[PATCH 2.6.21 review I] [22/25] x86_64: use memcpy_uncached_read() in RDMA interrupt handler to reduce packet loss

[Andi Kleen <ak@suse.de>]

From: "Bryan O'Sullivan" <bos@pathscale.com>

In cases where a large incoming RDMA is being received, we have to copy data
inside the interrupt handler before we can ACK each packet.  The source is
DMAed to by the hardware, which means that the CPU won't have it cached.  We
only read the source this one time; using normal load instructions pollutes
the dcache with useless data, reducing performance to the point where we can
lose a significant number of packets.

We use memcpy_uncached_read to try to not fill the dcache with useless data. 
Avoiding the cache refill penalty lets us keep up better with the sender,
resulting in many fewer dropped packets.

Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Andi Kleen <ak@suse.de>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---

 drivers/infiniband/hw/ipath/ipath_verbs.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux/drivers/infiniband/hw/ipath/ipath_verbs.c
===================================================================
--- linux.orig/drivers/infiniband/hw/ipath/ipath_verbs.c
+++ linux/drivers/infiniband/hw/ipath/ipath_verbs.c
@@ -167,7 +167,7 @@ void ipath_copy_sge(struct ipath_sge_sta
 		BUG_ON(len == 0);
 		if (len > length)
 			len = length;
-		memcpy(sge->vaddr, data, len);
+		memcpy_uncached_read(sge->vaddr, data, len);
 		sge->vaddr += len;
 		sge->length -= len;
 		sge->sge_length -= len;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/