[Qemu-devel] [PATCH] trace: drop LTTng Userspace Tracer backend

[Qemu-devel] [PATCH] trace: drop LTTng Userspace Tracer backend

[Stefan Hajnoczi]

The current LTTng Userspace Tracer backend does not build against modern
libraries.  LTTng has changed the library ABI several times, making it
difficult to support this backend.

Due to lack of users, remove this trace backend until someone turns up
who is willing to actively maintain it.

Details for anyone wishing to contribute an updated backend:

 * Header files have been renamed from <ust/*> to <lttng/*>
 * Tracepoint macros are no longer DECLARE_TRACE(name, TP_PROTO(args),
   TP_ARGS(argnames))
 * Probably more breakage

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 configure                        | 19 --------
 docs/tracing.txt                 | 11 +----
 scripts/tracetool/backend/ust.py | 93 ----------------------------------------
 3 files changed, 2 insertions(+), 121 deletions(-)
 delete mode 100644 scripts/tracetool/backend/ust.py

diff --git a/configure b/configure
index 05e16da..95a8a4f 100755
--- a/configure
+++ b/configure
@@ -3280,22 +3280,6 @@ if test "$?" -ne 0 ; then
 fi
 
 ##########################################
-# For 'ust' backend, test if ust headers are present
-if test "$trace_backend" = "ust"; then
-  cat > $TMPC << EOF
-#include <ust/tracepoint.h>
-#include <ust/marker.h>
-int main(void) { return 0; }
-EOF
-  if compile_prog "" "" ; then
-    LIBS="-lust -lurcu-bp $LIBS"
-    libs_qga="-lust -lurcu-bp $libs_qga"
-  else
-    error_exit "Trace backend 'ust' missing libust header files"
-  fi
-fi
-
-##########################################
 # For 'dtrace' backend, test if 'dtrace' command is present
 if test "$trace_backend" = "dtrace"; then
   if ! has 'dtrace' ; then
@@ -4191,9 +4175,6 @@ if test "$trace_backend" = "stderr"; then
   echo "CONFIG_TRACE_STDERR=y" >> $config_host_mak
   trace_default=no
 fi
-if test "$trace_backend" = "ust"; then
-  echo "CONFIG_TRACE_UST=y" >> $config_host_mak
-fi
 if test "$trace_backend" = "dtrace"; then
   echo "CONFIG_TRACE_DTRACE=y" >> $config_host_mak
   if test "$trace_backend_stap" = "yes" ; then
diff --git a/docs/tracing.txt b/docs/tracing.txt
index bfc261b..3d0d620 100644
--- a/docs/tracing.txt
+++ b/docs/tracing.txt
@@ -135,9 +135,8 @@ the following monitor command:
 
 The "tracetool" script automates tedious trace event code generation and also
 keeps the trace event declarations independent of the trace backend.  The trace
-events are not tightly coupled to a specific trace backend, such as LTTng or
-SystemTap.  Support for trace backends can be added by extending the "tracetool"
-script.
+events are not tightly coupled to a specific trace backend, such as SystemTap.
+Support for trace backends can be added by extending the "tracetool" script.
 
 The trace backend is chosen at configure time and only one trace backend can
 be built into the binary:
@@ -208,12 +207,6 @@ You must ensure that the same "trace-events" file was used to build QEMU,
 otherwise trace event declarations may have changed and output will not be
 consistent.
 
-=== LTTng Userspace Tracer ===
-
-The "ust" backend uses the LTTng Userspace Tracer library.  There are no
-monitor commands built into QEMU, instead UST utilities should be used to list,
-enable/disable, and dump traces.
-
 === SystemTap ===
 
 The "dtrace" backend uses DTrace sdt probes but has only been tested with
diff --git a/scripts/tracetool/backend/ust.py b/scripts/tracetool/backend/ust.py
deleted file mode 100644
index ea36995..0000000
--- a/scripts/tracetool/backend/ust.py
+++ /dev/null
@@ -1,93 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-"""
-LTTng User Space Tracing backend.
-"""
-
-__author__     = "Lluís Vilanova <vilanova@ac.upc.edu>"
-__copyright__  = "Copyright 2012, Lluís Vilanova <vilanova@ac.upc.edu>"
-__license__    = "GPL version 2 or (at your option) any later version"
-
-__maintainer__ = "Stefan Hajnoczi"
-__email__      = "stefanha@linux.vnet.ibm.com"
-
-
-from tracetool import out
-
-
-PUBLIC = True
-
-
-def c(events):
-    out('#include <ust/marker.h>',
-        '#undef mutex_lock',
-        '#undef mutex_unlock',
-        '#undef inline',
-        '#undef wmb',
-        '#include "trace.h"')
-
-    for e in events:
-        argnames = ", ".join(e.args.names())
-        if len(e.args) > 0:
-            argnames = ', ' + argnames
-
-            out('DEFINE_TRACE(ust_%(name)s);',
-                '',
-                'static void ust_%(name)s_probe(%(args)s)',
-                '{',
-                '    trace_mark(ust, %(name)s, %(fmt)s%(argnames)s);',
-                '}',
-                name = e.name,
-                args = e.args,
-                fmt = e.fmt,
-                argnames = argnames,
-                )
-
-        else:
-            out('DEFINE_TRACE(ust_%(name)s);',
-                '',
-                'static void ust_%(name)s_probe(%(args)s)',
-                '{',
-                '    trace_mark(ust, %(name)s, UST_MARKER_NOARGS);',
-                '}',
-                name = e.name,
-                args = e.args,
-                )
-
-    # register probes
-    out('',
-        'static void __attribute__((constructor)) trace_init(void)',
-        '{')
-
-    for e in events:
-        out('    register_trace_ust_%(name)s(ust_%(name)s_probe);',
-            name = e.name,
-            )
-
-    out('}')
-
-
-def h(events):
-    out('#include <ust/tracepoint.h>',
-        '#undef mutex_lock',
-        '#undef mutex_unlock',
-        '#undef inline',
-        '#undef wmb')
-
-    for e in events:
-        if len(e.args) > 0:
-            out('DECLARE_TRACE(ust_%(name)s, TP_PROTO(%(args)s), TP_ARGS(%(argnames)s));',
-                '#define trace_%(name)s trace_ust_%(name)s',
-                name = e.name,
-                args = e.args,
-                argnames = ", ".join(e.args.names()),
-                )
-
-        else:
-            out('_DECLARE_TRACEPOINT_NOARGS(ust_%(name)s);',
-                '#define trace_%(name)s trace_ust_%(name)s',
-                name = e.name,
-                )
-
-    out()
-- 
1.8.3.1

Re: [Qemu-devel] [lttng-dev] [PATCH] trace: drop LTTng Userspace Tracer backend

[Mohamad Gebai]

Hi Stefan,

I am actually using LTTng 2.x as a backend for UST to do some 
performance analysis and latency investigation using Qemu/KVM. I already 
have all the patches ready to replace the old 0.x interface, and I am 
preparing them for the merge upstream.

Re: [Qemu-devel] [lttng-dev] [PATCH] trace: drop LTTng Userspace Tracer backend

[Stefan Hajnoczi]

On Wed, Sep 25, 2013 at 12:34:26PM -0400, Mohamad Gebai wrote:
> I am actually using LTTng 2.x as a backend for UST to do some
> performance analysis and latency investigation using Qemu/KVM. I
> already have all the patches ready to replace the old 0.x interface,
> and I am preparing them for the merge upstream.

Excellent, I was hoping to find someone who wants to update the code.

Do you need the old 0.x code for your patches or is it cleaner if we
apply my patch to drop that first?  I guess you pretty much rewrote
the ./configure and tracetool pieces...

Stefan

Re: [Qemu-devel] [PATCH] trace: drop LTTng Userspace Tracer backend

[Alex Bennée]

stefanha@redhat.com writes:

> The current LTTng Userspace Tracer backend does not build against modern
> libraries.  LTTng has changed the library ABI several times, making it
> difficult to support this backend.

Looks good to me.

> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Reviewed by: Alex Bennée <alex@bennee.com>

-- 
Alex Bennée

Re: [Qemu-devel] [lttng-dev] [PATCH] trace: drop LTTng Userspace Tracer backend

[Mohamad Gebai]

Hi Stefan,
Sorry for the delay but I rewrote all the patches to make the merge as
non-intrusive as possible, which wasn't the case in the patches I wrote 
for my
personal use.
I submitted them in a different thread on the mailing list.

Mohamad

On 13-09-26 04:01 AM, Stefan Hajnoczi wrote:
> On Wed, Sep 25, 2013 at 12:34:26PM -0400, Mohamad Gebai wrote:
>> I am actually using LTTng 2.x as a backend for UST to do some
>> performance analysis and latency investigation using Qemu/KVM. I
>> already have all the patches ready to replace the old 0.x interface,
>> and I am preparing them for the merge upstream.
> Excellent, I was hoping to find someone who wants to update the code.
>
> Do you need the old 0.x code for your patches or is it cleaner if we
> apply my patch to drop that first?  I guess you pretty much rewrote
> the ./configure and tracetool pieces...
>
> Stefan