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