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