From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Lord Subject: [PATCH] 2.6.xx: sata_mv: another critical fix Date: Tue, 21 Mar 2006 00:00:36 -0500 Message-ID: <200603210000.36552.lkml@rtr.ca> References: <441F4F95.4070203@garzik.org> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <441F4F95.4070203@garzik.org> Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org To: Jeff Garzik Cc: Andrew Morton , Linus Torvalds , "linux-ide@vger.kernel.org" , Linux Kernel List-Id: linux-ide@vger.kernel.org This patch addresses a number of weird behaviours observed for the sata_mv driver, by fixing an "off by one" bug in processing of the EDMA response queue. Basically, sata_mv was looking in the wrong place for command results, and this produced a lot of unpredictable behaviour. Signed-off-by: Mark Lord --- linux-2.6.16/drivers/scsi/sata_mv.c 2006-03-20 00:53:29.000000000 -0500 +++ linux/drivers/scsi/sata_mv.c 2006-03-20 23:34:32.000000000 -0500 @@ -1098,36 +1098,39 @@ * Inherited from caller. */ static u8 mv_get_crpb_status(struct ata_port *ap) { void __iomem *port_mmio = mv_ap_base(ap); struct mv_port_priv *pp = ap->private_data; u32 out_ptr; + u8 ata_status; out_ptr = readl(port_mmio + EDMA_RSP_Q_OUT_PTR_OFS); /* the response consumer index should be the same as we remember it */ assert(((out_ptr >> EDMA_RSP_Q_PTR_SHIFT) & MV_MAX_Q_DEPTH_MASK) == pp->rsp_consumer); + ata_status = pp->crpb[pp->rsp_consumer].flags >> CRPB_FLAG_STATUS_SHIFT; + /* increment our consumer index... */ pp->rsp_consumer = mv_inc_q_index(&pp->rsp_consumer); /* and, until we do NCQ, there should only be 1 CRPB waiting */ assert(((readl(port_mmio + EDMA_RSP_Q_IN_PTR_OFS) >> EDMA_RSP_Q_PTR_SHIFT) & MV_MAX_Q_DEPTH_MASK) == pp->rsp_consumer); /* write out our inc'd consumer index so EDMA knows we're caught up */ out_ptr &= EDMA_RSP_Q_BASE_LO_MASK; out_ptr |= pp->rsp_consumer << EDMA_RSP_Q_PTR_SHIFT; writelfl(out_ptr, port_mmio + EDMA_RSP_Q_OUT_PTR_OFS); /* Return ATA status register for completed CRPB */ - return (pp->crpb[pp->rsp_consumer].flags >> CRPB_FLAG_STATUS_SHIFT); + return ata_status; } /** * mv_err_intr - Handle error interrupts on the port * @ap: ATA channel to manipulate * * In most cases, just clear the interrupt and move on. However,