[openib-general] [PATCH] mthca: fix clr_int calculation

[Michael S. Tsirkin]

----- Forwarded message from Leonid Keller <leonid at mellanox.co.il> -----

Subject: a bug ?
Date: Wed, 21 Sep 2005 21:22:05 +0300
From: "Leonid Keller" <leonid at mellanox.co.il>

in mthca_init_eq_table() there is code:
    dev->eq_table.clr_int = dev->clr_base + (dev->eq_table.inta_pin < 31 ? 4 : 0);

In VAPI i saw:

    dev->eq_table.clr_int = dev->clr_base + (dev->eq_table.inta_pin < 32 ? 4 : 0);

It's a bug or i'm wrong somewhere ?

----- End forwarded message -----

Roland, the following makes more sense, does it not?

---

Fix clr_int calculation.

Signed-off-by: Michael S. Tsirkin <mst at mellanox.co.il>

Index: linux-kernel/drivers/infiniband/hw/mthca/mthca_eq.c
===================================================================
--- linux-kernel.orig/drivers/infiniband/hw/mthca/mthca_eq.c	2005-09-26 17:17:01.000000000 +0300
+++ linux-kernel/drivers/infiniband/hw/mthca/mthca_eq.c	2005-09-26 17:17:08.000000000 +0300
@@ -838,7 +838,7 @@ int __devinit mthca_init_eq_table(struct
 		dev->eq_table.clr_mask =
 			swab32(1 << (dev->eq_table.inta_pin & 31));
 		dev->eq_table.clr_int  = dev->clr_base +
-			(dev->eq_table.inta_pin < 31 ? 4 : 0);
+			(dev->eq_table.inta_pin < 32 ? 4 : 0);
 	}
 
 	dev->eq_table.arm_mask = 0;

-- 
MST