Adaptec AHD U320 operating as only U160

Justin T. Gibbs gibbs at scsiguy.com
Sun Feb 17 15:52:25 UTC 2008


Niki Denev wrote:
> After some bonnie64 runs the first machine (8.0-CURRENT Dtrace
> snapshot with aic79xx_osm.c v1.30) spat out this :
> 
> ahd1: Recovery Initiated - Card was not paused
>>>>>>>>>>>>>>>>>>> Dump Card State Begins <<<<<<<<<<<<<<<<<
> ahd1: Dumping Card State at program address 0x38 Mode 0x11

...

> SAVED_MODE[0x11] DFFSTAT[0x31]:(CURRFIFO_1|FIFO0FREE|FIFO1FREE)
> SCSISIGI[0x0]:(P_DATAOUT) SCSIPHASE[0x0] SCSIBUS[0x0]
> LASTPHASE[0x1]:(P_DATAOUT|P_BUSFREE) SCSISEQ0[0x0]

According to the controller, the bus is idle.  However...

> Pending list:
> 325 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 403 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 439 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 380 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 495 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 353 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> 478 FIFO_USE[0x0] SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x17]
> Total 7

we have 7 commands still pending on the drive at ID 1, and the drive hasn't
provided status within the allowed timeout period.  There were lots of
problems early on with Seagate drives locking up like this when hit with
more than 31 (I think that was the magic number) concurrent commands with
WCE (write cache enable) set.  It was supposedly fixed in later firmware
releases.  Do you have the latest firmware for your drives?

You might be able to confirm this hypothesis by lowering the queued depth
via camcontrol and/or disabling WCE.

--
Justin


More information about the freebsd-current mailing list