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