Skip to content
  1. Feb 20, 2018
    • Yann Ylavic's avatar
      Merge r1823047, r1824454, r1824463, r1824464, r1824497, r1824862 from trunk: · 566b629a
      Yann Ylavic authored 
      mpm_event: move lingering close "sucker" from the listener to worker(s).

This was the last non-constant time action performed by the listener thread.

It's now handled by the worker thread directly after entering lingering close,
which should directly address the cases when the socket is already closed
remotely at that time, hence avoid more scheduling (it may be the common case
for some scenarios).

And it's only if the above would need blocking (i.e. more data to suck) that
the socket is added to the pollset for the listener to re-schedule a worker
later when ready. If no worker is available at that time then the socket is
forcibly closed (similarly to what's done for keepalive connections in this
case).

Also, since process_lingering_close() is now called by a worker thread and
with almost no depth in the call stack, we can grow the size of the "suck"
buffer from 2K to 32K to potentially call recv() up to sixteen times less.


mpm_event: follow up to r1823047.

Update clogged counter on read_request retry too.


mpm_event: follow up to r1823047: simplify "clogging" logic (reentrance).


mpm_event: follow up to r1823047: complete state validation after processing.


mpm_event: follow up to r1823047: CHANGES entry.


mpm_event: follow up to r1823047 and r1824464.

MMN bump for CONN_STATE_NUM, plus don't consider CONN_STATE_LINGER_* as valid
states returned process_connection (never have been).


Submitted by: ylavic
Reviewed by: ylavic, minfrin, jim


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x@1824868 13f79535-47bb-0310-9956-ffa450edef68
      566b629a
  3. Feb 19, 2018
  5. Feb 16, 2018
  7. Feb 15, 2018
  9. Feb 14, 2018
  11. Feb 13, 2018
  13. Feb 10, 2018
  15. Feb 09, 2018
  17. Jan 18, 2018
  19. Jan 13, 2018
  21. Jan 10, 2018
    • Yann Ylavic's avatar
      Merge r1818040 from trunk: · 01a4e02e
      Yann Ylavic authored 
      PR61891: looping over mostly full LDAP cache

  *) mod_ldap: Fix a case where a full LDAP cache would continually fail to
     purge old entries and log AH01323. PR61891.


Submitted By: Hendrik Harms <hendrik.harms gmail.com>
Committed By: covener
Reviewed By: covener, jim, ylavic


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x@1820800 13f79535-47bb-0310-9956-ffa450edef68
      01a4e02e
    • Yann Ylavic's avatar
      Merge r1818804, r1818951, r1818958, r1818960, r1819027, r1819214, r1820035 from trunk: · 6e5b7f3d
      Yann Ylavic authored 
      mpm_event: close connections not reported as handled by any module.

This avoids losing track of them and leaking scoreboard entries.
PR 61551.


mpm_event: follow up to r1818804.

Address corner case where connection is aborted due to ap_run_pre_connection()
failure, and update comment about ap_run_process_connection() expected return
status and state.


mpm_event: follow up to r1818804 and r1818951.

Align comment and fix typos.


mpm_event: follow up to r1818804.

Allow DONE as a successful ap_run_process_connection() return value, for
instance h2_conn_run() and h2_task_process_conn() uses it, third-party
modules may too...


mpm_event: follow up to r1818804 and r1818951.

Be more correct in comment about CONN_STATE_WRITE_COMPLETION.
We currently have/need no state to simply wait for readability on a socket,
so the previous comment was misleading. Write completion can't be used for
a simple "wait for read event and come back to process_connection hooks".


mpm_event: follow up to r1818804 and r1818960.

Align mod_http2 with expected returned state from process_connection hooks in
async MPMs.
When the master connection is handled, enter CONN_STATE_LINGER in any case.


Add missing APLOGNO


Submitted by: ylavic, jailletc36
Reviewed by: ylavic, icing, covener


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x@1820796 13f79535-47bb-0310-9956-ffa450edef68
      6e5b7f3d
    • Yann Ylavic's avatar
      Merge r1809881, r1809973, r1809976, r1812075 from trunk: · eb99f1c7
      Yann Ylavic authored 
      core: deregister all hooks before leaving pconf, otherwise some late cleanup
or function call (e.g. ap_log) may use one while DSOs are unloaded.

See PR 61558 (double/second fault).


core, MPMs unix: follow up to r1809881.

Deregister all hooks first (in pre_cleanup), by doing it last we could still
have had them run when DSOs were unloaded.

Likewise, avoid double faults when handling fatal signals by restoring the
default handler before pconf is cleared (we can't ap_log_error there).

Finally, we need to ignore sig_term/restart (do nothing) when the main
process is exiting (i.e. ap_pglobal is destroyed), since retained_data are
freed.

Aimed to fix all faults in PR 61558.


MPMs unix: follow up to r1809881 and r1809973.

unset_signals() is called when ap_pglobal is destroyed too.


Follow up to r1809881: CHANGES entry.


Reviewed by: ylavic, jim, covener


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/2.4.x@1820794 13f79535-47bb-0310-9956-ffa450edef68
      eb99f1c7
  23. Jan 04, 2018
  25. Dec 22, 2017
  27. Dec 21, 2017
  29. Dec 03, 2017
  31. Dec 02, 2017
  33. Nov 13, 2017