| 1 | Revision history for IO::AIO |
1 | Revision history for IO::AIO |
| 2 | |
2 | |
|
|
3 | TODO: scandir - some dirs mostly contain subdirs - invert logic? |
| 3 | TODO: aio_cptree/mvtree |
4 | TODO: aio_cptree/mvtree |
| 4 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
|
|
| 5 | TODO: vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
5 | TODO: vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
| 6 | TODO: aio_fcntl, at least for file-locking |
6 | TODO: getxattr, setxattr, listxattr, removexattr etc.? |
| 7 | TODO: aio_mincore? |
|
|
| 8 | TODO: getxattr etc.? |
|
|
| 9 | TODO: F_DUPFD_CLOEXEC |
|
|
| 10 | TODO: emulation for splice? |
7 | TODO: emulation for splice? |
| 11 | TODO: eio_mmap|mlock|munmap|splice... |
8 | TODO: syncfs/sync windows: |
|
|
9 | TODO: posix_fallocate when flags=0 |
|
|
10 | http://stackoverflow.com/questions/65170/how-to-get-name-associated-with-open-handle/5286888#5286888 |
|
|
11 | http://blogs.msdn.com/b/adioltean/archive/2005/04/16/408947.aspx |
|
|
12 | http://msdn.microsoft.com/en-us/library/aa366789%28v=vs.85%29.aspx |
|
|
13 | http://msdn.microsoft.com/en-us/library/windows/desktop/aa366789%28v=vs.85%29.aspx |
|
|
14 | http://msdn.microsoft.com/en-us/library/windows/desktop/aa364425%28v=vs.85%29.aspx |
|
|
15 | http://msdn.microsoft.com/en-us/library/windows/desktop/aa364963%28v=vs.85%29.aspx |
|
|
16 | http://msdn.microsoft.com/en-us/library/windows/desktop/aa364996%28v=vs.85%29.aspx |
|
|
17 | http://msdn.microsoft.com/en-us/library/windows/desktop/aa364994%28v=vs.85%29.aspx |
|
|
18 | TODO: extra socket/tcp constants &c? |
|
|
19 | TODO: name_to_handle_At + open_by_handle_at = clone fds |
|
|
20 | TODO: lchown |
|
|
21 | TODO: ioprio_get/set |
| 12 | |
22 | |
|
|
23 | TODO: rewrite rmtree et al. to support working directories (also speed them up) |
|
|
24 | TODO: maybe IO::AIO leaks fds when requests are cancelled? maybe initialise result to -1? |
|
|
25 | TODO: aio_wd should use O_PATH on linux, due to lacking O_SEARCH (http://comments.gmane.org/gmane.linux.file-systems/33611) |
|
|
26 | http://www.openwall.com/lists/musl/2013/02/23/4 |
|
|
27 | TODO: FIDEDUPERANGE, file_ioctl.c |
|
|
28 | TODO: mincore |
|
|
29 | TODO: SOCK_CLOEXEc etc. |
|
|
30 | TODO: unlinkat flags, linkat flags, statat, renameat2 flags |
|
|
31 | TODO: https://kernelnewbies.org/Linux_4.11 |
|
|
32 | TODO: define useful rules for [[$wd, $path], $path] &c and implement them |
|
|
33 | TODO: https://github.com/coreutils/coreutils/blob/master/src/stat.c maybe IO::AIO::FSID |
|
|
34 | TODO: mlock2, MLOCK_ONFAULT |
|
|
35 | TODO: dup3, pipe2, accept4 |
|
|
36 | TODO: dump config.log when automated testign and configure fails |
|
|
37 | TODO: ns time resolution in utime |
|
|
38 | TODO: statx |
|
|
39 | TODO: aio_fchdir/chdir? emulate chdirat? |
|
|
40 | TODO: rt-4.0.18-20109-1573826674-1229.126277-5-0@rt.cpan.org |
|
|
41 | TODO: aio_splice, aio_tee |
|
|
42 | TODO: copy_file_range |
|
|
43 | TODO: write more abpout the thread model, maybe bump some defaults? |
|
|
44 | TODO: io_uring, but it's so broken :/ |
|
|
45 | TODO: allow import of fds as aio_wd |
|
|
46 | TODO: close_range |
|
|
47 | TODO: move_mount, open_tree, fsmount, fspick, fsconfig, fsopen https://lwn.net/Articles/759499/ |
|
|
48 | TODO: aio_mmap |
|
|
49 | |
|
|
50 | - aio_copy now tries to preallocate the destination file. |
|
|
51 | - fexecve configure test used the wrong püointer type (patch by Paul Howarth). |
|
|
52 | |
|
|
53 | 4.80 Sat 01 Apr 2023 04:13:37 CEST |
|
|
54 | - remove long obsolete "paths must be absolute" text in aio_open/stat |
|
|
55 | descriptions. |
|
|
56 | - another workaround for the low-quality musl libc. |
|
|
57 | - test for umount seperately, as it is more portable. |
|
|
58 | - some low-quality posix attempts (openbsd, osx) declare the availability |
|
|
59 | of fexecve but then don't even bother to implemnent a stub. |
|
|
60 | |
|
|
61 | 4.79 Sun 25 Sep 2022 18:25:53 CEST |
|
|
62 | - the autoconf result of the mount check was not used, so it failed to |
|
|
63 | compile on most systems (reported by crackoff@tutanota.com). |
|
|
64 | - work around fedora not having a working C compiler setup by design |
|
|
65 | (reported by Paul Howarth). |
|
|
66 | |
|
|
67 | 4.78 Tue 06 Sep 2022 12:54:43 CEST |
|
|
68 | - add IO::AIO::mount and IO::AIO::umount. |
|
|
69 | - add a bunch of symbols from sys/mount.h. |
|
|
70 | |
|
|
71 | 4.77 Mon 05 Sep 2022 02:03:38 CEST |
|
|
72 | - add MFD_HUGETLB_2MB and MFD_HUGETLB_1GB constants. |
|
|
73 | - add fexecve. |
|
|
74 | - add lots of missing functions to @EXPORT_OK. |
|
|
75 | |
|
|
76 | 4.76 Tue Jul 27 20:45:01 CEST 2021 |
|
|
77 | - add autoconf test for siginfo_t, which is, of course, not |
|
|
78 | available on windows (reported by Peter Petky). |
|
|
79 | - disable syscalls on solaris, as perl seems to provide an |
|
|
80 | incompatible syscall prototype that clashes with sys/syscall.h. |
|
|
81 | - add MAP_FIXED_NOREPLACE, MAP_SHARED_VALIDATE, MAP_SYNC and |
|
|
82 | MAP_UNINITIALIZED. |
|
|
83 | |
|
|
84 | 4.75 Wed Dec 30 08:44:55 CET 2020 |
|
|
85 | - removed leftover debug output from treescan. |
|
|
86 | - added (but not documented) waitid-P_*, FSPICK_*, FSOPEN_*, |
|
|
87 | FSCONFIG_*, MOUNT_ATTR_* constants. |
|
|
88 | |
|
|
89 | 4.74 Tue Dec 29 16:19:28 CET 2020 |
|
|
90 | - added pidfd_open, pidfd_getfd and pidfd_send_signal functions. |
|
|
91 | - rework bin/treescan to avoid recursion. This fixes two bugs |
|
|
92 | where a deep directory traversal or a lot of command line |
|
|
93 | arguments could cause it to crash. |
|
|
94 | - added (but not documented) open_tree, AT_* and move_mount flags. |
|
|
95 | - support defining syscall numbers in gendef0. |
|
|
96 | |
|
|
97 | 4.73 Fri Dec 4 02:08:37 CET 2020 |
|
|
98 | - def0.h was not uptodate, running into musl problems |
|
|
99 | (reported by B Fraser). |
|
|
100 | - IO::AIO::splice and IO::AIO::tee didn't properly return |
|
|
101 | 64 bit values. |
|
|
102 | - added IO::AIO::accept4. |
|
|
103 | - add various F_SEAL related fcntl constants. |
|
|
104 | - removed experimental marker for fdlimit functions. |
|
|
105 | - fiemap now includes the last segment even if it overflows |
|
|
106 | the end offset, which is arguably the correct behaviour. |
|
|
107 | |
|
|
108 | 4.72 Wed Apr 3 04:56:03 CEST 2019 |
|
|
109 | - (libeio) if fd 0 is available do not use it for aio_wd, as it |
|
|
110 | collides with IO::AIO::CWD. |
|
|
111 | - added IO::AIO::memfd_create. |
|
|
112 | - correctly include <sys/uio.h> in the vmsplice test. |
|
|
113 | - reduce codesize by ~7% on amd64 by declaring more functions |
|
|
114 | as noinline. |
|
|
115 | - documentation fixes and updates. |
|
|
116 | - experimental and undocumented preliminary support for synchronous statx. |
|
|
117 | |
|
|
118 | 4.71 Sun Mar 10 13:09:26 CET 2019 |
|
|
119 | - due to an error in the linux manpages, the configure tests |
|
|
120 | for readahead, sync_file_range, splice etc. failed. This has been fixed. |
|
|
121 | |
|
|
122 | 4.7 Mon Mar 4 11:28:13 CET 2019 |
|
|
123 | - majorly speed up scandir for the very special case of a non-POSIX |
|
|
124 | filesystem that nevertheless reports valid dt_type information. |
|
|
125 | The only known filesystem of this type is currently btrfs, which |
|
|
126 | didn't get it's act together to implement POSIX semantics in all |
|
|
127 | these years, and I am tired of waiting. |
|
|
128 | - add IO::AIO::MCL_ONFAULT for mlockall, add IO::AIO::mlockall. |
|
|
129 | - neither sys/mkdev.h nor sys/sysmacros.h were included, even when they |
|
|
130 | were detected by autoconf. |
|
|
131 | |
|
|
132 | 4.6 Sat Aug 25 21:58:36 CEST 2018 |
|
|
133 | - add st_btime, st_btimesec, st_btimensec and st_gen accessors, |
|
|
134 | mostly for bsds which expose birthtime and st_gen members (netbsd, freebsd |
|
|
135 | but only st_gen on openbsd because they compiled their perl without support |
|
|
136 | for birthtimes). |
|
|
137 | - fix madvise and munlock to properly support negative offsets. |
|
|
138 | - allocate fd for the aio_close at boot time, to guarantee it working |
|
|
139 | later, rather than calling abort when it fails. also avoids |
|
|
140 | close-on-exec race issues after module load. |
|
|
141 | - #undef utime on win32, maybe this helps some reports of utime hangs. |
|
|
142 | - minor doc improvements. |
|
|
143 | - minor configure cleanups. |
|
|
144 | - use $Config{perllibs} instead of libs for configure. might help people |
|
|
145 | who didn't install all perl dependencies (might break things as well). |
|
|
146 | |
|
|
147 | 4.54 Tue Aug 14 13:54:48 CEST 2018 |
|
|
148 | - include sys/mkdev.h or sys/sysmacros.h if available. |
|
|
149 | - further tweaks to configure invocation for systems requiring --rpath. |
|
|
150 | - no longer rely on custom paths on win32 platforms. |
|
|
151 | - try to work around buggy PAGESIZE macro on solaris. |
|
|
152 | |
|
|
153 | 4.53 Tue Aug 14 11:27:50 CEST 2018 |
|
|
154 | - add $Config{libs} to LIBS for configure, to work around systems |
|
|
155 | with broken library dependencies (... openbsd). |
|
|
156 | |
|
|
157 | 4.52 Sun Aug 12 08:09:45 CEST 2018 |
|
|
158 | - config.h.in was missing in distribution. |
|
|
159 | |
|
|
160 | 4.51 Sun Aug 12 07:24:14 CEST 2018 |
|
|
161 | - complete rework of the autoconf framework: IO::AIO now uses its own |
|
|
162 | config.h, separate from libeio, and tries to test the actual perl |
|
|
163 | environment, not the standard system environment. |
|
|
164 | - provide nanosecond-accracy stat time accessors for both perl and IO::AIO |
|
|
165 | stat functions. |
|
|
166 | - removed non-portable C++ syntax from eio.c. |
|
|
167 | - try to fix readdir tests on cygwin spuriously failing. |
|
|
168 | |
|
|
169 | 4.5 Wed Aug 1 00:23:55 CEST 2018 |
|
|
170 | - aio_mtouch touch all pages as requested, not just the first page in most cases. |
|
|
171 | - new function: IO::AIO::mremap, linux-specific mremap, with constants MREMAP_MAYMOVE |
|
|
172 | and MREMAP_FIXED. |
|
|
173 | - add O_ACCMODE. |
|
|
174 | - add (undocumented) MSG_CMSG_CLOEXEC and SOCK_CLOEXEC constants. |
|
|
175 | |
|
|
176 | 4.4 Tue Feb 20 07:54:12 CET 2018 |
|
|
177 | - new aio_slurp request. |
|
|
178 | - add (experimental) IO::AIO::eventfd and timerfd* function. |
|
|
179 | - add (experimental) IO::AIO::get_fdlimit and IO::AIO::min_fdlimit. |
|
|
180 | - point out that aio_copy overwrites existing files. |
|
|
181 | - removed filesystem magic number table from aio_statvfs docs as |
|
|
182 | statvfs doesn't actually return this info at all (statfs does). |
|
|
183 | - add a bunch of (mostly linux-specific) constants for use in ioctls |
|
|
184 | (see aio_ioctl docs). |
|
|
185 | - treescan now has a proper manpage and useful --help output. |
|
|
186 | - new option --sync in treescan, to sync everything in a subtree. |
|
|
187 | - changed default for aio_msync flags to MSYNC_SYNC. |
|
|
188 | - document offset/length behaviour of mprotect/madvise. |
|
|
189 | - support linux's renameat2 (via aio_rename2). |
|
|
190 | - add aio_rename2, an aio_rename with flags. |
|
|
191 | - add F_DUPPFD_CLOEXEC, F_OFD_[SG]ETLKW? constants. |
|
|
192 | - add FALLOC_FL_INSERT_RANGE, FALLOC_FL_UNSHARE_RANGE. |
|
|
193 | - (libeio) use posix_close if available. |
|
|
194 | - (libeio) internal close() calls no longer disturb errno. |
|
|
195 | - add IO:AIO::FALLOC_FL_UNSHARE. |
|
|
196 | - update schmorp.h for removal of USE_SOCKETS_AS_HANDLES |
|
|
197 | and better eventfd detection. |
|
|
198 | |
|
|
199 | 4.34 Sun May 1 19:18:24 CEST 2016 |
|
|
200 | - def0.h was not properly generated during previous release, causing |
|
|
201 | compile errors on various platforms. |
|
|
202 | - major/minor were accidentally switched (reported by Alexander Lishenyuk). |
|
|
203 | - removed duplicate definition of MAP_HUGETLB, that was found due to |
|
|
204 | a bug in Perl::Tidy (testcase by alaska332@gmail.com). |
|
|
205 | - added (untested!) aio_fcntl, aio_ioctl requests. |
|
|
206 | - (libeio) names set via prctl are truncated to 15 chars + nul, not 16, |
|
|
207 | as manpages-dev originally claimed. |
|
|
208 | |
|
|
209 | 4.33 Mon Jan 18 12:50:10 CET 2016 |
|
|
210 | - add IO::AIO::pipe2 function. |
|
|
211 | - added support for FALLOC_FL_COLLAPSE_RANGE and FALLOC_FL_ZERO_RANGE |
|
|
212 | constants. |
|
|
213 | - added support for O_TMPFILE and O_PATH constants. |
|
|
214 | - added support for MAP_FIXED, MAP_GROWSDOWN,MAP_32BIT, MAP_HUGETLB, MAP_STACK |
|
|
215 | consdtants, whether they can be sensibly used or not. |
|
|
216 | - use NO_INIT where applicable. |
|
|
217 | - update libecb. |
|
|
218 | - added stability canary support. |
|
|
219 | - updated linux super magic table to 4.3.3. |
|
|
220 | |
|
|
221 | 4.32 Wed Feb 11 20:32:11 CET 2015 |
|
|
222 | - replace off_t by STRLEN where appropriate, should not result in |
|
|
223 | user-visible changes. |
|
|
224 | - update ecb.h for C11 compatibility. |
|
|
225 | |
|
|
226 | 4.31 Tue Jun 3 03:29:27 CEST 2014 |
|
|
227 | - work around more 5.20 bugs. backwards compatibility my ass. |
|
|
228 | |
|
|
229 | 4.3 Fri Apr 11 06:22:38 CEST 2014 |
|
|
230 | - perl5porters broke Async::Interrupt, BDB, EV, IO::AIO, OpenCL |
|
|
231 | without warning by switching the meaning of USE_SOCKETS_AS_HANDLES |
|
|
232 | in 5.18. What's so attractive about giving a shit about backwards |
|
|
233 | compatibility - I will never understand. |
|
|
234 | |
|
|
235 | 4.2 Sat Jan 25 01:13:14 CET 2014 |
|
|
236 | - aio_group could corrupt memory because it didn't restore |
|
|
237 | the stack after req_submit. |
|
|
238 | - be more careful on (e.g. permission) errors in bin/treescan. |
|
|
239 | - work around changes in ExtUtils::MakeMaker. |
|
|
240 | - (libeio) implement aio_realpath for win32. |
|
|
241 | - (xthread) work around compile time bugs in ptw32. |
|
|
242 | - added IO::AIO::pipesize. |
|
|
243 | - (libecb) insignificant update. |
|
|
244 | |
|
|
245 | 4.19 Sun Jan 6 12:47:26 CET 2013 |
|
|
246 | - avoid endless loop in fiemap with some XFS files. |
| 13 | - in aio_rename and aio_rmdir, specialcase the case of [$wd, "."] |
247 | - in aio_rename and aio_rmdir, specialcase the case of [$wd, "."] |
| 14 | and call rename/rmdir instead of renameat/unlinkat. |
248 | and call rename/rmdir instead of renameat/unlinkat. |
| 15 | |
249 | |
| 16 | 4.18 Thu Oct 11 07:01:26 CEST 2012 |
250 | 4.18 Thu Oct 11 07:01:26 CEST 2012 |
| 17 | - fix unintended xthread_create by intentionalising it :) |
251 | - fix unintended xthread_create by intentionalising it :) |
