Hi, I am bumping tags today. There have been a number of commits since the tag that should make it to release IMO. Actually, most of them should go in. Files marked with a [T] will have their tag bumped. Files marked with [-] won't. I have included the logs of the changes for your convenience. Lines marked RM: are lines with my commentary.
It would be nice if we could bump daedalus to the new tag and give this another 1-2 days on there. I am aiming for release on wednesday. By then it should be possible to have made a decission/have a solution on the atomics issue. Sander ===================================================================== httpd-2.0 ===================================================================== [T] CHANGES RM: Will be updated prior to retagging. --------------------------------------------------------------------- [T] Makefile.in (1.107) aaron Log: Don't install *.in config files. --------------------------------------------------------------------- [T] STATUS RM: Quite a few updates. I see no harm in bumping the tag on this. --------------------------------------------------------------------- [T] build/install-bindist.sh.in (1.7) trawick Log: get closer to having apxs work from a binary distribution we need to copy libtool and instdso.sh verbatim, and we need to edit config_vars.mk and apxs still broken: "httpd -l" as invoked by apxs can fail since support libraries (apr, aprutil, etc.) are probably not findable --------------------------------------------------------------------- [T] docs/manual/howto/ssi.html.en (1.11) [T] docs/manual/mod/mod_alias.xml (1.5) [T] docs/manual/vhosts/examples.html (1.10) RM: documentation changes always have to go in IMO. --------------------------------------------------------------------- [T] include/http_config.h (1.97) aaron Log: Style cleanup (remove tabs, fix alignment). --------------------------------------------------------------------- [T] modules/aaa/mod_auth_digest.c (1.62) fielding Log: kill a warning on Darwin for NONCE_LEN becoming a long int by math. --------------------------------------------------------------------- [T] modules/arch/win32/mod_isapi.c (1.64) wrowe Log: After review and testing against all of the PSDK examples (see http://www.apache.org/~wrowe/ for commentary on building the examples and making them work) ... this disable-optimization should no longer be required. --------------------------------------------------------------------- [T] modules/experimental/mod_ext_filter.c (1.27) wrowe Log: Trade one signedness mismatch for another, but choose the one that is known to be a positive value. --------------------------------------------------------------------- [T] modules/http/http_request.c (1.141) jerenkrantz Log: If a subreq added a filter (say INCLUDES) and the subreq was promoted via fast_redirect, the filter would still point at the subreq - rather than the original r. So, we must update any filters pointing at rr to be r. This would cause lots of problems with mod_include with mod_dir requests such as seen in PR 7966. mod_include would be unsetting the headers_out of rr instead of r. But, we disassociate rr->headers_out and r->headers_out. Therefore, the C-L header in r->headers_out would remain - even though it bears no relation to what we will be outputting - causing problems. This also now permits chunked-encoding of mod_dir/mod_include requests which could never happen before and fixes the content-length problem seen in PR 7966. As hinted at in PR 7966, there is a race condition - if for some reason the server stalls reading an included file (or even better, placing a sleep in the cgi-bin script!), the invalid C-L may get propogated to the client. (Note that internal_internal_redirect has this same code fragment.) PR: 7966 --------------------------------------------------------------------- [T] modules/mappers/mod_imap.c (1.39) stoddard Log: Fix spelling/typo brianp Log: Because mod_imap's handler runs on every request in the default configuration, rearrange the code to keep it from allocating a few pages worth of local variables on the stack on requests that don't use imagemaps --------------------------------------------------------------------- [T] modules/mappers/mod_userdir.c (1.46) brianp Log: Short-circuit out of mod_userdir's translation handler faster on non "/~*" requests --------------------------------------------------------------------- [T] modules/metadata/mod_unique_id.c (1.37) trawick Log: Allow mod_unique_id to work on systems with no IPv4 address corresponding to their host name. trawick Log: fix a compiler error with picky compilers that (correctly) don't let you add to void * --------------------------------------------------------------------- [T] modules/proxy/mod_proxy.h (1.81) [T] modules/proxy/proxy_ftp.c (1.119) [T] modules/proxy/proxy_http.c (1.148) [T] modules/proxy/proxy_util.c (1.90) trawick Log: get mod_proxy to build on EBCDIC machines trawick Log: get basic HTTP proxy working on EBCDIC machines (I would bet that there is more work to do throughout proxy, but at least it can serve some pages now) trawick Log: fix a compile break on EBCDIC machines --------------------------------------------------------------------- [T] server/core.c (1.172) gregames Log: remove a goto --------------------------------------------------------------------- [T] server/protocol.c (1.97) jerenkrantz Log: Simplify this little chunk of code since r is set to f->r earlier. --------------------------------------------------------------------- [T] server/util.c (1.127) bnicholes Log: Added a check to make sure that h_aliases is not NULL before we try to dereference it in the for(...) loop. Attempting to dereference a NULL pointer was causing a fault if there were no aliases found. brianp Log: Updated ap_getword_white() to use the same coding style (and optimizations) as ap_getword() --------------------------------------------------------------------- [T] server/mpm/experimental/leader/leader.c (1.15) brianp Log: Switch back from atomic_cas to mutexes, based on all the current portability challenges with atomics brianp Log: Fixed the creation of the worker threads' data structures brianp Log: Removed creation of unused struct --------------------------------------------------------------------- [T] server/mpm/experimental/threadpool/threadpool.c (1.7) aaron Log: When we signal a condition variable, we need to own the lock that is associated with that condition variable. This isn't necessary for Solaris, but for Posix it is. aaron Log: Some threadpool fixes: - We don't need the listener_blocked flag, just check if the element we're adding makes the stack non-empty. - When we are terminated, return APR_EOF; catch this event in the worker thread and the listener thread. - When shutting down, always signal a potential listener thread. - Wait to signal the listener thread until after we add the worker element to the stack. aaron Log: Clarify an incorrect statement about why we're purposfully putting the signal inside of a mutex. aaron Log: Some more threadpool MPM changes: - Add a "state" variable to the worker_wakeup_info struct. This is used to make sure that we act on the correct signal, and to know when to shut down a worker thread. - Fix the call when the worker thread waits for a connection to use the new state variable and use mutexes around the cond_wait() call. - Change the interrupt_all() call to set the WORKER_TERMINATED state. - Add two AP_DEBUG_ASSERT() to make sure that we aren't waking up a worker thread before it is idle. aaron Log: Remove unused variable. aaron Log: Rename the worker_stack_interrupt_all() function as worker_stack_terminate(). Also, broadcast a signal after we set the terminate flag on the worker stack. --------------------------------------------------------------------- [T] server/mpm/winnt/mpm_winnt.c (1.268) stoddard Log: my_generation is in the shared scoreboard now. stoddard Log: Fix a bug in the mod_status display caused by some uninitialized fields. Prep for taking mpm_winnt to multi-process... --------------------------------------------------------------------- [T] server/mpm/worker/fdqueue.c (1.22) [T] server/mpm/worker/fdqueue.h (1.19) [T] server/mpm/worker/worker.c (1.120) aaron Log: Convert the worker MPM's fdqueue from a LIFO back into a FIFO. Since elements in the queue represent accept()ed connections, we want them to be processed in the order that they were received. (I erroneously converted it to a LIFO quite awhile ago in the hopes that it would improve cache efficiency.) Remember to perform a make clean in the worker directory after this patch, since this patch changes the size of the fd_queue_t object (which is allocated in worker.c). aaron Log: Add a "queue_info" structure to the worker MPM. This is used to prevent the listener thread from accept()ing more connections than there are available workers. This prevents long-running requests from starving connections that have been accepted but not yet processed. The queue_info is a simple counter, mutex, and condition variable. Only the listener thread blocks on the condition, and only when there are no idle workers. In the fast path there is a mutex lock, integer decrement, and and unlock (among a few conditionals). The worker threads each notify the queue_info when they are about to block on the normal worker_queue waiting for some connection to process, which wakes up any sleeping listener thread to go perform another accept() in parallel. jerenkrantz Log: Preserve the original error, or if the unlock fails then use that error instead. brianp Log: Move a potentially expensive pool cleanup operation outside the mutex-protected critical path of ap_queue_pop() brianp Log: Added a missing pool deletion case and simplified the conditionals for the error-case return (thanks to Justin for suggesting the latter) brianp Log: Moved the recycled pool list from the queue to the queue_info structure. The advantage of doing this is that it enables us to guarantee that the number of ptrans pools in existence at once is no greater than the number of worker threads, and that we'll never have to delete ptrans pools. brianp Log: Removed the "not_full" condition variable and associated conditional logic from the file descriptor queue, now that we can rely on the idle worker reference count to keep the listener from trying to push connections into a full queue brianp Log: Removed some code from the mutex-protected block in ap_queue_pop(). If the worker thread synchronization is working properly, it's not necessary to set these fields to NULL after removing an element from the queue, and it's IMO more important to have a shorter code path leading up to the mutex_unlock. I left in support for NULLing the fields when debugging, though. aaron Log: Detect APR_EINTR from ap_queue_pop() and avoid calling ap_queue_info_set_idle() more than once at a time per worker thread. This fixes an assert coredump. ===================================================================== apr ===================================================================== [T] srclib/apr/CHANGES RM: Will be updated prior to retagging. --------------------------------------------------------------------- [T] srclib/apr/STATUS RM: No harm in retagging this one. --------------------------------------------------------------------- [-] srclib/apr/configure.in (1.436) jerenkrantz Log: Not all platforms can mmap /dev/zero, so we need to do an explicit check for that. If that were to fail, then make it appear as /dev/zero never existed in the first place. PR: 8537 RM: Justin adviced against including this in the tag. See: <[EMAIL PROTECTED]> --------------------------------------------------------------------- [T] srclib/apr/include/apr_signal.h (1.11) fielding Log: Avoid warnings on Darwin by working around a bug in sys/include.h. --------------------------------------------------------------------- [T] srclib/apr/locks/unix/thread_mutex.c (1.10) brianp Log: Optimization: rearranged the mutex lock/unlock code to eliminate a conditional branch on platforms where PTHREAD_SETS_ERRNO is not defined (at present, this means everything except OS/390) --------------------------------------------------------------------- [T] srclib/apr/network_io/unix/sendrecv.c (1.267) trawick Log: Reduce the number of apr_sendfile calls on AIX and OS/390 by remembering when the kernel tells us the next one will block. --------------------------------------------------------------------- [T] srclib/apr/network_io/win32/sendrecv.c (1.47) stoddard Log: A read with 0 bytes is an APR_EOF. Found this running ab. ===================================================================== apr-util ===================================================================== [T] srclib/apr-util/xml/expat/buildconf.sh jerenkrantz Log: Bring over change from APR to allow cleaning of the libtool files on each invocation of buildconf.sh. This allows switching from 1.3 to 1.4 series libtool possible without intervention.
