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