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