Hang after usb on EdgeRouter Lite

Joe Holden lists at rewt.org.uk
Fri Aug 30 14:15:37 UTC 2013


For a bit, anyway :P

Few select bins are dying with similar output to:

cpuid = 0
BAD_PAGE_FAULT: pid 1584 tid 100063 (sshd), uid 0: pc 0x16006b5f8 got a 
write fault (type 0x3) at 0x160375b80
Trapframe Register Dump:
         zero: 0 at: 0x160088000 v0: 0x120       v1: 0x160375b80
         a0: 0x160375b80 a1: 0   a2: 0x480       a3: 0x120
         a4: 0x160375b80 a5: 0   a6: 0x1c        a6: 0xe90
         t0: 0x63        t1: 0x70        t2: 0   t3: 0x16008c780
         t8: 0   t9: 0x16006b550 s0: 0x16007c000 s1: 0x1600752e0
         s2: 0x3 s3: 0x3 s4: 0x7fffffd040        s5: 0x480
         s6: 0   s7: 0x1fa000    k0: 0   k1: 0
         gp: 0x16008c780 sp: 0x7fffffcec0        s8: 0x7fffffcf00 
  ra: 0x1600671b4
         sr: 0x8084b3    mullo: 0x290    mulhi: 0        badvaddr: 
0x160375b80
         cause: 0xc      pc: 0x16006b5f8
Page table info for pc address 0x16006b5f8: pde = 0x9800000001fb2000, 
pte = 0x1400057d9a
Dumping 4 words starting at pc address 0x16006b5f8:
ac690000 64e7ffff 14e0fffd 64630004
Page table info for bad address 0x160375b80: pde = 0x9800000001fb2000, 
pte = 0
Failed to write core file for process sshd (error 14)
pid 1584 (sshd), uid 0: exited on signal 10

sshd, bsdtar, mostly ok otherwise, any suggestions (HEAD as of an hour ago)

Cheers,
Joe

On 28/08/2013 13:07, Joe Holden wrote:
> ok so, just built new HEAD with DDB, WITNESS etc to get a backtrace but
> alas, now it's booting fine... who unbroke it? :)
>
> On 19/08/2013 00:44, Ian Lepore wrote:
>> Olivier made some changes to unbreak the arm stuff right after that
>> commit.  Given that some of the mips code evolved from the arm code,
>> maybe there are clues in what he did for arm.
>>
>> -- Ian
>>
>> On Sun, 2013-08-18 at 15:18 -0700, Juli Mallett wrote:
>>> I looked at that change recently, and it changed how kstacks are
>>> allocated
>>> at least, and while it should have preserved the right alignment
>>> behaviour,
>>> I'm wondering now if there may be some subtler problem in the change
>>> there
>>> now.  There were other changes that could be implicated as part of that
>>> commit, but that's the bit that immediately set off some alarm for me,
>>> because I know MIPS is an unusual case in that regard.  (Requiring
>>> kstacks
>>> to fill up a whole double-TLB entry.)
>>>
>>> Thanks for tracking this down, and I'm hopeful that Jeff will have some
>>> insight.  Thanks to Adrian for following up with him.
>>>
>>> Juli.
>>>
>>>
>>> On Sun, Aug 18, 2013 at 3:01 PM, Adrian Chadd <adrian at freebsd.org>
>>> wrote:
>>>
>>>> Ok. I just replied to Jeff's commit and CC'ed you.
>>>>
>>>> Let's see if we can work with Jeff to figure out why the hell it
>>>> broke and
>>>> get it fixed up before 10.0-REL is cut.
>>>>
>>>> Would you please file a PR with all the relevant information that
>>>> you've
>>>> found?
>>>>
>>>> Thanks!
>>>>
>>>>
>>>>
>>>> -adrian
>>>>
>>>>
>>>> On 18 August 2013 14:33, Joe Holden <lists at rewt.org.uk> wrote:
>>>>
>>>>> Ok so r254025 is definitely the commit that broke it (both SMP *and*
>>>> USB) -
>>>>> looks like the commit is quite involved so won't be particularly
>>>>> easy to
>>>>> unbreak just mips (at least for me) :(
>>>>>
>>>>>> -----Original Message-----
>>>>>> From: owner-freebsd-mips at freebsd.org [mailto:owner-freebsd-
>>>>>> mips at freebsd.org] On Behalf Of Joe Holden
>>>>>> Sent: 18 August 2013 21:42
>>>>>> To: 'Adrian Chadd'
>>>>>> Cc: freebsd-mips at freebsd.org
>>>>>> Subject: RE: Hang after usb on EdgeRouter Lite
>>>>>>
>>>>>> ... and it boot with the second core disabled (usb still broken
>>>> though) -
>>>>>> something has gone horribly wrong :P
>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: owner-freebsd-mips at freebsd.org [mailto:owner-freebsd-
>>>>>>> mips at freebsd.org] On Behalf Of Joe Holden
>>>>>>> Sent: 18 August 2013 21:30
>>>>>>> To: 'Adrian Chadd'
>>>>>>> Cc: freebsd-mips at freebsd.org
>>>>>>> Subject: RE: Hang after usb on EdgeRouter Lite
>>>>>>>
>>>>>>> Sigh. it's actually getting worse - r254025 hangs after ULE setup.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> This may take longer than I first thought
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> From: adrian.chadd at gmail.com [mailto:adrian.chadd at gmail.com] On
>>>>>> Behalf
>>>>>>> Of Adrian Chadd
>>>>>>> Sent: 18 August 2013 20:27
>>>>>>> To: Joe Holden
>>>>>>> Cc: Warner Losh; freebsd-mips at freebsd.org
>>>>>>> Subject: Re: Hang after usb on EdgeRouter Lite
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> .. you can just disable those for now. You're debugging USB hangs,
>>>> not
>>>>>>> bootp/nfs compile problems. :0
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> -adiran
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On 18 August 2013 11:19, Joe Holden <lists at rewt.org.uk
>>>>>>> <mailto:lists at rewt.org.uk> > wrote:
>>>>>>>
>>>>>>> None yet - getting compilation problems when building with BOOTP and
>>>>>>> NFS root (undefined references)
>>>>>>>
>>>>>>>
>>>>>>>> -----Original Message-----
>>>>>>>> From: Warner Losh [mailto:wlosh at bsdimp.com
>>>>>>> <mailto:wlosh at bsdimp.com> ]
>>>>>>>> On
>>>>>>> Behalf Of Warner Losh
>>>>>>>> Sent: 18 August 2013 07:41
>>>>>>>> To: Adrian Chadd
>>>>>>>> Cc: Joe Holden; freebsd-mips at freebsd.org
>>>>>>>> <mailto:freebsd-mips at freebsd.org>
>>>>>>>
>>>>>>>> Subject: Re: Hang after usb on EdgeRouter Lite
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>> On Aug 16, 2013, at 12:12 PM, Adrian Chadd wrote:
>>>>>>>>> On 16 August 2013 01:56, Joe Holden <lists at rewt.org.uk
>>>>>>> <mailto:lists at rewt.org.uk> > wrote:
>>>>>>>>>> Not getting much closer but so far 253367 works, 253845 and
>>>> later
>>>>>>>>>> don't, I can't see any mips-specific changes that might break it
>>>>>>>>>> but I
>>>>>>> may
>>>>>>>> be wrong!
>>>>>>>>>
>>>>>>>>> Plenty of non-MIPS stuff does break MIPS. :)
>>>>>>>>
>>>>>>>> Any new news?
>>>>>>>>
>>>>>>>> Warner
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> freebsd-mips at freebsd.org mailing list
>>>>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
>>>>>>> To unsubscribe, send any mail to "
>>>> freebsd-mips-unsubscribe at freebsd.org
>>>>> "
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> freebsd-mips at freebsd.org mailing list
>>>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
>>>>>> To unsubscribe, send any mail to
>>>>>> "freebsd-mips-unsubscribe at freebsd.org
>>>> "
>>>>>
>>>>>
>>>>>
>>>> _______________________________________________
>>>> freebsd-mips at freebsd.org mailing list
>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
>>>> To unsubscribe, send any mail to "freebsd-mips-unsubscribe at freebsd.org"
>>>>
>>> _______________________________________________
>>> freebsd-mips at freebsd.org mailing list
>>> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
>>> To unsubscribe, send any mail to "freebsd-mips-unsubscribe at freebsd.org"
>>
>>
>> _______________________________________________
>> freebsd-mips at freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
>> To unsubscribe, send any mail to "freebsd-mips-unsubscribe at freebsd.org"
>>
>
> _______________________________________________
> freebsd-mips at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-mips
> To unsubscribe, send any mail to "freebsd-mips-unsubscribe at freebsd.org"



More information about the freebsd-mips mailing list