From: Jakub Kicinski <kuba@kernel.org>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, edumazet@google.com, pabeni@redhat.com,
andrew+netdev@lunn.ch, horms@kernel.org, almasrymina@google.com,
michael.chan@broadcom.com, tariqt@nvidia.com,
dtatulea@nvidia.com, hawk@kernel.org,
ilias.apalodimas@linaro.org, alexanderduyck@fb.com,
sdf@fomichev.me, Jakub Kicinski <kuba@kernel.org>
Subject: [PATCH net-next 12/15] eth: fbnic: allocate unreadable page pool for the payloads
Date: Tue, 19 Aug 2025 19:57:01 -0700 [thread overview]
Message-ID: <20250820025704.166248-13-kuba@kernel.org> (raw)
In-Reply-To: <20250820025704.166248-1-kuba@kernel.org>
Allow allocating a page pool with unreadable memory for the payload
ring (sub1). We need to provide the queue ID so that the memory provider
can match the PP, and use the appropriate page pool DMA sync helper.
While at it remove the define for page pool flags.
The rxq_idx is passed to fbnic_alloc_rx_qt_resources() explicitly
to make it easy to allocate page pools without NAPI (see the patch
after the next).
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
drivers/net/ethernet/meta/fbnic/fbnic_txrx.c | 31 +++++++++++++-------
1 file changed, 21 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_txrx.c b/drivers/net/ethernet/meta/fbnic/fbnic_txrx.c
index 7694b25ef77d..44d9f1598820 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_txrx.c
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_txrx.c
@@ -997,9 +997,8 @@ static void fbnic_add_rx_frag(struct fbnic_napi_vector *nv, u64 rcd,
FBNIC_BD_FRAG_SIZE;
/* Sync DMA buffer */
- dma_sync_single_range_for_cpu(nv->dev,
- page_pool_get_dma_addr_netmem(netmem),
- pg_off, truesize, DMA_BIDIRECTIONAL);
+ page_pool_dma_sync_netmem_for_cpu(qt->sub1.page_pool, netmem,
+ pg_off, truesize);
added = xdp_buff_add_frag(&pkt->buff, netmem, pg_off, len, truesize);
if (unlikely(!added)) {
@@ -1515,16 +1514,14 @@ void fbnic_free_napi_vectors(struct fbnic_net *fbn)
fbnic_free_napi_vector(fbn, fbn->napi[i]);
}
-#define FBNIC_PAGE_POOL_FLAGS \
- (PP_FLAG_DMA_MAP | PP_FLAG_DMA_SYNC_DEV)
-
static int
fbnic_alloc_qt_page_pools(struct fbnic_net *fbn, struct fbnic_napi_vector *nv,
- struct fbnic_q_triad *qt)
+ struct fbnic_q_triad *qt, unsigned int rxq_idx)
{
struct page_pool_params pp_params = {
.order = 0,
- .flags = FBNIC_PAGE_POOL_FLAGS,
+ .flags = PP_FLAG_DMA_MAP |
+ PP_FLAG_DMA_SYNC_DEV,
.pool_size = fbn->hpq_size + fbn->ppq_size,
.nid = NUMA_NO_NODE,
.dev = nv->dev,
@@ -1533,6 +1530,7 @@ fbnic_alloc_qt_page_pools(struct fbnic_net *fbn, struct fbnic_napi_vector *nv,
.max_len = PAGE_SIZE,
.napi = &nv->napi,
.netdev = fbn->netdev,
+ .queue_idx = rxq_idx,
};
struct page_pool *pp;
@@ -1553,10 +1551,23 @@ fbnic_alloc_qt_page_pools(struct fbnic_net *fbn, struct fbnic_napi_vector *nv,
return PTR_ERR(pp);
qt->sub0.page_pool = pp;
- page_pool_get(pp);
+ if (page_pool_rxq_wants_unreadable(&pp_params)) {
+ pp_params.flags |= PP_FLAG_ALLOW_UNREADABLE_NETMEM;
+ pp_params.dma_dir = DMA_FROM_DEVICE;
+
+ pp = page_pool_create(&pp_params);
+ if (IS_ERR(pp))
+ goto err_destroy_sub0;
+ } else {
+ page_pool_get(pp);
+ }
qt->sub1.page_pool = pp;
return 0;
+
+err_destroy_sub0:
+ page_pool_destroy(pp);
+ return PTR_ERR(pp);
}
static void fbnic_ring_init(struct fbnic_ring *ring, u32 __iomem *doorbell,
@@ -1961,7 +1972,7 @@ static int fbnic_alloc_rx_qt_resources(struct fbnic_net *fbn,
struct device *dev = fbn->netdev->dev.parent;
int err;
- err = fbnic_alloc_qt_page_pools(fbn, nv, qt);
+ err = fbnic_alloc_qt_page_pools(fbn, nv, qt, qt->cmpl.q_idx);
if (err)
return err;
--
2.50.1
next prev parent reply other threads:[~2025-08-20 2:57 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-20 2:56 [PATCH net-next 00/15] eth: fbnic: support queue API and zero-copy Rx Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 01/15] net: page_pool: add page_pool_get() Jakub Kicinski
2025-08-20 10:35 ` Jesper Dangaard Brouer
2025-08-20 10:58 ` Dragos Tatulea
2025-08-20 23:11 ` Mina Almasry
2025-08-20 2:56 ` [PATCH net-next 02/15] eth: fbnic: move page pool pointer from NAPI to the ring struct Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 03/15] eth: fbnic: move xdp_rxq_info_reg() to resource alloc Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 04/15] eth: fbnic: move page pool alloc to fbnic_alloc_rx_qt_resources() Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 05/15] eth: fbnic: use netmem_ref where applicable Jakub Kicinski
2025-08-20 23:22 ` Mina Almasry
2025-08-20 2:56 ` [PATCH net-next 06/15] eth: fbnic: request ops lock Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 07/15] eth: fbnic: split fbnic_disable() Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 08/15] eth: fbnic: split fbnic_flush() Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 09/15] eth: fbnic: split fbnic_enable() Jakub Kicinski
2025-08-20 2:56 ` [PATCH net-next 10/15] eth: fbnic: split fbnic_fill() Jakub Kicinski
2025-08-20 2:57 ` [PATCH net-next 11/15] net: page_pool: add helper to pre-check if PP will be unreadable Jakub Kicinski
2025-08-20 11:30 ` Dragos Tatulea
2025-08-20 14:52 ` Jakub Kicinski
2025-08-20 17:45 ` Dragos Tatulea
2025-08-20 2:57 ` Jakub Kicinski [this message]
2025-08-20 23:33 ` [PATCH net-next 12/15] eth: fbnic: allocate unreadable page pool for the payloads Mina Almasry
2025-08-21 0:45 ` Jakub Kicinski
2025-08-20 2:57 ` [PATCH net-next 13/15] eth: fbnic: defer page pool recycling activation to queue start Jakub Kicinski
2025-08-20 2:57 ` [PATCH net-next 14/15] eth: fbnic: don't pass NAPI into pp alloc Jakub Kicinski
2025-08-20 2:57 ` [PATCH net-next 15/15] eth: fbnic: support queue ops / zero-copy Rx Jakub Kicinski
2025-08-21 7:51 ` [PATCH net-next 00/15] eth: fbnic: support queue API and " Paolo Abeni
2025-08-21 14:28 ` Jakub Kicinski
2025-08-21 14:53 ` Taehee Yoo
2025-08-21 15:03 ` Jakub Kicinski
2025-08-21 15:22 ` Mina Almasry
2025-08-21 15:42 ` Jakub Kicinski
2025-08-21 15:02 ` Paolo Abeni
2025-08-21 15:20 ` patchwork-bot+netdevbpf
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=20250820025704.166248-13-kuba@kernel.org \
--to=kuba@kernel.org \
--cc=alexanderduyck@fb.com \
--cc=almasrymina@google.com \
--cc=andrew+netdev@lunn.ch \
--cc=davem@davemloft.net \
--cc=dtatulea@nvidia.com \
--cc=edumazet@google.com \
--cc=hawk@kernel.org \
--cc=horms@kernel.org \
--cc=ilias.apalodimas@linaro.org \
--cc=michael.chan@broadcom.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sdf@fomichev.me \
--cc=tariqt@nvidia.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.