aboutsummaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* specifiy mode for O_CREATHarald Hoyer2009-05-041-1/+2
|
* Don't re-init an existing device, it causes errors. (#484411, ↵Bill Nottingham2009-05-041-1/+1
| | | | <jpayne@redhat.com>)
* Man page updates for ipcalc(1)David Cantrell2009-05-042-7/+23
| | | | | Make sure the ipcalc man page explains the -c, -4, and -6 options. Update some other things. Shamelessly add my name.
* Add a address check capability for IPv4 and IPv6 addresses. (#464268)David Cantrell2009-05-041-0/+42
| | | | | | | | | | | | | Add the -c, --check argument which enables address validation. The ipcalc program will exit with 0 if the address is valid or 1 if it is invalid, for the specified family. Add the -4, --ipv4 argument to specify the IPv4 address family. Add the -6, --ipv6 argument to specify the IPv6 address family. Ensure that the IPv6 family argument is invalid with the IPv4-only show arguments (broadcast, network, netmask, prefix).
* Use asprintf() rather than malloc() and sprintf().David Cantrell2009-05-041-3/+6
| | | | | Use asprintf() for conversion and abort on memory allocation failures.
* Use inet_ntop() and inet_pton()David Cantrell2009-05-041-5/+27
| | | | | The inet_ntoa() and inet_aton() functions are deprecated and limited to AF_INET only.
* Use 'struct in_addr' for IPv4 addresses.David Cantrell2009-05-041-44/+54
| | | | | Change from using uint32_t to struct in_addr for IPv4 addresses. Get rid of the IPBITS and IPBYTES macros.
* Whitespace cleanups for src/ipcalc.cDavid Cantrell2009-05-041-135/+130
| | | | Consistently use 4 spaces for indentation throughout ipcalc.c
* Add an option for consoletype to always return 0. (#453798, ↵Bill Nottingham2009-05-042-3/+8
| | | | | | | <notting@redhat.com>) This allows tcsh -e to work, with lang.sh. It's a hack, but there really isn't a better way to solve it.
* https://bugzilla.redhat.com/show_bug.cgi?id=437110Harald Hoyer2008-09-161-6/+7
| | | | | | | | | | | | | | | | | | | | | | On System z, typical hardware setups can consist of up to several thousand I/O devices. The current udev rules for handling these devices and the ccw_init helper script are implemented in an insufficient way, leading to unnecessarily high usage of memory and cpu time during system startup. The attached patch against the initscripts-8.45.19.EL-1 source RPM reduces this resource usage drastically by a) fixing a number of coding problems in helper script ccw_init (the current version creates failure indications in /dev/.udev/failed for each device) b) restricting the ccw_init script to network devices (ccw_init is only relevant for network devices) Using these changes, memory consumption on an example system with ~1500 devices could be reduced by 80% (VSZ), resp. 50% (RSS) and the elapsed duration for udev could be reduced from 19s down to 6 seconds. Business case: with reduced memory and time consumption during startup, more Linux images can be run with the same amount of resources.
* fix genhostid on x86_64 (#306811)Bill Nottingham2008-01-101-1/+1
|
* Ignore any quotes around SUBCHANNELS and HWADDR, for admins that have ↵Bill Nottingham2008-01-091-2/+15
| | | | accidentally added them (#351291)
* - add a step to rename any temporarily renamed devices (#208740,#204803, ↵Bill Nottingham2006-11-281-2/+29
| | | | #214817)
* I can't countBill Nottingham2006-11-011-1/+1
|
* more SUBCHANNELS related fixage (#204803)r8-45-6-ELBill Nottingham2006-10-301-13/+48
|
* use SUBCHANNELS as the primary key for s390 network devices (#204803)Bill Nottingham2006-10-261-0/+21
|
* move ccwgroup initialization to a udev rule (#199139, #199655, #169161)Bill Nottingham2006-09-272-0/+43
|
* reset DEVPATH when renaming (#206884, <phil@fifi.org>)Bill Nottingham2006-09-181-0/+1
|
* name it something more obvious as a temporary deviceBill Nottingham2006-09-151-1/+1
|
* don't loop if there are multiple devices with the same hwaddr (#199242, ↵Bill Nottingham2006-07-191-6/+13
| | | | <markmc@redhat.com>)
* fix manpage ref (#198756)Bill Nottingham2006-07-131-1/+1
|
* ignore alias devices (#186355)Bill Nottingham2006-07-011-2/+9
| | | | | refresh the device name on collision, as it may have initialized since we initially scanned for devices
* duh.Bill Nottingham2006-05-231-2/+2
|
* link glib2 dynamically again now that it's in /libBill Nottingham2006-05-231-2/+2
|
* Fix typo.Miloslav Trmac2006-05-171-1/+1
|
* Fix handling of mount points with white space (#186713):Miloslav Trmac2006-05-083-1/+134
| | | | | | | | | | | | | - src/fstab-decode.c, src/fstab-decode.8: New files. - src/Makefile: Build and install fstab-decode. - initscripts.spec: Add fstab-decode to %files - rc.d/init.d/functions (fstab_decode_str): New function. (__umount_loop, __umount_loopback_loop): Factor out from copy&pasted copies. (action): Fix command quoting. - rc.d/rc.sysinit: Correctly quote and decode mount points. - rc.d/init.d/halt: Use __umount_* and fstab-decode. (action): Rename from runcmd. Fix command quoting. - rc.d/init.d/netfs: Use __umount_*.
* Update .cvsignoreMiloslav Trmac2006-04-181-0/+1
|
* addr8-32Bill Nottingham2006-04-101-0/+325
|
* apply changes from FC5 branch - automatically rename devices on startupBill Nottingham2006-04-101-2/+6
|
* - Fix one-byte buffer overflowMiloslav Trmac2006-03-081-2/+2
| | | | - Make theChild a pid_t
* Add man page for getkey.Miloslav Trmac2006-01-252-0/+81
|
* - Print message even if there is no timeout (#54481)Miloslav Trmac2006-01-251-43/+47
| | | | | | | | - Reject negative timeout (#54481) - Misc. cleanups: - Use _exit(), not exit() within a signal handler - Only install signal handlers after data they reference is initialized - Don't limit the number of allowed keys to 100
* Add genhostidMiloslav Trmac2006-01-251-0/+1
|
* udev now loads modules - use that, remove the kmodule stuffBill Nottingham2005-12-211-2/+1
| | | | don't set hotplug, it can't possibly work
* fix use-after-free (#171198)Bill Nottingham2005-10-271-1/+4
|
* don't probe for uninteresting devices. speeds things up (floppy probe is slow)Bill Nottingham2005-09-221-1/+1
|
* from popt.h: The argument array is malloc'd as a single area, so only argv ↵Bill Nottingham2005-08-041-3/+2
| | | | must be free'd. Oops. (#165033)
* fix glibc free errorDan Walsh2005-07-211-2/+4
|
* fix potential memory overread (#153685, <in-redhat@baka.org>)Bill Nottingham2005-04-051-1/+1
|
* free some of the more egregious memory leaks (#85935)Bill Nottingham2005-03-312-9/+25
|
* bring back initlog for third-party scripts. officially deprecate itBill Nottingham2005-03-313-36/+13
|
* flush stdout before forkingr8-02Bill Nottingham2005-01-031-1/+2
|
* simplifyBill Nottingham2004-12-221-41/+10
|
* in order to avoid probing twice on startup, add a daemon mode thatBill Nottingham2004-12-221-3/+72
| | | | | | waits for a connection on an abstract socket, and then dumps the probe information. Requires a corresponding kudzu change to read from said socket...
* remove initlog, minilogdBill Nottingham2004-12-221-9/+1
|
* no DDC probe, pleaseBill Nottingham2004-12-071-1/+1
|
* ftw? is slow.Bill Nottingham2004-11-231-21/+29
|
* fix various minilogd bogosities (uninitialized variable, handling of ↵Bill Nottingham2004-11-162-4/+8
| | | | wrong-protocol connections, don't check atime) (#106338)
* don't probe the serial busBill Nottingham2004-10-291-1/+1
|
* remove bogus include (#137470)Bill Nottingham2004-10-281-1/+0
|