Re: FTRACE_WARN_ON((rec->flags & ~FTRACE_FL_MASK) == 0))

[Steven Rostedt]

On Wed, 2013-10-02 at 10:16 -0400, Dave Jones wrote:
> On Tue, Oct 01, 2013 at 08:28:51AM -0400, Steven Rostedt wrote:
>  > On Tue, 2013-10-01 at 00:20 -0400, Dave Jones wrote:
>  > 
>  > > It seems like trace-cmd needs to be run as root. all hell will break loose if trinity gets root privs.
>  > 
>  > Then run this:
>  > 
>  > trace-cmd record -e syscalls -B trinity su davej -c 'trinity <args>'
> 
> I feel like I'm peeling an onion.
> 
> I keep hititng that other ftrace bug instead:
> 
> WARNING: CPU: 2 PID: 15596 at kernel/trace/ftrace.c:1701 ftrace_bug+0x206/0x270()
> Modules linked in: snd_seq_dummy(+) rfcomm hidp bnep nfnetlink scsi_transport_iscsi nfc caif_socket caif af_802154 phonet af_rxrpc can_raw can_bcm bluetooth rfkill can llc2 pppoe ppp
> ox ppp_generic slhc irda crc_ccitt rds ipt_ULOG af_key rose x25 atm netrom appletalk ipx p8023 psnap p8022 llc ax25 coretemp hwmon x86_pkg_temp_thermal kvm_intel xfs kvm libcrc32c snd_hda_codec_hdm
> i snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_seq crct10dif_pclmul snd_seq_device crc32c_intel ghash_clmulni_intel snd_pcm e1000e microcode serio_raw pcspkr snd_page_alloc shpch
> p usb_debug snd_timer ptp snd pps_core soundcore
> CPU: 2 PID: 15596 Comm: modprobe Not tainted 3.12.0-rc3+ #92 
>  ffffffff81a2a214 ffff880236abfc30 ffffffff8172a363 0000000000000000
>  ffff880236abfc68 ffffffff8105523d 0000000000000000 ffffffffa034a000
>  ffff880239d5b0e0 0000000000000000 ffff880234edb000 ffff880236abfc78
> Call Trace:
>  [<ffffffff8172a363>] dump_stack+0x4e/0x82
>  [<ffffffff8105523d>] warn_slowpath_common+0x7d/0xa0
>  [<ffffffffa034a000>] ? 0xffffffffa0349fff
>  [<ffffffff8105531a>] warn_slowpath_null+0x1a/0x20
>  [<ffffffff81120776>] ftrace_bug+0x206/0x270
>  [<ffffffffa034a000>] ? 0xffffffffa0349fff
>  [<ffffffff81120aec>] ftrace_process_locs+0x30c/0x640
>  [<ffffffff81120e5e>] ftrace_module_notify_enter+0x3e/0x50
>  [<ffffffff81738aa6>] notifier_call_chain+0x66/0x150
>  [<ffffffff8108b197>] __blocking_notifier_call_chain+0x67/0xc0
>  [<ffffffff8108b206>] blocking_notifier_call_chain+0x16/0x20
>  [<ffffffff810daddd>] load_module+0x1f7d/0x26a0

So trinity is loading a module?

>  [<ffffffff810d5f90>] ? store_uevent+0x40/0x40
>  [<ffffffff810db696>] SyS_finit_module+0x86/0xb0
>  [<ffffffff8173d4e4>] tracesys+0xdd/0xe2
> ---[ end trace d2b5caf2401c5667 ]---
> ftrace faulted on writing [<ffffffffa034a000>] dummy_free+0x0/0x10 [snd_seq_dummy]

Is this module compiled for this kernel? What happened was that the
address it tried to write to wasn't writable. Does function tracing
normally work? Speaking of which, the trace-cmd line I gave should not
have started function tracing, so something else must have done it.

-- Steve

> 
> 
> This locks up hard, and then I end up with 0 byte log files when I reboot :-/
> 
> 	Dave


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/