Re: [PATCH 1/1] aacraid: SCSI dma mapping failure case handling

[James Bottomley <jbottomley@parallels.com>]

On Tue, 2012-10-16 at 23:34 +0200, Tomas Henzl wrote:
> On 10/16/2012 10:59 PM, Mahesh Rajashekhara wrote:
> > This patch handles SCSI dma mapping failure case. Reporting error code to the upper layer instead of BUG_ON().
> >
> > This patch is created against current upstream kernel.
> >
> > Signed-off-by: Mahesh Rajashekhara <Mahesh_Rajashekhara@pmc-sierra.com>
> > ---
> >  drivers/scsi/aacraid/aachba.c  |   63 +++++++++++++++++++++++++++++++---------
> >  drivers/scsi/aacraid/aacraid.h |    2 +-
> >  2 files changed, 50 insertions(+), 15 deletions(-)
> >
> > diff --git a/drivers/scsi/aacraid/aachba.c b/drivers/scsi/aacraid/aachba.c
> > index d79457a..efa2900 100644
> > --- a/drivers/scsi/aacraid/aachba.c
> > +++ b/drivers/scsi/aacraid/aachba.c
> > @@ -971,6 +971,7 @@ static int aac_read_raw_io(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u3
> >  {
> >  	struct aac_dev *dev = fib->dev;
> >  	u16 fibsize, command;
> > +	unsigned long ret;
> >  
> >  	aac_fib_init(fib);
> >  	if (dev->comm_interface == AAC_COMM_MESSAGE_TYPE2 && !dev->sync_mode) {
> > @@ -982,7 +983,10 @@ static int aac_read_raw_io(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u3
> >  		readcmd2->byteCount = cpu_to_le32(count<<9);
> >  		readcmd2->cid = cpu_to_le16(scmd_id(cmd));
> >  		readcmd2->flags = cpu_to_le16(RIO2_IO_TYPE_READ);
> > -		aac_build_sgraw2(cmd, readcmd2, dev->scsi_host_ptr->sg_tablesize);
> > +		ret = aac_build_sgraw2(cmd, readcmd2,
> > +				dev->scsi_host_ptr->sg_tablesize);
> > +		if (ret < 0)
> > +			return ret;
> 
> Hi Mahesh, 'ret' is 'unsigned', the above test will not work.
> Tomas

Ping on this please Mahesh?  The criticism seems completely valid to me.

James