Qlogic (isp driver) gone mad on 6.0

Eric Anderson anderson at centtech.com
Sun Aug 21 02:51:14 GMT 2005


Matthew Jacob wrote:
> I surely don't understand why it's now not working for you, and no,
> it's not acceptable for 6.0.
> 
> I will be away this weekend but on Monday I certainly will take this
> more seriously.

Thanks for checking into this..

Eric



> On 8/19/05, Eric Anderson <anderson at centtech.com> wrote:
> 
>>Matthew Jacob wrote:
>>
>>>Sounds more like a problem with the target. Make the limit 255- not 256.
>>
>>No.  The problem is not in the target device at all.  I reverted to
>>5-STABLE and it works flawlessly again.  There's a bug in 6.0's
>>scsi/qlogic code.   Setting it to 255 will end up with the same output
>>as below, except s/256/255/g.
>>
>>This bug will prevent those using qlogic cards from using 6.0 - is that
>>acceptable for 6.0?
>>
>>Eric
>>
>>
>>
>>
>>
>>>>Well, the system boots faster as it doesn't have to probe 16k devices,
>>>>but there's still one problem - it sees the correct 3 luns per device,
>>>>but then adds a 'fake' device at lun 256 which is essentially the same
>>>>as the first lun.  Here's some more info:
>>>>
>>>># camcontrol devlist
>>>><PE/PV 1x3 SCSI BP 1.1>            at scbus0 target 6 lun 0 (pass0,ses0)
>>>><JetStor snappod1-vol1 R001>       at scbus2 target 0 lun 0 (pass1,da0)
>>>><JetStor snappod1-vol2 R001>       at scbus2 target 0 lun 1 (pass2,da1)
>>>><JetStor snappod1-vol3 R001>       at scbus2 target 0 lun 2 (pass3,da2)
>>>><JetStor snappod1-vol1 R001>       at scbus2 target 0 lun 256 (pass4,da3)
>>>><JetStor snappod2-vol1 R001>       at scbus2 target 2 lun 0 (pass5,da4)
>>>><JetStor snappod2-vol2 R001>       at scbus2 target 2 lun 1 (pass6,da5)
>>>><JetStor snappod2-vol3 R001>       at scbus2 target 2 lun 2 (pass7,da6)
>>>><JetStor snappod2-vol1 R001>       at scbus2 target 2 lun 256 (pass8,da7)
>>>><JetStor snappod3-vol1 R001>       at scbus2 target 3 lun 0 (pass9,da8)
>>>><JetStor snappod3-vol2 R001>       at scbus2 target 3 lun 1 (pass10,da9)
>>>><JetStor snappod3-vol3 R001>       at scbus2 target 3 lun 2 (pass11,da10)
>>>><JetStor snappod3-vol1 R001>       at scbus2 target 3 lun 256 (pass12,da11)
>>>><JetStor snappod1-vol1 R001>       at scbus3 target 0 lun 0 (pass13,da12)
>>>><JetStor snappod1-vol2 R001>       at scbus3 target 0 lun 1 (pass14,da13)
>>>><JetStor snappod1-vol3 R001>       at scbus3 target 0 lun 2 (pass15,da14)
>>>><JetStor snappod1-vol1 R001>       at scbus3 target 0 lun 256 (pass16,da15)
>>>><JetStor snappod2-vol1 R001>       at scbus3 target 1 lun 0 (pass17,da16)
>>>><JetStor snappod2-vol2 R001>       at scbus3 target 1 lun 1 (pass18,da17)
>>>><JetStor snappod2-vol3 R001>       at scbus3 target 1 lun 2 (pass19,da18)
>>>><JetStor snappod2-vol1 R001>       at scbus3 target 1 lun 256 (pass20,da19)
>>>><JetStor snappod3-vol1 R001>       at scbus3 target 3 lun 0 (pass21,da20)
>>>><JetStor snappod3-vol2 R001>       at scbus3 target 3 lun 1 (pass22,da21)
>>>><JetStor snappod3-vol3 R001>       at scbus3 target 3 lun 2 (pass23,da22)
>>>><JetStor snappod3-vol1 R001>       at scbus3 target 3 lun 256 (pass24,da23)
>>>>
>>>>
>>>>You can see that the 'lun 256' marked devices are duplicates.
>>>>
>>>>Eric
>>>>
>>>>
>>>>
>>>>
>>>>--
>>>>------------------------------------------------------------------------
>>>>Eric Anderson        Sr. Systems Administrator        Centaur Technology
>>>>Anything that works is better than anything that doesn't.
>>>>------------------------------------------------------------------------
>>>>
>>
>>
>>--
>>------------------------------------------------------------------------
>>Eric Anderson        Sr. Systems Administrator        Centaur Technology
>>Anything that works is better than anything that doesn't.
>>------------------------------------------------------------------------
>>


-- 
------------------------------------------------------------------------
Eric Anderson        Sr. Systems Administrator        Centaur Technology
Anything that works is better than anything that doesn't.
------------------------------------------------------------------------


More information about the freebsd-current mailing list