Pavol Kycina
12/18/2008 1:34 AM
post18912
|
On one of our systems we have experienced io-pkt crash.
I am attaching a file with core and list of pci devices.
Any idea what could have gone wrong?
Thanks, PKY
|
|
|
Sean Boudreau(deleted)
12/18/2008 9:21 AM
post18929
|
On Thu, Dec 18, 2008 at 01:35:46AM -0500, Pavol Kycina wrote:
> On one of our systems we have experienced io-pkt crash.
>
> I am attaching a file with core and list of pci devices.
>
> Any idea what could have gone wrong?
>
> Thanks, PKY
Can you also attach the io-pkt-v4-hc binary, any drivers /
dlls loaded therein (looks like devnp-shim.so and devn-e1000.so)
and your libc.so.
Thanks,
-seanb
|
|
|
Andrew Boyd(deleted)
12/18/2008 9:22 AM
post18931
|
> I am attaching a file with core
I unzipped it, and sure enough, all there was,
was a core file (and the pci info).
For us to examine a core file, we need:
1) your core file (got it)
2) your io-pkt binary (didn't get it)
3) ALL the libs loaded by io-pkt, which
includes the driver .so file, perhaps
qnet, and please libc as well
To figure out what libs you need to include
in the tarball, please run:
# pidin -p io-pkt-v4-hc mem
--
aboyd
|
|
|
Pavol Kycina
|
Re: RE: io-pktv4-hc crash
|
Pavol Kycina
01/07/2009 5:41 AM
post19504
|
Re: RE: io-pktv4-hc crash
I hope all the files needed are in the zip file.
I use devn-e1000.so. Is devnp-e1000 driver available? I couldn't find it.
Thanks, PKY
|
|
|
Hugh Brown
|
RE: RE: io-pktv4-hc crash
|
Hugh Brown
01/07/2009 7:45 AM
post19511
|
RE: RE: io-pktv4-hc crash
No it isn't available yet.
-----Original Message-----
From: Pavol Kycina [mailto:community-noreply@qnx.com]
Sent: Wednesday, January 07, 2009 5:42 AM
To: drivers-networking
Subject: Re: RE: io-pktv4-hc crash
I hope all the files needed are in the zip file.
I use devn-e1000.so. Is devnp-e1000 driver available? I couldn't find
it.
Thanks, PKY
_______________________________________________
Networking Drivers
http://community.qnx.com/sf/go/post19504
|
|
|
Andrew Boyd(deleted)
|
RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/07/2009 10:50 AM
post19544
|
RE: RE: io-pktv4-hc crash
There is no core file (eg io-pkt-v4-hc.core)
in the archive.
--
aboyd
|
|
|
Andrew Boyd(deleted)
|
RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/07/2009 10:58 AM
post19547
|
RE: RE: io-pktv4-hc crash
Never mind - I use the email interface to the
forums, which doesn't display previous posts.
I found the core file from 17 dec 08, and
am looking at it now.
--
aboyd
|
|
|
Andrew Boyd(deleted)
|
RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/07/2009 11:13 AM
post19552
|
RE: RE: io-pktv4-hc crash
Ok, if all the binaries are in sync, you
didn't die in the (e1000) driver at all.
Looking at the backtrace for the thread
that SIGSEGV'd, it blew up in a utility
function called b_to_q() which tried to
memcpy() to a null pointer. This was all
called by pppinput() - nothing to do with
the ethernet driver (or PCI) AFAIK.
Is this difficult to reproduce? I really
don't know squat about ppp (I know a little
about pppoe) but if you want to compile
io-pkt-v4-hc with the -g -O0 flags, and
send me another core, I can try to learn
a bit more - there's only so much I can
glean from disassembling x86.
Anything interesting about your ppp
configuration, or when this core occurs?
--
aboyd
|
|
|
Robert Craig
|
RE: RE: io-pktv4-hc crash
|
Robert Craig
01/07/2009 1:30 PM
post19576
|
RE: RE: io-pktv4-hc crash
We also need to know which version of the stack is being used. Is it a
custom build, a milestone build or are you running 6.4.0?
Thanks,
Robert.
-----Original Message-----
From: Andrew Boyd [mailto:community-noreply@qnx.com]
Sent: Wednesday, January 07, 2009 11:13 AM
To: drivers-networking
Subject: RE: RE: io-pktv4-hc crash
Ok, if all the binaries are in sync, you didn't die in the (e1000)
driver at all.
Looking at the backtrace for the thread
that SIGSEGV'd, it blew up in a utility
function called b_to_q() which tried to
memcpy() to a null pointer. This was all called by pppinput() - nothing
to do with the ethernet driver (or PCI) AFAIK.
Is this difficult to reproduce? I really don't know squat about ppp (I
know a little about pppoe) but if you want to compile io-pkt-v4-hc with
the -g -O0 flags, and send me another core, I can try to learn a bit
more - there's only so much I can glean from disassembling x86.
Anything interesting about your ppp
configuration, or when this core occurs?
--
aboyd
_______________________________________________
Networking Drivers
http://community.qnx.com/sf/go/post19552
|
|
|
Pavol Kycina
|
Re: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
01/08/2009 5:17 AM
post19612
|
Re: RE: RE: io-pktv4-hc crash
Its 6.4.0 build, all executables are from 6.4.0 DVD downloaded from your site.
|
|
|
Pavol Kycina
|
Re: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
01/08/2009 5:25 AM
post19613
|
Re: RE: RE: io-pktv4-hc crash
>
> Is this difficult to reproduce? I really
> don't know squat about ppp (I know a little
> about pppoe) but if you want to compile
> io-pkt-v4-hc with the -g -O0 flags, and
> send me another core, I can try to learn
> a bit more - there's only so much I can
> glean from disassembling x86.
>
> Anything interesting about your ppp
> configuration, or when this core occurs?
>
This problem had happened for the first time. The system uptime was about 1 week. So we don't know the way to reproduce
it at a request.
The system transfers some data with measurement station via tcp/ip/pppd/modem connection. But there is nothing special
with this pppd.
It's started in a script with following args:
speed localIP:remoteIP device nodetach lcp-echo-interval 10 lcp-echo-failure 5 connect "chat script"
After data transfer, pppd is killed
Regards, PKY
|
|
|
Andrew Boyd(deleted)
|
RE: RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/08/2009 1:33 PM
post19671
|
RE: RE: RE: io-pktv4-hc crash
|
|
|
Andrew Boyd(deleted)
|
Re: RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/08/2009 2:02 PM
post19677
|
Re: RE: RE: io-pktv4-hc crash
> The system uptime was about 1 week
For completeness, you should check and make sure that
you don't have a memory leak, and aren't running out of
memory on the machine.
For example, does "pidin in" show the number after
FreeMem steadily decreasing as the days go by? If so,
which process is leaking memory?
--
aboyd
|
|
|
Pavol Kycina
|
Re: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
01/20/2009 8:28 AM
post20369
|
Re: RE: RE: io-pktv4-hc crash
The crash happened again.
I can confirm, that there was enough free memory at the time of the crash.
PKY
> > The system uptime was about 1 week
>
> For completeness, you should check and make sure that
> you don't have a memory leak, and aren't running out of
> memory on the machine.
>
> For example, does "pidin in" show the number after
> FreeMem steadily decreasing as the days go by? If so,
> which process is leaking memory?
>
> --
> aboyd
|
|
|
Andrew Boyd(deleted)
|
RE: RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/20/2009 9:34 AM
post20378
|
RE: RE: RE: io-pktv4-hc crash
> there was enough free memory at the time of the crash.
ok, that rules out low memory as the problem.
Unless someone has a better idea, would you be willing
to run a custom io-pkt (that I can provide you) that I
have compiled "-g -O0" so that the next time it cores,
it produces a core file with a lot more information
for me to look at?
I wish I had a better suggestion :-(
--
aboyd
|
|
|
Pavol Kycina
|
Re: RE: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
01/21/2009 1:17 AM
post20468
|
Re: RE: RE: RE: io-pktv4-hc crash
Yes, I have no problem with running custom io-pkt.
Can I download it from somewhere?
Regards, PKY
>
> > there was enough free memory at the time of the crash.
>
> ok, that rules out low memory as the problem.
>
> Unless someone has a better idea, would you be willing
> to run a custom io-pkt (that I can provide you) that I
> have compiled "-g -O0" so that the next time it cores,
> it produces a core file with a lot more information
> for me to look at?
>
> I wish I had a better suggestion :-(
>
> --
> aboyd
|
|
|
Andrew Boyd(deleted)
|
RE: RE: RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
01/21/2009 11:06 AM
post20501
|
RE: RE: RE: RE: io-pktv4-hc crash
> Can I download it from somewhere?
www.pittspecials.com/etc/io-pkt-v4-hc
I just compiled it -g -O0. After you
download it, be sure that the size is
4,537,200 bytes, and the cksum is
14090642, then chmod it 777.
Thanks,
--
aboyd
|
|
|
Pavol Kycina
|
Re: RE: RE: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
03/09/2009 9:50 AM
post23890
|
Re: RE: RE: RE: RE: io-pktv4-hc crash
It took some time for a "special" io-pktv4-hc to crash.
I am attaching core file
Thanks, PKY
|
|
|
Andrew Boyd(deleted)
|
RE: RE: RE: RE: RE: io-pktv4-hc crash
|
Andrew Boyd(deleted)
03/09/2009 11:17 AM
post23898
|
RE: RE: RE: RE: RE: io-pktv4-hc crash
> attaching core file
After carefully verifying matching binaries - it's
been almost 2 months - it looks like a different
problem. The fault occurred as a result of
NWSIGUNLOCK_P()
call at line 765 of sys/receive.c, which according
to the backtrace ended up in libc.so.3 __assert()
which called abort() then raise() then SignalKill().
The first parameters to the NWSIGUNLOCK_P() macro
was &sctlp->pkt_ex which gdb says had an "owner" of
zero, and an apparently underflowed "count" of approximately
negative two billion and change.
I will pester sean about this - I seem to recall
something similar.
--
aboyd
|
|
|
Pavol Kycina
|
Re: RE: RE: RE: RE: RE: io-pktv4-hc crash
|
Pavol Kycina
03/10/2009 8:40 AM
post24000
|
Re: RE: RE: RE: RE: RE: io-pktv4-hc crash
|
|
|
Yao Zhao(deleted)
|
Re: RE: RE: RE: RE: RE: io-pktv4-hc crash
|
Yao Zhao(deleted)
07/16/2009 2:56 PM
post33981
|
Re: RE: RE: RE: RE: RE: io-pktv4-hc crash
(gdb) p/x stk_ctl
$9 = {iopkt = 0x819b018, do_cache = 0x1, recv_max = 0x21, recv_avail = 0x21, recv_iov = 0x819fe98,
recv_mbuf = 0x819ffb8, recv_start = 0x0, recv_loaded_proc = 0x81accf8, recv_procp = 0x81accf8,
chid = 0x1, coid = 0x40000002, dpp = 0x8196068, timer_int_id = 0xe, fastforward = 0x1,
stacksize = 0x1000, rx_prio = 0x15, proc_alloc_tot = 0x4, proc_alloc_used = 0x1, allprocs = 0x8198140,
proc0 = 0x81ac1e0, freeprocs = 0x81ac9a4, nprocs_cur_max = 0xf, nprocs_used = 0x4, nprocs_min = 0xf,
nprocs_incr = 0x19, nprocs_max = 0xc8, nthreads_core = 0x1, nthreads_core_max = 0x2, nthreads_oob = 0x0,
nthreads_oob_max = 0x2, nthreads_flow_max = 0x2, nthreads_cur = 0x1, nthreads_untracked = 0x1,
nwork_threads = 0x4, work_threads = 0x8194068, quiesce_count = 0x0, stack_ex = {count = 0x80000000,
owner = 0x0}, stack_inuse = 0x0, thread_inreceive = 0x1, msg_outstanding = 0x0, pkt_ex = {
count = 0x80000000, owner = 0x0}, pkt_rx_q = 0x0, work_thread_key = 0x0, proc_prio = {
prio_prios = 0x81a0058, prio_all = {head = 0x0, tail = 0x0}}, pstats = {p_ru = {ru_utime = {
tv_sec = 0x0, tv_usec = 0x0}, ru_stime = {tv_sec = 0x0, tv_usec = 0x0}, ru_maxrss = 0x0,
ru_ixrss = 0x0, ru_idrss = 0x0, ru_isrss = 0x0, ru_minflt = 0x0, ru_majflt = 0x0, ru_nswap = 0x0,
ru_inblock = 0x0, ru_oublock = 0x0, ru_msgsnd = 0x74be6c, ru_msgrcv = 0x12ef4cb, ru_nsignals = 0x0,
ru_nvcsw = 0x0, ru_nivcsw = 0x0}, p_cru = {ru_utime = {tv_sec = 0x0, tv_usec = 0x0}, ru_stime = {
tv_sec = 0x0, tv_usec = 0x0}, ru_maxrss = 0x0, ru_ixrss = 0x0, ru_idrss = 0x0, ru_isrss = 0x0,
ru_minflt = 0x0, ru_majflt = 0x0, ru_nswap = 0x0, ru_inblock = 0x0, ru_oublock = 0x0,
ru_msgsnd = 0x0, ru_msgrcv = 0x0, ru_nsignals = 0x0, ru_nvcsw = 0x0, ru_nivcsw = 0x0}, p_timer = {{
it_interval = {tv_sec = 0x0, tv_usec = 0x0}, it_value = {tv_sec = 0x0, tv_usec = 0x0}}, {
it_interval = {tv_sec = 0x0, tv_usec = 0x0}, it_value = {tv_sec = 0x0, tv_usec = 0x0}}, {
it_interval = {tv_sec = 0x0, tv_usec = 0x0}, it_value = {tv_sec = 0x0, tv_usec = 0x0}}}, p_prof = {
pr_base = 0x0, pr_size = 0x0, pr_off = 0x0, pr_scale = 0x0, pr_addr = 0x0, pr_ticks = 0x0},
p_start = {tv_sec = 0x0, tv_usec = 0x0}}, plimit = {pl_rlimit = {{rlim_cur = 0x0, rlim_cur_hi = 0x0,
rlim_max = 0x0, rlim_max_hi = 0x0}, {rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0,
rlim_max_hi = 0x0}, {rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0, rlim_max_hi = 0x0}, {
rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0, rlim_max_hi = 0x0}, {rlim_cur = 0x0,
rlim_cur_hi = 0x0, rlim_max = 0x0, rlim_max_hi = 0x0}, {rlim_cur = 0x7fffffff, rlim_cur_hi = 0x0,
rlim_max = 0x7fffffff, rlim_max_hi = 0x0}, {rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0,
rlim_max_hi = 0x0}, {rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0, rlim_max_hi = 0x0}, {
rlim_cur = 0x7fffffff, rlim_cur_hi = 0x0, rlim_max = 0x7fffffff, rlim_max_hi = 0x0}, {
rlim_cur = 0x0, rlim_cur_hi = 0x0, rlim_max = 0x0, rlim_max_hi = 0x0}}, pl_corename = 0x0,
p_lflags = 0x0, p_refcnt = 0x1}}
(gdb) p stk_ctl.work_threads[0]@4
$3 = {0x81a2880, 0x0, 0x0, 0x0}
(gdb) p *(struct nw_work_thread*)0x81a2880
$7 = {wt_critical = -1, wt_intr_pending = 0, intr_sighot = 1, am_stack = 0, blocking = 0, inreceive = 0,
tidx_wt = 0, tidx_irupt = 0, flowctl = 0x81a2b70, wt_bru = {wt_bru_brctl = 0x81950a8,
wt_bru_bridx = 0x81950a8}, wt_oob = 0x0, wt_mzones = {{p = 0x87bd500, avail = 32, max = 32}, {
p = 0x82f1f00, avail = 32, max = 32}}, jp = 0x0, intr_stack_tos = 0x81a2820 "",
saved_sp = 133831884, tls = 0x7fc6fc0, rx_loop_ctxt = {rx_loop_jmp_buf = {{__jmpbuf_un = {__savearea = {
0 <repeats 13 times>}, __alignment = 0}, __flg = 0, __msk = {0, 0}}}, rx_loop_ctxt = {cpu = {
edi = 0, esi = 0, ebp = 0, exx = 0, ebx =...
View Full Message
|
|
|
|