Kernel Panic on BeagleBone Black

[Ian Lepore]

On Wed, 2014-07-30 at 13:43 -0400, Shawn Webb wrote:
> On Jul 30, 2014 11:30 AM -0600, Ian Lepore wrote:
> > On Wed, 2014-07-30 at 08:19 -0400, Shawn Webb wrote:
> > > Hey All,
> > > 
> > > I've just updated to a recent HEAD (r269240). I get a kernel panic
> > > almost immediately on boot. Below is the log.
> > > 
> > > ==== Start of dump ====
> > > cpsw0: <3-port Switch Ethernet Subsystem> mem 0x4a100000-0x4a103fff irq 40,41,42,43 on simplebus0
> > > cpsw0: CPSW SS Version 1.12 (0)
> > > cpsw0: Initial queue size TX=128 RX=384
> > > cpsw0: Ethernet address: 1c:ba:8c:e4:6d:6a
> > > cpsw0: Failed to read from PHY.
> > > cpsw0: attaching PHYs failed
> > > 
> > > vm_fault(0xc070b9e0, 0, 1, 0) -> 1
> > > Fatal kernel mode data abort: 'Translation Fault (S)'
> > > trapframe: 0xc080eb28
> > > FSR=00000005, FAR=00000018, spsr=80000193
> > > r0 =c266f280, r1 =00000000, r2 =00000019, r3 =60000193
> > > r4 =00000000, r5 =c266f280, r6 =00000006, r7 =c05c99b4
> > > r8 =c266f280, r9 =c26ca28c, r10=c26c80c8, r11=c080eb88
> > > r12=00000000, ssp=c080eb78, slr=c05ee1cc, pc =c03d2614
> > > 
> > > [ thread pid 0 tid 100000 ]
> > > Stopped at      device_delete_child+0x14:       ldr     r1, [r4, #0x018]
> > > db> bt
> > > Tracing pid 0 tid 100000 td 0xc070b6d0
> > [...]
> > > Unable to unwind further
> > > ==== End of dump ====
> > > 
> > > If there's anything I can do, let me know.
> > > 
> > > Thanks,
> > > 
> > > Shawn
> > 
> > Hmm, I can't reproduce this on my BB White at r269302.  The real error
> > is "Failed to read from PHY".  The kernel abort was just accidental
> > fallout from trying to clean up and detach the ethernet driver since it
> > failed to init properly (error paths never get tested enough).
> > 
> > What was the prior release you were on that worked okay?
> > 
> > Since I just made a series of changes to armv6 busdma I'm tempted to
> > suspect them, even though there shouldn't be any DMA involved in talking
> > to the PHY.  Still, it would be interesting to know if backing off to
> > r269134 makes the problem go away.
> > 
> > -- Ian
> > 
> > 
> 
> The release prior worked okay. I did get occasional kernel panics
> (though, I didn't have time to do a backtrace and post here about it).
> I'm unsure what svn rev the prior build was at. I can try reverting that
> revision you mentioned tonight and report back if you'd like.
> 
> Thanks,
> 
> Shawn

Don't revert that single revision I mentioned, revert the whole source
tree to that rev (or to speed things up, really just the sys tree, and
rebuild just the kernel).

-- Ian