amd64/139156: [ata] MCP55 SATA Disks fail to initialise, leading to failure to boot [regression]

[Lee Packham]

The following reply was made to PR amd64/139156; it has been noted by GNATS.

From: Lee Packham <lpackham at leenux.org.uk>
To: bug-followup at FreeBSD.org,
 lpackham at leenux.org.uk
Cc:  
Subject: Re: amd64/139156: [ata] MCP55 SATA Disks fail to initialise, leading to failure to boot [regression]
Date: Sun, 27 Sep 2009 10:06:06 +0100

 I just want to add that I appear to have replicated the dmesg errors  
 on 7.2 when hot swapping this disk:
 
 Sep 27 09:57:40 smegbsd kernel:
 Sep 27 09:57:44 smegbsd kernel: ad10: WARNING - SETFEATURES SET  
 TRANSFER MODE taskqueue timeout - completing request directly
 Sep 27 09:57:44 smegbsd kernel: unknown: WARNING - ATA_IDENTIFY  
 freeing taskqueue zombie request
 Sep 27 09:57:48 smegbsd kernel: ad10: WARNING - SETFEATURES SET  
 TRANSFER MODE taskqueue timeout - completing request directly
 Sep 27 09:57:48 smegbsd kernel: ad10: WARNING - SETFEATURES SET  
 TRANSFER MODE freeing taskqueue zombie request
 Sep 27 09:57:52 smegbsd kernel: ad10: WARNING - SETFEATURES ENABLE  
 RCACHE taskqueue timeout - completing request directly
 Sep 27 09:57:52 smegbsd kernel: ad10: WARNING - SETFEATURES SET  
 TRANSFER MODE freeing taskqueue zombie request
 Sep 27 09:57:56 smegbsd kernel: ad10: WARNING - SETFEATURES ENABLE  
 WCACHE taskqueue timeout - completing request directly
 Sep 27 09:57:56 smegbsd kernel: ad10: WARNING - SETFEATURES ENABLE  
 RCACHE freeing taskqueue zombie request
 Sep 27 09:58:00 smegbsd kernel: ad10: WARNING - SET_MULTI taskqueue  
 timeout - completing request directly
 Sep 27 09:58:00 smegbsd kernel: ad10: 152627MB <GB0160CAABV HPG1> at  
 ata5-master SATA150
 Sep 27 09:58:00 smegbsd kernel: ad10: WARNING - SETFEATURES ENABLE  
 WCACHE freeing taskqueue zombie request
 Sep 27 09:58:00 smegbsd kernel: ad10: WARNING - SET_MULTI freeing  
 taskqueue zombie request
 
 The disk works fine when in the machine from startup on 7.2 - but not  
 if hot swapped in. Could it be that the 8.0 kernel is behaving like  
 the disk is hot swapped in as part of initialising it? Just thought I  
 would post this in case it helps.