1 2004-02-04 multipath-016
2 * add devmap_name proggy for udev to name devmaps as per their
3 internal DM name and not only by their sysfs enum name (dm-*)
4 The corresponding udev.rules line is :
5 KERNEL="dm-[0-9]*", PROGRAM="/sbin/devmap_name %M %m", NAME="%k", SYMLINK="%c"
6 * remove make_dm_node fn & call. Rely on udev for this.
7 * don't rely on the linux symlink in the udev/klibc dir since
8 udev build doesn't use it anymore. This corrects build breakage
9 2004-01-19 multipath-013
10 * update the DM target synthax to the 2.6.0-udm5 style
11 2003-12-29 multipath-012
12 * check hotplug event refers to a block device; if not exit early
14 * add the uninstall target in Makefile
15 2003-12-22 multipath-010
16 * tweak the install target in Makefile
17 * stop passing fds as argument : this change enable a strict
18 segregation of ugly 2.4 code
19 * sysfs version of get_lun_strings()
20 * be careful about the return of get_unique_id() since errors
21 formerly caught up by if(open()) in the caller fn are now returned
23 * send get_serial() in unused.c
24 * introduce dm-simplecmd for RESUME & SUSPEND requests
25 * split add_map() in setup_map() & dm-addmap()
26 * setup_map() correctly submits "SUSPEND-RELOAD-RESUME or CREATE"
27 sequences instead of the bogus "RELOAD or CREATE"
28 * don't print .sg_dev if equal to .dev (2.6) in print_path()
29 * since the kernel code handles defective paths, remove all
30 code to cope with them :
31 * move do_tur() to unused.c
32 * remove .state from path struct
33 * remove .state settings & conditionals
34 * add a cmdline switch to force maps to failover mode,
35 ie 1 path per priority group
36 * add default policies to the whitelist array (spread io ==
37 MULTIBUS / io forced to 1 path == FAILOVER)
38 * move get_disk_size() call out of add_map() to coalesce()
39 * comment tricky coalesce() fn
40 * bogus unsused.c file renamed to unused.c
41 2003-12-20 multipath-010
42 * big ChangeLog update
43 * start to give a little control over target params :
44 introduce cmdline arg -i to control polling interval
45 * cope with hotplug-style calling convention :
46 ie "multipath scsi $DEVPATH" ... to avoid messing with
47 online maps not concerned by an event
48 * example hotplug agent to drop in /etc/hotplug.d/scsi
49 * revert the run & resched patch : unless someone proves me
50 wrong, this was overdesigned
51 * move commented out functions in unused.c
52 * update multipath target params to "udm[23] style"
53 * mp target now supports nr_path == 1, so do we
54 * add gratuitous free()
55 * push version forward
56 2003-12-15 multipath-009
57 * Make the HW-specific get_unique_id switch pretty
58 * Prepare to field-test by whitelisting all known fibre array,
59 try to fetch WWID from the standard EVPD 0x83 off 8 for everyone
60 * configure the multipath target with round-robin path selector and
61 conservative default for a start (udm1 style) :
62 yes it makes this release the firstreally useful one.
63 * temporarily disable map creation for single path device
64 due to current restrictive defaults in the kernel target.
65 Sistina should work it out.
66 * correct the strncmp logic in blacklist function.
67 * update the Makefiles to autodetect libgcc.a & gcc includes
68 "ulibc-style". Factorisation of udevdirs & others niceties
69 * drop a hint about absent /dev/sd? on failed open()
70 * implement a reschedule flag in /var/run.
71 Last thing the prog do before exit is check if a call to multipath
72 was done (but canceled by /var/run/multipath.run check) during its
73 execution. If so restart themain loop.
74 * implement a blacklist of sysfs bdev to not bother with for now
75 (hd,md, dm, sr, scd, ram, raw).
76 This avoid sending SG_IO to unappropiate devices.
77 * Adds a /var/run/multipath.run handling to avoid simultaneous runs.
78 * Remove a commented-out "printf"
79 * drop a libdevmapper copy in extras/multipath;
80 maybe discussions w/Sistina folks will bring a better solution in the future.
81 * drop a putchar usage in libdevmapper to compile cleanly with klibc
82 * drop another such usage of my own in main.c
83 * massage the Makefile to compile libdevmapper against klibc
84 * use "ld" to produce the binary rather than "gcc -static"
85 * stop being stupid w/ uneeded major, minor & dev in main.c:dm_mk_node()
86 * reverse to creating striped target for now because the multipath target
87 is more hairy than expected initialy
88 * push the version code to 009 to be in synch w/ udev
89 2003-11-27 multipath-007
90 * removes sg_err.[ch] deps
91 * makes sure the core code play nice with klibc
92 * port the sysfs calls to dlist helpers
93 * links against udev's sysfs (need libsysfs.a & dlist.a)
94 * finally define DM_TARGET as "multipath" as Joe posted the code today (not tested yet)
95 * push version forward (do you want it in sync with udev version?)
97 * merged in udev-006 tree
98 2003-09-18 Christophe Varoqui <christophe.varoqui@free.fr>
99 * multipath 0.0.1 released.