| Message ID | 1a39bf0efb8c2832245216d7ccd41582c408e9f4.1604686496.git.lorenzo@kernel.org (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | xdp: introduce bulking for page_pool tx return path | expand |
On Fri, 6 Nov 2020 19:19:08 +0100 Lorenzo Bianconi <lorenzo@kernel.org> wrote: > diff --git a/net/core/page_pool.c b/net/core/page_pool.c > index ef98372facf6..31dac2ad4a1f 100644 > --- a/net/core/page_pool.c > +++ b/net/core/page_pool.c > @@ -11,6 +11,8 @@ > #include <linux/device.h> > > #include <net/page_pool.h> > +#include <net/xdp.h> > + > #include <linux/dma-direction.h> > #include <linux/dma-mapping.h> > #include <linux/page-flags.h> > @@ -362,8 +364,9 @@ static bool pool_page_reusable(struct page_pool *pool, struct page *page) > * If the page refcnt != 1, then the page will be returned to memory > * subsystem. > */ > -void page_pool_put_page(struct page_pool *pool, struct page *page, > - unsigned int dma_sync_size, bool allow_direct) > +static struct page * > +__page_pool_put_page(struct page_pool *pool, struct page *page, > + unsigned int dma_sync_size, bool allow_direct) > { > /* This allocator is optimized for the XDP mode that uses > * one-frame-per-page, but have fallbacks that act like the > @@ -379,15 +382,12 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, > page_pool_dma_sync_for_device(pool, page, > dma_sync_size); > > - if (allow_direct && in_serving_softirq()) > - if (page_pool_recycle_in_cache(page, pool)) > - return; > + if (allow_direct && in_serving_softirq() && > + page_pool_recycle_in_cache(page, pool)) > + return NULL; > > - if (!page_pool_recycle_in_ring(pool, page)) { > - /* Cache full, fallback to free pages */ > - page_pool_return_page(pool, page); > - } > - return; > + /* page is candidate for bulking */ I don't like this comment. This function is also used by non-bulking case. Reaching this point means the page is ready or fulfill the requirement for being recycled into the ptr_ring. I suggest (as before): /* Page found as candidate for recycling */ > + return page; > } > /* Fallback/non-XDP mode: API user have elevated refcnt. > * > @@ -405,9 +405,55 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, > /* Do not replace this with page_pool_return_page() */ > page_pool_release_page(pool, page); > put_page(page); > + > + return NULL; > +} > + > +void page_pool_put_page(struct page_pool *pool, struct page *page, > + unsigned int dma_sync_size, bool allow_direct) > +{ > + page = __page_pool_put_page(pool, page, dma_sync_size, allow_direct); > + if (page && !page_pool_recycle_in_ring(pool, page)) { > + /* Cache full, fallback to free pages */ > + page_pool_return_page(pool, page); > + } > } > EXPORT_SYMBOL(page_pool_put_page); > > +/* Caller must not use data area after call, as this function overwrites it */ > +void page_pool_put_page_bulk(struct page_pool *pool, void **data, > + int count) > +{ > + int i, bulk_len = 0, pa_len = 0; > + > + for (i = 0; i < count; i++) { > + struct page *page = virt_to_head_page(data[i]); > + > + page = __page_pool_put_page(pool, page, -1, false); > + /* Approved for bulk recycling in ptr_ring cache */ > + if (page) > + data[bulk_len++] = page; > + } > + > + if (!bulk_len) > + return; > + > + /* Bulk producer into ptr_ring page_pool cache */ > + page_pool_ring_lock(pool); > + for (i = 0; i < bulk_len; i++) { > + if (__ptr_ring_produce(&pool->ring, data[i])) > + data[pa_len++] = data[i]; > + } > + page_pool_ring_unlock(pool); > + > + /* ptr_ring cache full, free pages outside producer lock since > + * put_page() with refcnt == 1 can be an expensive operation > + */ > + for (i = 0; i < pa_len; i++) > + page_pool_return_page(pool, data[i]); > +} > +EXPORT_SYMBOL(page_pool_put_page_bulk); Rest looks okay :-)
> On Fri, 6 Nov 2020 19:19:08 +0100 > Lorenzo Bianconi <lorenzo@kernel.org> wrote: > [...] > > I don't like this comment. This function is also used by non-bulking > case. Reaching this point means the page is ready or fulfill the > requirement for being recycled into the ptr_ring. > > I suggest (as before): > /* Page found as candidate for recycling */ ack, I will fix it in v5. Regards, Lorenzo > > > + return page; > > } > > /* Fallback/non-XDP mode: API user have elevated refcnt. > > * > > @@ -405,9 +405,55 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, > > /* Do not replace this with page_pool_return_page() */ > > page_pool_release_page(pool, page); > > put_page(page); > > + > > + return NULL; > > +} > > + > > +void page_pool_put_page(struct page_pool *pool, struct page *page, > > + unsigned int dma_sync_size, bool allow_direct) > > +{ > > + page = __page_pool_put_page(pool, page, dma_sync_size, allow_direct); > > + if (page && !page_pool_recycle_in_ring(pool, page)) { > > + /* Cache full, fallback to free pages */ > > + page_pool_return_page(pool, page); > > + } > > } > > EXPORT_SYMBOL(page_pool_put_page); > > > > +/* Caller must not use data area after call, as this function overwrites it */ > > +void page_pool_put_page_bulk(struct page_pool *pool, void **data, > > + int count) > > +{ > > + int i, bulk_len = 0, pa_len = 0; > > + > > + for (i = 0; i < count; i++) { > > + struct page *page = virt_to_head_page(data[i]); > > + > > + page = __page_pool_put_page(pool, page, -1, false); > > + /* Approved for bulk recycling in ptr_ring cache */ > > + if (page) > > + data[bulk_len++] = page; > > + } > > + > > + if (!bulk_len) > > + return; > > + > > + /* Bulk producer into ptr_ring page_pool cache */ > > + page_pool_ring_lock(pool); > > + for (i = 0; i < bulk_len; i++) { > > + if (__ptr_ring_produce(&pool->ring, data[i])) > > + data[pa_len++] = data[i]; > > + } > > + page_pool_ring_unlock(pool); > > + > > + /* ptr_ring cache full, free pages outside producer lock since > > + * put_page() with refcnt == 1 can be an expensive operation > > + */ > > + for (i = 0; i < pa_len; i++) > > + page_pool_return_page(pool, data[i]); > > +} > > +EXPORT_SYMBOL(page_pool_put_page_bulk); > > Rest looks okay :-) > -- > Best regards, > Jesper Dangaard Brouer > MSc.CS, Principal Kernel Engineer at Red Hat > LinkedIn: http://www.linkedin.com/in/brouer >
On Fri, 6 Nov 2020 19:19:08 +0100 Lorenzo Bianconi <lorenzo@kernel.org> wrote: > Introduce the capability to batch page_pool ptr_ring refill since it is > usually run inside the driver NAPI tx completion loop. > > Suggested-by: Jesper Dangaard Brouer <brouer@redhat.com> > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> > --- > include/net/page_pool.h | 26 ++++++++++++++++ > net/core/page_pool.c | 66 ++++++++++++++++++++++++++++++++++------- > net/core/xdp.c | 9 ++---- > 3 files changed, 84 insertions(+), 17 deletions(-) My own suggestion to create __page_pool_put_page() for code sharing, cause a performance regression, because compiler chooses not to inline the function. This cause unnecessary code for in_serving_softirq() and function call overhead. I'm currently testing with __always_inline and likely/unlikely annotation to get compiler to layout code better for I-cache. This shows improvements in my benchmarks. Details in[1]. [1] https://github.com/xdp-project/xdp-project/blob/master/areas/mem/xdp_bulk_return01.org > diff --git a/net/core/page_pool.c b/net/core/page_pool.c > index ef98372facf6..31dac2ad4a1f 100644 > --- a/net/core/page_pool.c > +++ b/net/core/page_pool.c [...] > @@ -362,8 +364,9 @@ static bool pool_page_reusable(struct page_pool *pool, struct page *page) > * If the page refcnt != 1, then the page will be returned to memory > * subsystem. > */ > -void page_pool_put_page(struct page_pool *pool, struct page *page, > - unsigned int dma_sync_size, bool allow_direct) > +static struct page * > +__page_pool_put_page(struct page_pool *pool, struct page *page, > + unsigned int dma_sync_size, bool allow_direct) > { > /* This allocator is optimized for the XDP mode that uses > * one-frame-per-page, but have fallbacks that act like the > @@ -379,15 +382,12 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, > page_pool_dma_sync_for_device(pool, page, > dma_sync_size); > > - if (allow_direct && in_serving_softirq()) > - if (page_pool_recycle_in_cache(page, pool)) > - return; > + if (allow_direct && in_serving_softirq() && > + page_pool_recycle_in_cache(page, pool)) > + return NULL; > > - if (!page_pool_recycle_in_ring(pool, page)) { > - /* Cache full, fallback to free pages */ > - page_pool_return_page(pool, page); > - } > - return; > + /* page is candidate for bulking */ > + return page; > } > /* Fallback/non-XDP mode: API user have elevated refcnt. > * > @@ -405,9 +405,55 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, > /* Do not replace this with page_pool_return_page() */ > page_pool_release_page(pool, page); > put_page(page); > + > + return NULL; > +} > + > +void page_pool_put_page(struct page_pool *pool, struct page *page, > + unsigned int dma_sync_size, bool allow_direct) > +{ > + page = __page_pool_put_page(pool, page, dma_sync_size, allow_direct); > + if (page && !page_pool_recycle_in_ring(pool, page)) { > + /* Cache full, fallback to free pages */ > + page_pool_return_page(pool, page); > + } > } > EXPORT_SYMBOL(page_pool_put_page); > > +/* Caller must not use data area after call, as this function overwrites it */ > +void page_pool_put_page_bulk(struct page_pool *pool, void **data, > + int count) > +{ > + int i, bulk_len = 0, pa_len = 0; > + > + for (i = 0; i < count; i++) { > + struct page *page = virt_to_head_page(data[i]); > + > + page = __page_pool_put_page(pool, page, -1, false); > + /* Approved for bulk recycling in ptr_ring cache */ > + if (page) > + data[bulk_len++] = page; > + } > + > + if (!bulk_len) > + return; > + > + /* Bulk producer into ptr_ring page_pool cache */ > + page_pool_ring_lock(pool); > + for (i = 0; i < bulk_len; i++) { > + if (__ptr_ring_produce(&pool->ring, data[i])) > + data[pa_len++] = data[i]; > + } > + page_pool_ring_unlock(pool); > + > + /* ptr_ring cache full, free pages outside producer lock since > + * put_page() with refcnt == 1 can be an expensive operation > + */ > + for (i = 0; i < pa_len; i++) > + page_pool_return_page(pool, data[i]); > +} > +EXPORT_SYMBOL(page_pool_put_page_bulk);
diff --git a/include/net/page_pool.h b/include/net/page_pool.h index 81d7773f96cd..b5b195305346 100644 --- a/include/net/page_pool.h +++ b/include/net/page_pool.h @@ -152,6 +152,8 @@ struct page_pool *page_pool_create(const struct page_pool_params *params); void page_pool_destroy(struct page_pool *pool); void page_pool_use_xdp_mem(struct page_pool *pool, void (*disconnect)(void *)); void page_pool_release_page(struct page_pool *pool, struct page *page); +void page_pool_put_page_bulk(struct page_pool *pool, void **data, + int count); #else static inline void page_pool_destroy(struct page_pool *pool) { @@ -165,6 +167,11 @@ static inline void page_pool_release_page(struct page_pool *pool, struct page *page) { } + +static inline void page_pool_put_page_bulk(struct page_pool *pool, void **data, + int count) +{ +} #endif void page_pool_put_page(struct page_pool *pool, struct page *page, @@ -215,4 +222,23 @@ static inline void page_pool_nid_changed(struct page_pool *pool, int new_nid) if (unlikely(pool->p.nid != new_nid)) page_pool_update_nid(pool, new_nid); } + +static inline void page_pool_ring_lock(struct page_pool *pool) + __acquires(&pool->ring.producer_lock) +{ + if (in_serving_softirq()) + spin_lock(&pool->ring.producer_lock); + else + spin_lock_bh(&pool->ring.producer_lock); +} + +static inline void page_pool_ring_unlock(struct page_pool *pool) + __releases(&pool->ring.producer_lock) +{ + if (in_serving_softirq()) + spin_unlock(&pool->ring.producer_lock); + else + spin_unlock_bh(&pool->ring.producer_lock); +} + #endif /* _NET_PAGE_POOL_H */ diff --git a/net/core/page_pool.c b/net/core/page_pool.c index ef98372facf6..31dac2ad4a1f 100644 --- a/net/core/page_pool.c +++ b/net/core/page_pool.c @@ -11,6 +11,8 @@ #include <linux/device.h> #include <net/page_pool.h> +#include <net/xdp.h> + #include <linux/dma-direction.h> #include <linux/dma-mapping.h> #include <linux/page-flags.h> @@ -362,8 +364,9 @@ static bool pool_page_reusable(struct page_pool *pool, struct page *page) * If the page refcnt != 1, then the page will be returned to memory * subsystem. */ -void page_pool_put_page(struct page_pool *pool, struct page *page, - unsigned int dma_sync_size, bool allow_direct) +static struct page * +__page_pool_put_page(struct page_pool *pool, struct page *page, + unsigned int dma_sync_size, bool allow_direct) { /* This allocator is optimized for the XDP mode that uses * one-frame-per-page, but have fallbacks that act like the @@ -379,15 +382,12 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, page_pool_dma_sync_for_device(pool, page, dma_sync_size); - if (allow_direct && in_serving_softirq()) - if (page_pool_recycle_in_cache(page, pool)) - return; + if (allow_direct && in_serving_softirq() && + page_pool_recycle_in_cache(page, pool)) + return NULL; - if (!page_pool_recycle_in_ring(pool, page)) { - /* Cache full, fallback to free pages */ - page_pool_return_page(pool, page); - } - return; + /* page is candidate for bulking */ + return page; } /* Fallback/non-XDP mode: API user have elevated refcnt. * @@ -405,9 +405,55 @@ void page_pool_put_page(struct page_pool *pool, struct page *page, /* Do not replace this with page_pool_return_page() */ page_pool_release_page(pool, page); put_page(page); + + return NULL; +} + +void page_pool_put_page(struct page_pool *pool, struct page *page, + unsigned int dma_sync_size, bool allow_direct) +{ + page = __page_pool_put_page(pool, page, dma_sync_size, allow_direct); + if (page && !page_pool_recycle_in_ring(pool, page)) { + /* Cache full, fallback to free pages */ + page_pool_return_page(pool, page); + } } EXPORT_SYMBOL(page_pool_put_page); +/* Caller must not use data area after call, as this function overwrites it */ +void page_pool_put_page_bulk(struct page_pool *pool, void **data, + int count) +{ + int i, bulk_len = 0, pa_len = 0; + + for (i = 0; i < count; i++) { + struct page *page = virt_to_head_page(data[i]); + + page = __page_pool_put_page(pool, page, -1, false); + /* Approved for bulk recycling in ptr_ring cache */ + if (page) + data[bulk_len++] = page; + } + + if (!bulk_len) + return; + + /* Bulk producer into ptr_ring page_pool cache */ + page_pool_ring_lock(pool); + for (i = 0; i < bulk_len; i++) { + if (__ptr_ring_produce(&pool->ring, data[i])) + data[pa_len++] = data[i]; + } + page_pool_ring_unlock(pool); + + /* ptr_ring cache full, free pages outside producer lock since + * put_page() with refcnt == 1 can be an expensive operation + */ + for (i = 0; i < pa_len; i++) + page_pool_return_page(pool, data[i]); +} +EXPORT_SYMBOL(page_pool_put_page_bulk); + static void page_pool_empty_ring(struct page_pool *pool) { struct page *page; diff --git a/net/core/xdp.c b/net/core/xdp.c index 66ac275a0360..ff7c801bd40c 100644 --- a/net/core/xdp.c +++ b/net/core/xdp.c @@ -393,16 +393,11 @@ EXPORT_SYMBOL_GPL(xdp_return_frame_rx_napi); void xdp_flush_frame_bulk(struct xdp_frame_bulk *bq) { struct xdp_mem_allocator *xa = bq->xa; - int i; - if (unlikely(!xa)) + if (unlikely(!xa || !bq->count)) return; - for (i = 0; i < bq->count; i++) { - struct page *page = virt_to_head_page(bq->q[i]); - - page_pool_put_full_page(xa->page_pool, page, false); - } + page_pool_put_page_bulk(xa->page_pool, bq->q, bq->count); bq->count = 0; } EXPORT_SYMBOL_GPL(xdp_flush_frame_bulk);
Introduce the capability to batch page_pool ptr_ring refill since it is usually run inside the driver NAPI tx completion loop. Suggested-by: Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> --- include/net/page_pool.h | 26 ++++++++++++++++ net/core/page_pool.c | 66 ++++++++++++++++++++++++++++++++++------- net/core/xdp.c | 9 ++---- 3 files changed, 84 insertions(+), 17 deletions(-)