Could this be made inline?

> +{
> + return page_ref_count(page) == 1 && !page_is_pfmemalloc(page);
> +}
> +
> /* If the page refcnt == 1, this will try to recycle the page.
> * if PP_FLAG_DMA_SYNC_DEV is set, we'll try to sync the DMA area for
> * the configured size min(dma_sync_size, pool->max_len).
> @@ -678,7 +683,7 @@ __page_pool_put_page(struct page_pool *pool, struct page *page,
> * page is NOT reusable when allocated when system is under
> * some pressure. (page_is_pfmemalloc)
> */
> - if (likely(page_ref_count(page) == 1 && !page_is_pfmemalloc(page))) {
> + if (likely(__page_pool_page_can_be_recycled(page))) {
> /* Read barrier done in page_ref_count / READ_ONCE */
>
> if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
> @@ -793,7 +798,7 @@ static struct page *page_pool_drain_frag(struct page_pool *pool,
> if (likely(page_pool_unref_page(page, drain_count)))
> return NULL;
>
> - if (page_ref_count(page) == 1 && !page_is_pfmemalloc(page)) {
> + if (__page_pool_page_can_be_recycled(page)) {
> if (pool->p.flags & PP_FLAG_DMA_SYNC_DEV)
> page_pool_dma_sync_for_device(pool, page, -1);
>