eudev/test/udev-test.pl

1474 lines
36 KiB
Perl
Raw Normal View History

#!/usr/bin/perl
# udev-test
#
# Provides automated testing of the udev binary.
# The whole test is self contained in this file, except the matching sysfs tree.
# Simply extend the @tests array, to add a new test variant.
#
# Every test is driven by its own temporary config file.
# This program prepares the environment, creates the config and calls udev.
#
# udev parses the rules, looks at the provided sysfs and
# first creates and then removes the device node.
# After creation and removal the result is checked against the
# expected value and the result is printed.
#
# Kay Sievers <kay.sievers@vrfy.org>, 2003
# Leann Ogasawara <ogasawara@osdl.org>, 2004
use warnings;
use strict;
my $PWD = $ENV{PWD};
my $sysfs = "sys/";
my $udev_bin = "../udev";
my $udev_root = "udev-root/"; # !!! directory will be removed !!!
my $udev_db = ".udevdb";
my $udev_conf = "udev-test.conf";
my $udev_rules = "udev-test.rules";
2005-04-27 13:15:56 +07:00
# uncomment following line to run udev with valgrind.
# Should make this a runtime option to the script someday...
#my $udev_bin = "valgrind --tool=memcheck --leak-check=yes ../udev";
my @tests = (
{
desc => "label test of scsi disc",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "boot_disk" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME="boot_disk%n"
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "label test of scsi partition",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda1",
exp_name => "boot_disk1" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME="boot_disk%n"
EOF
},
{
desc => "label test of pattern match",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda1",
exp_name => "boot_disk1" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="?IBM-ESXS", NAME="boot_disk%n-1"
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS?", NAME="boot_disk%n-2"
BUS=="scsi", SYSFS{vendor}=="IBM-ES??", NAME="boot_disk%n"
BUS=="scsi", SYSFS{vendor}=="IBM-ESXSS", NAME="boot_disk%n-3"
EOF
},
{
desc => "label test of multiple sysfs files",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda1",
exp_name => "boot_disk1" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", SYSFS{model}=="ST336605LW !#", NAME="boot_diskX%n"
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", SYSFS{model}=="ST336605LW !#", NAME="boot_disk%n"
EOF
},
{
desc => "label test of max sysfs files",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda1",
exp_name => "boot_disk1" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", SYSFS{model}=="ST336605LW !#", SYSFS{scsi_level}=="4", SYSFS{rev}=="B245", SYSFS{type}=="2", SYSFS{queue_depth}=="32", NAME="boot_diskXX%n"
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", SYSFS{model}=="ST336605LW !#", SYSFS{scsi_level}=="4", SYSFS{rev}=="B245", SYSFS{type}=="0", NAME="boot_disk%n"
EOF
},
{
desc => "catch device by *",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor/0" ,
rules => <<EOF
KERNEL=="ttyUSB*", NAME="visor/%n"
EOF
},
{
desc => "catch device by * - take 2",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor/0" ,
rules => <<EOF
KERNEL=="*USB1", NAME="bad"
KERNEL=="*USB0", NAME="visor/%n"
EOF
},
{
desc => "catch device by ?",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor/0" ,
rules => <<EOF
KERNEL=="ttyUSB??*", NAME="visor/%n-1"
KERNEL=="ttyUSB??", NAME="visor/%n-2"
KERNEL=="ttyUSB?", NAME="visor/%n"
EOF
},
{
desc => "catch device by character class",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor/0" ,
rules => <<EOF
KERNEL=="ttyUSB[A-Z]*", NAME="visor/%n-1"
KERNEL=="ttyUSB?[0-9]", NAME="visor/%n-2"
KERNEL=="ttyUSB[0-9]*", NAME="visor/%n"
EOF
},
{
desc => "replace kernel name",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "Handle comment lines in config file (and replace kernel name)",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
# this is a comment
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "Handle comment lines in config file with whitespace (and replace kernel name)",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
# this is a comment with whitespace before the comment
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "Handle whitespace only lines (and replace kernel name)",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "whitespace" ,
rules => <<EOF
# this is a comment with whitespace before the comment
KERNEL=="ttyUSB0", NAME="whitespace"
EOF
},
{
desc => "Handle empty lines in config file (and replace kernel name)",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "Handle backslashed multi lines in config file (and replace kernel name)",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
KERNEL=="ttyUSB0", \\
NAME="visor"
EOF
},
{
desc => "preserve backslashes, if they are not for a newline",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "aaa",
rules => <<EOF
KERNEL=="ttyUSB0", PROGRAM=="/bin/echo -e \\101", RESULT=="A", NAME="aaa"
EOF
},
{
desc => "Handle stupid backslashed multi lines in config file (and replace kernel name)",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
#
\\
\\\\
#\\
KERNEL=="ttyUSB0", \\
NAME="visor"
EOF
},
{
desc => "subdirectory handling",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "sub/direct/ory/visor" ,
rules => <<EOF
KERNEL=="ttyUSB0", NAME="sub/direct/ory/visor"
EOF
},
{
desc => "place on bus of scsi partition",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "first_disk3" ,
rules => <<EOF
BUS=="scsi", ID=="0:0:0:0", NAME="first_disk%n"
EOF
},
{
desc => "test NAME substitution chars",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "Major:8:minor:3:kernelnumber:3:bus:0:0:0:0" ,
rules => <<EOF
BUS=="scsi", ID=="0:0:0:0", NAME="Major:%M:minor:%m:kernelnumber:%n:bus:%b"
EOF
},
{
desc => "test NAME substitution chars (with length limit)",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "M8-m3-n3-b0:0-sIBM" ,
rules => <<EOF
BUS=="scsi", ID=="0:0:0:0", NAME="M%M-m%m-n%n-b%3b-s%3s{vendor}"
EOF
},
{
desc => "sustitution of sysfs value (%s{file})",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "disk-IBM-ESXS-sda" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME="disk-%s{vendor}-%k"
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "program result substitution",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "special-device-3" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n special-device", RESULT=="-special-*", NAME="%c-1-%n"
BUS=="scsi", PROGRAM=="/bin/echo -n special-device", RESULT=="special--*", NAME="%c-2-%n"
BUS=="scsi", PROGRAM=="/bin/echo -n special-device", RESULT=="special-device-", NAME="%c-3-%n"
BUS=="scsi", PROGRAM=="/bin/echo -n special-device", RESULT=="special-devic", NAME="%c-4-%n"
BUS=="scsi", PROGRAM=="/bin/echo -n special-device", RESULT=="special-*", NAME="%c-%n"
EOF
},
{
desc => "program result substitution (no argument should be subsystem)",
subsys => "block",
devpath => "/block/sda/sda3",
exp_name => "subsys_block" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo", RESULT=="block", NAME="subsys_block"
[PATCH] netdev - udevdb+dev.d changes Here is a patch to change the netdev handling in the database and for the dev.d/ calls. I applies on top of the udevd.patch, cause klibc has no sysinfo(). o netdev's are also put into our database now. I want this for the udevruler gui to get a list of all handled devices. All devices in the db are stamped with the system uptime value at the creation time. 'udevinfo -d' prints it. o the DEVPATH value is the key for udevdb, but if we rename a netdev, the name is replaced in the kernel, so we add the changed name to the db to match with the remove event. NOTE: The dev.d/ scripts still get the original name from the hotplug call. Should we replace DEVPATH with the new name too? o We now only add a device to the db, if we have successfully created the main node or successfully renamed a netdev. This is the main part of the patch, cause I needed to clean the retval passing trough all the functions used for node creation. o DEVNODE sounds a bit ugly for netdev's so I exported DEVNAME too. Can we change the name? o I've added a UDEV_NO_DEVD to possibly skip the script execution and used it in udev-test.pl. udevstart is the same horror now, if you have scripts with logging statements in dev.d/ it takes minutes to finish, can we skip the scripts here too? o The get_device_type() function is changed to be more strict, cause 'udevinfo -a -p /block/' gets a class device for it and tries to print the major/minor values. o bugfix, the RESULT value has now a working newline removal and a test for this case.
2004-04-01 14:12:57 +07:00
EOF
},
{
desc => "program result substitution (newline removal)",
subsys => "block",
devpath => "/block/sda/sda3",
exp_name => "newline_removed" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo test", RESULT=="test", NAME="newline_removed"
EOF
},
{
desc => "program result substitution",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "test-0:0:0:0" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n test-%b", RESULT=="test-0:0*", NAME="%c"
EOF
},
{
desc => "program with escaped format char (tricky: callout returns format char!)",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "escape-3" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n escape-%%n", KERNEL=="sda3", NAME="%c"
EOF
},
{
desc => "program with lots of arguments",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "foo9" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n foo3 foo4 foo5 foo6 foo7 foo8 foo9", KERNEL=="sda3", NAME="%c{7}"
EOF
},
{
desc => "program with subshell",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "bar9" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/sh -c 'echo foo3 foo4 foo5 foo6 foo7 foo8 foo9 | sed s/foo9/bar9/'", KERNEL=="sda3", NAME="%c{7}"
EOF
},
{
desc => "program arguments combined with apostrophes",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "foo7" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n 'foo3 foo4' 'foo5 foo6 foo7 foo8'", KERNEL=="sda3", NAME="%c{5}"
EOF
},
{
desc => "characters before the %c{N} substitution",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "my-foo9" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n foo3 foo4 foo5 foo6 foo7 foo8 foo9", KERNEL=="sda3", NAME="my-%c{7}"
EOF
},
{
desc => "substitute the second to last argument",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda/sda3",
exp_name => "my-foo8" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n foo3 foo4 foo5 foo6 foo7 foo8 foo9", KERNEL=="sda3", NAME="my-%c{6}"
EOF
},
{
desc => "invalid program for device with no bus",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/console",
exp_name => "TTY" ,
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n foo", RESULT=="foo", NAME="foo"
KERNEL=="console", NAME="TTY"
EOF
},
{
desc => "valid program for device with no bus",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/console",
exp_name => "foo" ,
rules => <<EOF
PROGRAM=="/bin/echo -n foo", RESULT=="foo", NAME="foo"
KERNEL=="console", NAME="TTY"
EOF
},
{
desc => "invalid label for device with no bus",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/console",
exp_name => "TTY" ,
rules => <<EOF
BUS=="foo", SYSFS{dev}=="5:1", NAME="foo"
KERNEL=="console", NAME="TTY"
EOF
},
{
desc => "valid label for device with no bus",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/console",
exp_name => "foo" ,
rules => <<EOF
SYSFS{dev}=="5:1", NAME="foo"
KERNEL=="console", NAME="TTY"
EOF
},
{
desc => "program and bus type match",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "scsi-0:0:0:0" ,
rules => <<EOF
BUS=="usb", PROGRAM=="/bin/echo -n usb-%b", NAME="%c"
BUS=="scsi", PROGRAM=="/bin/echo -n scsi-%b", NAME="%c"
BUS=="foo", PROGRAM=="/bin/echo -n foo-%b", NAME="%c"
EOF
},
{
desc => "create all possible partitions",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "boot_disk15" ,
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME{all_partitions}="boot_disk"
EOF
},
{
desc => "sysfs parent hierarchy",
subsys => "tty",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/tty/ttyUSB0",
exp_name => "visor" ,
rules => <<EOF
SYSFS{idProduct}=="2008", NAME="visor"
EOF
},
{
desc => "name test with ! in the name",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/rd!c0d0",
exp_name => "rd/c0d0" ,
rules => <<EOF
BUS=="scsi", NAME="%k"
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "name test with ! in the name, but no matching rule",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/rd!c0d0",
exp_name => "rd/c0d0" ,
rules => <<EOF
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "name test with ! in the name for a partition",
subsys => "block",
devpath => "/block/cciss!c0d0/cciss!c0d0p1",
exp_name => "cciss/c0d0p1" ,
rules => <<EOF
BUS=="scsi", NAME="%k"
KERNEL=="ttyUSB0", NAME="visor"
EOF
},
{
desc => "ID rule",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "scsi-0:0:0:0",
rules => <<EOF
BUS=="usb", ID=="0:0:0:0", NAME="not-scsi"
BUS=="scsi", ID=="0:0:0:1", NAME="no-match"
BUS=="scsi", ID==":0", NAME="short-id"
BUS=="scsi", ID=="/0:0:0:0", NAME="no-match"
BUS=="scsi", ID=="0:0:0:0", NAME="scsi-0:0:0:0"
EOF
},
{
desc => "ID wildcard all",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "scsi-0:0:0:0",
rules => <<EOF
BUS=="scsi", ID=="*:1", NAME="no-match"
BUS=="scsi", ID=="*:0:1", NAME="no-match"
BUS=="scsi", ID=="*:0:0:1", NAME="no-match"
BUS=="scsi", ID=="*", NAME="scsi-0:0:0:0"
BUS=="scsi", ID=="0:0:0:0", NAME="bad"
EOF
},
{
desc => "ID wildcard partial",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "scsi-0:0:0:0",
rules => <<EOF
BUS=="scsi", ID=="*:0", NAME="scsi-0:0:0:0"
BUS=="scsi", ID=="0:0:0:0", NAME="bad"
EOF
},
{
desc => "ID wildcard partial 2",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "scsi-0:0:0:0",
rules => <<EOF
BUS=="scsi", ID=="*:0:0:0", NAME="scsi-0:0:0:0"
BUS=="scsi", ID=="0:0:0:0", NAME="bad"
EOF
},
{
desc => "ignore SYSFS attribute whitespace",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "ignored",
rules => <<EOF
BUS=="scsi", SYSFS{whitespace_test}=="WHITE SPACE", NAME="ignored"
EOF
},
{
desc => "do not ignore SYSFS attribute whitespace",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "matched-with-space",
rules => <<EOF
BUS=="scsi", SYSFS{whitespace_test}=="WHITE SPACE ", NAME="wrong-to-ignore"
BUS=="scsi", SYSFS{whitespace_test}=="WHITE SPACE ", NAME="matched-with-space"
EOF
},
{
desc => "permissions USER=bad GROUP=name",
subsys => "tty",
devpath => "/class/tty/tty33",
exp_name => "tty33",
exp_perms => "0:0:0660",
rules => <<EOF
KERNEL=="tty33", NAME="tty33", OWNER="bad", GROUP="name"
EOF
},
{
desc => "permissions OWNER=5000",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "node",
exp_perms => "5000::0660",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OWNER="5000"
EOF
},
{
desc => "permissions GROUP=100",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => ":100:0660",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", GROUP="100"
EOF
},
{
desc => "textual user id",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => "nobody::0660",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OWNER="nobody"
EOF
},
{
desc => "textual group id",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => ":daemon:0660",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", GROUP="daemon"
EOF
},
{
desc => "textual user/group id",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => "root:mail:0660",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OWNER="root", GROUP="mail"
EOF
},
{
desc => "permissions MODE=0777",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => "::0777",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", MODE="0777"
EOF
},
{
desc => "permissions OWNER=5000 GROUP=100 MODE=0777",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_perms => "5000:100:0777",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OWNER="5000", GROUP="100", MODE="0777"
EOF
},
{
desc => "permissions OWNER to 5000",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "5000::",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", OWNER="5000"
EOF
},
{
desc => "permissions GROUP to 100",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => ":100:0660",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", GROUP="100"
EOF
},
{
desc => "permissions MODE to 0060",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "::0060",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", MODE="0060"
EOF
},
{
desc => "permissions OWNER, GROUP, MODE",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "5000:100:0777",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", OWNER="5000", GROUP="100", MODE="0777"
EOF
},
{
desc => "permissions only rule",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "5000:100:0777",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", OWNER="5000", GROUP="100", MODE="0777"
KERNEL=="ttyUSX[0-9]*", OWNER="5001", GROUP="101", MODE="0444"
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n"
EOF
},
{
desc => "multiple permissions only rule",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "3000:4000:0777",
rules => <<EOF
SUBSYSTEM=="tty", OWNER="3000"
SUBSYSTEM=="tty", GROUP="4000"
SUBSYSTEM=="tty", MODE="0777"
KERNEL=="ttyUSX[0-9]*", OWNER="5001", GROUP="101", MODE="0444"
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n"
EOF
},
{
desc => "permissions only rule with override at NAME rule",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "ttyUSB0",
exp_perms => "3000:8000:0777",
rules => <<EOF
SUBSYSTEM=="tty", OWNER="3000"
SUBSYSTEM=="tty", GROUP="4000"
SUBSYSTEM=="tty", MODE="0777"
KERNEL=="ttyUSX[0-9]*", OWNER="5001", GROUP="101", MODE="0444"
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", GROUP="8000"
EOF
},
{
desc => "major/minor number test",
subsys => "block",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/block/sda",
exp_name => "node",
exp_majorminor => "8:0",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node"
EOF
},
{
desc => "big minor number test",
subsys => "i2c-dev",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/i2c-dev/i2c-300",
exp_name => "node",
exp_majorminor => "89:300",
rules => <<EOF
KERNEL=="i2c-300", NAME="node"
EOF
},
{
desc => "big major number test",
subsys => "i2c-dev",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/i2c-dev/i2c-fake1",
exp_name => "node",
exp_majorminor => "4095:1",
rules => <<EOF
KERNEL=="i2c-fake1", NAME="node"
EOF
},
{
desc => "big major and big minor number test",
subsys => "i2c-dev",
[PATCH] hmm, handle net devices with udev? Hmm, Arndt Bergmann sent a patch like this one a few weeks ago and I want to bring the question back, if we want to handle net device naming with udev. With this patch it is actually possible to specify something like this in udev.rules: KERNEL="dummy*", SYSFS{address}="00:00:00:00:00:00", SYSFS{features}="0x0", NAME="blind%n" KERNEL="eth*", SYSFS{address}="00:0d:60:77:30:91", NAME="private" and you will get: [root@pim udev.kay]# cat /proc/net/dev Inter-| Receive | Transmit face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed lo: 1500 30 0 0 0 0 0 0 1500 30 0 0 0 0 0 0 private: 278393 1114 0 0 0 0 0 0 153204 1468 0 0 0 0 0 0 sit0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 blind0: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The udevinfo program is also working: [root@pim udev.kay]# ./udevinfo -a -p /sys/class/net/private looking at class device '/sys/class/net/private': SYSFS{addr_len}="6" SYSFS{address}="00:0d:60:77:30:91" SYSFS{broadcast}="ff:ff:ff:ff:ff:ff" SYSFS{features}="0x3a9" SYSFS{flags}="0x1003" SYSFS{ifindex}="2" SYSFS{iflink}="2" SYSFS{mtu}="1500" SYSFS{tx_queue_len}="1000" SYSFS{type}="1" follow the class device's "device" looking at the device chain at '/sys/devices/pci0000:00/0000:00:1e.0/0000:02:01.0': BUS="pci" ID="0000:02:01.0" SYSFS{class}="0x020000" SYSFS{detach_state}="0" SYSFS{device}="0x101e" SYSFS{irq}="11" SYSFS{subsystem_device}="0x0549" SYSFS{subsystem_vendor}="0x1014" SYSFS{vendor}="0x8086" The matching device will be renamed to the given name. The device name will not be put into the udev database, cause the kernel renames the device and the sysfs name disappears. I like it, cause it plugs in nicely. We have all the naming features and sysfs queries and walks inside of udev. The sysfs timing races are already solved and the management tools are working for net devices too. nameif can only match the MAC address now. udev can match any sysfs value of the device tree the net device is connected to. But right, net devices do not have device nodes :)
2004-03-25 14:19:39 +07:00
devpath => "/class/i2c-dev/i2c-fake2",
exp_name => "node",
exp_majorminor => "4094:89999",
rules => <<EOF
KERNEL=="i2c-fake2", NAME="node"
EOF
},
{
desc => "multiple symlinks with format char",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "symlink2-ttyUSB0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="symlink1-%n symlink2-%k symlink3-%b"
EOF
},
{
desc => "symlink creation (same directory)",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "visor0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="visor%n"
EOF
},
{
desc => "symlink creation (relative link forward)",
subsys => "block",
devpath => "/block/sda/sda2",
exp_name => "1/2/symlink" ,
exp_target => "a/b/node",
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME="1/2/a/b/node", SYMLINK="1/2/symlink"
EOF
},
{
desc => "symlink creation (relative link back and forward)",
subsys => "block",
devpath => "/block/sda/sda2",
exp_name => "1/2/c/d/symlink" ,
exp_target => "../../a/b/node",
rules => <<EOF
BUS=="scsi", SYSFS{vendor}=="IBM-ESXS", NAME="1/2/a/b/node", SYMLINK="1/2/c/d/symlink"
EOF
},
{
desc => "multiple symlinks",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "second-0" ,
exp_target => "visor" ,
rules => <<EOF
KERNEL=="ttyUSB0", NAME="visor", SYMLINK="first-%n second-%n third-%n"
EOF
},
{
desc => "symlink only rule",
subsys => "block",
devpath => "/block/sda",
exp_name => "symlink-only2",
exp_target => "link",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", SYMLINK="symlink-only1"
BUS=="scsi", KERNEL=="sda", SYMLINK="symlink-only2"
BUS=="scsi", KERNEL=="sda", NAME="link", SYMLINK="symlink0"
EOF
},
{
desc => "symlink name '.'",
subsys => "block",
devpath => "/block/sda",
exp_name => ".",
exp_target => "link",
exp_add_error => "yes",
exp_rem_error => "yes",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="link", SYMLINK="."
EOF
},
{
desc => "symlink node to itself",
subsys => "tty",
devpath => "/class/tty/tty0",
exp_name => "link",
exp_target => "link",
exp_rem_error => "yes",
option => "clear",
rules => <<EOF
KERNEL=="tty0", NAME="link", SYMLINK="link"
EOF
},
{
desc => "symlink %n substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "symlink0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="symlink%n"
EOF
},
{
desc => "symlink %k substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "symlink-ttyUSB0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="symlink-%k"
EOF
},
{
desc => "symlink %M:%m substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "major-188:0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="major-%M:%m"
EOF
},
{
desc => "symlink %b substitution",
subsys => "block",
devpath => "/block/sda",
exp_name => "symlink-0:0:0:0",
exp_target => "node",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", SYMLINK="symlink-%b"
EOF
},
{
desc => "symlink %c substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "test",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", PROGRAM=="/bin/echo test" NAME="ttyUSB%n", SYMLINK="%c"
EOF
},
{
desc => "symlink %c{N} substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "test",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", PROGRAM=="/bin/echo symlink test this" NAME="ttyUSB%n", SYMLINK="%c{2}"
EOF
},
{
desc => "symlink %c{N+} substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "this",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", PROGRAM=="/bin/echo symlink test this" NAME="ttyUSB%n", SYMLINK="%c{2+}"
EOF
},
{
desc => "symlink only rule with %c{N+}",
subsys => "block",
devpath => "/block/sda",
exp_name => "test",
exp_target => "link",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", PROGRAM=="/bin/echo link test this" SYMLINK="%c{2+}"
BUS=="scsi", KERNEL=="sda", NAME="link", SYMLINK="symlink0"
EOF
},
{
desc => "symlink %s{filename} substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "188:0",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="%s{dev}"
EOF
},
{
desc => "symlink %Ns{filename} substitution",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "188",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="%3s{dev}"
EOF
},
{
desc => "symlink with '%' in name",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "percent%sign",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="percent%%sign"
EOF
},
{
desc => "symlink with '%' in name",
subsys => "tty",
devpath => "/class/tty/ttyUSB0",
exp_name => "%ttyUSB0_name",
exp_target => "ttyUSB0",
rules => <<EOF
KERNEL=="ttyUSB[0-9]*", NAME="ttyUSB%n", SYMLINK="%%%k_name"
EOF
},
{
desc => "program result substitution (numbered part of)",
subsys => "block",
devpath => "/block/sda/sda3",
exp_name => "link1",
exp_target => "node",
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n node link1 link2", RESULT=="node *", NAME="%c{1}", SYMLINK="%c{2} %c{3}"
EOF
},
{
desc => "program result substitution (numbered part of+)",
subsys => "block",
devpath => "/block/sda/sda3",
exp_name => "link4",
exp_target => "node",
rules => <<EOF
BUS=="scsi", PROGRAM=="/bin/echo -n node link1 link2 link3 link4", RESULT=="node *", NAME="%c{1}", SYMLINK="%c{2+}"
EOF
},
{
desc => "enumeration char test (single test)",
subsys => "block",
devpath => "/block/sda",
exp_name => "cdrom",
rules => <<EOF
KERNEL=="sda", NAME="cdrom%e"
EOF
},
{
desc => "enumeration char test sequence 1/5 (keep)",
subsys => "block",
devpath => "/block/sda",
exp_name => "cdrom",
option => "keep",
rules => <<EOF
KERNEL=="sda", NAME="cdrom%e"
EOF
},
{
desc => "enumeration char test sequence 2/5 (keep)",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "enum",
option => "keep",
rules => <<EOF
KERNEL=="sda1", NAME="enum%e"
EOF
},
{
desc => "enumeration char test sequence 3/5 (keep)",
subsys => "block",
devpath => "/block/sda/sda2",
exp_name => "cdrom1",
option => "keep",
rules => <<EOF
KERNEL=="sda2", NAME="cdrom%e"
EOF
},
{
desc => "enumeration char test sequence 4/5 (keep)",
subsys => "block",
devpath => "/block/sda/sda3",
exp_name => "enum1",
option => "keep",
rules => <<EOF
KERNEL=="sda3", NAME="enum%e"
EOF
},
{
desc => "enumeration char test sequence 5/5 (clean)",
subsys => "block",
devpath => "/block/sda/sda4",
exp_name => "cdrom2",
option => "clear",
rules => <<EOF
KERNEL=="sda4", NAME="cdrom%e"
EOF
},
{
desc => "enumeration char test after cleanup (single test)",
subsys => "block",
devpath => "/block/sda",
exp_name => "cdrom",
rules => <<EOF
KERNEL=="sda", NAME="cdrom%e"
EOF
},
{
desc => "ignore rule test",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_add_error => "yes",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OPTIONS="ignore"
EOF
},
{
desc => "all_partitions, option-only rule",
subsys => "block",
devpath => "/block/sda",
exp_name => "node6",
rules => <<EOF
SUBSYSTEM=="block", OPTIONS="all_partitions"
BUS=="scsi", KERNEL=="sda", NAME="node"
EOF
},
{
desc => "all_partitions, option-only rule (fail on partition)",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "node6",
exp_add_error => "yes",
rules => <<EOF
SUBSYSTEM=="block", OPTIONS="all_partitions"
BUS=="scsi", KERNEL=="sda", NAME="node"
EOF
},
{
desc => "ignore remove event test",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
exp_rem_error => "yes",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OPTIONS="ignore_remove"
EOF
},
{
desc => "ignore remove event test (with all partitions)",
subsys => "block",
devpath => "/block/sda",
exp_name => "node14",
exp_rem_error => "yes",
option => "clear",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="node", OPTIONS="ignore_remove, all_partitions"
EOF
},
{
desc => "SUBSYSTEM match test",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="should_not_match", SUBSYSTEM=="vc"
BUS=="scsi", KERNEL=="sda", NAME="node", SUBSYSTEM=="block"
BUS=="scsi", KERNEL=="sda", NAME="should_not_match2", SUBSYSTEM=="vc"
EOF
},
{
desc => "DRIVER match test",
subsys => "block",
devpath => "/block/sda",
exp_name => "node",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="should_not_match", DRIVER=="sd-wrong"
BUS=="scsi", KERNEL=="sda", NAME="node", DRIVER=="sd"
EOF
},
{
desc => "temporary node creation test",
subsys => "block",
devpath => "/block/sda",
2005-02-11 09:28:17 +07:00
exp_name => "node",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", PROGRAM=="/usr/bin/test -b %N" NAME="node"
EOF
},
{
desc => "devpath substitution test",
subsys => "block",
devpath => "/block/sda",
exp_name => "sda",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", PROGRAM=="/bin/echo %p", RESULT=="/block/sda" NAME="%k"
EOF
},
{
desc => "parent node name substitution test sequence 1/2 (keep)",
subsys => "block",
devpath => "/block/sda",
exp_name => "main_device",
option => "keep",
rules => <<EOF
BUS=="scsi", KERNEL=="sda", NAME="main_device"
EOF
},
{
desc => "parent node name substitution test sequence 2/2 (clean)",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "main_device-part-1",
option => "clean",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", NAME="%P-part-1"
EOF
},
{
desc => "udev_root substitution",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "start-udev-root-end",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", NAME="start-%r-end"
EOF
},
{
desc => "last_rule option",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "last",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", SYMLINK="last", OPTIONS="last_rule"
BUS=="scsi", KERNEL=="sda1", NAME="very-last"
EOF
},
{
desc => "negation KERNEL!=",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "match",
rules => <<EOF
BUS=="scsi", KERNEL!="sda1", NAME="matches-but-is-negated"
BUS=="scsi", KERNEL!="xsda1", NAME="match"
BUS=="scsi", KERNEL=="sda1", NAME="wrong"
EOF
},
{
desc => "negation SUBSYSTEM!=",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "not-anything",
rules => <<EOF
BUS=="scsi", SUBSYSTEM=="block", KERNEL!="sda1", NAME="matches-but-is-negated"
BUS=="scsi", SUBSYSTEM!="anything", NAME="not-anything"
BUS=="scsi", KERNEL=="sda1", NAME="wrong"
EOF
},
{
desc => "negation PROGRAM!= exit code",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "nonzero-program",
rules => <<EOF
KERNEL=="sda1", PROGRAM!="/bin/false", NAME="nonzero-program"
BUS=="scsi", KERNEL=="sda1", NAME="wrong"
EOF
},
{
desc => "test for whitespace between the operator",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "true",
rules => <<EOF
KERNEL == "sda1" , NAME = "true"
BUS=="scsi", KERNEL=="sda1", NAME="wrong"
EOF
},
{
desc => "ENV{} test",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "true",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="go", NAME="wrong"
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="test", NAME="true"
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="bad", NAME="bad"
EOF
},
{
desc => "ENV{} test",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "true",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="go", NAME="wrong"
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="yes", ENV{ACTION}=="add", ENV{DEVPATH}=="/block/sda/sdax1", NAME="no"
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="test", ENV{ACTION}=="add", ENV{DEVPATH}=="/block/sda/sda1", NAME="true"
BUS=="scsi", KERNEL=="sda1", ENV{ENV_KEY_TEST}=="bad", NAME="bad"
EOF
},
{
desc => "untrusted string sanitize",
subsys => "block",
devpath => "/block/sda/sda1",
exp_name => "sane",
rules => <<EOF
BUS=="scsi", KERNEL=="sda1", PROGRAM=="/bin/echo -e name; (/sbin/badprogram)", RESULT="name_ _/sbin/badprogram_", NAME="sane"
EOF
},
{
desc => "read sysfs value from device down in the chain",
subsys => "block",
devpath => "/class/tty/ttyUSB0",
exp_name => "serial-0000:00:09.0",
rules => <<EOF
KERNEL=="ttyUSB*", NAME="serial-%s{serial}"
EOF
},
);
# set env
$ENV{ENV_KEY_TEST} = "test";
$ENV{SYSFS_PATH} = $sysfs;
$ENV{UDEV_CONFIG_FILE} = $udev_conf;
[PATCH] netdev - udevdb+dev.d changes Here is a patch to change the netdev handling in the database and for the dev.d/ calls. I applies on top of the udevd.patch, cause klibc has no sysinfo(). o netdev's are also put into our database now. I want this for the udevruler gui to get a list of all handled devices. All devices in the db are stamped with the system uptime value at the creation time. 'udevinfo -d' prints it. o the DEVPATH value is the key for udevdb, but if we rename a netdev, the name is replaced in the kernel, so we add the changed name to the db to match with the remove event. NOTE: The dev.d/ scripts still get the original name from the hotplug call. Should we replace DEVPATH with the new name too? o We now only add a device to the db, if we have successfully created the main node or successfully renamed a netdev. This is the main part of the patch, cause I needed to clean the retval passing trough all the functions used for node creation. o DEVNODE sounds a bit ugly for netdev's so I exported DEVNAME too. Can we change the name? o I've added a UDEV_NO_DEVD to possibly skip the script execution and used it in udev-test.pl. udevstart is the same horror now, if you have scripts with logging statements in dev.d/ it takes minutes to finish, can we skip the scripts here too? o The get_device_type() function is changed to be more strict, cause 'udevinfo -a -p /block/' gets a class device for it and tries to print the major/minor values. o bugfix, the RESULT value has now a working newline removal and a test for this case.
2004-04-01 14:12:57 +07:00
$ENV{UDEV_NO_DEVD} = "yes";
$ENV{UDEV_NO_HOTPLUGD} = "yes";
sub udev {
my ($action, $subsys, $devpath, $rules) = @_;
$ENV{DEVPATH} = $devpath;
# create temporary rules
open CONF, ">$udev_rules" || die "unable to create rules file: $udev_rules";
print CONF $$rules;
close CONF;
$ENV{ACTION} = $action;
system("$udev_bin $subsys");
}
my $error = 0;
sub permissions_test {
my($rules, $uid, $gid, $mode) = @_;
my $wrong = 0;
my $userid;
my $groupid;
$rules->{exp_perms} =~ m/^(.*):(.*):(.*)$/;
if ($1 ne "") {
if (defined(getpwnam($1))) {
$userid = int(getpwnam($1));
} else {
$userid = $1;
}
if ($uid != $userid) { $wrong = 1; }
}
if ($2 ne "") {
if (defined(getgrnam($2))) {
$groupid = int(getgrnam($2));
} else {
$groupid = $2;
}
if ($gid != $groupid) { $wrong = 1; }
}
if ($3 ne "") {
if (($mode & 07777) != oct($3)) { $wrong = 1; };
}
if ($wrong == 0) {
print "permissions: ok\n";
} else {
printf " expected permissions are: %s:%s:%#o\n", $1, $2, oct($3);
printf " created permissions are : %i:%i:%#o\n", $uid, $gid, $mode & 07777;
print "permissions: error\n";
$error++;
}
}
sub major_minor_test {
my($rules, $rdev) = @_;
my $major = ($rdev >> 8) & 0xfff;
my $minor = ($rdev & 0xff) | (($rdev >> 12) & 0xfff00);
my $wrong = 0;
$rules->{exp_majorminor} =~ m/^(.*):(.*)$/;
if ($1 ne "") {
if ($major != $1) { $wrong = 1; };
}
if ($2 ne "") {
if ($minor != $2) { $wrong = 1; };
}
if ($wrong == 0) {
print "major:minor: ok\n";
} else {
printf " expected major:minor is: %i:%i\n", $1, $2;
printf " created major:minor is : %i:%i\n", $major, $minor;
print "major:minor: error\n";
$error++;
}
}
sub symlink_test {
my ($rules) = @_;
my $output = `ls -l $PWD/$udev_root$rules->{exp_name}`;
if ($output =~ m/(.*)-> (.*)/) {
if ($2 eq $rules->{exp_target}) {
print "symlink: ok\n";
} else {
print " expected symlink from: \'$rules->{exp_name}\' to \'$rules->{exp_target}\'\n";
print " created symlink from: \'$rules->{exp_name}\' to \'$2\'\n";
print "symlink: error";
if ($rules->{exp_add_error}) {
print " as expected\n";
} else {
print "\n";
$error++;
}
}
} else {
print " expected symlink from: \'$rules->{exp_name}\' to \'$rules->{exp_target}\'\n";
print "symlink: not created";
if ($rules->{exp_add_error}) {
print " as expected\n";
} else {
print "\n";
$error++;
}
}
}
sub run_test {
my ($rules, $number) = @_;
print "TEST $number: $rules->{desc}\n";
if ($rules->{exp_target}) {
print "device \'$rules->{devpath}\' expecting symlink '$rules->{exp_name}' to node \'$rules->{exp_target}\'\n";
} else {
print "device \'$rules->{devpath}\' expecting node \'$rules->{exp_name}\'\n";
}
udev("add", $rules->{subsys}, $rules->{devpath}, \$rules->{rules});
if ((-e "$PWD/$udev_root$rules->{exp_name}") ||
(-l "$PWD/$udev_root$rules->{exp_name}")) {
my ($dev, $ino, $mode, $nlink, $uid, $gid, $rdev, $size,
$atime, $mtime, $ctime, $blksize, $blocks) = stat("$PWD/$udev_root$rules->{exp_name}");
if (defined($rules->{exp_perms})) {
permissions_test($rules, $uid, $gid, $mode);
}
if (defined($rules->{exp_majorminor})) {
major_minor_test($rules, $rdev);
}
if (defined($rules->{exp_target})) {
symlink_test($rules);
}
print "add: ok\n";
} else {
print "add: error";
if ($rules->{exp_add_error}) {
print " as expected\n";
} else {
print "\n";
system("tree $udev_root");
print "\n";
$error++;
}
}
if (defined($rules->{option}) && $rules->{option} eq "keep") {
print "\n\n";
return;
}
udev("remove", $rules->{subsys}, $rules->{devpath}, \$rules->{rules});
if ((-e "$PWD/$udev_root$rules->{exp_name}") ||
(-l "$PWD/$udev_root$rules->{exp_name}")) {
print "remove: error";
if ($rules->{exp_rem_error}) {
print " as expected\n";
} else {
print "\n";
system("tree $udev_root");
print "\n";
$error++;
}
} else {
print "remove: ok\n";
}
print "\n";
if (defined($rules->{option}) && $rules->{option} eq "clear") {
system("rm -rf $udev_db");
system("rm -rf $udev_root");
mkdir($udev_root) || die "unable to create udev_root: $udev_root\n";
}
}
# only run if we have root permissions
# due to mknod restrictions
if (!($<==0)) {
print "Must have root permissions to run properly.\n";
exit;
}
# prepare
system("rm -rf $udev_root");
mkdir($udev_root) || die "unable to create udev_root: $udev_root\n";
# create config file
open CONF, ">$udev_conf" || die "unable to create config file: $udev_conf";
print CONF "udev_root=\"$udev_root\"\n";
print CONF "udev_db=\"$udev_db\"\n";
print CONF "udev_rules=\"$udev_rules\"\n";
close CONF;
my $test_num = 1;
if ($ARGV[0]) {
# only run one test
$test_num = $ARGV[0];
if (defined($tests[$test_num-1]->{desc})) {
print "udev-test will run test number $test_num only:\n\n";
run_test($tests[$test_num-1], $test_num);
} else {
print "test does not exist.\n";
}
} else {
# test all
print "\nudev-test will run ".($#tests + 1)." tests:\n\n";
foreach my $rules (@tests) {
run_test($rules, $test_num);
$test_num++;
}
}
print "$error errors occured\n\n";
# cleanup
system("rm -rf $udev_db");
system("rm -rf $udev_root");
unlink($udev_rules);
unlink($udev_conf);