NetBSD's PTHREAD_KEYS_MAX is set to 256, which is low compared to
other systems like Linux (1024) or MacOS X (512). As a result some
setups tested on Linux will exhibit problems on NetBSD because of
pthread_keys usage beyond the limit. This happens for instance on
Apache with various module loaded, and in this case no particular
developper can be blamed for going beyond the limit, since several
modules from different sources contribute to the problem.
This patch makes the limit conigurable through the PTHREAD_KEYS_MAX
environement variable. If undefined, the default remains unchanged
(256). In any case, the value cannot be lowered below POSIX-mandated
_POSIX_THREAD_KEYS_MAX (128).
While there:
- use EXIT_FAILURE instead of 1 when calling err(3) in libpthread.
- Reset _POSIX_THREAD_KEYS_MAX to POSIX mandated 128, instead of 256.
checks. Compared to OpenBSD's reallocarray, makes it easier to avoid memory
leaks on allocation failures and it doesn't depend on malloc(0) != NULL
for correct error checking. Compared to plain realloc, it also avoids
the problem of intermediate integer overflows. The trade-off is the use
of void * to side step C type system with regard to generic pointer to
pointer.
unistd.h and libc, and add a man page.
Allow wiggle room in the man page for implementations of fallocate
that either (a) don't fully unwind on failure, leaving new blocks
allocated without changing the file size, or (b) create only in-memory
transient reservations that disappear when crashing or rebooting.
Also, add crossreference to fdiscard from ftruncate(2), and remove the
old BUGS entry from there that called for a generalized version that
allows discarding ranges of a file: that's what fdiscard is.
that uses rpc needs to define _REENTRANT to work properly. We'll keep
compatibility but setting the old variable in the library, but we are not
advertising it anymore.
consttime_memequal is the same as the old consttime_bcmp.
explicit_memset is to memset as explicit_bzero was to bcmp.
Passes amd64 release and i386/ALL, but I'm sure I missed some spots,
so please let me know.
Provide LC_GLOBAL_LOCALE in a way that works with all locale functions.
Merge constant data used by the initial global locale and the C locale.
Drop function call layer for _current_locale() and directly return the
locale reference, not a pointer to it. Use protected access for global
variables, so that libc references can avoid the GOT overhead.
called. Use this to implement newlocale and provide duplocale/freelocale
as well. Based on patches by Takehiko Nozaki with simplications and fix
for the init order by myself.
retrieve and return if there has been a conversion error, so that
it can keep performing char-by-char processing if a multi-byte
conversion occurred. Also some more code cleanups in the "extra"
processing.
and willy nilly people kept adding headers appropriate to the new
functionality So we reached where we are today, this header causes
lots of namespace conflicts (for example you cannot use util.h from
a lex file, because ECHO is macro used in lex, and ECHO is defined
to 8 in termios). Roll back time, and remove all the unnecessary
includes, leaving it to the program to include the ancillary headers
they need.
was deleted from the filesystem to the disk driver, commonly
known as "discard" or "trim".
fs/driver support is in ffs and ata wd for now.
This is what was posted here:
http://mail-index.netbsd.org/tech-kern/2012/02/28/msg012813.html
with minor cleanup, and the global switch replaced by a mount option.
abd userland, as proposed on tech-security, with explicit_bzero using
a volatile function pointer as suggested by Alan Barrett.
Both do what the name says. For userland, both are prefixed by "__"
to keep them out of the user namespace.
Change some memset/memcmp uses to the new functions where it makes
sense -- these are just some examples, more to come.
depend on new devname_r(3) as heart. Add /dev/pts magic directly to
devname(3). While it can lead to returning non-existing paths, the
behavior is more consistent that way. Drop caching layer in devname(3),
it doesn't buy anything for the common case of having access to the
database. Teach devname(3) proper fallback behavior of scanning /dev.
Create both old-style and new-style database for now in /etc/rc.d/sysdb.
copying the passed string (which is not ToG compliant), instead of using
it directly in the environment arrat as it should. Needs to be pulled up
to NetBSd-6.
vector. This can normally be found via __ps_strings, but libc is
initialised too early when linked dynamically and doesn't have access to
it yet, so provide an alternative mechanism via ld.elf_so.
Bump libc minor.
unsigned. They are signed. (While unsigned might have been a better
choice, it doesn't really matter and the majority of preexisting uses
were signed. And consistency is good.)
filesystem, instead of blocks and files together.
This results in fetching each FS-level quota entry twice and therefore
doing slightly more work, but (1) quota access isn't a critical path
and (2) after fetching the block values the file values will be hot in
the cache, so it won't add much total time.
Also move more of the FS-independent defintions from <quota.h> to
<sys/quota.h> so we can use them internally.
Step 4 of 5 for QUOTACTL_GET.
Note: this change requires a kernel version bump.
state of the stream. Change argument of the seek function to funopen() from
fpos_t to off_t. Make f{g,s}etpos() use the new fpos_t struct, while providing
backwards compatible entry points. Approved by releng@
This is the user/application API and does not include the kernel
interface. Definitions shared with the kernel are found in
<sys/quota.h>.
Since I've been working on a local branch I have taken the liberty of
moving the latest-so-far version of <quota.h> to the front of the
patch queue. This loses (a small amount of) history but has the
advantage of avoiding unnecessary version bumps of libquota.so.
- Add missing stuff from Issue 6 (some of it commented out)
- Fix _POSIX_THREAD -> PTHREAD variable confusion
- Amend PTHREAD_KEYS_MAX to 256 since this is what libpthread defined internally
(I found no instances where three arguments were actually
used, as per the claim in the comment. If there are any
in third party code then they will need to use a cast)
http://mail-index.netbsd.org/tech-userlevel/2011/08/25/msg005404.html
This is used by disk tools such as disklabel(8) to dynamically decide is
the undelyling platform uses a disklabel-in-mbr-partition or not
(instead of using a compile-time list of ports).
getlabelusesmbr() reads the sysctl kern.labelusesmbr, takes its value from the
machdep #define LABELUSESMBR.
For evbmips, make LABELUSESMBR 1 if the platform uses pmon
as bootloader, and 0 (the previous value) otherwise.
unimplemented, except enough of linkat(2) to hardlink to a symlink.
Everything new in headers is guarded #ifdef _INCOMPLETE_XOPEN_C063 since
some software (e.g.: xcvs in our own tree) will assume they can use openat(2)
when AT_FDCWD is defined. _INCOMPLETE_XOPEN_C063 will go away once support
will be completed.
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.
set close-on-exec on creation (http://udrepper.livejournal.com/20407.html).
- Add F_DUPFD_CLOEXEC to fcntl(2).
- Add MSG_CMSG_CLOEXEC to recvmsg(2) for unix file descriptor passing.
- Add dup3(2) syscall with a flags argument for O_CLOEXEC, O_NONBLOCK.
- Add pipe2(2) syscall with a flags argument for O_CLOEXEC, O_NONBLOCK.
- Add flags SOCK_CLOEXEC, SOCK_NONBLOCK to the socket type parameter
for socket(2) and socketpair(2).
- Add new paccept(2) syscall that takes an additional sigset_t to alter
the sigmask temporarily and a flags argument to set SOCK_CLOEXEC,
SOCK_NONBLOCK.
- Add new mode character 'e' to fopen(3) and popen(3) to open pipes
and file descriptors for close on exec.
- Add new kqueue1(2) syscall with a new flags argument to open the
kqueue file descriptor with O_CLOEXEC, O_NONBLOCK.
* Fix the system calls that take socklen_t arguments to actually do so.
* Don't include userland header files (signal.h) from system header files
(rump_syscallargs.h).
* Bump libc version for the new syscalls.
for UFS1).
Remove kernel option for EA backing store autocreation and do it by
default. Add a sysctl so that autocreated attriutr size can be modified.
plunky