[Click] oops on click-uninstall
Beyers Cronje
bcronje at gmail.com
Mon Feb 26 04:29:10 EST 2007
Hi Eddie,
Update, it wasnt frame pointers, I enabled that and it's still working, will
let you know about the other options...
Beyers
On 2/26/07, Eddie Kohler <kohler at cs.ucla.edu> wrote:
>
> I wonder if Click's configuration/compilation strategy was NOT using frame
> pointers or soemthing. Do let us know what the results of your experiment
> are! Thanks as always,
>
> Eddie
>
>
>
> Beyers Cronje wrote:
> > Hi Eddie,
> >
> > Sorry for only giving feedback on this issue now, been moving house over
> > 800km from where I used to stay. Now I have a nice view from my office
> > to the beatiful South African ocean :)
> >
> > I got it working in the end. I started with a fresh Click CVS version,
> > and fresh 2.6.16.13 <http://2.6.16.13> kernel and disable the following
> > kernel Debug options that was enabled before:
> >
> > Debug Memory Allocations
> > Debug VM
> > Compile the kernel with frame pointers
> >
> > After that everything worked perfectly as before. I'm not sure if it was
> > the fresh click installation or disabling the above debug options that
> > solved the problem, but I'm fairly certain click wasnt using an older
> > proclikefs when I was strugling as I remember I manually delete
> > proclikefs together with a make clean on click at some stage during the
> > struggles. Once we've settled into the new home I will verify if
> > enabling the above debug options introduces the problem.
> >
> > Cheers
> >
> > Beyers
> >
> >
> > On 2/21/07, * Eddie Kohler* <kohler at cs.ucla.edu
> > <mailto:kohler at cs.ucla.edu>> wrote:
> >
> > Hi Beyers,
> >
> > Are you still getting this error? I have not seen anything like it,
> at
> > least on my laptop, and others haven't seen problems either. I
> wonder
> > if there could possibly be something like using an older proclikefs
> > with
> > a newer kernel module. Almost certainly not but....
> >
> > Eddie
> >
> >
> > Beyers Cronje wrote:
> > > Hi guys,
> > >
> > > This is a start on an attempt to narrow down crashes on
> > > click-uninstall with proclikefs. It seems I can reproduce this
> crash
> > > now. Unfortunately I dont have a console connection to the server
> so
> > > I'm left with what's reported in syslog. The one time I did have
> > > console access it seemed the offending function was -
> > > proclikefs_put_super
> > >
> > > Currently running 2.6.19.2 <http://2.6.19.2>, last night's Click
> > CVS version, e1000-7.x
> > > polling driver. Unfortunately the click 2.6.19.2
> > <http://2.6.19.2> patch does not patch
> > > files enabled by CONFIG_LOCKDEP, I will correct this and enable
> full
> > > spinlock debugging.
> > >
> > > As a start this is what was logged in syslog:
> > >
> > > Feb 16 10:38:46 crovanto2850 kernel: Slab corruption:
> > start=edc229d0, len=64
> > > Feb 16 10:38:46 crovanto2850 kernel: Redzone:
> 0x5a2cf071/0x5a2cf071.
> > > Feb 16 10:38:46 crovanto2850 kernel: Last user:
> > > [<fae85a1e>](_ZdaPv+0x6a/0x9a [click])
> > > Feb 16 10:38:46 crovanto2850 kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 64 81 7f f4
> > > Feb 16 10:38:46 crovanto2850 kernel: Prev obj: start=edc22984,
> len=64
> > > Feb 16 10:38:46 crovanto2850 kernel: Redzone:
> 0x170fc2a5/0x170fc2a5.
> > > Feb 16 10:38:46 crovanto2850 kernel: Last user:
> > > [<fae85ec1>](_Znwj+0x21/0x70 [click])
> > > Feb 16 10:38:46 crovanto2850 kernel: 000: 7f 3f ff ff 58 58 58 69
> 10
> > > 00 00 00 d4 e5 ea ed
> > > Feb 16 10:38:46 crovanto2850 kernel: 010: 30 fc ed ee 01 00 00 00
> 20
> > > 00 00 00 14 00 00 00
> > > Feb 16 10:38:46 crovanto2850 kernel: Next obj: start=edc22a1c,
> len=64
> > > Feb 16 10:38:46 crovanto2850 kernel: Redzone:
> 0x5a2cf071/0x5a2cf071.
> > > Feb 16 10:38:46 crovanto2850 kernel: Last user:
> > > [<fae85b3c>](click_lfree+0x22/0x24 [click])
> > > Feb 16 10:38:46 crovanto2850 kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:38:46 crovanto2850 kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:40:40 crovanto2850 kernel: e1000_poll_off
> > > Feb 16 10:40:40 crovanto2850 kernel: Slab corruption:
> > start=edc229d0, len=64
> > > Feb 16 10:40:40 crovanto2850 kernel: Redzone:
> 0x5a2cf071/0x5a2cf071.
> > > Feb 16 10:40:40 crovanto2850 kernel: Last user:
> > > [<fae858de>](_ZdlPv+0x6a/0x9a [click])
> > > Feb 16 10:40:40 crovanto2850 kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 90 db 8d f0
> > > Feb 16 10:40:40 crovanto2850 kernel: Prev obj: start=edc22984,
> > len=64
> > > Feb 16 10:40:40 crovanto2850 kernel: Redzone:
> 0x5a2cf071/0x5a2cf071.
> > > Feb 16 10:40:40 crovanto2850 kernel: Last user:
> > > [<c01a1981>](free_rb_tree_fname+0x41/0x82)
> > > Feb 16 10:40:40 crovanto2850 kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:40:40 crovanto2850 kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:40:40 crovanto2850 kernel: Next obj: start=edc22a1c,
> len=64
> > > Feb 16 10:40:40 crovanto2850 kernel: Redzone:
> 0x5a2cf071/0x5a2cf071.
> > > Feb 16 10:40:40 crovanto2850 kernel: Last user:
> > > [<fae858de>](_ZdlPv+0x6a/0x9a [click])
> > > Feb 16 10:40:40 crovanto2850 kernel: 000: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:40:40 crovanto2850 kernel: 010: 6b 6b 6b 6b 6b 6b 6b 6b
> 6b
> > > 6b 6b 6b 6b 6b 6b 6b
> > > Feb 16 10:40:40 crovanto2850 kernel: device eth2 left promiscuous
> > mode
> > > Feb 16 10:40:40 crovanto2850 kernel: click: stopping router
> > thread pid 14206
> > > Feb 16 10:40:40 crovanto2850 kernel: poll fb05c3a0:
> > 680928/1952766090
> > > freed, 661510/1857091501 allocated
> > > Feb 16 10:40:40 crovanto2850 kernel: click module exiting
> > > Feb 16 10:40:41 crovanto2850 kernel: list_del corruption.
> prev->next
> > > should be f5dac820, but was 6b6b6b6b
> > >
> > > Cheers
> > >
> > > Beyers
> > > _______________________________________________
> > > click mailing list
> > > click at amsterdam.lcs.mit.edu <mailto:click at amsterdam.lcs.mit.edu>
> > > https://amsterdam.lcs.mit.edu/mailman/listinfo/click
> >
> >
>
>
More information about the click
mailing list