| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
|
|
|
| |
consistency with the C++ headers.
|
|
|
|
|
|
|
|
|
|
|
|
| |
FreeBSD, by default the VM will sync dirty pages periodically, even if
the pages do not need to be reclaimed. Since in this case the contents of
the map will contain sensitive information, try to avoid that.
The FreeBSD 7.0 man page for mmap also warns that severe file fragmentation
can result from using MAP_NOSYNC with sparse files, which we are currently
using. It recommends instead explicitly writing out zero bytes. Since
typically the full range of pages in the file will be eventually accessed
and modified, this doesn't seem like a problem.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
completely wrong, and it is rather disturbing I made that mistake. The
poll() function will gather up to the number of bytes passed as its argument;
it will do this by opening up files and reading up to 1024 bytes from each
one. So we might open between 256 and 256*1024 files, depending on the sizes
of them (ignoring empty files, which count for zero bytes). (Idea: also
include the name of the file in the output? Is that useful?)
Move the read_buf out of the loop to minimize allocator thrashing.
|
|
|
|
| |
That seems excessive. Reduce to 32 Kb and 256 Kb (resp).
|
| |
|
| |
|
|
|
|
| |
to rng.h (eventually base.h will be split up entirely and go away)
|
|\
| |
| |
| |
| |
| | |
85596a0b1fbee6696261d36def7dad742375f7d9)
to branch 'net.randombit.botan.remove-libstate' (head faf75eec422a3eb6108748315724fb6ac217dede)
|
| |\
| | |
| | |
| | |
| | |
| | | |
bcae6a19d8d96ebde33c832c76a130ea02bc923a)
to branch 'net.randombit.botan.remove-libstate' (head f640991b6e433a264f5d22a08338cfc9c24c82da)
|
| | |\
| | | |
| | | |
| | | |
| | | |
| | | | |
6afe2db1f710f75bc27e189bb8bdb23613ce1ca3)
to branch 'net.randombit.botan.remove-libstate' (head e40f0dbdfd847024c30fa0092c2acefc19a550b8)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
timer with an unspecified update rate and epoch. It is only used
inside the entropy sources to provide some timing-dependent
randomness. However, it is easier and basically 'as good' to treat the
timers as entropy sources in their own right and feed their output
directly into an entropy pool.
This commit removes Library_State::system_clock and all calls to that
function.
|
|/ / /
| | |
| | |
| | |
| | | |
sources, etc), but drivers for x86-64/2.6 are not in evidence these days. To test
this I would probably have to get a 32-bit machine running ~RH9/2.4 kernel...
|
| | |
| | |
| | |
| | | |
new build.h macro BOTAN_MEM_POOL_CHUNK_SIZE
|
|/ / |
|
| |
| |
| |
| | |
instead passing those values as arguments.
|
| |
| |
| |
| |
| | |
source will default to using the PROV_RSA_FULL provider if an empty string
is passed to the constructor.
|
| |
| |
| |
| |
| |
| | |
failures.
Seemingly from a bad merge around Christmas?
|
|/
|
|
|
|
| |
the actual copyright holders. For rationale, see my post to botan-devel
on April 9, subject 'Changing license to directly reflect contributors'
(http://www.randombit.net/pipermail/botan-devel/2008-April/000527.html)
|
| |
|
| |
|
|
|
|
|
| |
updated dates on files that have actually changed this year. This makes
the diff across versions readable again.
|
|
|
|
|
|
|
|
| |
takes advantage of unaligned reads/writes being legal for some extra performance,
but should be rewritten to use SSE2 and non-termporal writes.
Most of the functions in bit_ops.cpp are implemented by x86-64, just not
easily accessible from C++
|
|
|
|
| |
harmless, to avoid a valgrind warning
|
|
|
|
|
|
|
| |
with the last one being both one of the input values and the output carry
register, since almost always they were in fact the same variable.
Also update the x86 and x86-64 modules.
|
|
|
|
|
|
|
| |
but might as well keep it up to date. And it's easier to do it once with
a 'perl -pi' command than to update each file over time.
Apologies to anyone looking at diffs.
|
|
|
|
| |
for struct stat
|
|
|
|
|
|
|
| |
Has not been tested on all of the systems listed, and on many of them it
won't be relevant anyway since /dev/random and company only exist on some
of them (Linux, BSDs, recent Solaris, and it looks like recent AIX and
HP-UX as well).
|
|
|
|
|
|
| |
explicit :: (it is unfortunate that there is no good way to detect all
of such calls in an automated manner). Also use new-style casts in parts
of the zlib code.
|
|
|
|
|
| |
iostreams, it uses unbuffered Unix I/O syscalls and is careful to avoid
blocking for more than short amounts of time.
|
|
|
|
|
|
| |
by Joel Low on the mailing list, the STL container types have only a
single version of push_back(), along with variations of insert() for
handling range-based appending.
|
|
|
|
| |
Change all callers in the library and self-test code.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
use the generic variable reading routines).
Instead of hardcoding the module sets (historically, 'unix', 'beos', and
'win32') into the script, have each module specify which group(s) (if any) it
should be considered a member of in its modinfo.txt file.
Add a new module set compression which contains (currently) the zlib and bzip2
modules.
|
|
|
|
| |
reinterpret_cast
|
|
|
|
| |
included by bit_ops.h
|
|
|
|
|
|
|
|
|
| |
Where loadstor.h was needed but only implicitly included via bit_ops.h,
include it directly
Add endian reversal functions to bit_ops.h
Remove some unneeded includes in big_ops2.cpp and a few other files.
|
|
|
|
|
|
|
| |
All are now specified through the config. The new default is just /bin,
/sbin, /usr/bin, and /usr/sbin. Formerly /usr/ucb, /usr/etc, and /etc were
also searched. If you want this behavior again you have to explicitly set
the rng/unix_path configuration setting.
|
|
|
|
| |
Use u32bit instead of uint32_t
|
|
|
|
|
|
| |
programs, the fast poll will just call a handful of simple Unix/POSIX
functions like getpid, getuid, getrusage, etc. Identifying further useful
sources would probably be helpful.
|
| |
|
| |
|
|
|
|
| |
hurt and might help. From Yves Jerschow.
|
|\
| |
| |
| |
| |
| | |
63c067b5454498ee48466101aedb54613615a983)
to branch 'net.randombit.botan' (head 2444e9d08f77852cf2df58c34ca38145c2f416ba)
|
| |
| |
| |
| | |
folks (I think this was also required under QNX, actually).
|
| |
| |
| |
| |
| | |
It is not amazingly optimized, about 5% faster than what GCC 4.1.1 does
on my Core2 with the normal C++ code, but it's a start.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
stack. At least SuSE and Gentoo are using a patch for this in their trees,
probably others are as well.
I still have not had a chance to check the portability aspects of this,
especially on Solaris (the only ELF-based x86/amd64 operating system that
I know of that does not use the GNU toolchain).
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Now three classes are defined: 'request', 'auto', and 'asm_ok'. The 'auto'
class is loaded automatically if the platform support matches up with
what we are building for (this is the former default). The 'request' mode
means it is only loaded if specifically requested by name. The 'asm_ok'
module is marked for all modules that use any assembler (including inline
assembler). This normally functions like 'auto', unless --debug is passed
to configure, in which case it is treated as 'request'.
Modules which do not specify a load behavior are given a default of
'request'.
|