Chops another ~10% off create/join in a loop on i386.
- Disable low level debugging as this is stable. Improves benchmarks
across the board by a small percentage. Uncontested mutex acquire
and release in a loop becomes about 8% quicker.
- Minor cleanup.
it possible to get the pid, euid and egid of the process at the remote
end at the time it did bind() or connect().
Add a new libc function, getpeereid() to easily get at the euid and egid.
As a consequence, bump libc's minor number.
Document the LOCAL_PEEREID socket option in unix(4).
Based on contribution by Arne H. Juul, minor modifications by myself.
of pending atexit handlers before the structure is reused. This prevents
__cxa_finalize from going into an infinite loop when an atexit handler
register a new atexit handler as in:
#include <stdlib.h>
void two(void) {
}
void one(void) {
atexit(two);
}
int main(void) {
atexit(one);
return 0;
}
hint pointer, but do so in a way that remains compatible with older
pthread libraries. This can be used to wake another thread before the
calling thread goes asleep, saving at least one syscall + involuntary
context switch. This turns out to be a fairly large win on the condvar
benchmarks that I have tried.
detach/join.
- Make mutex acquire spin for a short time, as done with spinlocks.
- Make the number of spins controllable with the env var PTHREAD_NSPINS.
- Reduce the amount of time that libpthread internal spinlocks are held.
- Rely more on the barrier effects of park/unpark to avoid taking spinlocks.
- Simplify the locking around pthreads and the global queues.
- Align per-thread sync data on a 128 byte boundary.
- Offset thread stacks by a small amount to try and reduce cache thrash.
null character outside of the byte region [sbuf, sbuf + len).
1) If the length of the buffer is 0, do not write a null character
at all. Previously, sockaddr_snprintf() may have been able to
overwrite sbuf[-1] if len was 0.
2) If the length of the buffer, len, is greater than 0, then write
the null at sbuf[len - 1]. Previously, sockaddr_snprintf()
wrote the null at buf[len - 1], where `buf' was a "cursor" that
did not necessarily equal `sbuf', the start of the buffer. Now,
sockaddr_snprintf() always writes the null at sbuf[len - 1].
the maximum request size after mount. Calling mount(MNT_GETARGS)
from the file server is currently not kosher, as it vrele()s the
root vnode, potentially causing an inactive, which the file server
cannot handle while it itself is inactive in the kernel (deadlock).
Add an additional call with MNT_GETARGS to retrieve the modified
information instead of relying on the kernel code abusing the mount
interface during mount.
* in addition add/remove, allow enable/disable, which can be used
to control events for descriptors without having to remove all the
data associated with them
* add directsend/receive, which can be used to pass the same buffer
from the caller to read/writeframe and back again
* add flags to enqueue functions and allow urgent buffers to be
processed as the next PDU
- if the process does not exist anymore, return NULL
- if the process has created more lwps retry
XXX[1]: We should fix sysctl_kern_lwp() to return the number of lwps it
needs in oldlenp when ENOMEM. The we can avoid calling sysctl
twice (by starting let's say with always 1 lwp, or 10).
XXX[2]: We should fix kvm_getlwps to never spit errors.
- Silence new lint warnings
o If a long option looks like an ambiguous abbreviation of two or more long
options, but all the possible interpretations would return the same
value, then just return that value without complaining that it's
ambiguous.
o If a long option could be interpreted either as an exact match for one
long option, or as an abbreviation for one or more other long options,
then treat it as the exact match.
These changes align NetBSD's getopt_long(3) with the current behavior of
GNU getopt_long(3), the de facto standard, and FreeBSD's getopt_long(3).
file always after creation to cache the inode number given by the
backend file system. Otherwise we would not find a newly created
node from incore and create another one. In practise this was
pretty well hidden by the kernel name cache.
and destroyed, but not yet reclaimed. This prevents puffs_pn_nodewalk()
from returning stale entries. Make nullfs use this (some file
systems are a bit too happy with recycling inode numbers).
don't create if it is determined that the system in question doesn't need
them. This cleans up build warnings on systems with err.h (HAVE_ERR_H)
for example Mac OS X 10.3.
information, enabled if MULTITHREADED_REFUSE is defined at build
time. This is not enabled by default since libpuffs is not
threadsafe yet.
librefuse is still not threadsafe, but this gets rid of a whole lot of
problems in the meantime.
doesn't obtain the ports, gain and balance related parameters.
Those generally require reading from the hardware and therefore are much
more expensive to obtain. Modify OSS emulation to use the new ioctl
where possible.
This reduces CPU usage of mplayer during mp3 playback with my Thinkpad
from 20% to < 1% and from 50% to 20% during Xvid playback.
Review and comments from jmcneill@
include path: the normal header files now include the "SSP" ones (which one
should note are not really named right: SSP and FORTIFY_SOURCE are independent
features).
Disable USE_SSP on targets where the compiler doesn't support it at all
(mips, alpha) or it's known broken (sh3). But enable FORTIFY_SOURCE,
without SSP, on those platforms -- tested on mipsel.
(what other systems keep in libssp, we already have in libc) into libc
to match what other systems with FORTIFY_SOURCE do. Goodbye, libssp
dependency in libraries and executables. Discussed with christos and
mrg; Christos will merge the headers to get us the rest of the way to a
FORTIFY_SOURCE implementation that works as others' code expects.
has been set in the environment, this prevents people using MKDEBUGLIB
getting more than they bargained for.
Tidied up the debug settings in the Makefile to reflect the above change,
we no longer need to have FULL_DEBUG since nothing is written by default.
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.
This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.
This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.
Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.
turn a reference to puffs_framebuf in the file system from a
cb/justsend operation to a cc wait, should the file system find
itself desiring the result.
equal, larger, respectively instead of 0/1 for non/equal. This
will allow sorting the buffers for faster matching in libpuffs.
While here, change the name from respcmp to framecmp, as that better
reflects the purpose.
NOTE! there is no obvious way to make compilation fail for file
systems which may already be using this feature (although I don't
think there are any outside our tree, as the feature is two weeks
old). Nevertheless, non-updated file systems will fail very quickly.
reclaiming and the network/server is slow, we might have thousands
of buffers allocated at the same time causing the process to run
out of vm space. Rate limiting the number of outstanding ops would
be a nicer choice, but that requires more complex changes.
loop function might generate some results. and this is still "after"
event handling (except for the first call, but I'm not too keen on
optimizing for that)
* don't be such a baby about EINTR from kevent(). if we get it, suck
it up and continue instead of quitting
shouldn't release resources. As a quick hack to prevent double
frees etc, introduce "dead" variable into fuse channel to signal
if the backing puffs resource was already freed. XXX: the check
is only in exit/unmount currently.
Proper fix really depends on some puffs features I still haven't
gotten around to doing.
instead of puffs_start(). Get completely rid of puffs_start(), as
everything it used to do is now handled by the mount routine.
Introduce an optional pre-mount call puffs_setrootinfo() for setting
non-default root node information. As the old puffs_mount() is
now virtually useless, say byebye to it and rename the old
puffs_domount() to puffs_mount(), but add a root cookie parameter
to compensate for the late puffs_start().
- make sure that the args array is NULL terminated.
- make sure argc is initialized and argv is set to NULL after freeing.
- make the deep copy function an opt interface function.
up.
- Use malloc and free instead of macros, delete macros.
- Merge the two copies of deep copy and free args that were slightly buggy
into one and use that one.
XXX: Lots of ints should be size_t.
XXX: Bailing on error is not a good thing for a library.
XXX: Defs.h should be eliminated completely. Not all files need all the headers.
determines the API.
Provide a default for FUSE_USE_VERSION if it's not set, and use this
value to determine the number of arguments given to fuse_main().
the end of the argument vector for an argument that doesn't begin with '-',
and use it - this improves upon the previous naive approach of using the
last argument, which could have been placed there by one of the argument
appending functions.
Be consistent with the mount name processing, and always do the
puffs:refuse:filesystem substitution once and once only.
% priv mount.ntfs-3g ntfs-refuse.img /mnt
% df /mnt
Filesystem 1K-blocks Used Avail Capacity Mounted on
puffs:refuse:ntfs-3g 1000943 66309 934634 6% /mnt
% priv umount /mnt
% priv /usr/src/share/examples/refuse/id3fs/id3fs /mnt &
% df /mnt
Filesystem 1K-blocks Used Avail Capacity Mounted on
puffs:refuse:id3fs 0 0 0 100% /mnt
% priv umount /mnt
%
support removal and addition of i/o file descriptors on the fly.
* detect closed file descriptors
* automatically free waiters of a dead file descriptor
* give the file server the possibility to specify a callback which
notifies of a dead file descriptor
* move loop function to be a property of the mainloop instead of
framebuf (doesn't change effective behaviour)
* add the possibility to configure a timespec parameter which
attempts to call the loop function periodically
* move the event loop functions from the puffs_framebuf namespace
to puffs_framev to differential between pure memory management
functions
framebuf event loop to take n i/o fd's as parameters and also allow
dynamic add/remove of fd's. (not tested except for one fd, but more
changes coming soon)
stack instead of the continuation stack. This is for lib/36011, where
pthread gets confused since we aren't running on the regular stack.
I'm not really sure which direction to go to with this quite yet, so
make the hack hard to enable on purpose. The whole request dispatch
code needs cleaning anyway.
and event handling mechanisms required in file servers with blocking
I/O backends. puffs_framebuf is built on the concept of puffs_cc
and uses those to multiplex execution where needed.
File systems are required to implement three methods:
* read frame
* write frame
* compare if frame is a response to the given one
Memory management is provided by puffs_framebuf, but the file
systems must still, of course, interpret the protocol and do e.g.
byte order conversion.
As always, puffs_framebuf is work in progress. Current users are
mount_psshfs and mount_9p.
Refuse doessn't have a fuse_common.h, so include fuse_opt.h explicitly
from fuse.h. To avoid circular dependencies, don't include fuse.h
from fuse_opt.h.
This means that we no longer have to modify packages which use the argument
and option parsing routines to include fuse_opt.h.
It was only ever non-zero for zombies, and now doesn't exist.
The 'pstats' field has been valid for zombies for a while (it isn't in
the uarea any more).
Fixes part of build breakage.
- Add POSIX defined system variables and constants of AIO_LISTIO_MAX and
AIO_MAX values. Both with _POSIX_ASYNCHRONOUS_IO, provide them in
sysconf(3) and getconf(1) interfaces.
- Clean up sysconf(3) for handling sysctl nodes dynamically.
Seems to be quite stable. Some work still left to do.
Please note, that syscalls are not yet MP-safe, because
of the file and vnode subsystems.
Reviewed by: <tech-kern>, <ad>
+ put some meat into the add_arg and insert_arg functions
+ implement the free_args function
+ use 0 and 1 return values, instead of EXIT_SUCCESS and EXIT_FAILURE, in
internal functions (they bear no relation to shell exit values)
+ deep copy argument structures in refuse.c - we need to do this because
some FUSE file systems attempt to obliterate^Wcloak the argument vector
"for security reasons"
+ a side benefit of this is that we get the specific refuse file system
name back again:
[10:48:43] agc@inspiron1300 ...src/lib/librefuse 38 > df
Filesystem 1K-blocks Used Avail Capacity Mounted on
/dev/dk0 28101396 19318530 7377798 72% /
kernfs 1 1 0 100% /kern
procfs 4 4 0 100% /proc
ptyfs 1 1 0 100% /dev/pts
puffs:refuse:id3fs 0 0 0 100% /mnt
[10:48:51] agc@inspiron1300 ...src/lib/librefuse 39 >
only take the bare essentials, which currently means removing
"maxreqlen" from the argument list (all current callers I'm aware
of set it as 0 anyway). Introduce puffs_init(), which provides a
context for setting various parameters and puffs_domount(), which
can be used to mount the file system. Keep puffs_mount() as a
shortcut for the above two for simple file systems.
Bump development ABI version to 13. After all, it's Friday the 13th.
Watch out! Bad things can happen on Friday the 13th. --No carrier--
ad