hpsa: disable volume status check for older controller

Commit Message

Hannes Reinecke Aug. 4, 2017, 4:40 p.m. UTC

On 08/04/2017 06:36 PM, Hannes Reinecke wrote:
> On 08/04/2017 06:26 PM, Meelis Roos wrote:
>>> Older Controller might not support volume status, so disable
>>> it if 'hpsa_allow_any' is present.
>>
>> It makes my DL380G4 boot with hpsa.hpsa_allow_any=1 and it seems to 
>> work. Howevere, there are multiole verbose dmesg entries about failing 
>> things:
>>
> Hehe. How cool is that?
> 
>>     0.267224] HP HPSA Driver (v 3.4.18-0)
>> [    0.267351] hpsa 0000:04:03.0: unsupported board ID: 0x40910e11
>> [    0.267494] hpsa 0000:04:03.0: unsupported board ID: 0x40910e11
>> [    0.267695] hpsa 0000:04:03.0: PCI IRQ 51 -> rerouted to legacy IRQ 19
>> [    0.267889] hpsa 0000:04:03.0: Physical aborts not supported
>> [    0.268036] hpsa 0000:04:03.0: Logical aborts not supported
>> [    0.268163] hpsa 0000:04:03.0: HP SSD Smart Path aborts not supported
>> [    0.448240] scsi host0: hpsa
>> [    0.452228] hpsa 0000:04:03.0: report luns requested format 2, got 0
>> [    0.453950] hpsa 0000:04:03.0: hpsa0: hpsa_update_device_info: can't get device id for host 0:C0:T0:L0       Direct-Access           MAN3735MC       
>> [    0.454312] hpsa 0000:04:03.0: SCSI status: LUN:0000000000000000 CDB:26000000000015000801000000000000
>> [    0.454525] hpsa 0000:04:03.0: SCSI Status = 02, Sense key = 0x05, ASC = 0x3a, ASCQ = 0x00
> Curious. CDB 26 is a vendor specific command, and the controller
> responds with 'Medium not present'.
> Will be checking what's happening here.
> 
>> Among these, the following show up red in dmesg:
>>
>> report luns requested format 2, got 0
>> hpsa0: hpsa_update_device_info: can't get device id for host 0:C0:T0:L0       Direct-Access           MAN3735MC
>> hpsa0: hpsa_update_device_info: can't get device id for host 0:C0:T-1:L-1     Direct-Access           MAP3735NC
>> report luns requested format 2, got 0
>>
> Hmm. Looking into it.
> 
> But thanks for testing!
> 
You should be able to suppress the "can't get device id' messages with:

%d:C0:T%d:L%d\t%s\t%.16s\n",
                        h->ctlr, __func__,

Cheers,

Hannes

Comments

Meelis Roos Aug. 4, 2017, 7:58 p.m. UTC | #1

> You should be able to suppress the "can't get device id' messages with:

Yes, these messages are gone and it still works.

Patch

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index 2cf6cccf067f..56fd654c1ecd 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -3813,7 +3813,7 @@  static int hpsa_update_device_info(struct
ctlr_info *h,
        memset(this_device->device_id, 0,
                sizeof(this_device->device_id));
        if (hpsa_get_device_id(h, scsi3addr, this_device->device_id, 8,
-               sizeof(this_device->device_id)))
+               sizeof(this_device->device_id)) < 0)
                dev_err(&h->pdev->dev,
                        "hpsa%d: %s: can't get device id for host