From: Tejun Heo <htejun@gmail.com>
Date: Wed, 23 May 2007 09:22:15 +0000 (+0200)
Subject: libata: don't consider 0xff as port empty if SStatus is available
X-Git-Tag: v2.6.22-rc3~10^2~2
X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=fd7fe701612e42fb8780d7bf61fbb0467a488c9b;p=~shefty%2Frdma-dev.git

libata: don't consider 0xff as port empty if SStatus is available

Some SATA controllers (sata_sil) use 0xff to indicate port not ready
status, not port empty.  As libata interprets 0xff as port empty, this
causes unnecessary reset failure and retry.  Don't consider 0xff as
port empty if SStatus is available and indicates that port is online.

Signed-off-by: tejun Heo <htejun@gmail.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Indan Zupancic <indan@nul.nu>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
---

diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index a6de57e0204..5788a9dd044 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -3022,7 +3022,7 @@ int ata_wait_ready(struct ata_port *ap, unsigned long deadline)
 
 		if (!(status & ATA_BUSY))
 			return 0;
-		if (status == 0xff)
+		if (!ata_port_online(ap) && status == 0xff)
 			return -ENODEV;
 		if (time_after(now, deadline))
 			return -EBUSY;