Overview
Request 144266 accepted
- Update to new upstream release 0.36
* Work around a kernel change in FreeBSD 9.0-STABLE where /dev/klog
has become pollable via kevent(), but returns -EINVAL if you try
to register an EVFILT_WRITE filter, which confused the
iv_fd_register_try() logic in iv_fd_kqueue.
* Rework iv_timer to use less memory when only a small number of
timers is in use, and allow adding radix tree levels as needed
to remove the previous ~1M timer limit.
- Update to new upstream release 0.35
* Fix an abort on iv_work worker thread timeout introduced by the
race condition fix optimization that went into 0.32.
* Two iv_event race condition fixes.
Request History
jengelh created request
- Update to new upstream release 0.36
* Work around a kernel change in FreeBSD 9.0-STABLE where /dev/klog
has become pollable via kevent(), but returns -EINVAL if you try
to register an EVFILT_WRITE filter, which confused the
iv_fd_register_try() logic in iv_fd_kqueue.
* Rework iv_timer to use less memory when only a small number of
timers is in use, and allow adding radix tree levels as needed
to remove the previous ~1M timer limit.
- Update to new upstream release 0.35
* Fix an abort on iv_work worker thread timeout introduced by the
race condition fix optimization that went into 0.32.
* Two iv_event race condition fixes.
factory-auto accepted review
Builds for repo openSUSE_Factory
saschpe accepted review
factory-auto added a reviewer
Please review sources
licensedigger approved review
{"approve": "preliminary, version number changed"}
licensedigger accepted review
{"approve": "preliminary, version number changed"}
namtrac accepted request
- Update to new upstream release 0.36
* Work around a kernel change in FreeBSD 9.0-STABLE where /dev/klog
has become pollable via kevent(), but returns -EINVAL if you try
to register an EVFILT_WRITE filter, which confused the
iv_fd_register_try() logic in iv_fd_kqueue.
* Rework iv_timer to use less memory when only a small number of
timers is in use, and allow adding radix tree levels as needed
to remove the previous ~1M timer limit.
- Update to new upstream release 0.35
* Fix an abort on iv_work worker thread timeout introduced by the
race condition fix optimization that went into 0.32.
* Two iv_event race condition fixes.
