[ltt-dev] [PATCH] LTTng: Make mark-control work in debugfs
Mathieu Desnoyers
compudj at krystal.dyndns.org
Wed Feb 11 00:32:21 EST 2009
* Gui Jianfeng (guijianfeng at cn.fujitsu.com) wrote:
> Mathieu Desnoyers wrote:
> >
> > Hrm, I'm playing a bit with the interface, and I think I should wait
> > before I merge it. There are a few things to fix.
> >
> > First, when enabling a marker with the old ltt-marker-control.ko, if we
> > rmmod the module containing this marker and after modprobe the module
> > again, the marker stays enabled. With this debugfs interface, it seems
> > to become disabled.
>
> Hi Mathieu,
>
> When enabling a marker with either way(old proc or debugfs), if this marker
> belongs a module, I'm afraid this module can't be unloaded any more.
> For example:
> #modprobe fs_trace
> #echo "connect fs close default" > /proc/ltt
> #rmmod fs_trace
> We'll get:
> ERROR: Module fs_trace is in use
>
> I tested under lttng-0.91, do i miss something?
>
That's expected when a marker is a trace_mark_tp. That's the simplest
way I found to deal with references. As I recall, a trace_mark() (not
_tp) should not have this side-effect. So basically, it's mostly the
lttng probes which have this refcounting.
> >
> > Also, when I connect to a marker, there is some module refcounting
> > involved to make sure the marked modules stays loaded. For some reason
> > I've been able to unload a module with armed markers.. this should
> > probably be looked at.
>
> Do you mean even if the marker is enabled(by using debugfs or proc interface),
> the module which includes this marker are still able to be unloaded?
>
See the above paragraph about explanation around trace_mark_tp(). So
basically, I expect that lttng module containing trace_mark_tp() should
not be unloadable when armed.
> >
> > Thirdly, when I do
> >
> > cd /mnt/debugfs/ltt/markers
> > for a in */*/enable; do echo $a; echo 1 > $a; done
> >
> > I get :
> > metadata/core_marker_format
> > -su: echo: write error: File exists
> > metadata/core_marker_id
> > -su: echo: write error: File exists
> >
> > Which is expected, because those marker are activated within the kernel
> > and should not have their state modified by the user-visible API.
> >
> > Then I do :
> > for a in */*/enable; do echo $a; echo 0 > $a; done
> >
> > And do the first command again :
> > for a in */*/enable; do echo $a; echo 1 > $a; done
> >
> > This shows that userspace is able to alter the state of builtin "core"
> > markers, which I disallow in ltt-marker-control.ko. The same would be
> > good in the debugfs interface.
>
> I tried on lttng-0.91
> #echo "disconnect metadata core_marker_format" > /proc/ltt
> #cat /proc/ltt | grep core_marker_format
> i got the following:
> channel: metadata marker: core_marker_format format: "channel %s name %s format %s" state: 0 event_id: 1 call: 0xc0460b7c probe single : 0xc045f904
>
> It seems core_marker_format has been disabled by old interface. Do i miss something again?
>
> ltt-armall do the restriction, consider the follwing piece of code
> MARKERS=`cat /proc/ltt|grep -v %k|awk '{print $2 " " $4}'|sort -u|grep -v ^metadata
> | grep -v ^locking|grep -v ^lockdep`
> Actually, this block of code do the restriction from userspace.
>
Ah, this is possible. Then we have to make sure the core event IDs are
never allocated dynamically.
I am pulling your new interface for 2.6.29-rc4, we'll start from there.
As you say, it provides the same functionnality as the old interface. We
might want to test the cases I spotted. They might well be problems that
were there in the previous interface as well.
Thanks,
Mathieu
> >
> > I'm providing the patch modified to fit with checkpatch below.
> >
> > Thanks,
> >
> > Mathieu
> >
> >
> > lttng-make-mark-control-work-in-debugfs
> >
> > This patch makes mark-control work in debugfs
> > Control files are organized as following.
> > ltt/
> > |-- markers
> > | |-- fs
> > | | |-- buffer_wait_end
> > | | | |-- enable
> > | | | `-- info
> > | | |-- buffer_wait_start
> > | | | |-- enable
> > | | | `-- info
> > | | |-- close
> > | | | |-- enable
> > | | | `-- info
> > | | |-- exec
> > | | | |-- enable
> > | | | `-- info
> > ...
> >
> > One can enable or disable a marker by using the following command
> > # echo 1 > ltt/markers/fs/close/enable
> > # echo 0 > ltt/markers/fs/close/enable
> >
> > Show the marker information as following
> > # cat ltt/markers/fs/close/info
> > format: "fd %u"
> > state: 1
> > event_id: 2
> > call: 0xc0459118
> > probe single : 0xc04f7b08
> >
> > Mathieu :
> >
> > enabled -> enable to follow the semantic of the rest of the debugfs API.
> > checkpatch.pl cleanup.
> >
> > Signed-off-by: Gui Jianfeng <guijianfeng at cn.fujitsu.com>
> > Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers at polymtl.ca>
> > ---
> > ltt/ltt-trace-control.c | 298 +++++++++++++++++++++++++++++++++++++++++++++++-
> > 1 file changed, 294 insertions(+), 4 deletions(-)
> >
> > Index: linux-2.6-lttng/ltt/ltt-trace-control.c
> > ===================================================================
> > --- linux-2.6-lttng.orig/ltt/ltt-trace-control.c 2009-02-04 00:56:05.000000000 -0500
> > +++ linux-2.6-lttng/ltt/ltt-trace-control.c 2009-02-04 01:01:46.000000000 -0500
> > @@ -1,8 +1,10 @@
> > /*
> > * LTT trace control module over debugfs.
> > *
> > - * Copyright 2008 -
> > - * Zhaolei <zhaolei at cn.fujitsu.com>
> > + * Copyright 2008 - Zhaolei <zhaolei at cn.fujitsu.com>
> > + *
> > + * Copyright 2009 - Gui Jianfeng <guijianfeng at cn.fujitsu.com>
> > + * Make mark-control work in debugfs
> > */
> >
> > /*
> > @@ -16,14 +18,17 @@
> > #include <linux/uaccess.h>
> > #include <linux/debugfs.h>
> > #include <linux/ltt-tracer.h>
> > +#include <linux/notifier.h>
> >
> > #define LTT_CONTROL_DIR "control"
> > +#define MARKERS_CONTROL_DIR "markers"
> > #define LTT_SETUP_TRACE_FILE "setup_trace"
> > #define LTT_DESTROY_TRACE_FILE "destroy_trace"
> >
> > #define LTT_WRITE_MAXLEN (128)
> >
> > -struct dentry *ltt_control_dir, *ltt_setup_trace_file, *ltt_destroy_trace_file;
> > +struct dentry *ltt_control_dir, *ltt_setup_trace_file, *ltt_destroy_trace_file,
> > + *markers_control_dir;
> >
> > /*
> > * the traces_lock nests inside control_lock.
> > @@ -692,6 +697,270 @@ static struct file_operations ltt_destro
> > .write = destroy_trace_write,
> > };
> >
> > +static int marker_enable_open(struct inode *inode, struct file *filp)
> > +{
> > + filp->private_data = inode->i_private;
> > + return 0;
> > +}
> > +
> > +static ssize_t marker_enable_read(struct file *filp, char __user *ubuf,
> > + size_t cnt, loff_t *ppos)
> > +{
> > + struct marker *marker;
> > + char *buf;
> > + int len;
> > +
> > + marker = (struct marker *)filp->private_data;
> > + buf = kmalloc(1024, GFP_KERNEL);
> > +
> > + len = sprintf(buf, "%d\n", _imv_read(marker->state));
> > +
> > + len = simple_read_from_buffer(ubuf, cnt, ppos, buf, len);
> > + kfree(buf);
> > +
> > + return len;
> > +}
> > +
> > +static ssize_t marker_enable_write(struct file *filp, const char __user *ubuf,
> > + size_t cnt, loff_t *ppos)
> > +{
> > + char buf[NAME_MAX];
> > + int buf_size;
> > + int err = 0;
> > + struct marker *marker;
> > +
> > + marker = (struct marker *)filp->private_data;
> > + buf_size = min(cnt, sizeof(buf) - 1);
> > + err = copy_from_user(buf, ubuf, buf_size);
> > + if (err)
> > + return err;
> > +
> > + buf[buf_size] = 0;
> > +
> > + switch (buf[0]) {
> > + case 'Y':
> > + case 'y':
> > + case '1':
> > + err = ltt_marker_connect(marker->channel, marker->name,
> > + "default");
> > + if (err)
> > + return err;
> > + break;
> > + case 'N':
> > + case 'n':
> > + case '0':
> > + err = ltt_marker_disconnect(marker->channel, marker->name,
> > + "default");
> > + if (err)
> > + return err;
> > + break;
> > + default:
> > + return -EPERM;
> > + }
> > +
> > + return cnt;
> > +}
> > +
> > +static const struct file_operations enable_fops = {
> > + .open = marker_enable_open,
> > + .read = marker_enable_read,
> > + .write = marker_enable_write,
> > +};
> > +
> > +static int marker_info_open(struct inode *inode, struct file *filp)
> > +{
> > + filp->private_data = inode->i_private;
> > + return 0;
> > +}
> > +
> > +static ssize_t marker_info_read(struct file *filp, char __user *ubuf,
> > + size_t cnt, loff_t *ppos)
> > +{
> > + struct marker *marker;
> > + char *buf;
> > + int len;
> > +
> > + marker = (struct marker *)filp->private_data;
> > + buf = kmalloc(1024, GFP_KERNEL);
> > +
> > + len = sprintf(buf, "format: \"%s\"\nstate: %d\n"
> > + "event_id: %hu\n"
> > + "call: 0x%p\n"
> > + "probe %s : 0x%p\n",
> > + marker->format, _imv_read(marker->state),
> > + marker->event_id, marker->call, marker->ptype ?
> > + "multi" : "single", marker->ptype ?
> > + (void *)marker->multi : (void *)marker->single.func);
> > +
> > + len = simple_read_from_buffer(ubuf, cnt, ppos, buf, len);
> > + kfree(buf);
> > +
> > + return len;
> > +}
> > +
> > +static const struct file_operations info_fops = {
> > + .open = marker_info_open,
> > + .read = marker_info_read,
> > +};
> > +
> > +static int build_marker_file(struct marker *marker)
> > +{
> > + struct dentry *channel_d, *marker_d, *enable_d, *info_d;
> > + int err;
> > +
> > + channel_d = dir_lookup(markers_control_dir, marker->channel);
> > + if (!channel_d) {
> > + channel_d = debugfs_create_dir(marker->channel,
> > + markers_control_dir);
> > + if (IS_ERR(channel_d) || !channel_d) {
> > + printk(KERN_ERR
> > + "%s: build channel dir of %s failed\n",
> > + __func__, marker->channel);
> > + err = -ENOMEM;
> > + goto err_build_fail;
> > + }
> > + }
> > +
> > + marker_d = dir_lookup(channel_d, marker->name);
> > + if (!marker_d) {
> > + marker_d = debugfs_create_dir(marker->name, channel_d);
> > + if (IS_ERR(marker_d) || !marker_d) {
> > + printk(KERN_ERR
> > + "%s: marker dir of %s failed\n",
> > + __func__, marker->name);
> > + err = -ENOMEM;
> > + goto err_build_fail;
> > + }
> > + }
> > +
> > + enable_d = dir_lookup(marker_d, "enable");
> > + if (!enable_d) {
> > + enable_d = debugfs_create_file("enable", 0644, marker_d,
> > + marker, &enable_fops);
> > + if (IS_ERR(enable_d) || !enable_d) {
> > + printk(KERN_ERR
> > + "%s: create file of %s failed\n",
> > + __func__, "enable");
> > + err = -ENOMEM;
> > + goto err_build_fail;
> > + }
> > + }
> > +
> > + info_d = dir_lookup(marker_d, "info");
> > + if (!info_d) {
> > + info_d = debugfs_create_file("info", 0444, marker_d,
> > + marker, &info_fops);
> > + if (IS_ERR(info_d) || !info_d) {
> > + printk(KERN_ERR
> > + "%s: create file of %s failed\n",
> > + __func__, "enable");
> > + err = -ENOMEM;
> > + goto err_build_fail;
> > + }
> > + }
> > +
> > + return 0;
> > +
> > +err_build_fail:
> > + return err;
> > +}
> > +
> > +static int build_marker_control_files(void)
> > +{
> > + struct marker_iter iter;
> > + int err;
> > +
> > + err = 0;
> > + if (!markers_control_dir)
> > + return -EEXIST;
> > +
> > + marker_iter_reset(&iter);
> > + marker_iter_start(&iter);
> > + for (; iter.marker != NULL; marker_iter_next(&iter)) {
> > + err = build_marker_file(iter.marker);
> > + if (err)
> > + goto err_build_fail;
> > + }
> > + marker_iter_stop(&iter);
> > + return 0;
> > +
> > +err_build_fail:
> > + return err;
> > +}
> > +
> > +static int remove_marker_control_dir(struct marker *marker)
> > +{
> > + struct dentry *channel_d, *marker_d;
> > +
> > + channel_d = dir_lookup(markers_control_dir, marker->channel);
> > + if (!channel_d)
> > + return -ENOENT;
> > +
> > + marker_d = dir_lookup(channel_d, marker->name);
> > + if (!marker_d)
> > + return -ENOENT;
> > +
> > + debugfs_remove_recursive(marker_d);
> > + if (list_empty(&channel_d->d_subdirs))
> > + debugfs_remove(channel_d);
> > +
> > + return 0;
> > +}
> > +
> > +static void cleanup_control_dir(struct marker *begin, struct marker *end)
> > +{
> > + struct marker *iter;
> > +
> > + if (!markers_control_dir)
> > + return;
> > +
> > + for (iter = begin; iter < end; iter++)
> > + remove_marker_control_dir(iter);
> > +
> > + return;
> > +}
> > +
> > +static void build_control_dir(struct marker *begin, struct marker *end)
> > +{
> > + struct marker *iter;
> > + int err;
> > +
> > + err = 0;
> > + if (!markers_control_dir)
> > + return;
> > +
> > + for (iter = begin; iter < end; iter++) {
> > + err = build_marker_file(iter);
> > + if (err)
> > + goto err_build_fail;
> > + }
> > +
> > + return;
> > +err_build_fail:
> > + cleanup_control_dir(begin, end);
> > +}
> > +
> > +static int module_notify(struct notifier_block *self,
> > + unsigned long val, void *data)
> > +{
> > + struct module *mod = data;
> > +
> > + switch (val) {
> > + case MODULE_STATE_COMING:
> > + build_control_dir(mod->markers,
> > + mod->markers + mod->num_markers);
> > + break;
> > + case MODULE_STATE_GOING:
> > + cleanup_control_dir(mod->markers,
> > + mod->markers + mod->num_markers);
> > + break;
> > + }
> > + return NOTIFY_DONE;
> > +}
> > +
> > +static struct notifier_block module_nb = {
> > + .notifier_call = module_notify,
> > +};
> >
> > static int __init ltt_trace_control_init(void)
> > {
> > @@ -733,8 +1002,27 @@ static int __init ltt_trace_control_init
> > goto err_create_destroy_trace_file;
> > }
> >
> > - return 0;
> > + markers_control_dir = debugfs_create_dir(MARKERS_CONTROL_DIR,
> > + ltt_root_dentry);
> > + if (IS_ERR(markers_control_dir) || !markers_control_dir) {
> > + printk(KERN_ERR
> > + "ltt_channel_control_init: create dir of %s failed\n",
> > + MARKERS_CONTROL_DIR);
> > + err = -ENOMEM;
> > + goto err_create_marker_control_dir;
> > + }
> >
> > + if (build_marker_control_files())
> > + goto err_build_fail;
> > +
> > + if (!register_module_notifier(&module_nb))
> > + return 0;
> > +
> > +err_build_fail:
> > + debugfs_remove_recursive(markers_control_dir);
> > + markers_control_dir = NULL;
> > +err_create_marker_control_dir:
> > + debugfs_remove(ltt_destroy_trace_file);
> > err_create_destroy_trace_file:
> > debugfs_remove(ltt_setup_trace_file);
> > err_create_setup_trace_file:
> > @@ -759,6 +1047,8 @@ static void __exit ltt_trace_control_exi
> > debugfs_remove(ltt_setup_trace_file);
> > debugfs_remove(ltt_destroy_trace_file);
> > debugfs_remove_recursive(ltt_control_dir);
> > + debugfs_remove_recursive(markers_control_dir);
> > + unregister_module_notifier(&module_nb);
> > }
> >
> > module_init(ltt_trace_control_init);
> >
> >
>
> --
> Regards
> Gui Jianfeng
>
>
>
> _______________________________________________
> ltt-dev mailing list
> ltt-dev at lists.casi.polymtl.ca
> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>
--
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68
More information about the lttng-dev
mailing list