[PATCH] HAL: Ignore internal DM devices with new DM udev rules

[PATCH] HAL: Ignore internal DM devices with new DM udev rules

[Milan Broz]

With new device-mapper udev rules are /dev/mapper/* symlinks
to basic device name /dev/dm-X.
(Change requested by udev upstream.)

This change breaks temporary-cryptsetup workaround inside hal.

With new dm-udev rules (uncluded since device-mapper 1.02.39)
there is DM_UDEV_DISABLE_OTHER_RULES_FLAG variable
which controls that scan should be ignored for this device
(it is set for all internal devices, including temporary cryptsetup,
internal parts of lvm devices etc.)

Ignore device if this flag is set.

See bugs
https://bugzilla.redhat.com/show_bug.cgi?ida3909
http://bugs.debian.org/cgi-bin/bugreport.cgi?bugX6286

Signed-off-by: Milan Broz <mbroz@redhat.com>
---
 hald/linux/osspec.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/hald/linux/osspec.c b/hald/linux/osspec.c
index a8233fd..6d616bc 100644
--- a/hald/linux/osspec.c
+++ b/hald/linux/osspec.c
@@ -215,6 +215,11 @@ hald_udev_data (GIOChannel *source, GIOCondition condition, gpointer user_data)
 				g_free (str);
 			}
 			g_free (dstr);
+		} else if (strncmp(key, "DM_UDEV_DISABLE_OTHER_RULES_FLAG=", 33) = 0) {
+			if (strtoul(&key[33], NULL, 10) = 1) {
+				HAL_INFO (("ignoring device requested by DM udev rules"));
+				goto invalid;
+			}
 		}
 	}
 
-- 
1.7.1

Re: [PATCH] HAL: Ignore internal DM devices with new DM udev rules

[Robby Workman]

[-- Attachment #1: Type: text/plain, Size: 1501 bytes --]

On Tue, 13 Jul 2010 14:08:14 +0200
Milan Broz <mbroz@redhat.com> wrote:

> With new device-mapper udev rules are /dev/mapper/* symlinks
> to basic device name /dev/dm-X.
> (Change requested by udev upstream.)
> 
> This change breaks temporary-cryptsetup workaround inside hal.
> 
> With new dm-udev rules (uncluded since device-mapper 1.02.39)
> there is DM_UDEV_DISABLE_OTHER_RULES_FLAG variable
> which controls that scan should be ignored for this device
> (it is set for all internal devices, including temporary cryptsetup,
> internal parts of lvm devices etc.)
> 
> Ignore device if this flag is set.
> 
> See bugs
> https://bugzilla.redhat.com/show_bug.cgi?id=613909
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=586286


Perhaps this is a problem with how Xfce's automount implementation
calls everything (although it worked fine prior to our move to
udev-enabled LVM2 and newer cryptsetup (1.1.x)), but I still don't
get an encrypted usb thumbdrive mounted when plugging it in after
applying this patch and rebuilding hal.  

I get two separate dialog windows to enter the passphrase
(the first appears, I enter the passphrase, and then the 
second appears), and after entering the passphrase for the 
second time, I get this:

  Failed to setup the encrypted volume.
  /dev/sdb is already setup?

Just for grins and giggles, I tested an old mmc card, and it
works as expected.

I'm attaching a debug log created by running hald in verbose
mode; any ideas on where to go from here?

-RW

[-- Attachment #2: debug --]
[-- Type: application/octet-stream, Size: 10163 bytes --]

[5822]: 04:04:08.000 [I] addon-storage.c:356: Checking whether device /dev/sdb is locked on HAL
[5822]: 04:04:08.000 [I] addon-storage.c:365: ... device /dev/sdb is not locked on HAL
[5822]: 04:04:08.001 [D] addon-storage.c:548: Media insertion detected on /dev/sdb
04:04:08.001 [D] hald_dbus.c:1659: udi=/org/freedesktop/Hal/devices/storage_serial_SanDisk_Cruzer_Micro_SNDK5D1BF51877607705_0_0, key=storage.removable.media_available
04:04:08.001 [I] hald_dbus.c:3115: entering, local_interface=1
04:04:08.002 [D] hald_dbus.c:3124: udi=/org/freedesktop/Hal/devices/storage_serial_SanDisk_Cruzer_Micro_SNDK5D1BF51877607705_0_0
04:04:08.002 [I] blockdev.c:1780: blockdev_rescan_device: udi=/org/freedesktop/Hal/devices/storage_serial_SanDisk_Cruzer_Micro_SNDK5D1BF51877607705_0_0
Run started hald-probe-storage (20000) (0) 
!  full path is '/usr/libexec/hald-probe-storage', program_dir is '/usr/libexec'
woohoo
[5825]: 04:04:08.009 [D] probe-storage.c:142: Doing probe-storage for /dev/sdb (bus usb) (drive_type disk) (udi=/org/freedesktop/Hal/devices/storage_serial_SanDisk_Cruzer_Micro_SNDK5D1BF51877607705_0_0) (--only-check-for-fs==1)
[5825]: 04:04:08.010 [D] probe-storage.c:440: Checking for file system on /dev/sdb
[5825]: 04:04:08.010 [D] probe-storage.c:446: Doing open ("/dev/sdb", O_RDONLY)
[5825]: 04:04:08.010 [D] probe-storage.c:454: Returned from open(2)
[5825]: 04:04:08.010 [D] probe-storage.c:470: look for existing partitions for sdb
04:04:08.023 [I] hald_dbus.c:1371: storage.removable.media_available -> True
04:04:08.023 [I] hald_dbus.c:1355: storage.removable.media_size -> 1024966656
04:04:08.023 [I] hald_dbus.c:1339: storage.partitioning_scheme -> none
pid 5825: rc=2 signaled=0: /usr/libexec/hald-probe-storage
04:04:08.024 [I] blockdev.c:1736: hald-probe-storage --only-check-for-media returned 2 (exit_type=0)
04:04:08.024 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:08.024 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:18.543 [I] hald_dbus.c:5198: OK for method 'Setup' with signature 's' on interface 'org.freedesktop.Hal.Device.Volume.Crypto' for UDI '/org/freedesktop/Hal/devices/volume_uuid_10660ad9_8d12_4c12_bd3e_c45f4960bd43' and execpath 'hal-luks-setup'
04:04:18.543 [I] hald_dbus.c:4082: no need to enqueue
Run started hal-luks-setup (0) (1) 
!  full path is '/usr/libexec/scripts/hal-luks-setup', program_dir is '/usr/libexec/scripts'
04:04:18.550 [I] access-check.c:72: uid for caller is 0
04:04:18.551 [I] ci-tracker.c:212: Removing CICallerInfo object for :1.200
04:04:18.551 [I] device.c:1894: Removing locks from ':1.200'
04:04:18.554 [I] access-check.c:72: uid for caller is 0
04:04:18.555 [I] ci-tracker.c:212: Removing CICallerInfo object for :1.201
04:04:18.555 [I] device.c:1894: Removing locks from ':1.201'
04:04:18.560 [I] device.c:1894: Removing locks from ':1.202'
04:04:19.530 [I] osspec.c:256: SEQNUM=2241, ACTION=add, SUBSYSTEM=bdi, DEVPATH=/sys/devices/virtual/bdi/253:6, DEVNAME=, IFINDEX=0
04:04:19.530 [D] hotplug.c:476: checking ADD event /sys/devices/virtual/bdi/253:6
04:04:19.530 [D] hotplug.c:397: event /sys/devices/virtual/bdi/253:6: skip ourselves and all later events
04:04:19.530 [I] hotplug.c:145: /sys/devices/virtual/bdi/253:6 is a device (subsystem)
04:04:19.530 [I] device.c:4997: add_dev: subsys=bdi sysfs_path=/sys/devices/virtual/bdi/253:6 dev= parent_dev=0x00000000
04:04:19.530 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:19.530 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:19.530 [I] osspec.c:220: ignoring device requested by DM udev rules
04:04:19.531 [I] osspec.c:220: ignoring device requested by DM udev rules
04:04:19.532 [I] osspec.c:220: ignoring device requested by DM udev rules
04:04:19.561 [I] osspec.c:220: ignoring device requested by DM udev rules
04:04:19.563 [I] osspec.c:256: SEQNUM=2245, ACTION=remove, SUBSYSTEM=bdi, DEVPATH=/sys/devices/virtual/bdi/253:6, DEVNAME=, IFINDEX=0
04:04:19.563 [D] hotplug.c:478: checking REMOVE event /sys/devices/virtual/bdi/253:6
04:04:19.563 [D] hotplug.c:397: event /sys/devices/virtual/bdi/253:6: skip ourselves and all later events
04:04:19.563 [I] device.c:5078: remove_dev: subsys=bdi sysfs_path=/sys/devices/virtual/bdi/253:6
04:04:19.563 [W] device.c:5082: Error removing device
04:04:19.563 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:19.563 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:19.564 [I] osspec.c:256: SEQNUM=2246, ACTION=remove, SUBSYSTEM=block, DEVPATH=/sys/devices/virtual/block/dm-6, DEVNAME=dm-6, IFINDEX=0
04:04:19.564 [D] hotplug.c:478: checking REMOVE event /sys/devices/virtual/block/dm-6
04:04:19.564 [D] hotplug.c:397: event /sys/devices/virtual/block/dm-6: skip ourselves and all later events
04:04:19.564 [I] blockdev.c:1663: block_rem: sysfs_path=/sys/devices/virtual/block/dm-6
04:04:19.564 [W] blockdev.c:1667: Device is not in the HAL database
04:04:19.564 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:19.564 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:19.700 [I] osspec.c:220: ignoring device requested by DM udev rules
04:04:19.701 [I] osspec.c:256: SEQNUM=2249, ACTION=add, SUBSYSTEM=bdi, DEVPATH=/sys/devices/virtual/bdi/253:6, DEVNAME=, IFINDEX=0
04:04:19.701 [D] hotplug.c:476: checking ADD event /sys/devices/virtual/bdi/253:6
04:04:19.701 [D] hotplug.c:397: event /sys/devices/virtual/bdi/253:6: skip ourselves and all later events
04:04:19.701 [I] hotplug.c:145: /sys/devices/virtual/bdi/253:6 is a device (subsystem)
04:04:19.701 [I] device.c:4997: add_dev: subsys=bdi sysfs_path=/sys/devices/virtual/bdi/253:6 dev= parent_dev=0x00000000
04:04:19.701 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:19.701 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:19.735 [I] osspec.c:256: SEQNUM=2247, ACTION=change, SUBSYSTEM=block, DEVPATH=/sys/devices/pci0000:00/0000:00:1d.7/usb2/2-1/2-1:1.0/host9/target9:0:0/9:0:0:0/block/sdb, DEVNAME=/dev/sdb, IFINDEX=0
04:04:19.735 [D] hotplug.c:480: checking event /sys/devices/pci0000:00/0000:00:1d.7/usb2/2-1/2-1:1.0/host9/target9:0:0/9:0:0:0/block/sdb, action: 4
04:04:19.735 [D] hotplug.c:397: event /sys/devices/pci0000:00/0000:00:1d.7/usb2/2-1/2-1:1.0/host9/target9:0:0/9:0:0:0/block/sdb: skip ourselves and all later events
04:04:19.735 [I] blockdev.c:1716: block_change: sysfs_path=/sys/devices/pci0000:00/0000:00:1d.7/usb2/2-1/2-1:1.0/host9/target9:0:0/9:0:0:0/block/sdb
04:04:19.735 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:19.735 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:20.064 [D] osspec.c:249: Found a dm-device (/dev/dm-6), mark it
04:04:20.064 [I] osspec.c:256: SEQNUM=2250, ACTION=change, SUBSYSTEM=block, DEVPATH=/sys/devices/virtual/block/dm-6, DEVNAME=/dev/dm-6, IFINDEX=0
04:04:20.064 [D] hotplug.c:480: checking event /sys/devices/virtual/block/dm-6, action: 4
04:04:20.064 [D] hotplug.c:397: event /sys/devices/virtual/block/dm-6: skip ourselves and all later events
04:04:20.065 [I] blockdev.c:904: block_add: sysfs_path=/sys/devices/virtual/block/dm-6 dev=/dev/dm-6 is_part=0, parent=0x00000000
04:04:20.065 [I] blockdev.c:958: Looking in /sys/devices/virtual/block/dm-6/slaves for Device Mapper
04:04:20.065 [I] blockdev.c:969:  /sys/devices/virtual/block/dm-6/slaves/. -> /sys/devices/virtual/block/dm-6/slaves
04:04:20.065 [I] blockdev.c:1003: Couldn't find slave volume in devices
04:04:20.065 [I] blockdev.c:969:  /sys/devices/virtual/block/dm-6/slaves/.. -> /sys/devices/virtual/block/dm-6
04:04:20.065 [I] blockdev.c:1003: Couldn't find slave volume in devices
04:04:20.065 [I] blockdev.c:969:  /sys/devices/virtual/block/dm-6/slaves/sdb -> /sys/devices/pci0000:00/0000:00:1d.7/usb2/2-1/2-1:1.0/host9/target9:0:0/9:0:0:0/block/sdb
04:04:20.065 [I] blockdev.c:1009: Done looking in /sys/devices/virtual/block/dm-6/slaves
04:04:20.065 [I] blockdev.c:1016: Ignoring hotplug event - no parent
04:04:20.065 [W] blockdev.c:1587: Not adding device object
04:04:20.065 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:20.065 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:20.087 [I] device.c:1894: Removing locks from ':1.203'
pid 5826: rc=0 signaled=0: /usr/libexec/scripts/hal-luks-setup
04:04:20.087 [I] hald_dbus.c:4106: No more methods in queue
04:04:20.087 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:20.087 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:04:59.627 [I] hald_dbus.c:5198: OK for method 'Setup' with signature 's' on interface 'org.freedesktop.Hal.Device.Volume.Crypto' for UDI '/org/freedesktop/Hal/devices/volume_uuid_10660ad9_8d12_4c12_bd3e_c45f4960bd43' and execpath 'hal-luks-setup'
04:04:59.627 [I] hald_dbus.c:4082: no need to enqueue
Run started hal-luks-setup (0) (1) 
!  full path is '/usr/libexec/scripts/hal-luks-setup', program_dir is '/usr/libexec/scripts'
04:04:59.632 [I] access-check.c:72: uid for caller is 0
04:04:59.633 [I] ci-tracker.c:212: Removing CICallerInfo object for :1.204
04:04:59.633 [I] device.c:1894: Removing locks from ':1.204'
04:04:59.637 [I] access-check.c:72: uid for caller is 0
04:04:59.638 [I] ci-tracker.c:212: Removing CICallerInfo object for :1.205
04:04:59.638 [I] device.c:1894: Removing locks from ':1.205'
04:04:59.642 [I] device.c:1894: Removing locks from ':1.206'
pid 5866: rc=1 signaled=0: /usr/libexec/scripts/hal-luks-setup
04:04:59.646 [I] hald_dbus.c:4106: No more methods in queue
04:04:59.646 [I] hald_dbus.c:4169: failed with 'org.freedesktop.Hal.Device.Volume.Crypto.SetupError' '/dev/sdb is already setup?'
04:04:59.646 [D] hotplug.c:500: events queued = 0, events in progress = 0
04:04:59.646 [D] hotplug.c:505: Hotplug-queue empty now ... no hotplug events in progress
04:05:18.890 [I] ci-tracker.c:212: Removing CICallerInfo object for :1.199
04:05:18.890 [I] device.c:1894: Removing locks from ':1.199'
04:05:18.970 [I] device.c:1894: Removing locks from ':1.207'

Re: [PATCH] HAL: Ignore internal DM devices with new DM udev rules

[Milan Broz]

On 07/14/2010 11:12 AM, Robby Workman wrote:
> On Tue, 13 Jul 2010 14:08:14 +0200
> Milan Broz <mbroz@redhat.com> wrote:

> I get two separate dialog windows to enter the passphrase
> (the first appears, I enter the passphrase, and then the 
> second appears), and after entering the passphrase for the 
> second time, I get this:

sigh. If udev rule set this flag, it is not top-level device
and nobody should touch it.

Top level device is then visible, that's the one you want to mount...

if you see the log, event comes there for top-level:

04:04:20.064 [D] osspec.c:249: Found a dm-device (/dev/dm-6), mark it
04:04:20.064 [I] osspec.c:256: SEQNUM"50, ACTION=change, SUBSYSTEM=block, DEVPATH=/sys/devices/virtual/block/dm-6, DEVNAME=/dev/dm-6, IFINDEX=0

so I do not think it is problem with this patch.

But in fact I have no idea how the whole desktop machinery above
that works, I expect that there is some race with another auto mount
system - see dmsetup table if the device is not mapped elsewhere already:-)

Milan