Skip to contentSkip to conversation

D-BUS patch for udev-008

PATCH David Zeuthen 3 participants 5 messages Dec 7, 2003 – Dec 10, 2003 Kbuild

lore
Share & tools
lore
#0 lore raw

Hi,

Attached is a patch against udev-008 to send out a D-BUS message when a device node is added or removed.

Using D-BUS lingo, udev acquires the org.kernel.udev service and sends out a NodeCreated or NodeDeleted signal on the org.kernel.udev.NodeMonitor interface. Each signal carries two
parameters: the node in question and the corresponding sysfs path.

[Note: the D-BUS concepts of service, interface, object can be a bit confusing at first glance]

An example program listening for these messages looks like this

#!/usr/bin/python

import dbus
import gtk

def udev_signal_received(dbus_iface, member, service, object_path, message):
[filename, sysfs_path] = message.get_args_list()
if member=='NodeCreated':
print 'Node %s created for %s'%(filename, sysfs_path)
elif member=='NodeDeleted':
print 'Node %s deleted for %s'%(filename, sysfs_path)

def main():
bus = dbus.Bus(dbus.Bus.TYPE_SYSTEM) bus.add_signal_receiver(udev_signal_received,
'org.kernel.udev.NodeMonitor', # interface
'org.kernel.udev', # service
'/org/kernel/udev/NodeMonitor') # object

gtk.mainloop()

if __name__ == '__main__':

    main()

and this is the output when hot-plugging some usb-storage.

[david@laptop udev-008]$ ~/node_monitor.py
Node /udev/sda created for /block/sda
Node /udev/sda1 created for /block/sda/sda1
Node /udev/sda1 deleted for /block/sda/sda1
Node /udev/sda deleted for /block/sda

The patch requires D-BUS 0.20 or later while the python example program requires D-BUS from CVS as I only recently applied a patch against the python bindings.

Cheers,
David

Makefile renamed from orig/udev-008/Makefile +27 -2
diff -Nru ../orig/udev-008/Makefile ./Makefile
@@ -20,6 +20,10 @@
 # Leave this set to `false' for production use.
 DEBUG = false
 
+# Set the following to `true' to make udev emit a D-BUS signal when a
+# new node is created.
+USE_DBUS = true
+
 
 ROOT =		udev
 VERSION =	008
@@ -34,6 +38,7 @@
 sbindir =	${exec_prefix}/sbin
 mandir =	${prefix}/usr/share/man
 hotplugdir =	${etcdir}/hotplug.d/default
+dbusdir =	${etcdir}/dbus-1/system.d
 configdir =	${etcdir}/udev/
 srcdir = .
 
@@ -133,6 +138,13 @@
 	LDFLAGS = --static 
 endif
 
+ifeq ($(USE_DBUS), true)
+	CFLAGS += -DUSE_DBUS
+	CFLAGS += $(shell pkg-config --cflags dbus-1)
+	LIB_OBJS += $(shell pkg-config --libs-only-l dbus-1)
+endif
+
+
 all: $(ROOT)
 
 $(ROOT): $(LIBC)
@@ -225,7 +237,20 @@
 	@echo "Built $(RELEASE_NAME).tar.gz"
 
 
-install: all
+ifeq ($(USE_DBUS), true)
+install-dbus-policy:
+	$(INSTALL) -d $(DESTDIR)$(dbusdir)
+	$(INSTALL_DATA) udev_sysbus_policy.conf $(DESTDIR)$(dbusdir)
+uninstall-dbus-policy:
+	- rm $(DESTDIR)$(dbusdir)/udev_sysbus_policy.conf
+else
+install-dbus-policy:
+	-
+uninstall-dbus-policy:
+	-
+endif
+
+install: install-dbus-policy all
 	$(INSTALL) -d $(DESTDIR)$(udevdir)
 	$(INSTALL) -d $(DESTDIR)$(configdir)
 	$(INSTALL) -d $(DESTDIR)$(hotplugdir)
@@ -237,7 +262,7 @@
 	- rm -f $(DESTDIR)$(hotplugdir)/udev.hotplug
 	- ln -s $(sbindir)/$(ROOT) $(DESTDIR)$(hotplugdir)/udev.hotplug
 
-uninstall:
+uninstall: uninstall-dbus-policy
 	- rm $(hotplugdir)/udev.hotplug
 	- rm $(configdir)/udev.permissions
 	- rm $(configdir)/udev.rules
TODO renamed from orig/udev-008/TODO +0 -1
diff -Nru ../orig/udev-008/TODO ./TODO
@@ -21,7 +21,6 @@
   will have an upgrade path.
 - do early boot logic (putting udev into initramfs, handle pivot-root,
   etc.)
-- add hooks to call D-BUS when new node is created or removed
 - lots of other stuff...
 - actually use the BUS= value to determine where the LABEL rule should look
   (right now it's ignored, and we only look in the current sysfs directory.)
udev-add.c renamed from orig/udev-008/udev-add.c +42 -0

54 lines not on this page · raw message

udev.c renamed from orig/udev-008/udev.c +64 -0

82 lines not on this page · raw message

udev.h renamed from orig/udev-008/udev.h +9 -0
diff -Nru ../orig/udev-008/udev.h ./udev.h
@@ -92,4 +92,13 @@
 extern char udev_rules_filename[PATH_MAX+NAME_MAX];
 extern char default_mode_str[NAME_MAX];
 
+#ifdef USE_DBUS
+
+#define DBUS_API_SUBJECT_TO_CHANGE
+#include <dbus/dbus.h>
+
+extern DBusConnection* sysbus_connection;
+
+#endif /* USE_DBUS */
+
 #endif
udev-remove.c renamed from orig/udev-008/udev-remove.c +44 -2

59 lines not on this page · raw message

udev_sysbus_policy.conf renamed from orig/udev-008/udev_sysbus_policy.conf +23 -0
diff -Nru ../orig/udev-008/udev_sysbus_policy.conf ./udev_sysbus_policy.conf
@@ -0,0 +1,23 @@
+<!DOCTYPE busconfig PUBLIC
+ "-//freedesktop//DTD D-BUS Bus Configuration 1.0//EN"
+ "http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd">
+<busconfig>
+
+  <!-- This configuration file specifies the required security policies
+       for udev to work. -->
+
+  <!-- Only root can own the udev service and only root can use the
+       org.kernel.udev.NodeMonitor interface to send signals -->
+  <policy user="root">
+    <allow own="org.kernel.udev"/>
+
+    <allow send_interface="org.kernel.udev.NodeMonitor"/>
+  </policy>
+
+  <!-- Allow anyone to listen to the org.kernel.udev.NodeMonitor interface 
+       for messages send from the owner of the org.kernel.udev service -->
+  <policy context="default">
+    <allow receive_interface="org.kernel.udev.NodeMonitor"/>
+  </policy>
+</busconfig>
+
#1 lore raw
Quoted 2 lines · David Zeuthen

Attached is a patch against udev-008 to send out a D-BUS message when a device node is added or removed.

Rock!

Good job, David.

Rob Love

-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id78&alloc_id371&op=click
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

#2 lore raw
In reply to David Zeuthen
Quoted 4 lines · David Zeuthen

Hi,

Attached is a patch against udev-008 to send out a D-BUS message when a device node is added or removed.

Very nice. I've applied your patch, and then made the following changes:
- fixed the formatting to use tabs instead of spaces
- removed all of the #ifdef USE_DBUS statments out of the code (I've been writing kernel code too long, #ifdef within .c files looks ugly to me...) and created a new file udev_dbus.c to hold these functions.
- changed the Makefile variable to DBUS instead of USE_DBUS
- turned the default option to not build the DBUS support as I can't get it to link properly on a number of different OSs.

How did you test this? I get lots of errors at link time using a Fedora 1 machine. And I can't even build DBUS on an old Red Hat 7.3 box due to automake issues :(

I've checked all of this into the bk tree if anyone else wants to play with it.

thanks,

greg k-h

-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id78&alloc_id371&op=click
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

#3 lore raw
In reply to David Zeuthen
Quoted 6 lines · Greg KH
Quoted 4 lines · David Zeuthen

Hi,

Attached is a patch against udev-008 to send out a D-BUS message when a device node is added or removed.

Quoted 11 lines

Very nice. I've applied your patch, and then made the following changes:
- fixed the formatting to use tabs instead of spaces
- removed all of the #ifdef USE_DBUS statments out of the code (I've been writing kernel code too long, #ifdef within .c files looks ugly to me...) and created a new file udev_dbus.c to hold these functions.
- changed the Makefile variable to DBUS instead of USE_DBUS
- turned the default option to not build the DBUS support as I can't get it to link properly on a number of different OSs.

Many thanks for this.

Quoted 5 lines

How did you test this? I get lots of errors at link time using a Fedora 1 machine. And I can't even build DBUS on an old Red Hat 7.3 box due to automake issues :(

I'm using Fedora 1 with D-BUS from CVS installed in /usr/local. I do get a few message from the linker because of static linking, see below, but it does link. What's the reason for linking statically?

Btw, Fedora 1 ships with D-BUS 0.13 which is way different from D-BUS in CVS, so you might need to set PKG_CONFIG_PATH to point to the correct dbus-1.pc and install udev_sysbus_policy.conf into the proper location.

Oh, and even though you have done this, you might still link with the old D-BUS because I made a mistake in the Makefile. You need to change

  LIB_OBJS += $(shell pkg-config --libs-only-l dbus-1)

to something like

  LDFLAGS += $(shell pkg-config --libs dbus-1)

yes? (I was tweaking the Makefile quite a lot myself; sorry about that)

I've also found a problem with my patch: There's an issue where when two instances of udev is running at the same time (processing /block/sda and /block/sda/sda1 when plugging in usb-storage).

One of the copies become a secondary owner of the org.kernel.udev service and this means that the emitted D-BUS signal doesn't appear to come from the owner of that service as I thought it would. I'm not sure if this is the intended behavior of D-BUS, it might be..

The only reason it works in python is that, uhm, the python bindings are a bit broken. It only works in C if you ignore filtering on service owner.

Cheers,
David

[david@laptop udev-008]$ touch udev.c
[david@laptop udev-008]$ make
gcc -pipe -Wall -Wshadow -Wstrict-prototypes -Wmissing-prototypes
-Wmissing-declarations -Os -fomit-frame-pointer -D_GNU_SOURCE
-I/usr/lib/gcc-lib/i386-redhat-linux/3.3.2/include -DUSE_DBUS
-I/usr/local/include/dbus-1.0 -I/usr/local/lib/dbus-1.0/include -c
-o udev.o udev.c
udev.c:75: warning: function declaration isn't a prototype
udev.c:85: warning: function declaration isn't a prototype
gcc --static -o udev udev.o udev_config.o udev-add.o udev-remove.o
udevdb.o logging.o namedev.o namedev_parse.o libsysfs/sysfs_bus.o
libsysfs/sysfs_class.o libsysfs/sysfs_device.o libsysfs/sysfs_dir.o
libsysfs/sysfs_driver.o libsysfs/sysfs_utils.o libsysfs/dlist.o
tdb/tdb.o tdb/spinlock.o -lc -ldbus-1
udev-add.o(.text+0x30d): In function `create_node':
: warning: Using 'getgrnam' in statically linked applications requires
at runtime the shared libraries from the glibc version used for linking
udev-add.o(.text+0x2cb): In function `create_node':
: warning: Using 'getpwnam' in statically linked applications requires
at runtime the shared libraries from the glibc version used for linking
/usr/local/lib/libdbus-1.a(dbus-sysdeps.o)(.text+0x14e0): In function
`fill_user_info':
/home/david/xdg-hacking/dbus/dbus/dbus-sysdeps.c:1504: warning: Using
'getgrouplist' in statically linked applications requires at runtime the
shared libraries from the glibc version used for linking
/usr/local/lib/libdbus-1.a(dbus-sysdeps.o)(.text+0x1482):/home/david/xdg-hacking/dbus/dbus/dbus-sysdeps.c:1467: warning: Using 'getpwuid' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking /usr/local/lib/libdbus-1.a(dbus-sysdeps.o)(.text+0x6e3): In function
`_dbus_connect_tcp_socket':
/home/david/xdg-hacking/dbus/dbus/dbus-sysdeps.c:617: warning: Using
'gethostbyname' in statically linked applications requires at runtime
the shared libraries from the glibc version used for linking
strip -s --remove-section=.note --remove-section=.comment udev

-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id78&alloc_id371&op=click
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

#4 lore raw
In reply to David Zeuthen
Quoted 3 lines · David Zeuthen

I'm using Fedora 1 with D-BUS from CVS installed in /usr/local. I do get a few message from the linker because of static linking, see below, but it does link. What's the reason for linking statically?

udev will need to be able to go into initramfs for early boot. We will not have access to glibc at that time (well, we could, but it would be a lot of overkill...)

Quoted 3 lines

Btw, Fedora 1 ships with D-BUS 0.13 which is way different from D-BUS in CVS, so you might need to set PKG_CONFIG_PATH to point to the correct dbus-1.pc and install udev_sysbus_policy.conf into the proper location.

Ok, thanks. I'll go try that later.

Quoted 11 lines

Oh, and even though you have done this, you might still link with the old D-BUS because I made a mistake in the Makefile. You need to change

  LIB_OBJS += $(shell pkg-config --libs-only-l dbus-1)

to something like

  LDFLAGS += $(shell pkg-config --libs dbus-1)

yes? (I was tweaking the Makefile quite a lot myself; sorry about that)

Thanks, I've applied this to the repo.

Quoted 4 lines

I've also found a problem with my patch: There's an issue where when two instances of udev is running at the same time (processing /block/sda and /block/sda/sda1 when plugging in usb-storage).

Quoted 8 lines

One of the copies become a secondary owner of the org.kernel.udev service and this means that the emitted D-BUS signal doesn't appear to come from the owner of that service as I thought it would. I'm not sure if this is the intended behavior of D-BUS, it might be..

The only reason it works in python is that, uhm, the python bindings are a bit broken. It only works in C if you ignore filtering on service owner.

Hm, eventually udev will turn into a daemon, and only have 1 version running at a time. The events will be queued up into the udev daemon so you will not have this problem.

thanks,

greg k-h

-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id78&alloc_id371&op=click
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel