IB/mlx4: Add support for masked atomic operations

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

From: Linux Kernel Mailing List

Subject: IB/mlx4: Add support for masked atomic operations

Date: Thursday, May 20, 2010 - 10:59 am

Gitweb:     http://git.kernel.org/linus/6fa8f719844b8455033e295f720e739c1dc3804a
Commit:     6fa8f719844b8455033e295f720e739c1dc3804a
Parent:     5e80ba8ff0bd33ff4af2365969a231cbdb98cafb
Author:     Vladimir Sokolovsky <vlad@mellanox.co.il>
AuthorDate: Wed Apr 14 17:23:39 2010 +0300
Committer:  Roland Dreier <rolandd@cisco.com>
CommitDate: Wed Apr 21 16:37:49 2010 -0700

    IB/mlx4: Add support for masked atomic operations
    
    Add support for masked atomic operations (masked compare and swap,
    masked fetch and add).
    
    Signed-off-by: Vladimir Sokolovsky <vlad@mellanox.co.il>
    Signed-off-by: Roland Dreier <rolandd@cisco.com>
---
 drivers/infiniband/hw/mlx4/cq.c   |    8 ++++++
 drivers/infiniband/hw/mlx4/main.c |    1 +
 drivers/infiniband/hw/mlx4/qp.c   |   50 ++++++++++++++++++++++++++++--------
 include/linux/mlx4/device.h       |    4 +-
 include/linux/mlx4/qp.h           |    7 +++++
 5 files changed, 57 insertions(+), 13 deletions(-)

diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index cc2ddd2..5a219a2 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -661,6 +661,14 @@ repoll:
 			wc->opcode    = IB_WC_FETCH_ADD;
 			wc->byte_len  = 8;
 			break;
+		case MLX4_OPCODE_MASKED_ATOMIC_CS:
+			wc->opcode    = IB_WC_MASKED_COMP_SWAP;
+			wc->byte_len  = 8;
+			break;
+		case MLX4_OPCODE_MASKED_ATOMIC_FA:
+			wc->opcode    = IB_WC_MASKED_FETCH_ADD;
+			wc->byte_len  = 8;
+			break;
 		case MLX4_OPCODE_BIND_MW:
 			wc->opcode    = IB_WC_BIND_MW;
 			break;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 01f2a3f..3905141 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -139,6 +139,7 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 	props->local_ca_ack_delay  = dev->dev->caps.local_ca_ack_delay;
 	props->atomic_cap	   = dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_ATOMIC ?
 		IB_ATOMIC_HCA : IB_ATOMIC_NONE;
+	props->masked_atomic_cap   = IB_ATOMIC_HCA;
 	props->max_pkeys	   = dev->dev->caps.pkey_table_len[1];
 	props->max_mcast_grp	   = dev->dev->caps.num_mgms + dev->dev->caps.num_amgms;
 	props->max_mcast_qp_attach = dev->dev->caps.num_qp_per_mgm;
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 5643f4a..6a60827 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -74,17 +74,19 @@ enum {
 };
 
 static const __be32 mlx4_ib_opcode[] = {
-	[IB_WR_SEND]			= cpu_to_be32(MLX4_OPCODE_SEND),
-	[IB_WR_LSO]			= cpu_to_be32(MLX4_OPCODE_LSO),
-	[IB_WR_SEND_WITH_IMM]		= cpu_to_be32(MLX4_OPCODE_SEND_IMM),
-	[IB_WR_RDMA_WRITE]		= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE),
-	[IB_WR_RDMA_WRITE_WITH_IMM]	= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM),
-	[IB_WR_RDMA_READ]		= cpu_to_be32(MLX4_OPCODE_RDMA_READ),
-	[IB_WR_ATOMIC_CMP_AND_SWP]	= cpu_to_be32(MLX4_OPCODE_ATOMIC_CS),
-	[IB_WR_ATOMIC_FETCH_AND_ADD]	= cpu_to_be32(MLX4_OPCODE_ATOMIC_FA),
-	[IB_WR_SEND_WITH_INV]		= cpu_to_be32(MLX4_OPCODE_SEND_INVAL),
-	[IB_WR_LOCAL_INV]		= cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL),
-	[IB_WR_FAST_REG_MR]		= cpu_to_be32(MLX4_OPCODE_FMR),
+	[IB_WR_SEND]				= cpu_to_be32(MLX4_OPCODE_SEND),
+	[IB_WR_LSO]				= cpu_to_be32(MLX4_OPCODE_LSO),
+	[IB_WR_SEND_WITH_IMM]			= cpu_to_be32(MLX4_OPCODE_SEND_IMM),
+	[IB_WR_RDMA_WRITE]			= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE),
+	[IB_WR_RDMA_WRITE_WITH_IMM]		= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM),
+	[IB_WR_RDMA_READ]			= cpu_to_be32(MLX4_OPCODE_RDMA_READ),
+	[IB_WR_ATOMIC_CMP_AND_SWP]		= cpu_to_be32(MLX4_OPCODE_ATOMIC_CS),
+	[IB_WR_ATOMIC_FETCH_AND_ADD]		= cpu_to_be32(MLX4_OPCODE_ATOMIC_FA),
+	[IB_WR_SEND_WITH_INV]			= cpu_to_be32(MLX4_OPCODE_SEND_INVAL),
+	[IB_WR_LOCAL_INV]			= cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL),
+	[IB_WR_FAST_REG_MR]			= cpu_to_be32(MLX4_OPCODE_FMR),
+	[IB_WR_MASKED_ATOMIC_CMP_AND_SWP]	= cpu_to_be32(MLX4_OPCODE_MASKED_ATOMIC_CS),
+	[IB_WR_MASKED_ATOMIC_FETCH_AND_ADD]	= cpu_to_be32(MLX4_OPCODE_MASKED_ATOMIC_FA),
 };
 
 static struct mlx4_ib_sqp *to_msqp(struct mlx4_ib_qp *mqp)
@@ -1407,6 +1409,9 @@ static void set_atomic_seg(struct mlx4_wqe_atomic_seg *aseg, struct ib_send_wr *
 	if (wr->opcode == IB_WR_ATOMIC_CMP_AND_SWP) {
 		aseg->swap_add = cpu_to_be64(wr->wr.atomic.swap);
 		aseg->compare  = cpu_to_be64(wr->wr.atomic.compare_add);
+	} else if (wr->opcode == IB_WR_MASKED_ATOMIC_FETCH_AND_ADD) {
+		aseg->swap_add = cpu_to_be64(wr->wr.atomic.compare_add);
+		aseg->compare  = cpu_to_be64(wr->wr.atomic.compare_add_mask);
 	} else {
 		aseg->swap_add = cpu_to_be64(wr->wr.atomic.compare_add);
 		aseg->compare  = 0;
@@ -1414,6 +1419,15 @@ static void set_atomic_seg(struct mlx4_wqe_atomic_seg *aseg, struct ib_send_wr *
 
 }
 
+static void set_masked_atomic_seg(struct mlx4_wqe_masked_atomic_seg *aseg,
+				  struct ib_send_wr *wr)
+{
+	aseg->swap_add		= cpu_to_be64(wr->wr.atomic.swap);
+	aseg->swap_add_mask	= cpu_to_be64(wr->wr.atomic.swap_mask);
+	aseg->compare		= cpu_to_be64(wr->wr.atomic.compare_add);
+	aseg->compare_mask	= cpu_to_be64(wr->wr.atomic.compare_add_mask);
+}
+
 static void set_datagram_seg(struct mlx4_wqe_datagram_seg *dseg,
 			     struct ib_send_wr *wr)
 {
@@ -1567,6 +1581,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			switch (wr->opcode) {
 			case IB_WR_ATOMIC_CMP_AND_SWP:
 			case IB_WR_ATOMIC_FETCH_AND_ADD:
+			case IB_WR_MASKED_ATOMIC_FETCH_AND_ADD:
 				set_raddr_seg(wqe, wr->wr.atomic.remote_addr,
 					      wr->wr.atomic.rkey);
 				wqe  += sizeof (struct mlx4_wqe_raddr_seg);
@@ -1579,6 +1594,19 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 
 				break;
 
+			case IB_WR_MASKED_ATOMIC_CMP_AND_SWP:
+				set_raddr_seg(wqe, wr->wr.atomic.remote_addr,
+					      wr->wr.atomic.rkey);
+				wqe  += sizeof (struct mlx4_wqe_raddr_seg);
+
+				set_masked_atomic_seg(wqe, wr);
+				wqe  += sizeof (struct mlx4_wqe_masked_atomic_seg);
+
+				size += (sizeof (struct mlx4_wqe_raddr_seg) +
+					 sizeof (struct mlx4_wqe_masked_atomic_seg)) / 16;
+
+				break;
+
 			case IB_WR_RDMA_READ:
 			case IB_WR_RDMA_WRITE:
 			case IB_WR_RDMA_WRITE_WITH_IMM:
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index e92d1bf..7a7f9c1 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -123,8 +123,8 @@ enum {
 	MLX4_OPCODE_RDMA_READ		= 0x10,
 	MLX4_OPCODE_ATOMIC_CS		= 0x11,
 	MLX4_OPCODE_ATOMIC_FA		= 0x12,
-	MLX4_OPCODE_ATOMIC_MASK_CS	= 0x14,
-	MLX4_OPCODE_ATOMIC_MASK_FA	= 0x15,
+	MLX4_OPCODE_MASKED_ATOMIC_CS	= 0x14,
+	MLX4_OPCODE_MASKED_ATOMIC_FA	= 0x15,
 	MLX4_OPCODE_BIND_MW		= 0x18,
 	MLX4_OPCODE_FMR			= 0x19,
 	MLX4_OPCODE_LOCAL_INVAL		= 0x1b,
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 9f29d86..7abe643 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -285,6 +285,13 @@ struct mlx4_wqe_atomic_seg {
 	__be64			compare;
 };
 
+struct mlx4_wqe_masked_atomic_seg {
+	__be64			swap_add;
+	__be64			compare;
+	__be64			swap_add_mask;
+	__be64			compare_mask;
+};
+
 struct mlx4_wqe_data_seg {
 	__be32			byte_count;
 	__be32			lkey;
--
To unsubscribe from this list: send the line "unsubscribe git-commits-head" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:

IB/mlx4: Add support for masked atomic operations, Linux Kernel Mailing ..., (Thu May 20, 10:59 am)


linux-kernel:
Mel Gorman	Re: [PATCH 1/4] vmstat: remove zone->lock from walk_zones_in_node
Guenter Roeck	Re: [lm-sensors] Location for thermal drivers
David Woodhouse	Re: RFC: Moving firmware blobs out of the kernel.
Siddha, Suresh B	Re: [PATCH 2.6.21 review I] [11/25] x86: default to physical mode on hotplug CPU k...
Peter Zijlstra	Re: [patch 4/6] mm: merge populate and nopage into fault (fixes nonlinear)
git-commits-head:
Linux Kernel Mailing List	[MIPS] Fix potential latency problem due to non-atomic cpu_wait.
Linux Kernel Mailing List	USB: rename USB_SPEED_VARIABLE to USB_SPEED_WIRELESS
Linux Kernel Mailing List	lib/vsprintf.c: fix bug omitting minus sign of numbers (module_param)
Linux Kernel Mailing List	[Bluetooth] Initiate authentication during connection establishment
Linux Kernel Mailing List	[POWERPC] 4xx: Add ppc40x_defconfig
linux-netdev:
MERCEDES	Your mail id has won 950,000.00 in the MERCEDES Benz Online Promo.for claims send:
David Miller	Re: [PATCH] xen/netfront: do not mark packets of length < MSS as GSO
David Miller	Re: skb_segment() questions
Shan Wei	[RFC PATCH net-next 2/5]IPv6:netfilter: Send an ICMPv6 "Fragment Reassembly Timeou...
Stanislaw Gruszka	[PATCH 1/4] bnx2x: use smp_mb() to keep ordering of read write operations
git:
Nicolas Sebrecht	git-svn died of signal 11 (was "3 failures on test t9100 (svn)")
Junio C Hamano	Re: [PATCH 2/2] Add url.<base>.pushInsteadOf: URL rewriting for push only
Martin Langhoff	Re: [PATCH] GIT commit statistics.
Alexandre Julliard	[PATCH] gitweb: Put back shortlog instead of graphiclog in the project list.
Josh Triplett	[PATCH 2/2] Add url.<base>.pushInsteadOf: URL rewriting for push only
openbsd-misc:
Taisto Qvist XX	Re: AMD GEODE LX-800 just works with kernel from install42.iso and kernelpanics wi...
Nico Meijer	Re: gOS Develop Kit with VIA pc-1 Processor Platform VIA C7-D
Andreas Bihlmaier	Re: jetway board sensors (Fintek F71805F)
admin	Drive a 2009 car from R799p/m
Antti Harri	Re: how to create a sha256 hash