[PATCH] mlx4/IB: Add set_4k_mtu module parameter.

[PATCH] mlx4/IB: Add set_4k_mtu module parameter.

[Vladimir Sokolovsky]

It control Infiniband link MTU for all IB ports in a host.

Signed-off-by: Vladimir Sokolovsky <vlad-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
 drivers/net/mlx4/port.c |   10 ++++++++++
 1 files changed, 10 insertions(+), 0 deletions(-)

diff --git a/drivers/net/mlx4/port.c b/drivers/net/mlx4/port.c
index 606aa58..0e9d745 100644
--- a/drivers/net/mlx4/port.c
+++ b/drivers/net/mlx4/port.c
@@ -37,6 +37,10 @@
 
 #include "mlx4.h"
 
+static int mlx4_ib_set_4k_mtu;
+module_param_named(set_4k_mtu, mlx4_ib_set_4k_mtu, int, 0444);
+MODULE_PARM_DESC(set_4k_mtu, "attempt to set 4K MTU to all ConnectX ports");
+
 #define MLX4_MAC_VALID		(1ull << 63)
 #define MLX4_MAC_MASK		0xffffffffffffULL
 
@@ -308,6 +312,12 @@ int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port)
 
 	memset(mailbox->buf, 0, 256);
 
+	if (mlx4_ib_set_4k_mtu)
+		((__be32 *) mailbox->buf)[0] |= cpu_to_be32((1 << 22) |
+							    (1 << 21) |
+							    (5 << 12) |
+							    (2 << 4));
+
 	((__be32 *) mailbox->buf)[1] = dev->caps.ib_port_def_cap[port];
 	err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
 		       MLX4_CMD_TIME_CLASS_B);
-- 
1.5.4.3

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] mlx4/IB: Add set_4k_mtu module parameter.

[Roland Dreier]

I'm not thrilled about this patch.

 > +module_param_named(set_4k_mtu, mlx4_ib_set_4k_mtu, int, 0444);
 > +MODULE_PARM_DESC(set_4k_mtu, "attempt to set 4K MTU to all ConnectX ports");

What does "attempt to set 4K MTU" mean?  Try to set the MTU cap to 4K?
What happens if it fails?  Is there any reason not to just let the MTU
cap be 4K all time?

Having this be a module parameter really seems like a bad idea for the
future....

 > +	if (mlx4_ib_set_4k_mtu)
 > +		((__be32 *) mailbox->buf)[0] |= cpu_to_be32((1 << 22) |
 > +							    (1 << 21) |
 > +							    (5 << 12) |
 > +							    (2 << 4));

Is there any way for this to write this with fewer unexplained magic numbers?

 - R.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html