Re: regression on quad Xeon: no SCSI-disks

2007-05-04 Thread Wolfgang Erig
On Fri, May 04, 2007 at 03:18:54AM +0400, Ivan Kokshaysky wrote: > On Thu, May 03, 2007 at 10:41:41AM +0200, Wolfgang Erig wrote: > > I am prepared to do tweaks to your small patch, but I need your help. > > My own blindly experiments failed miserably. > > I don't think that patch did anything wro

Re: regression on quad Xeon: no SCSI-disks

2007-05-03 Thread Ivan Kokshaysky
On Thu, May 03, 2007 at 10:41:41AM +0200, Wolfgang Erig wrote: > I am prepared to do tweaks to your small patch, but I need your help. > My own blindly experiments failed miserably. I don't think that patch did anything wrong, most likely it just triggered a bug elsewhere. These two lines from you

Re: regression on quad Xeon: no SCSI-disks

2007-05-03 Thread Wolfgang Erig
Hello Ivan, On Wed, May 02, 2007 at 09:12:15PM +0400, Ivan Kokshaysky wrote: > On Wed, May 02, 2007 at 04:47:16PM +0200, Wolfgang Erig wrote: > > On Tue, May 01, 2007 at 12:01:56PM -0400, Chuck Ebbert wrote: > > > Output from kernel with CONFIG_PCI_DEBUG could be useful. > > > > > Linux version 2

Re: regression on quad Xeon: no SCSI-disks

2007-05-02 Thread Ivan Kokshaysky
On Wed, May 02, 2007 at 04:47:16PM +0200, Wolfgang Erig wrote: > On Tue, May 01, 2007 at 12:01:56PM -0400, Chuck Ebbert wrote: > > Output from kernel with CONFIG_PCI_DEBUG could be useful. > > > Linux version 2.6.21-gde46c337 ([EMAIL PROTECTED]) (gcc version 4.1.2 > 20061115 (prerelease) (Debian

Re: regression on quad Xeon: no SCSI-disks

2007-05-02 Thread Wolfgang Erig
On Tue, May 01, 2007 at 12:01:56PM -0400, Chuck Ebbert wrote: > Wolfgang Erig wrote: > > Sorry > > for detecting this 2 year old regression so late. > > > > 2.6.13 or later is bad. > > 2.6.12 is good, > > > > "git bisect" worked fine and points to the attached patch. > > The patch is pretty small

Re: regression on quad Xeon: no SCSI-disks

2007-05-01 Thread Chuck Ebbert
Wolfgang Erig wrote: > Sorry > for detecting this 2 year old regression so late. > > 2.6.13 or later is bad. > 2.6.12 is good, > > "git bisect" worked fine and points to the attached patch. > The patch is pretty small. The problem seemed to be dependant > on the PCI-architecture of these particul

regression on quad Xeon: no SCSI-disks

2007-05-01 Thread Wolfgang Erig
Sorry for detecting this 2 year old regression so late. 2.6.13 or later is bad. 2.6.12 is good, "git bisect" worked fine and points to the attached patch. The patch is pretty small. The problem seemed to be dependant on the PCI-architecture of these particular systems. Tried this on two machines