Age | Commit message (Collapse) | Author |
|
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Almost all packagers have always patched the stripping out,
cause the buildsystems takes care of it.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
The rules files are parsed only once at daemon startup. Every udev
event process will be fork()'d from udevd without exec()'ing the udev
binary. The in-memory rules will be inherited from the daemon itself.
If inotify is available, udevd will reload all rules if any change in
/etc/udev/rules.d/ happens. Otherwise -HUP or "udevcontrol reload_rules"
can be used.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
FAT32 volumes should never have a cluster count, that fits
into a 16 bit value, but mkdosfs can create such volumes.
No sane formatter or Windows will ever do this, but the
Linux kernel as Windows can read/write it.
Thanks to Ted Ts'o <tytso@mit.edu> for convincing me.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
|
|
Recent changes to the sysfs layout introduced class-devices pointed to
by a symlink, instead of the real object at that location.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Taken from Red Hat CVS.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Netlink events get lost when the kernel creates thousends of events
faster than udevd reads it. The default is 128 KB, which can carry
app. 500 events. Set it to 16 MB now.
I have 4000 fibrechannel LUNs connected to my system. There are two
paths to the devices and two ports on the host connected via a switch.
This gives 16000 when probed.
I have had problems getting all of the entries in /dev created.
-- Mark Haverkamp <markh@osdl.org>
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
UDEVD_EVENT_TIMEOUT=0 didn't work directly after udevd startup.
The whole event timeout handling is not needed since we use netlink.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
|
|
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
|
|
Thanks to:
Marco d'Itri <md@Linux.IT>
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
After beeing hit by proprietary applications which statically
link the LGPL'd libusb, which needs a patch to reflect the recent
kernel changes, I decided not to provide LGPL code anymore.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Thanks to Theodore Ts'o <tytso@mit.edu> for the hint and a test image.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
/tmp is not writable on most systems, so just use /dev to create a
temporary node. If called from a udev rule, "-d $tempnode" should be
used and udev will create a temporary node and pass the name before
calling scsi_id.
Also remove gen_scsi_id_udev_rules.sh per Patrick's request, as it's
no longer needed with the persistent disk links.
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Based on:
"Linux Persistent Device Naming", 2004, Hannes Reinecke <hare@suse.de>
|
|
Opticals like that and should be ok for disks too, as long as we
don't read().
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
The attached patch adds a check to edd_id.c to verify that the MBR
signature on the device node passed to the program is unique to only
that disk.
Signed-off-by: John Hull <john_hull@dell.com>
|
|
|
|
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
This patch is to enable the use of scsi_id to derive a UID for a SCSI-2
device which is not compliant with the page 83 inquiry reply format for
either SPC-2 or SPC-3. In this case, the page 83 reply does not
contain a list of Identification descriptors but a single binary encoded
hexa-decimal Vendor Specified Identifier.
The update is being driven by the need for scsi_id to support older
model EMC Symmetrix hardware, that is, models 4, 5, and 6.
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|
|
Signed-off-by: Kay Sievers <kay.sievers@suse.de>
|