DPDK patches and discussions
 help / color / mirror / Atom feed
From: Andrew Rybchenko <arybchenko@solarflare.com>
To: <dev@dpdk.org>
Subject: [dpdk-dev] [PATCH 06/23] net/sfc: factor out function to push Rx doorbell
Date: Thu, 19 Apr 2018 12:36:49 +0100	[thread overview]
Message-ID: <1524137826-5675-7-git-send-email-arybchenko@solarflare.com> (raw)
In-Reply-To: <1524137826-5675-1-git-send-email-arybchenko@solarflare.com>

The function may be shared by different Rx datapath implementations.

Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Reviewed-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Reviewed-by: Andy Moreton <amoreton@solarflare.com>
---
 drivers/net/sfc/sfc_ef10.h    | 31 +++++++++++++++++++++++++++++++
 drivers/net/sfc/sfc_ef10_rx.c | 33 +++------------------------------
 2 files changed, 34 insertions(+), 30 deletions(-)

diff --git a/drivers/net/sfc/sfc_ef10.h b/drivers/net/sfc/sfc_ef10.h
index ace6a1d..865359f 100644
--- a/drivers/net/sfc/sfc_ef10.h
+++ b/drivers/net/sfc/sfc_ef10.h
@@ -79,6 +79,37 @@ sfc_ef10_ev_present(const efx_qword_t ev)
 	       ~EFX_QWORD_FIELD(ev, EFX_DWORD_1);
 }
 
+
+/**
+ * Alignment requirement for value written to RX WPTR:
+ * the WPTR must be aligned to an 8 descriptor boundary.
+ */
+#define SFC_EF10_RX_WPTR_ALIGN	8u
+
+static inline void
+sfc_ef10_rx_qpush(volatile void *doorbell, unsigned int added,
+		  unsigned int ptr_mask)
+{
+	efx_dword_t dword;
+
+	/* Hardware has alignment restriction for WPTR */
+	RTE_BUILD_BUG_ON(SFC_RX_REFILL_BULK % SFC_EF10_RX_WPTR_ALIGN != 0);
+	SFC_ASSERT(RTE_ALIGN(added, SFC_EF10_RX_WPTR_ALIGN) == added);
+
+	EFX_POPULATE_DWORD_1(dword, ERF_DZ_RX_DESC_WPTR, added & ptr_mask);
+
+	/* DMA sync to device is not required */
+
+	/*
+	 * rte_write32() has rte_io_wmb() which guarantees that the STORE
+	 * operations (i.e. Rx and event descriptor updates) that precede
+	 * the rte_io_wmb() call are visible to NIC before the STORE
+	 * operations that follow it (i.e. doorbell write).
+	 */
+	rte_write32(dword.ed_u32[0], doorbell);
+}
+
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/drivers/net/sfc/sfc_ef10_rx.c b/drivers/net/sfc/sfc_ef10_rx.c
index 7d6b64e..92e1ef0 100644
--- a/drivers/net/sfc/sfc_ef10_rx.c
+++ b/drivers/net/sfc/sfc_ef10_rx.c
@@ -30,12 +30,6 @@
 	SFC_DP_LOG(SFC_KVARG_DATAPATH_EF10, ERR, dpq, __VA_ARGS__)
 
 /**
- * Alignment requirement for value written to RX WPTR:
- * the WPTR must be aligned to an 8 descriptor boundary.
- */
-#define SFC_EF10_RX_WPTR_ALIGN	8
-
-/**
  * Maximum number of descriptors/buffers in the Rx ring.
  * It should guarantee that corresponding event queue never overfill.
  * EF10 native datapath uses event queue of the same size as Rx queue.
@@ -88,29 +82,6 @@ sfc_ef10_rxq_by_dp_rxq(struct sfc_dp_rxq *dp_rxq)
 }
 
 static void
-sfc_ef10_rx_qpush(struct sfc_ef10_rxq *rxq)
-{
-	efx_dword_t dword;
-
-	/* Hardware has alignment restriction for WPTR */
-	RTE_BUILD_BUG_ON(SFC_RX_REFILL_BULK % SFC_EF10_RX_WPTR_ALIGN != 0);
-	SFC_ASSERT(RTE_ALIGN(rxq->added, SFC_EF10_RX_WPTR_ALIGN) == rxq->added);
-
-	EFX_POPULATE_DWORD_1(dword, ERF_DZ_RX_DESC_WPTR,
-			     rxq->added & rxq->ptr_mask);
-
-	/* DMA sync to device is not required */
-
-	/*
-	 * rte_write32() has rte_io_wmb() which guarantees that the STORE
-	 * operations (i.e. Rx and event descriptor updates) that precede
-	 * the rte_io_wmb() call are visible to NIC before the STORE
-	 * operations that follow it (i.e. doorbell write).
-	 */
-	rte_write32(dword.ed_u32[0], rxq->doorbell);
-}
-
-static void
 sfc_ef10_rx_qrefill(struct sfc_ef10_rxq *rxq)
 {
 	const unsigned int ptr_mask = rxq->ptr_mask;
@@ -120,6 +91,8 @@ sfc_ef10_rx_qrefill(struct sfc_ef10_rxq *rxq)
 	void *objs[SFC_RX_REFILL_BULK];
 	unsigned int added = rxq->added;
 
+	RTE_BUILD_BUG_ON(SFC_RX_REFILL_BULK % SFC_EF10_RX_WPTR_ALIGN != 0);
+
 	free_space = rxq->max_fill_level - (added - rxq->completed);
 
 	if (free_space < rxq->refill_threshold)
@@ -178,7 +151,7 @@ sfc_ef10_rx_qrefill(struct sfc_ef10_rxq *rxq)
 
 	SFC_ASSERT(rxq->added != added);
 	rxq->added = added;
-	sfc_ef10_rx_qpush(rxq);
+	sfc_ef10_rx_qpush(rxq->doorbell, added, ptr_mask);
 }
 
 static void
-- 
2.7.4

  parent reply	other threads:[~2018-04-19 11:37 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-04-19 11:36 [dpdk-dev] [PATCH 00/23] net/sfc: support equal stride super-buffer Rx mode Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 01/23] net/sfc/base: update autogenerated MCDI and TLV headers Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 02/23] net/sfc/base: make RxQ type data an union Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 03/23] net/sfc/base: detect equal stride super-buffer support Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 04/23] net/sfc/base: support equal stride super-buffer Rx mode Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 05/23] net/sfc/base: add equal stride super-buffer prefix layout Andrew Rybchenko
2018-04-19 11:36 ` Andrew Rybchenko [this message]
2018-04-19 11:36 ` [dpdk-dev] [PATCH 07/23] net/sfc: prepare EF10 Rx event parser to be reused Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 08/23] net/sfc: move EF10 Rx event parser to shared header Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 09/23] net/sfc: conditionally compile support for tunnel packets Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 10/23] net/sfc: allow one Rx queue entry carry many packet buffers Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 11/23] net/sfc: allow to take mbuf pool into account when sizing Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 12/23] net/sfc: support equal stride super-buffer Rx mode Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 13/23] net/sfc: support callback to check if mempool is supported Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 14/23] net/sfc: check mempool when equal stride super-buffer used Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 15/23] net/sfc: support DPDK firmware variant Andrew Rybchenko
2018-04-19 11:36 ` [dpdk-dev] [PATCH 16/23] net/sfc: add Rx descriptor wait timeout Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 17/23] net/sfc: support flow marks in equal stride super-buffer Rx Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 18/23] net/sfc/base: get actions MARK and FLAG support Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 19/23] net/sfc/base: support MARK and FLAG actions in filters Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 20/23] net/sfc/base: get max supported value for action MARK Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 21/23] net/sfc: make processing of flow rule actions more uniform Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 22/23] net/sfc: support MARK and FLAG actions in flow API Andrew Rybchenko
2018-04-19 11:37 ` [dpdk-dev] [PATCH 23/23] doc: advertise equal stride super-buffer Rx mode support in net/sfc Andrew Rybchenko
2018-04-26 22:47 ` [dpdk-dev] [PATCH 00/23] net/sfc: support equal stride super-buffer Rx mode Ferruh Yigit

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=1524137826-5675-7-git-send-email-arybchenko@solarflare.com \
    --to=arybchenko@solarflare.com \
    --cc=dev@dpdk.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;
as well as URLs for NNTP newsgroup(s).