1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
|
Copyright (c) 2004-2006 The Trustees of Indiana University and Indiana
University Research and Technology
Corporation. All rights reserved.
Copyright (c) 2004-2005 The Regents of the University of California.
All rights reserved.
Copyright (c) 2006-2010 Cisco Systems, Inc. All rights reserved.
$COPYRIGHT$
Additional copyrights may follow
$HEADER$
===========================================================================
This file contains the main features as well as overviews of specific
bug fixes (and other actions) for each version of PLPA since
version 1.0.
1.3.2
-----
- Removed a clause from AMD's license notice in plpa_map.c by
direction from Barry S. Newberger, Ph.D., Assistant General Counsel,
Advanced Micro Devices, Inc. per a Red Hat licensing concern
(initiated at https://bugzilla.redhat.com/show_bug.cgi?id=530230).
Removing this clause alleviates Red Hat's concern, even though this
resolution not shown on the above-cited bugzilla ticket.
- Julian Seward noticed that PLPA's use of the Valgrind API in the
affinity API probe was incorrect and suggested an improvement. Even
though we're ramping down PLPA development, I couldn't really ignore
the Valgrind author's suggestion!
1.3.1
-----
- Fixed bug identified by Eugene Loh at Sun; plpa-taskset -c ranges
were accidentally excluding the processor ID at the top of the
range.
- For standalone PLPA builds, the .so library is now versioned
according to the GNU Libtool guidelines.
1.3
---
- Added --with-valgrind support that allows building support for
Valgrind into PLPA (i.e., tell Valgrind to not warn about things
that we know are ok).
1.2
---
- Allow mapping from simple/contiguous processor, socket, and core
numbers to the corresponding back-end Linux ID (and clean up
documentation references between "X number" and "X ID", where X is
one of processor, socket, core).
- Support CPU hot plugging.
- Add explicit control over the PLPA topology information cache.
1.1.1
-----
- Ensure that --enable-included also disables executables by default.
Thanks to Brian Curtis for identifying the problem and supplying a
patch.
- Add ENABLE_EXECUTABLES m4 macro to allow behavior in m4
configuration analogous to "--enable-included --enable-executables".
1.1
---
- New command: plpa-taskset. It is intended as a drop-in replacement
for the "taskset" command, except it also understands sockets and
cores. See "plpa-taskset --help" for more details
- Renamed "plpa_info" to "plpa-info".
- Added PLPA_{MAJOR|MINOR|RELEASE}_VERSION integer macros in plpa.h.
This release, they have the values of 1, 1, and 0, respectively.
- Add new API functions to map from (socket,core) back and forth from
the Linux virtual processor ID. Thanks to AMD for the initial code
contribution that made this possible. See the documentation for
plpa_map_to_processor_id() as a starting point for more
information.
- Various fixes to "included" mode.
1.0.5
-----
- Fix an issue where the PLPA get/set affinity functions would only
operate on the current process; the PID argument to these functions
was effectively ignored.
1.0.4
-----
- Fix some 64 bit issues. Thanks to David Creasy for reporting the
problems.
- Fix plpa.h to be C++-friendly. Thanks to Emmanuel Paris for
pointing out this problem.
1.0.3
-----
- Resolve some sizing / units ambiguities that mistakenly did not make
it into 1.0.1. Thanks to Bert Wesarg for pointing these problems out.
1.0.2
-----
- Ensure that plpa_sched_[set|get]affinity() returns 0 upon success.
Thanks to David Creasy for bringing this to our attention.
1.0.1
-----
- Specify that cpusetsize should be in units of bytes and add some
missing "* 8"'s in plpa_dispatch.c.
1.0
---
- Initial release.
0.9a2
-----
- Change the back-end type of the plpa_bitmask_t to unsigned long
instead of unsigned char to avoid an endian problem.
- No longer attempt to set the affinity in plpa_api_probe() to avoid a
possible race condition (setting it after getting it).
- Fix PLPA_CPU_ZERO macro. Thanks to Bogdan Costescu for pointing
this out.
0.9a1
-----
Initial public release.
|
