NetBSD-Bugs archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: lib/50485: dtrace doesn't work (fails on loading a script, psinfo.d)
The following reply was made to PR lib/50485; it has been noted by GNATS.
From: Richard PALO <richard%netbsd.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc:
Subject: Re: lib/50485: dtrace doesn't work (fails on loading a script, psinfo.d)
Date: Tue, 02 Aug 2016 07:43:55 +0200
Le 10/12/15 09:20, Ryota Ozaki a écrit :
> The following reply was made to PR lib/50485; it has been noted by GNATS.
>
> From: Ryota Ozaki <ozaki-r%netbsd.org@localhost>
> To: Christos Zoulas <christos%zoulas.com@localhost>
> Cc: "gnats-bugs%NetBSD.org@localhost" <gnats-bugs%netbsd.org@localhost>, lib-bug-people%netbsd.org@localhost,
> gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost
> Subject: Re: lib/50485: dtrace doesn't work (fails on loading a script, psinfo.d)
> Date: Thu, 10 Dec 2015 17:13:57 +0900
>
> On Sat, Nov 28, 2015 at 10:21 PM, Christos Zoulas <christos%zoulas.com@localhost> wrote:
> > On Nov 28, 9:45am, ozaki-r%netbsd.org@localhost (ozaki-r%netbsd.org@localhost) wrote:
> > -- Subject: lib/50485: dtrace doesn't work (fails on loading a script, psinfo
> >
> > | >Number: 50485
> > | >Category: lib
> > | >Synopsis: dtrace doesn't work (fails on loading a script, psinfo.d)
> > | >Confidential: no
> > | >Severity: serious
> > | >Priority: medium
> > | >Responsible: lib-bug-people
> > | >State: open
> > | >Class: sw-bug
> > | >Submitter-Id: net
> > | >Arrival-Date: Sat Nov 28 09:45:00 +0000 2015
> > | >Originator: Ryota Ozaki
> > | >Release: NetBSD-current
> > | >Organization:
> > | >Environment:
> > | NetBSD kvm 7.99.22 NetBSD 7.99.22 (KVM) #876: Sat Nov 28 18:22:04 JST 2015 ozaki-r@rangeley:(hidden) amd64
> > | >Description:
> > | dtrace(1M) fails with the following error:
> > |
> > | # dtrace -n BEGIN
> > | dtrace: invalid probe specifier BEGIN: "/usr/lib/dtrace/psinfo.d", line 46: syntax error near "u_int"
> > |
> > | psinfo.d seems to contain syntax error(s). If I remove the file and try again, it works fine.
> > | >How-To-Repeat:
> > | modload dtrace
> > | dtrace -n BEGIN
> >
> > You need to build a kernel with CTF support.
>
> (Forgot to reply. Sorry for that.)
>
> You're right. It works with CTF. Thanks!
>
> ozaki-r
>
>
I'm curious whether this needs to be reopened as I'm experiencing similar on recent current
with both default and slightly modified GENERIC kernel.
> bash-4.3$ uname -a
> NetBSD nb32.baou.fr 7.99.34 NetBSD 7.99.34 (GEN.nb32) #0: Mon Aug 1 00:06:10 CEST 2016 richard%nb32.xxxx.fr@localhost:/usr/obj/sys/arch/i386/compile/GEN.nb32 i386
> bash-4.3$ sudo modstat |grep dtra
> cyclic misc filesys a 1 4252 dtrace
> dtrace driver filesys a 4 110162 solaris
> dtrace_fbt misc filesys - 0 5564 dtrace
> dtrace_profile misc filesys - 0 2158 dtrace,cyclic
> dtrace_sdt misc filesys - 0 1562 dtrace
> bash-4.3$ diff /usr/src/sys/arch/i386/conf/GENERIC /usr/src/sys/arch/i386/conf/GEN.nb32
> 32a33
>> makeoptions CPUFLAGS="-march=pentium-mmx"
> 581c582
> <
> ---
>> options DRM_DEBUG
> 1224c1225
> < ulpt* at uhub? port ? configuration ? interface ?
> ---
>> #ulpt* at uhub? port ? configuration ? interface ?
> bash-4.3$ grep -E 'CTF|DTRA' /etc/mk.conf
> MKDTRACE=yes
> MKCTF=yes
> bash-4.3$ sudo dtrace -l |head
> ID PROVIDER MODULE FUNCTION NAME
> 1 dtrace BEGIN
> 2 dtrace END
> 3 dtrace ERROR
> 4 proc kernel exec-failure
> 5 proc kernel exec-success
> 6 proc kernel exec
> 7 proc kernel exit
> 8 proc kernel create
> 9 proc kernel lwp-exit
> bash-4.3$ sudo dtrace -l |wc -l
> 51650
> bash-4.3$ cat hello.d
> BEGIN
> {
> trace("Hello world");
> exit(0);
> }
> bash-4.3$ sudo dtrace -s hello.d
> dtrace: failed to compile script hello.d: "/usr/lib/dtrace/psinfo.d", line 46: syntax error near "u_int"
> bash-4.3$ sudo dtrace -n BEGIN
> dtrace: invalid probe specifier BEGIN: "/usr/lib/dtrace/psinfo.d", line 46: syntax error near "u_int"
I've noticed some noise in the build that I worked around (patches available for inspection) but there
seems to be some ctf issues in the libproc processing where psinfo.d comes into play.
Anybody else try or notice lately?
What object should contain the ctf data here in question?
--
Richard PALO
Home |
Main Index |
Thread Index |
Old Index