Fix a couple hpsa bugs related to device discovery
Subsystems (2)
cover Fix a couple hpsa bugs related to device discovery
This series fixes a couple of bugs related to discovering changes to devices. 1) We should not consider a firmware change to mean the device identity has changed. Just because a logical drive's firmware changes doesn't mean it's not the same drive with the same data, so it should not be removed and re-added to the scsi mid layer.
2) Likewise if the RAID level is observed to change. This can happen for instance by doing a RAID level migration from with the Array Configuration Utility.
---
Stephen M. Cameron (2):
hpsa: do not consider firmware revision when looking for device changes.
hpsa: do not consider RAID level to be part of device identity
drivers/scsi/hpsa.c | 11 ----------- drivers/scsi/hpsa.h | 1 - 2 files changed, 0 insertions(+), 12 deletions(-)
Signature · 1 line
-- steve
PATCH 1/2 hpsa: do not consider firmware revision when looking for device changes.
From: Stephen M. Cameron <[email protected]>
The firmware may have been updated, in which case, it's the same device, and in that case, we do not want to remove and add the device, we want to let it continue as is.
--- drivers/scsi/hpsa.c | 9 --------- drivers/scsi/hpsa.h | 1 - 2 files changed, 0 insertions(+), 10 deletions(-)
drivers/scsi/hpsa.c +0 -9
@@ -641,11 +641,6 @@ static void fixup_botched_add(struct ctlr_info *h,static inline int device_is_the_same (struct hpsa_scsi_dev_t *dev1,struct hpsa_scsi_dev_t *dev2) { -if ((is_logical_dev_addr_mode (dev1->scsi3addr) || - (dev1->lun !=-1 && dev2->lun !=-1 )) && - dev1->devtype !=0x0C ) -return (memcmp (dev1, dev2,sizeof (*dev1)) ==0 ); -/* we compare everything except lun and target as these * are not yet assigned. Compare parts likely * to differ first
@@ -660,8 +655,6 @@ static inline int device_is_the_same(struct hpsa_scsi_dev_t *dev1,return 0 ;if (memcmp (dev1->vendor, dev2->vendor,sizeof (dev1->vendor)) !=0 )return 0 ; -if (memcmp (dev1->revision, dev2->revision,sizeof (dev1->revision)) !=0 ) -return 0 ;if (dev1->devtype != dev2->devtype)return 0 ;if (dev1->raid_level != dev2->raid_level)
@@ -1477,8 +1470,6 @@ static int hpsa_update_device_info(struct ctlr_info *h, sizeof(this_device->vendor));memcpy (this_device->model, &inq_buff[16 ],sizeof (this_device->model)); -memcpy (this_device->revision, &inq_buff[32 ], -sizeof (this_device->revision));memset (this_device->device_id,0 ,sizeof (this_device->device_id)); hpsa_get_device_id(h, scsi3addr, this_device->device_id,
drivers/scsi/hpsa.h +0 -1
@@ -45,7 +45,6 @@ struct hpsa_scsi_dev_t {unsigned char device_id[16 ];/* from inquiry pg. 0x83 */ unsigned char vendor[8 ];/* bytes 8-15 of inquiry data */ unsigned char model[16 ];/* bytes 16-31 of inquiry data */ -unsigned char revision[4 ];/* bytes 32-35 of inquiry data */ unsigned char raid_level;/* from inquiry page 0xC1 */ };
PATCH 2/2 hpsa: do not consider RAID level to be part of device identity
In reply to Stephen M. CameronFrom: Stephen M. Cameron <StephenM.Cameron>
Otherwise, after doing a RAID level migration, the disk will be disruptively removed and re-added as a different disk on rescan.
--- drivers/scsi/hpsa.c | 2 -- 1 files changed, 0 insertions(+), 2 deletions(-)
drivers/scsi/hpsa.c +0 -2
@@ -657,8 +657,6 @@ static inline int device_is_the_same(struct hpsa_scsi_dev_t *dev1,return 0 ;if (dev1->devtype != dev2->devtype)return 0 ; -if (dev1->raid_level != dev2->raid_level) -return 0 ;if (dev1->bus != dev2->bus)return 0 ;return 1 ;