LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [net-next PATCH] octeontx2-af: cn10k: Set cache lines for NPA batch alloc
@ 2021-08-25  5:35 Geetha sowjanya
  2021-08-25 11:10 ` patchwork-bot+netdevbpf
  0 siblings, 1 reply; 2+ messages in thread
From: Geetha sowjanya @ 2021-08-25  5:35 UTC (permalink / raw)
  To: netdev, linux-kernel
  Cc: kuba, davem, sgoutham, lcherian, gakula, jerinj, sbhatta, hkelam

Set NPA batch allocation engine to process 35 cache lines
per turn on CN10k platform.

Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
---
 drivers/net/ethernet/marvell/octeontx2/af/mbox.h    |  1 +
 drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c | 11 +++++++++++
 drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h |  1 +
 3 files changed, 13 insertions(+)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
index add4a39edced..d8f5e61f0304 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
@@ -594,6 +594,7 @@ struct npa_lf_alloc_rsp {
 	u32 stack_pg_ptrs;  /* No of ptrs per stack page */
 	u32 stack_pg_bytes; /* Size of stack page */
 	u16 qints; /* NPA_AF_CONST::QINTS */
+	u8 cache_lines; /*BATCH ALLOC DMA */
 };
 
 /* NPA AQ enqueue msg */
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
index 24c2bfdfec4e..f046f2e4256a 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
@@ -419,6 +419,10 @@ int rvu_mbox_handler_npa_lf_alloc(struct rvu *rvu,
 	rsp->stack_pg_ptrs = (cfg >> 8) & 0xFF;
 	rsp->stack_pg_bytes = cfg & 0xFF;
 	rsp->qints = (cfg >> 28) & 0xFFF;
+	if (!is_rvu_otx2(rvu)) {
+		cfg = rvu_read64(rvu, block->addr, NPA_AF_BATCH_CTL);
+		rsp->cache_lines = (cfg >> 1) & 0x3F;
+	}
 	return rc;
 }
 
@@ -478,6 +482,13 @@ static int npa_aq_init(struct rvu *rvu, struct rvu_block *block)
 #endif
 	rvu_write64(rvu, block->addr, NPA_AF_NDC_CFG, cfg);
 
+	/* For CN10K NPA BATCH DMA set 35 cache lines */
+	if (!is_rvu_otx2(rvu)) {
+		cfg = rvu_read64(rvu, block->addr, NPA_AF_BATCH_CTL);
+		cfg &= ~0x7EULL;
+		cfg |= BIT_ULL(6) | BIT_ULL(2) | BIT_ULL(1);
+		rvu_write64(rvu, block->addr, NPA_AF_BATCH_CTL, cfg);
+	}
 	/* Result structure can be followed by Aura/Pool context at
 	 * RES + 128bytes and a write mask at RES + 256 bytes, depending on
 	 * operation type. Alloc sufficient result memory for all operations.
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
index 960ee1c2e178..4600c31b336b 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h
@@ -156,6 +156,7 @@
 #define NPA_AF_AQ_DONE_INT_W1S          (0x0688)
 #define NPA_AF_AQ_DONE_ENA_W1S          (0x0690)
 #define NPA_AF_AQ_DONE_ENA_W1C          (0x0698)
+#define NPA_AF_BATCH_CTL		(0x06a0)
 #define NPA_AF_LFX_AURAS_CFG(a)         (0x4000 | (a) << 18)
 #define NPA_AF_LFX_LOC_AURAS_BASE(a)    (0x4010 | (a) << 18)
 #define NPA_AF_LFX_QINTS_CFG(a)         (0x4100 | (a) << 18)
-- 
2.17.1


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2021-08-25 11:10 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-25  5:35 [net-next PATCH] octeontx2-af: cn10k: Set cache lines for NPA batch alloc Geetha sowjanya
2021-08-25 11:10 ` patchwork-bot+netdevbpf

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).