dynamic handlers mod_fastcgi, mod_scgi, and mod_proxy can now read
response from backend prior to finishing sending request body.
If the backend closes the connections (or shuts down socket write
end so that lighttpd read() 0 to indicate EOF), then lighttpd will
abort attempting to send request body to backend.
x-ref:
"mod_fastcgi should handle "quick" responses"
https://redmine.lighttpd.net/issues/2566
"FastCGI FCGI_STDOUT before FCGI_STDIN bug"
https://redmine.lighttpd.net/issues/131
remove handle_joblist hook and remove the hooks defined in
mod_fastcgi and mod_scgi. The calls made to fdevent management
are redundant. If the calls were actually needed, then
mod_proxy would have needed a handle_joblist handler, too.
to detect client disconnect. Do so even when waiting on backend,
and not polling for POLLRD or POLLWR on client connection.
This reduces unnecessary load on backends when backends are slow
to respond and client has given up waiting.
x-ref:
"https://redmine.lighttpd.net/issues/399"
FastCGI performance on high load
read request body in dynamic handlers supporting request body
(mod_cgi, mod_fastcgi, mod_proxy, mod_scgi, mod_webdav)
(In the future, each dynamic handler might choose whether or not to
buffer request body or to stream request body to backend as request
body is received.)
modify mod_webdav to mark request in handle_physical hook, and move
the main logic to handle_subrequest hook, where the main logic is
for other dynamic handlers.
connection_handle_read()
connection_handle_read_ssl()
connection_handle_read_post_state()
no code changes besides making connection_handle_read() public
(by removing 'static' and adding to connections.h)
read request body right before calling subrequest handler,
allowing request to be handled prior to reading request body,
e.g. to send 401 Unauthorized response when authentication is required
(In the future, this might move into each dynamic handler which supports
request body (mod_cgi, mod_fastcgi, mod_proxy, mod_scgi, mod_webdav) so
that each dynamic handler can choose whether or not to buffer request
body or to stream request body to backend as request body is received.)
keep-alive is disabled if request body has not been completely read
prior to sending response
x-ref:
"HTTP 401 Unauthorized only sent back after full POST request is read"
https://redmine.lighttpd.net/issues/2541
for dynamic handlers mod_cgi, mod_fastcgi, mod_scgi, mod_proxy
(mod_cgi control flow logic simplification began in a prior commit)
- connection state machine calls the subrequest handler
- subrequest handler sets up the connection to the backend
and registers fdevent handler to handle backend events
- fdevent handler handles backend events and then schedules
a call to connection state machine
- when retrying an alternate backend, backend state is reset
and then response state is reset so that the connection state
machine will call back into the subrequest handler to retry
Dynamic handlers no longer directly modify connection state
(calls to connection_set_state() from dynamic handlers were removed)
Dynamic handlers no longer reset con->physical.path, and they
preserve con->mode when retrying alternate backends. This is done
to skip repeated processing in response.c:http_response_prepare()
While this patch increases consistency in control flow handling,
there is more work to be done that can further improve upon this.
x-ref:
"handle-req time too long"
https://redmine.lighttpd.net/issues/1149
Previous code would fail on partial write, EINTR, and ENOSPC.
Upon any of the above errors, this patch tries next tempdir in list,
if list of tempdirs provided by config option server.upload-dirs
x-ref:
"Problem when uploading large files"
https://redmine.lighttpd.net/issues/2588
github:
Closes #54
See doc/config/lighttpd.conf for explanation of listen() backlog queue
Additionally, mod_fastcgi and mod_scgi backend servers can now also be
configured with separate listen-backlog settings per server
x-ref:
"add server.listen-backlog option instead of hard-coded value (128 * 8) for listen()"
https://redmine.lighttpd.net/issues/2116
"Don't disable backend when overloaded"
https://redmine.lighttpd.net/issues/1825
github:
Closes #50
http_chunk_append_file() opens fd when appending file to chunkqueue.
Defers calculation of content length until response is finished.
This reduces race conditions pertaining to stat() and then (later)
open(), when the result of the stat() was used for Content-Length
or to generate chunked headers.
Note: this does not change how lighttpd handles files that are modified
in-place by another process after having been opened by lighttpd --
don't do that. This *does* improve handling of files that are
frequently modified via a temporary file and then atomically renamed
into place.
mod_fastcgi has been modified to use http_chunk_append_file_range() with
X-Sendfile2 and will open the target file multiple times if there are
multiple ranges.
Note: (future todo) not implemented for chunk.[ch] interfaces used by
range requests in mod_staticfile or by mod_ssi. Those uses could lead
to too many open fds. For mod_staticfile, limits should be put in place
for max number of ranges accepted by mod_staticfile. For mod_ssi,
limits would need to be placed on the maximum number of includes, and
the primary SSI file split across lots of SSI directives should either
copy the pieces or perhaps chunk.h could be extended to allow for an
open fd to be shared across multiple chunks. Doing either of these
would improve the performance of SSI since they would replace many file
opens on the pieces of the SSI file around the SSI directives.
x-ref:
"Serving a file that is getting updated can cause an empty response or incorrect content-length error"
https://redmine.lighttpd.net/issues/2655
github:
Closes #49
save physical path to environment (PATH_TRANSLATED_DIRINDEX)
when index generator is full path (not relative path)
fix uri.path when index generator is full path (not relative path)
Summary:
A new SSI directive, "ssi.conditional-requests", allows to inform
lighttpd which SSI pages should be considered as cacheable and which
should not. In particular, the "ETag" & "Last-Modified" headers will
only be sent for those SSI pages for which the directive is enabled.
Long description:
"ETag" and "Last-Modified" headers were being sent for all SSI pages,
regardless of whether they were cacheable or not. And yet, there was
no cache validation at all for any SSI page.
This commit fixes these two minor issues by adding a new directive,
"ssi.conditional-requests", which allows to specify which SSI pages
are cacheable and which are not, and by adding cache validation to
those SSI pages which are cacheable. And since sending ETags for
non-cacheable documents is not appropriate, they are no longuer
computed nor sent for those SSI pages which are not cacheable.
Regarding the "Last-Modified" header for non-cacheable documents,
the standards allow to either send the current date and time for
that header or to simply skip it. The approach chosen is to not send
it for non-cacheable SSI pages. "ETag" and "Last-Modified" headers
are therefore only sent for an SSI page if ssi.conditional-requests
is enabled for that page.
The ssi.conditional-requests directive can be enabled or disabled
globally and/or in any context. It is disabled by default.
An index.shtml which only includes deterministic SSI commands such as:
<!--#echo var="LAST_MODIFIED"-->
is a trivial example of a dynamic SSI page that is cacheable.
remove Cache-Control and Expires headers before handling error docs
(caching headers may have been set by mod_expire before http status
was determined to be an error)
x-ref:
"mod_expires sends headers on 404 responses"
https://redmine.lighttpd.net/issues/1919
bug #1499 was mostly fixed in 05858f6c
This patch additionally resets the cond_cache since scheme might change
x-ref:
"HTTPS env var should be "on" when using mod_extforward and the X-Forwarded-Proto header is set."
https://redmine.lighttpd.net/issues/1499
error if unix socket path is duplicated
(does not check across modules, but will detect duplicated unix socket
paths within fastcgi directives, and separately, duplicated unix socket
paths within scgi directives)
Common case is on non-empty files, and doing fstat() after open()
eliminates ToC-ToU between stat() and then open(). While file size
of the target file might still change, the fstat() is on the opened
file, and can not be on different file (which was possible with stat())
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3137 152afb58-edef-0310-8abb-c4023f1b3aa9
also change passing of fixed-sized arrays: need to pass pointer to array
as otherwise size does not get enforced
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3135 152afb58-edef-0310-8abb-c4023f1b3aa9
(still not supporting Depth: Infinity on directories)
patch by mstorsjo submitted as part of feature request #1953
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3133 152afb58-edef-0310-8abb-c4023f1b3aa9
For backwards compatibility with existing lighttpd configs, default is
mimetype.xattr-name = "Content-Type"
Those who wish to use the freedesktop.org definition of xattr mimetype
can set the following in the global lighttpd config:
mimetype.xattr-name = "user.mime_type"
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3131 152afb58-edef-0310-8abb-c4023f1b3aa9
lighttpd -t loads config file and performs syntax check
lighttpd -tt (new) performs preflight startup checks,
including loading and initializing modules, but skipping any
potentially destructive actions which might affect an already
running server (separate instance). These currently include:
- skipping pidfile modification
- skipping bind() to network sockets
- skipping open of error and access logs
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3130 152afb58-edef-0310-8abb-c4023f1b3aa9
lighttpd does not currently support request body transfer-codings
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3128 152afb58-edef-0310-8abb-c4023f1b3aa9
Warn at startup if any dirs in server.upload-dirs do not exist.
Take server.chroot into account, if set.
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3125 152afb58-edef-0310-8abb-c4023f1b3aa9
This is a proposal to add to lighttpd the famous SSI variables
SCRIPT_URI and SCRIPT_URL (known to Apache users), as well as a bonus
ENV variable called REQUEST_SCHEME.
SCRIPT_URI and SCRIPT_URL will be available as SSI variables from
within documents handled by mod_ssi.
They can be used like any other SSI var with the "#echo var" command:
<!--#echo var="SCRIPT_URI"-->
<!--#echo var="SCRIPT_URL"-->
Webmasters willing to display links to the W3C Validator will be able
to use:
<a href="http://validator.w3.org/check?uri=<!--#echo var="SCRIPT_URI"-->">…</a>
instead of the generic http://validator.w3.org/check?uri=referer link
which does not work on some (most?) browsers which do not send
referers when the link itself resides in a document sent through
https.
REQUEST_SCHEME will be available both as an environment variable. It
is defined as "http" or "https", depending on the scheme of the
connection. It is safe to use this name as it does not conflict with
any existing variable on Apache or Nginx. This is slightly different
from the HTTPS var which is often added by webadmins on their server's
configuration. EDIT: Some Apache modules also define REQUEST_SCHEME
with the same possible values as this proposal.
From: fbrosson <fbrosson@users.noreply.github.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3124 152afb58-edef-0310-8abb-c4023f1b3aa9
There actually was one bug: if the input consisted only of spaces,
it would read one byte too much.
`pre` was splitted into `pre2` and (already existing) `pre1` - the two
characters which were read before the current one in `c`.
Restructuring the loop eliminated some code before the loop, which was
similar to the one at the end of the loop.
From: Stefan Bühler <stbuehler@web.de>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3120 152afb58-edef-0310-8abb-c4023f1b3aa9
fbrosson
