Re: [BUG] OOPS 2.6.24.2 raid5 write with ioatdma

["Dan Williams" <dan.j.williams@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 334 bytes --]

On Fri, Feb 15, 2008 at 9:19 AM, Laurent CORBES
<laurent.corbes@smartjog.com> wrote:
> Hi all,
>
>  I got a raid5 oops when trying to write on a raid 5 array, with ioatdma loaded
>  and without DCA activated in bios:
>

At first glance I believe the attached patch may fix the issue, I'll
try to reproduce this locally.

Regards,
Dan

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: ioat-init-ack.patch --]
[-- Type: text/x-patch; name=ioat-init-ack.patch, Size: 861 bytes --]

ioat: fix 'ack' handling, driver must ensure that 'ack' is zero

From: Dan Williams <dan.j.williams@intel.com>

Initialize 'ack' to zero in case the descriptor has been recycled.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---

 drivers/dma/ioat_dma.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)


diff --git a/drivers/dma/ioat_dma.c b/drivers/dma/ioat_dma.c
index 45e7b46..8cf542b 100644
--- a/drivers/dma/ioat_dma.c
+++ b/drivers/dma/ioat_dma.c
@@ -726,6 +726,7 @@ static struct dma_async_tx_descriptor *ioat1_dma_prep_memcpy(
 
 	if (new) {
 		new->len = len;
+		new->async_tx.ack = 0;
 		return &new->async_tx;
 	} else
 		return NULL;
@@ -749,6 +750,7 @@ static struct dma_async_tx_descriptor *ioat2_dma_prep_memcpy(
 
 	if (new) {
 		new->len = len;
+		new->async_tx.ack = 0;
 		return &new->async_tx;
 	} else
 		return NULL;