[PATCH] udevd - cleanup and better timeout handling
Here is the next revision for udevd:
o Small cleanups all over the place.
o Swich to the usual linked list format "list.h".
o Better timeout handling.
We store a timestamp in in every queued event, so we don't wait longer
than the timeout specified, if the hole in the list is not shrinking.
o ignore udevd target if klibc is used
[PATCH] udev - next round of udev event order daemon
Here is the next round of udevd/udevsend:
udevsend - If the IPC message we send is not catched by a receiver we fork
the udevd daemon to process this and the following events
udevd - We reorder the events we receive and execute our current udev for
every event. If one or more events are missing, we wait
10 seconds and then go ahead in the queue.
If the queue is empty and we don't receive any event for the next
30 seconds, the daemon exits.
The next incoming event will fork the daemon again.
config - The path's to the executable are specified in udevd.h
Now they are pointing to the current directory only.
I don't like daemons hiding secrets (and mem leaks :)) inside,
so I want to try this model. It should be enough logic to get all possible
hotplug events executed in the right order.
If no event, then no daemon! So everybody should be happy :)
Here we see:
1. the daemon fork,
2. the udev work,
3. the 10 sec timeout and the skipped events,
4. the udev work,
...,
5. and the 30 sec timeout and exit.
udevinfo is now capable to print "all" attributes along the device chain
of a sysfs device. Just like udev itself it walks the chain upwards and
prints all usable attributes in the udev key format.
So it should be easy to find unique attributes to compose a rule.
All multiline attribute values and values containing non printable
characters are skipped now. I hope nothing useful gets lost with this :)
NOTE:
The BUS value corresponding with the attributes is printed for every
device. Don't specify BUS= in a rule and mix SYSFS_attributes from
different busses, the rule can't match.
./udevinfo /sys/block/sda/sda1
device '/sys/block/sda/sda1' has major:minor 8:1
looking at class device '/sys/block/sda/sda1':
SYSFS_dev="8:1"
SYSFS_start="32"
SYSFS_size="160"
SYSFS_stat=" 0 0 0 0"
follow the class device's "device"
looking at the device chain at '/sys/devices/pci0000:00/0000:00:1d.0/usb1/1-1/1-1.3/1-1.3:1.0/host32/32:0:0:0':
BUS="scsi"
ID="32:0:0:0"
SYSFS_detach_state="0"
SYSFS_type="0"
SYSFS_model="USB 2 HS-CF"
SYSFS_vendor="SMSC "
SYSFS_max_sectors="240"
SYSFS_device_blocked="0"
SYSFS_queue_depth="1"
SYSFS_scsi_level="3"
SYSFS_rev="1.25"
SYSFS_online="1"
looking at the device chain at '/sys/devices/pci0000:00/0000:00:1d.0/usb1/1-1/1-1.3/1-1.3:1.0/host32':
BUS=""
ID="host32"
SYSFS_detach_state="0"
looking at the device chain at '/sys/devices/pci0000:00/0000:00:1d.0/usb1/1-1/1-1.3/1-1.3:1.0':
BUS="usb"
ID="1-1.3:1.0"
SYSFS_detach_state="0"
SYSFS_bInterfaceNumber="00"
SYSFS_bAlternateSetting=" 0"
SYSFS_bNumEndpoints="02"
SYSFS_bInterfaceClass="08"
SYSFS_bInterfaceSubClass="06"
SYSFS_bInterfaceProtocol="50"
SYSFS_iInterface="00"
looking at the device chain at '/sys/devices/pci0000:00/0000:00:1d.0/usb1/1-1/1-1.3':
BUS="usb"
ID="1-1.3"
SYSFS_detach_state="0"
SYSFS_bNumConfigurations="1"
SYSFS_bNumInterfaces=" 1"
SYSFS_bConfigurationValue="1"
SYSFS_bmAttributes="80"
SYSFS_bMaxPower=" 96mA"
SYSFS_idVendor="0424"
SYSFS_idProduct="20fc"
SYSFS_bcdDevice="0125"
SYSFS_bDeviceClass="00"
SYSFS_bDeviceSubClass="00"
SYSFS_bDeviceProtocol="00"
SYSFS_speed="12"
SYSFS_manufacturer="SMSC"
SYSFS_product="USB 2 Flash Media Device"
SYSFS_serial="0305037000C2"
looking at the device chain at '/sys/devices/pci0000:00/0000:00:1d.0/usb1/1-1':
BUS="usb"
ID="1-1"
SYSFS_detach_state="0"
SYSFS_bNumConfigurations="1"
SYSFS_bNumInterfaces=" 1"
SYSFS_bConfigurationValue="1"
SYSFS_bmAttributes="e0"
SYSFS_bMaxPower=" 64mA"
SYSFS_idVendor="03eb"
SYSFS_idProduct="3301"
SYSFS_bcdDevice="0300"
SYSFS_bDeviceClass="09"
SYSFS_bDeviceSubClass="00"
SYSFS_bDeviceProtocol="00"
SYSFS_speed="12"
SYSFS_product="Standard USB Hub"
On Thu, Jan 22, 2004 at 01:27:45AM +0100, Kay Sievers wrote:
> On Wed, Jan 21, 2004 at 02:38:25PM +0100, Kay Sievers wrote:
> > On Thu, Jan 15, 2004 at 01:45:10PM -0800, Greg KH wrote:
> > > On Thu, Jan 15, 2004 at 10:36:25PM +0800, Ling, Xiaofeng wrote:
> > > > Hi, Greg
> > > > I wrote a simple implementation for the two pieces
> > > > of send and receive hotplug event,
> > > > use a message queue and a list for the out of order
> > > > hotplug event. It also has a timeout timer of 3 seconds.
> > > > They are now separate program. the file nseq is the test script.
> > > > Could you have a look to see wether it is feasible?
> > > > If so, I'll continue to merge with udev.
> > >
> > > Yes, very nice start. Please continue on.
> > >
> > > One minor comment, please stick with the kernel coding style when you
> > > are writing new code for udev.
> >
> > I took the code from Xiaofeng, cleaned the whitespace, renamed some bits,
> > tweaked the debugging, added the udev exec and created a patch for the current tree.
> >
> > It seems functional now, by simply executing our current udev (dirty hack).
> > It reorders the incoming events and if one is missing it delays the
> > execution of the following ones up to a maximum of 10 seconds.
> >
> > Test script is included, but you can't mix hotplug sequence numbers and
> > test script numbers, it will result in waiting for the missing numbers :)
>
> Hey, nobody want's to play with me?
> So here I'm chatting with myself :)
>
> This is the next version with signal handling for resetting the expected
> signal number. I changed the behaviour of the timeout to skip all
> missing events at once and to proceed with the next event in the queue.
>
> So it's now possible to use the test script at any time, cause it resets
> the daemon, if real hotplug event coming in later all missing nimbers will
> be skipped after a timeout of 10 seconds and the queued events are applied.
Here is the next updated updated version to apply to the lastet udev.
I've added infrastructure for getting the state of the IPC queue in the
sender and set the program to exec by the daemon. Also the magic key id
is replaced by the usual key generation by path/nr.
It looks promising, I use it on my machine and my 4in1 USB-flash-reader
connect/disconnect emits the events "randomly" but udevd is able to
reorder it and calls our normal udev in the right order.
[PATCH] udev - program to query all device attributes to build a rule
On Sun, Jan 18, 2004 at 07:30:50AM +0100, Kay Sievers wrote:
> Hi,
> here is a small program to query all attributes of a device and
> print these in the udev key format. It may help to get the keys to
> define a rule.
Fixed a typo - the sysfs_device is right now.
USB FLash Reader:
kay@pim:~/src/udev.kay$ extras/udevinfo/udevinfo /sys/block/sda/sda1
device '/sys/block/sda/sda1' has major:minor 8:1
looking at class device '/sys/block/sda/sda1':
SYSFS_dev="8:1"
SYSFS_start="32"
SYSFS_size="160"
SYSFS_stat=" 0 0 0 0"
follow class device's "device" link '/sys/block/sda':
BUS="scsi"
ID="57:0:0:0"
SYSFS_detach_state="0"
SYSFS_type="0"
SYSFS_device_blocked="0"
SYSFS_queue_depth="1"
SYSFS_scsi_level="3"
SYSFS_vendor="SMSC "
SYSFS_model="USB 2 HS-CF"
SYSFS_rev="1.25"
SYSFS_online="1"
I've edited the man page today, so this is alreay included :)
Also a few more trivials:
o added the defaults to udev.conf.in
o removed class_dev from get_default_mode(), to match with Hanna's
o changed size of mode_str to MODE_SIZE
o changed a few char compares from from 0x00 to '\0'
This patch fixes a bug where the udev database stored empty strings
for Owner and Group if they were default. This patch stores the default
value into the database if not set otherwise. See example output:
crw------- 1 root root 4, 65 Jan 16 11:13 ttyS1
P: /class/tty/ttyS1
N: ttyS1
S:
O: root
G: root
This is a bit of a hack. However, until udev supports setting the
o/g values they will be root/root anyway so the database might as
well reflect the truth instead of empty strings.
On Thu, Jan 15, 2004 at 05:14:16AM +0100, Kay Sievers wrote:
> On Wed, Jan 14, 2004 at 01:10:43PM -0800, Greg KH wrote:
> > On Wed, Jan 14, 2004 at 02:34:26PM -0600, Clay Haapala wrote:
> > > On Wed, 14 Jan 2004, Chris Friesen spake thusly:
> > > >
> > > > Maybe for ones with a matching rule, you could print something like:
> > > >
> > > >
> > > Is the act of printing/syslogging a rule in an of itself?
> >
> > No, as currently the only way stuff ends up in the syslog is if
> > DEBUG=true is used on the build line.
> >
> > But it's sounding like we might want to change that... :)
>
> How about this in the syslog after connect/disconnect?
>
> Jan 15 05:07:45 pim udev[28007]: configured rule in '/etc/udev/udev.rules' at line 17 applied, 'video*' becomes 'video/webcam%n'
> Jan 15 05:07:45 pim udev[28007]: creating device node '/udev/video/webcam0'
> Jan 15 05:07:47 pim udev[28015]: removing device node '/udev/video/webcam0'
Here is a slightly better version. I've created a logging.h file and
moved the debug macros from udev.h in there.
If you type:
'make' - you will get a binary that prints one or two lines to syslog
if a device node is created or deleted
'make LOG=false' - you get a binary that prints asolutely nothing
'make DEBUG=true' - the same as today, it will print all debug lines
I've attached a patch that adds a few rules to udev.rules.devfs, making it
look a little more like devfs on my system. (I have the sysfs patches from
2.6.1-rc1-mm2) I added rules for oss, misc, floppy, and input devices. The
oss rules look like trouble with a wildcard at the end of each name, but I'm
not sure how I can make it any better.
Devfs has a bunch of other devices in the floppy directory for floppies
formatted in unusual ways, but I don't see them in udev. Not that I ever used
them, since they're usually automatically detected.
On Mon, Jan 12, 2004 at 05:04:45PM -0800, Greg KH wrote:
> Very nice, applied. But I did have to make one small change to get the
> code to build properly with klibc:
>
> > +static void print_record(char *path, struct udevice *dev)
> > +{
> > + printf("P: %s\n", path);
> > + printf("N: %s\n", dev->name);
> > + printf("S: %s\n", dev->symlink);
> > + printf("O: %s\n", dev->owner);
> > + printf("G: %s\n", dev->group);
> > + printf("\n");
> > +}
>
> Turns out that gcc likes to convert single character printf() calls to
> putchar() which is only defined in klibc as a macro :(
Just for information. This seems to fix the gcc with klibc :)
> Hi,
> as promised yesterday, here is a patch that drops the explicit methods
> given in the udev config and implement only one type of rule.
>
> A rule now consists only of a number of keys to match. All known keys
> are valid in any combination. The former configs should work with a few
> changes:
>
> o the "<METHOD>, " at the beginning of the line should be removed
>
> o the result of the externel program is matched with RESULT= instead if ID=
> the PROGRAM= key is only valid if the program exits with zero
> (just exit with nozero in a script if the rule should not match)
>
> o rules are processed in order they appear in the file, no priority
>
> o if NAME="" is given, udev is instructed to ignore this device,
> no node will be created
>
>
> EXAMPLE:
>
> # combined BUS, SYSFS and KERNEL
> BUS="usb", KERNEL="video*", SYSFS_model="Creative Labs WebCam*", NAME="test/webcam%n"
>
> # exec script only for the first ide drive (hda), all other will be skipped
> BUS="ide", KERNEL="hda*", PROGRAM="/home/kay/src/udev.kay/extras/ide-devfs.sh %k %b %n", RESULT="hd*", NAME="%1c", SYMLINK="%2c %3c"
>
>
> The udev-test.pl and test.block works fine here.
> Please adapt your config and give it a try.
>
Here is a slightly better version of the patch.
After a conversation with Patrick, we are now able to execute the PROGRAM
and also match in all following rules with the RESULT value from this exec.
EXAMPLE:
We have 7 rules with RESULT and 2 with PROGRAM.
Only the 5th rule matches with the callout result from the exec in the 4th rule.
RESULT:
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: process rule
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check PROGRAM
Jan 11 23:36:52 pim udev[26050]: execute_program: executing '/bin/echo abc'
Jan 11 23:36:52 pim udev[26050]: execute_program: result is 'abc'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: PROGRAM returned successful
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for RESULT dev->result='no_match', udev->program_result='abc'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: RESULT is not matching
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: process rule
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for KERNEL dev->kernel='video*' class_dev->name='video0'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: KERNEL matches
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for RESULT dev->result='123', udev->program_result='abc'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: RESULT is not matching
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: process rule
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for KERNEL dev->kernel='video*' class_dev->name='video0'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: KERNEL matches
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for RESULT dev->result='123', udev->program_result='abc'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: RESULT is not matching
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: process rule
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check PROGRAM
Jan 11 23:36:52 pim udev[26050]: execute_program: executing '/bin/echo 123'
Jan 11 23:36:52 pim udev[26050]: execute_program: result is '123'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: PROGRAM returned successful
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for RESULT dev->result='no_match', udev->program_result='123'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: RESULT is not matching
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: process rule
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for KERNEL dev->kernel='video*' class_dev->name='video0'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: KERNEL matches
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: check for RESULT dev->result='123', udev->program_result='123'
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: RESULT matches
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: found matching rule, 'video*' becomes ''
Jan 11 23:36:52 pim udev[26050]: namedev_name_device: name, 'web-yes' is going to have owner='', group='', mode = 0600
This patch improves the user options for udev.
It is possible now to query for the name, the symlinks or owner/group.
If asked for the name of the node we are able to prepend the udev_root
with the -r option.
SAMPLE:
kay@pim:~/src/udev.test$ ./udev -V
udev, version 012_bk
kay@pim:~/src/udev.test$ ./udev -h
Usage: [-qrVh]
-q <name> query database for the specified value
-p <path> device path used for query
-r print udev root
-V print udev version
-h print this help text
kay@pim:~/src/udev.test$ ./udev -r
/udev/
kay@pim:~/src/udev.test$ ./udev -q name -p /class/video4linux/video0
video/webcam0
I had too much time during the holidays, so I played a bit with udev. The
changes are like last time mostly on the init stuff. I'm sending you this as
a great diff which is just for comments.
What it does:
-fix a typo in Makefile
-use only one "grep -v" instead of many
-don't include BK-Files into release (shrinks the stuff to 30%!)
-add a new init script which is LSB compliant
-add some flags to choose which one to use
-use /etc/udev/udev.conf in Redhat init script as the source for the udev
directory. If this is not done then the init script may create a directory
which udev itself isn't using (I changed /udev to /Udev to avoid collisions
with /usr and ran into this)
-first check for sysfs_dir before creating udev_root (maybe someone else has
already fixed this, I saw this discussion on lkml)
Hey, this is funny.
I couldn't resist to give it a try and we need a few changes:
- it's %2c otherwise nearly all my CD's are "good", but sure I also have bad ones :)
- remove the node first, cause get_cddb() dies and leaves the old one there
- remove spaces in name, cause this is our separator
Hello,
without this patch:
wolf@duel:/tmp/ud/udev-012>gcc -v
Reading specs from /usr/lib/gcc-lib/i386-slackware-linux/2.95.3/specs
gcc version 2.95.3 20010315 (release)
cannot compile udev.
On Wed, Dec 31, 2003 at 11:24:53AM -0800, Greg KH wrote:
> > There should be a possibility to tell udev not to create a device node.
> >
> > device-mapper: Usually set up by libdevmapper (or EVMS tools) which
> > creates the device node on its own under /dev/mapper/<name>.
> >
> > With udev a second device is created named /dev/dm-<minor> which is not
> > really needed.
>
> Good point. Ok, I'll agree with you. Care to make up a patch for this
> kind of feature?
Yes, I can try.
There was no way to tell not to do anything so I created one. Errors
are signalled via negative return values, so I thought that a positive,
non-zero one could mean to ignore the device. I don't like it but
perhaps you have a better solution.
This one is nothing important, just add some quotes to be more consistent
with the rest and make sure that the return value is positive (since the
error return values are negative). Hmm?
[PATCH] make udev user callable to query the database
Here is a slightly better version that prints the usage if a unknown option is given:
kay@pim:~/src/udev.kay$ ./udev -x
./udev: invalid option -- x
Usage: [-qrVh]
-q arg query database
-r print udev root
-V print udev version
-h print this help text
> Here is a patch that makes it possible to call udev with options on the command line.
> Valid options are for now:
>
> -V for the udev version:
> kay@pim:~/src/udev.kay$ ./udev -V
> udev, version 011_bk
>
> -r for the udev root:
> kay@pim:~/src/udev.kay$ ./udev -r
> /udev/
>
> -q to query the database with the sysfs path for the name of the node:
> kay@pim:~/src/udev.kay$ ./udev -q /class/video4linux/video0
> test/video/webcam0
[PATCH] make symlink work properly if there is already a file in its place
If a file that is not a symlink (node, socket, fifo, etc) already
exist where udev need to create a symlink, symlink() fails. This
patch basically test for an existing file, and unlink it.
Attached is a patch for ide-devfs.sh,
The script is merged with the one from Martin Schlemmer, and cleaned up
by him, to create both types of symlinks with one single rule:
projects / systemd / log