[fm-discuss] EFT Diagnosis Engine unable to produce a diagnosis

[Sean Ye]

It seems there were PCIe error reported in the U40 M2 system but FMD
failed to produce a diagnosis.

If you're not using any error injector, there ought to be a bug.

On Thu, Feb 21, 2008 at 09:57:33PM +1000, James C. McPherson wrote:
> On Thu, 21 Feb 2008 19:49:34 +0800
> Sean Ye <Sean.Ye at Sun.COM> wrote:
> 
> > Hi, James,
> > 
> > What is the OS version?
> > Would you please provide the output of:
> > 
> > fmdump -eVu 6aa8faed-aa89-6c7e-9b80-fe223c600829
> > 
> > 
> > On Thu, Feb 21, 2008 at 09:22:27PM +1000, James C. McPherson wrote:
> ...
> > > $ fmdump
> > > TIME                 UUID
> > > SUNW-MSG-ID Feb 21 08:35:14.8078
> > > 8862971a-4f95-c40a-f384-c31cbb40a1e2 SUNOS-8000-1L Feb 21
> > > 19:49:55.9692 6aa8faed-aa89-6c7e-9b80-fe223c600829 SUNOS-8000-1L
> 
> 
> Hi Sean,
> I'm running Solaris Nevada build snv_83:
> 
> 
> # uname -a
> SunOS blinder 5.11 snv_83 i86pc i386 i86pc
> 
> 
> bash-3.2# fmdump -eVu 8862971a-4f95-c40a-f384-c31cbb40a1e2 
> TIME                           CLASS
> Feb 21 2008 08:35:09.766775965 ereport.io.pci.sec-ma
> nvlist version: 0
>         class = ereport.io.pci.sec-ma
>         ena = 0x47415b319fc00801
>         detector = (embedded nvlist)
>         nvlist version: 0
>                 version = 0x0
>                 scheme = dev
>                 device-path = /pci at 0,0/pci10de,377 at f
>         (end detector)
> 
>         pci-sec-status = 0x2000
>         pci-bdg-ctrl = 0xf
>         __ttl = 0x1
>         __tod = 0x47bcab1d 0x2db4129d
> 
> Feb 21 2008 08:35:09.766775625 ereport.io.pciex.dl.rto
> nvlist version: 0
>         class = ereport.io.pciex.dl.rto
>         ena = 0x47415b319fc00801
>         detector = (embedded nvlist)
>         nvlist version: 0
>                 version = 0x0
>                 scheme = dev
> 		device-path = /pci at 0,0/pci10de,377 at f
> 	(end detector)
> 
> 	dev-status = 0x1
> 	ce-status = 0x1000
> 	__ttl = 0x1
> 	__tod = 0x47bcab1d 0x2db41149
> 
> bash-3.2# fmdump -eVu 6aa8faed-aa89-6c7e-9b80-fe223c600829 
> TIME                           CLASS
> Feb 21 2008 19:49:50.941095961 ereport.io.pci.sec-ma
> nvlist version: 0
> 	class = ereport.io.pci.sec-ma
> 	ena = 0x945ceeca58400801
> 	detector = (embedded nvlist)
> 	nvlist version: 0
> 		version = 0x0
> 		scheme = dev
> 		device-path = /pci at 0,0/pci10de,377 at f
> 	(end detector)
> 
> 	pci-sec-status = 0x2000
> 	pci-bdg-ctrl = 0xf
> 	__ttl = 0x1
> 	__tod = 0x47bd493e 0x3817fc19
> 
> Feb 21 2008 19:49:50.941096506 ereport.io.pciex.dl.rto
> nvlist version: 0
> 	class = ereport.io.pciex.dl.rto
> 	ena = 0x945ceeca58400801
> 	detector = (embedded nvlist)
> 	nvlist version: 0
> 		version = 0x0
> 		scheme = dev
> 		device-path = /pci at 0,0/pci10de,377 at f
> 	(end detector)
> 
> 	dev-status = 0x1
> 	ce-status = 0x1000
> 	__ttl = 0x1
> 	__tod = 0x47bd493e 0x3817fe3a
> 
> 
> That device path goes to my framebuffer:
> 
> 
> bash-3.2# grep "pci at 0,0/pci10de,377 at f" /etc/path_to_inst
> "/pci at 0,0/pci10de,377 at f" 5 "pcie_pci"
> "/pci at 0,0/pci10de,377 at f/display at 0" 0 "nvidia"
> "/pci at 0,0/pci10de,377 at f/display at 0" 0 "vgatext"
> 
> 
> 
> 
> Thanks,
> James C. McPherson
> --
> Senior Kernel Software Engineer, Solaris
> Sun Microsystems
> http://blogs.sun.com/jmcp	http://www.jmcp.homeunix.com/blog