diff options
| author | Hannes Reinecke <hare@suse.de> | 2006-03-08 13:00:07 +0100 |
|---|---|---|
| committer | James Bottomley <jejb@mulgrave.il.steeleye.com> | 2006-03-12 09:06:57 -0600 |
| commit | 76744f429a178ad2df06b722e5e7bbbf1b5245ee (patch) | |
| tree | 92406bc62b0316803746c57e2e461721bc454578 /drivers/scsi/aic7xxx | |
| parent | 986a25f96accc8272059ae8b3eb63055f58cadb0 (diff) | |
| download | kernel_samsung_smdk4412-76744f429a178ad2df06b722e5e7bbbf1b5245ee.zip kernel_samsung_smdk4412-76744f429a178ad2df06b722e5e7bbbf1b5245ee.tar.gz kernel_samsung_smdk4412-76744f429a178ad2df06b722e5e7bbbf1b5245ee.tar.bz2 | |
[SCSI] aic79xx: Avoid renegotiation on inquiry
Because of some quirk in the SCSI spec the aic79xx driver chose to
force a renegotiation when sending an inquiry. This should better
be handled by the upper layers if required at all.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Diffstat (limited to 'drivers/scsi/aic7xxx')
| -rw-r--r-- | drivers/scsi/aic7xxx/aic79xx_osm.c | 24 |
1 files changed, 0 insertions, 24 deletions
diff --git a/drivers/scsi/aic7xxx/aic79xx_osm.c b/drivers/scsi/aic7xxx/aic79xx_osm.c index 0fbd1ec..bcced0a 100644 --- a/drivers/scsi/aic7xxx/aic79xx_osm.c +++ b/drivers/scsi/aic7xxx/aic79xx_osm.c @@ -1526,30 +1526,6 @@ ahd_linux_run_command(struct ahd_softc *ahd, struct ahd_linux_device *dev, if ((tstate->auto_negotiate & mask) != 0) { scb->flags |= SCB_AUTO_NEGOTIATE; scb->hscb->control |= MK_MESSAGE; - } else if (cmd->cmnd[0] == INQUIRY - && (tinfo->curr.offset != 0 - || tinfo->curr.width != MSG_EXT_WDTR_BUS_8_BIT - || tinfo->curr.ppr_options != 0) - && (tinfo->curr.ppr_options & MSG_EXT_PPR_IU_REQ)==0) { - /* - * The SCSI spec requires inquiry - * commands to complete without - * reporting unit attention conditions. - * Because of this, an inquiry command - * that occurs just after a device is - * reset will result in a data phase - * with mismatched negotiated rates. - * The core already forces a renegotiation - * for reset events that are visible to - * our controller or that we initiate, - * but a third party device reset or a - * hot-plug insertion can still cause this - * issue. Therefore, we force a re-negotiation - * for every inquiry command unless we - * are async. - */ - scb->flags |= SCB_NEGOTIATE; - scb->hscb->control |= MK_MESSAGE; } if ((dev->flags & (AHD_DEV_Q_TAGGED|AHD_DEV_Q_BASIC)) != 0) { |