Willy Tarreau [Fri, 6 Jan 2017 16:41:29 +0000 (17:41 +0100)]
MEDIUM: server: split the address and the port into two different fields
Keeping the address and the port in the same field causes a lot of problems,
specifically on the DNS part where we're forced to cheat on the family to be
able to keep the port. This causes some issues such as some families not being
resolvable anymore.
This patch first moves the service port to a new field "svc_port" so that the
port field is never used anymore in the "addr" field (struct sockaddr_storage).
All call places were adapted (there aren't that many).
Willy Tarreau [Fri, 6 Jan 2017 18:18:32 +0000 (19:18 +0100)]
BUG/MEDIUM: server: consider AF_UNSPEC as a valid address family
The DNS code is written so as to support AF_UNSPEC to decide on the
server family based on responses, but unfortunately snr_resolution_cb()
considers it as invalid causing a DNS storm to happen when a server
arrives with this family.
This situation is not supposed to happen as long as unresolved addresses
are forced to AF_INET, but this will change with the upcoming fixes and
it's possible that it's not granted already when changing an address on
the CLI.
Willy Tarreau [Fri, 6 Jan 2017 15:46:22 +0000 (16:46 +0100)]
BUG/MINOR: tools: fix off-by-one in port size check
port_to_str() checks that the port size is at least 5 characters instead
of at least 6. While in theory it could permit a buffer overflow, it's
harmless because all callers have at least 6 characters here.
This fix needs to be backported to 1.7, 1.6 and 1.5.
Willy Tarreau [Fri, 6 Jan 2017 11:21:38 +0000 (12:21 +0100)]
BUG/MINOR: config: emit a warning if http-reuse is enabled with incompatible options
http-reuse should normally not be used in conjunction with the proxy
protocol or with "usesrc clientip". While there's nothing fundamentally
wrong with this, whenever these options are used, the server expects the
IP address to be the source address for all requests, which doesn't make
sense with http-reuse.
Willy Tarreau [Wed, 4 Jan 2017 13:44:46 +0000 (14:44 +0100)]
BUG/MAJOR: http: fix risk of getting invalid reports of bad requests
Commits 5f10ea3 ("OPTIM: http: improve parsing performance of long
URIs") and 0431f9d ("OPTIM: http: improve parsing performance of long
header lines") introduced a bug in the HTTP parser : when a partial
request is read, the first part ends up on a 8-bytes boundary (or 4-byte
on 32-bit machines), the end lies in the header field value part, and
the buffer used to contain a CR character exactly after the last block,
then the parser could be confused and read this CR character as being
part of the current request, then switch to a new state waiting for an
LF character. Then when the next part of the request appeared, it would
read the character following what was erroneously mistaken for a CR,
see that it is not an LF and fail on a bad request. In some cases, it
can even be worse and the header following the hole can be improperly
indexed causing all sort of unexpected behaviours like a content-length
being ignored or a header appended at the wrong position. The reason is
that there's no control of and of parsing just after breaking out of the
loop.
Willy Tarreau [Sun, 4 Dec 2016 23:10:57 +0000 (00:10 +0100)]
MINOR: tools: add a generic hexdump function for debugging
debug_hexdump() prints to the requested output stream (typically stdout
or stderr) an hex dump of the blob passed in argument. This is useful
to help debug binary protocols.
Willy Tarreau [Thu, 5 Jan 2017 18:58:24 +0000 (19:58 +0100)]
BUILD: scripts: automatically update the branch in version.h when releasing
The stats page proudly displays "Updates (v1.5)". This version is inherited
from version.h which has not been updated since 1.5, so let's teach the
create-release script about it.
This must be backported to 1.7. 1.6 now uses the same script (externally)
for the release and will automatically benefit from it.
Willy Tarreau [Wed, 4 Jan 2017 13:51:22 +0000 (14:51 +0100)]
BUG/MINOR: http: report real parser state in error captures
Error captures almost always report a state 26 (MSG_ERROR) making it
very hard to know what the parser was expecting. The reason is that
we have to switch to MSG_ERROR to trigger the dump, and then during
the dump we capture the current state which is already MSG_ERROR. With
this change we now copy the current state into an err_state field that
will be reported as the faulty state.
This patch looks a bit large because the parser doesn't update the
current state until it runs out of data so the current state is never
known when jumping to ther error label! Thus the code had to be updated
to take copies of the current state before switching to MSG_ERROR based
on the switch/case values.
As a bonus, it now shows the current state in human-readable form and
not only in numeric form ; in the past it was not an issue since it was
always 26 (MSG_ERROR).
At least now we can get exploitable invalid request/response reports :
[05/Jan/2017:19:28:57.095] frontend f (#2): invalid request
backend <NONE> (#-1), server <NONE> (#-1), event #1
src 127.0.0.1:39894, session #4, session flags 0x00000080
HTTP msg state MSG_RQURI(4), msg flags 0x00000000, tx flags 0x00000000
HTTP chunk len 0 bytes, HTTP body len 0 bytes
buffer flags 0x00908002, out 0 bytes, total 20 bytes
pending 20 bytes, wrapping at 16384, error at position 5:
00000 GET /\e HTTP/1.0\r\n
00017 \r\n
00019 \n
[05/Jan/2017:19:28:33.827] backend b (#3): invalid response
frontend f (#2), server s1 (#1), event #0
src 127.0.0.1:39718, session #0, session flags 0x000004ce
HTTP msg state MSG_HDR_NAME(17), msg flags 0x00000000, tx flags 0x08300000
HTTP chunk len 0 bytes, HTTP body len 0 bytes
buffer flags 0x80008002, out 0 bytes, total 59 bytes
pending 59 bytes, wrapping at 16384, error at position 31:
BUG/MAJOR: channel: Fix the definition order of channel analyzers
It is important to defined analyzers (AN_REQ_* and AN_RES_*) in the same order
they are evaluated in process_stream. This order is really important because
during analyzers evaluation, we run them in the order of the lower bit to the
higher one. This way, when an analyzer adds/removes another one during its
evaluation, we know if it is located before or after it. So, when it adds an
analyzer which is located before it, we can switch to it immediately, even if it
has already been called once but removed since.
With the time, and introduction of new analyzers, this order was broken up. the
main problems come from the filter analyzers. We used values not related with
their evaluation order. Furthermore, we used same values for request and response
analyzers.
So, to fix the bug, filter analyzers have been splitted in 2 distinct lists to
have different analyzers for the request channel than those for the response
channel. And of course, we have moved them to the right place.
Some other analyzers have been reordered to respect the evaluation order:
* AN_REQ_HTTP_TARPIT has been moved just before AN_REQ_SRV_RULES
* AN_REQ_PRST_RDP_COOKIE has been moved just before AN_REQ_STICKING_RULES
* AN_RES_STORE_RULES has been moved just after AN_RES_WAIT_HTTP
Note today we have 29 analyzers, all stored into a 32 bits bitfield. So we can
still add 4 more analyzers before having a problem. A good way to fend off the
problem for a while could be to have a different bitfield for request and
response analyzers.
[wt: all of this must be backported to 1.7, and part of it must be backported
to 1.6 and 1.5]
Olivier Doucet [Mon, 2 Jan 2017 10:48:57 +0000 (11:48 +0100)]
BUG/MINOR: option prefer-last-server must be ignored in some case
when using "option prefer-last-server", we may not always stay on
the same backend if option balance told us otherwise.
For example, backend may change in the following cases:
balance hdr()
balance rdp-cookie
balance source
balance uri
balance url_param
David Carlier [Mon, 21 Nov 2016 21:25:58 +0000 (21:25 +0000)]
MEDIUM: regex: pcre2 support
this adds a support of the newest pcre2 library,
more secure than its older sibling in a cost of a
more complex API.
It works pretty similarly to pcre's part to keep
the overall change smooth, except :
- we define the string class supported at compile time.
- after matching the ovec data is properly sized, althought
we do not take advantage of it here.
- the lack of jit support is treated less 'dramatically'
as pcre2_jit_compile in this case is 'no-op'.
In systemd mode (-Ds), the master haproxy process is waiting for each
child to exit in a specific order. If a process die when it's not his
turn, it will become a zombie process until every processes exit.
The master is now waiting for any process to exit in any order.
This patch should be backported to 1.7, 1.6 and 1.5.
Willy Tarreau [Thu, 22 Dec 2016 20:58:38 +0000 (21:58 +0100)]
BUG/MEDIUM: ssl: for a handshake when server-side SNI changes
Calling SSL_set_tlsext_host_name() on the current SSL ctx has no effect
if the session is being resumed because the hostname is already stored
in the session and is not advertised again in subsequent connections.
It's visible when enabling SNI and health checks at the same time because
checks do not send an SNI and regular traffic reuses the same connection,
resulting in no SNI being sent.
The only short-term solution is to reset the reused session when the
SNI changes compared to the previous one. It can make the server-side
performance suffer when SNIs are interleaved but it will work. A better
long-term solution would be to keep a small cache of a few contexts for
a few SNIs.
Now with SSL_set_session(ctx, NULL) it works. This needs to be double-
checked though. The man says that SSL_set_session() frees any previously
existing context. Some people report a bit of breakage when calling
SSL_set_session(NULL) on openssl 1.1.0a (freed session not reusable at
all though it's not an issue for now).
Marcin Deranek [Thu, 22 Dec 2016 15:21:08 +0000 (16:21 +0100)]
BUG/MINOR: backend: nbsrv() should return 0 if backend is disabled
According to nbsrv() documentation this fetcher should return "an
integer value corresponding to the number of usable servers".
In case backend is disabled none of servers is usable, so I believe
fetcher should return 0.
Willy Tarreau [Thu, 22 Dec 2016 22:12:01 +0000 (23:12 +0100)]
CLEANUP: ssl: move most ssl-specific global settings to ssl_sock.c
Historically a lot of SSL global settings were stored into the global
struct, but we've reached a point where there are 3 ifdefs in it just
for this, and others in haproxy.c to initialize it.
This patch moves all the private fields to a new struct "global_ssl"
stored in ssl_sock.c. This includes :
char *crt_base;
char *ca_base;
char *listen_default_ciphers;
char *connect_default_ciphers;
int listen_default_ssloptions;
int connect_default_ssloptions;
int tune.sslprivatecache; /* Force to use a private session cache even if nbproc > 1 */
unsigned int tune.ssllifetime; /* SSL session lifetime in seconds */
unsigned int tune.ssl_max_record; /* SSL max record size */
unsigned int tune.ssl_default_dh_param; /* SSL maximum DH parameter size */
int tune.ssl_ctx_cache; /* max number of entries in the ssl_ctx cache. */
The "tune" part was removed (useless here) and the occasional "ssl"
prefixes were removed as well. Thus for example instead of
global.tune.ssl_default_dh_param
we now have :
global_ssl.default_dh_param
A few initializers were present in the constructor, they could be brought
back to the structure declaration.
A few other entries had to stay in global for now. They concern memory
calculationn (used in haproxy.c) and stats (used in stats.c).
The code is already much cleaner now, especially for global.h and haproxy.c
which become readable.
Willy Tarreau [Thu, 22 Dec 2016 21:46:15 +0000 (22:46 +0100)]
CLEANUP: ssl: move tlskeys_finalize_config() to a post_check callback
tlskeys_finalize_config() was the only reason for haproxy.c to still
require ifdef and includes for ssl_sock. This one fits perfectly well
in the late initializers so it was changed to be registered with
hap_register_post_check().
Willy Tarreau [Thu, 22 Dec 2016 20:16:08 +0000 (21:16 +0100)]
MINOR: ssl_sock: implement and use prepare_srv()/destroy_srv()
Now we can simply check the transport layer at run time and decide
whether or not to initialize or destroy these entries. This removes
other ifdefs and includes from cfgparse.c, haproxy.c and hlua.c.
Willy Tarreau [Thu, 22 Dec 2016 19:44:00 +0000 (20:44 +0100)]
CLEANUP: connection: remove all direct references to raw_sock and ssl_sock
Now we exclusively use xprt_get(XPRT_RAW) instead of &raw_sock or
xprt_get(XPRT_SSL) for &ssl_sock. This removes a bunch of #ifdef and
include spread over a number of location including backend, cfgparse,
checks, cli, hlua, log, server and session.
Willy Tarreau [Thu, 22 Dec 2016 19:25:26 +0000 (20:25 +0100)]
MINOR: connection: add a minimal transport layer registration system
There are still a lot of #ifdef USE_OPENSSL in the code (still 43
occurences) because we never know if we can directly access ssl_sock
or not. This patch attacks the problem differently by providing a
way for transport layers to register themselves and for users to
retrieve the pointer. Unregistered transport layers will point to NULL
so it will be easy to check if SSL is registered or not. The mechanism
is very inexpensive as it relies on a two-entries array of pointers,
so the performance will not be affected.
Willy Tarreau [Thu, 22 Dec 2016 17:38:00 +0000 (18:38 +0100)]
MINOR: server: move the use_ssl field out of the ifdef USE_OPENSSL
Having it in the ifdef complicates certain operations which require
additional ifdefs just to access a member which could remain zero in
non-ssl cases. Let's move it out, it will not even increase the
struct size on 64-bit machines due to alignment.
Instead of hard-coding all SSL destruction in cfgparse.c and haproxy.c,
we now register this new function as the transport layer's destroy_bind_conf()
and call it only when defined. This removes some non-obvious SSL-specific
code and #ifdefs from cfgparse.c and haproxy.c
Willy Tarreau [Thu, 22 Dec 2016 16:30:20 +0000 (17:30 +0100)]
MINOR: connection: add a new destroy_bind_conf() entry to xprt_ops
This one will be set by the transport layers which want to destroy
a bind_conf. It will typically be used by SSL to release certificates,
CAs and so on.
Instead of hard-coding all SSL preparation in cfgparse.c, we now register
this new function as the transport layer's prepare_bind_conf() and call it
only when definied. This removes some non-obvious SSL-specific code from
cfgparse.c as well as a #ifdef.
Willy Tarreau [Thu, 22 Dec 2016 16:19:24 +0000 (17:19 +0100)]
MINOR: connection: add a new prepare_bind_conf() entry to xprt_ops
This one will be set by the transport layers which want to initialize
a bind_conf. It will typically be used by SSL to load certificates, CAs
and so on.
Willy Tarreau [Thu, 22 Dec 2016 16:08:28 +0000 (17:08 +0100)]
MEDIUM: ssl: remote the proxy argument from most functions
Most of the SSL functions used to have a proxy argument which was mostly
used to be able to emit clean errors using Alert(). First, many of them
were converted to memprintf() and don't require this pointer anymore.
Second, the rare which still need it also have either a bind_conf argument
or a server argument, both of which carry a pointer to the relevant proxy.
So let's now get rid of it, it needlessly complicates the API and certain
functions already have many arguments.
Willy Tarreau [Wed, 21 Dec 2016 23:13:31 +0000 (00:13 +0100)]
MEDIUM: move listener->frontend to bind_conf->frontend
Historically, all listeners have a pointer to the frontend. But since
the introduction of SSL, we now have an intermediary layer called
bind_conf corresponding to a "bind" line. It makes no sense to have
the frontend on each listener given that it's the same for all
listeners belonging to a same bind_conf. Also certain parts like
SSL can only operate on bind_conf and need the frontend.
This patch fixes this by moving the frontend pointer from the listener
to the bind_conf. The extra indirection is quite cheap given and the
places were this is used are very scarce.
Willy Tarreau [Wed, 21 Dec 2016 21:04:54 +0000 (22:04 +0100)]
MINOR: listener: move the transport layer pointer to the bind_conf
A mistake was made when the socket layer was cut into proto and
transport, the transport was attached to the listener while all
listeners in a single "bind" line always have exactly the same
transport. It doesn't seem obvious but this is the reason why there
are so many #ifdefs USE_OPENSSL in cfgparse : a lot of operations
have to be open-coded because cfgparse only manipulates bind_conf
and we don't have the information of the transport layer here.
Very little code makes use of the transport layer, mainly session
setup and log. These places can afford an extra pointer indirection
(the listener points to the bind_conf). This change is thus very small,
it saves a little bit of memory (8B per listener) and makes the code
more flexible.
Willy Tarreau [Thu, 22 Dec 2016 14:59:02 +0000 (15:59 +0100)]
MEDIUM: spoe: don't create a dummy listener for outgoing connections
The code currently creates a listener only to ensure that sess->li is
properly populated, and to retrieve the frontend (which is also available
directly from the session).
It turns out that the current infrastructure (for a large part) already
supports not having any listener on a session (since Lua does the same),
except for the following places which were not yet converted :
- session_count_new() : used by session_accept_fd, ie never for spoe
- session_accept_fd() : never used here, an applet initiates the session
- session_prepare_log_prefix() : embryonic sessions only, thus unused
- session_kill_embryonic() : same
- conn_complete_session() : same
- build_log_line() for fields %cp, %fp and %ft : unused here but may change
- http_wait_for_request() and subsequent functions : unused here
Thus for now it's as safe to run SPOE without listener as it is for Lua,
and this was an obstacle against some cleanups of the listener code. The
places above should be plugged so that it becomes save over the long term
as well.
An alternative in the future might be to create a dummy listener that
outgoing connections could use just to avoid keeping a null here.
Willy Tarreau [Thu, 22 Dec 2016 17:14:41 +0000 (18:14 +0100)]
MINOR: tcp-rules: check that the listener exists before updating its counters
The tcp rules may be applied to a TCP stream initiated by applets (spoe,
lua, peers, later H2). These ones do not necessarily have a valid listener
so we must verify the field is not null before updating the stats. For now
there's no way to trigger this bug because lua and peers don't have analysers,
h2 is not implemented and spoe has a dummy listener. But this threatens to
break at any instant.
Thierry FOURNIER [Mon, 19 Dec 2016 15:50:42 +0000 (16:50 +0100)]
BUG/MINOR: stats: fix be/sessions/current out in typed stats
"scur" was typed as "limit" (FO_CONFIG) and "config value" (FN_LIMIT).
The real types of "scur" are "metric" (FO_METRIC) and "gauge"
(FN_GAUGE). FO_METRIC and FN_GAUGE are the value 0.
Willy Tarreau [Thu, 22 Dec 2016 16:57:46 +0000 (17:57 +0100)]
BUG/MEDIUM: ssl: avoid double free when releasing bind_confs
ssl_sock functions don't mark pointers as NULL after freeing them. So
if a "bind" line specifies some SSL settings without the "ssl" keyword,
they will get freed at the end of check_config_validity(), then freed
a second time on exit. Simply mark the pointers as NULL to fix this.
This fix needs to be backported to 1.7 and 1.6.
Willy Tarreau [Thu, 22 Dec 2016 20:54:21 +0000 (21:54 +0100)]
BUG/MEDIUM: ssl: properly reset the reused_sess during a forced handshake
We have a bug when SSL reuse is disabled on the server side : we reset
the context but do not set it to NULL, causing a multiple free of the
same entry. It seems like this bug cannot appear as-is with the current
code (or the conditions to get it are not obvious) but it did definitely
strike when trying to fix another bug with the SNI which forced a new
handshake.
This fix should be backported to 1.7, 1.6 and 1.5.
Willy Tarreau [Thu, 22 Dec 2016 18:46:17 +0000 (19:46 +0100)]
MEDIUM: compression: move the zlib-specific stuff from global.h to compression.c
This finishes to clean up the zlib-specific parts. It also unbreaks recent
commit b97c6fb ("CLEANUP: compression: use the build options list to report
the algos") which broke USE_ZLIB due to MAXWBITS not being defined anymore
in haproxy.c.
It's worth mentionning that some of them used to have incorrect sign
checks possibly resulting in some negative values being used. All of
them are now checked for being positive.
Willy Tarreau [Wed, 21 Dec 2016 21:44:46 +0000 (22:44 +0100)]
MINOR: cfgparse: move parsing of "ca-base" and "crt-base" to ssl_sock
This removes 2 #ifdefs and makes the code much cleaner. The controls
are still there and the two parsers have been merged into a single
function ssl_parse_global_ca_crt_base().
It's worth noting that there's still a check to prevent a change when
the value was already specified. This test seems useless and possibly
counter-productive, it may have to be revisited later, but for now it
was implemented identically.
Willy Tarreau [Wed, 21 Dec 2016 21:41:44 +0000 (22:41 +0100)]
MINOR: cfgparse: add two new functions to check arguments count
We already had alertif_too_many_args{,_idx}(), but these ones are
specifically designed for use in cfgparse. Outside of it we're
trying to avoid calling Alert() all the time so we need an
equivalent using a pointer to an error message.
These new functions called too_many_args{,_idx)() do exactly this.
They don't take the file name nor the line number which they have
no use for but instead they take an optional pointer to an error
message and the pointer to the error code is optional as well.
With (NULL, NULL) they'll simply check the validity and return a
verdict. They are quite convenient for use in isolated keyword
parsers.
These two new functions as well as the previous ones have all been
exported.
Willy Tarreau [Wed, 21 Dec 2016 20:25:06 +0000 (21:25 +0100)]
CLEANUP: da: move global settings out of the global section
We replaced global.deviceatlas with global_deviceatlas since there's no need
to store all this into the global section. This removes the last #ifdefs,
and now the code is 100% self-contained in da.c. The file da.h was now
removed because it was only used to load dac.h, which is more easily
loaded directly from da.c. It provides another good example of how to
integrate code in the future without touching the core parts.
Willy Tarreau [Wed, 21 Dec 2016 20:18:44 +0000 (21:18 +0100)]
CLEANUP: 51d: move global settings out of the global section
We replaced global._51degrees with global_51degrees since there's no need
to store all this into the global section. This removes the last #ifdefs,
and now the code is 100% self-contained in 51d.c. The file 51d.h was now
removed because it was only used to load 51Degrees.h, which is more easily
loaded from 51d.c. It provides a good example of how to integrate code in
the future without touching the core parts.
Willy Tarreau [Wed, 21 Dec 2016 13:57:34 +0000 (14:57 +0100)]
CLEANUP: wurfl: move global settings out of the global section
We replaced global.wurfl with global_wurfl since there's no need to store
all this into the global section. This removes the last #ifdefs, and now
the code is 100% self-contained in wurfl.c. It provides a good example of
how to integrate code in the future without touching the core parts.
Willy Tarreau [Wed, 21 Dec 2016 19:59:01 +0000 (20:59 +0100)]
CLEANUP: 51d: register the deinitialization function
deinit_51degrees() is not called anymore from haproxy.c, removing
2 #ifdefs and one include. The function was made static. The include
file still includes 51Degrees.h which is needed by global.h and 51d.c
so it was not touched beyond this last function removal.
Willy Tarreau [Wed, 21 Dec 2016 19:52:38 +0000 (20:52 +0100)]
CLEANUP: wurfl: register the deinit function via the dedicated list
By registering the deinit function we avoid another #ifdef in haproxy.c.
The ha_wurfl_deinit() function has been made static and unexported. Now
proto/wurfl.h is totally empty, the code being self-contained in wurfl.c,
so the useless .h has been removed.
Willy Tarreau [Wed, 21 Dec 2016 19:46:26 +0000 (20:46 +0100)]
MINOR: haproxy: add a registration for post-deinit functions
The 3 device detection engines stop at the same place in deinit()
with the usual #ifdefs. Similar to the other functions we can have
some late deinitialization functions. These functions do not return
anything however so we have to use a different type.
Willy Tarreau [Wed, 21 Dec 2016 19:39:16 +0000 (20:39 +0100)]
CLEANUP: da: make use of the late init registration code
Instead of having a #ifdef in the main init code we now use the registered
init functions. Doing so also enables error checking as errors were previously
reported as alerts but ignored. Also they were incorrect as the 'status'
variable was hidden by a second one and was always reporting DA_SYS (which
is apparently an error) in every case including the case where no file was
loaded. The init_deviceatlas() function was unexported since it's not used
outside of this place anymore.
Willy Tarreau [Wed, 21 Dec 2016 19:30:05 +0000 (20:30 +0100)]
CLEANUP: 51d: make use of the late init registration
This removes some #ifdefs from the main haproxy code path. Function
init_51degrees() now returns ERR_* instead of exit(1) on error, and
this function was made static and is not exported anymore.
Willy Tarreau [Wed, 21 Dec 2016 19:20:17 +0000 (20:20 +0100)]
CLEANUP: wurfl: make use of the late init registration
This removes some #ifdefs from the main haproxy code path and enables
error checking. The current code only makes use of warnings even for
some errors that look serious. While this choice is questionnable, it
has been kept as-is, and only the return codes were adapted to ERR_WARN
to at least report that some warnings were emitted. ha_wurfl_init() was
unexported as it's not needed anymore.
Willy Tarreau [Wed, 21 Dec 2016 19:04:48 +0000 (20:04 +0100)]
CLEANUP: checks: make use of the post-init registration to start checks
Instead of calling the checks directly from the init code, we now
register the start_checks() function to be run at this point. This
also allows to unexport the check init function and to remove one
include from haproxy.c.
Willy Tarreau [Wed, 21 Dec 2016 18:57:00 +0000 (19:57 +0100)]
MINOR: haproxy: add a registration for post-check functions
There's a significant amount of late initialization calls which are
performed after the point where we exit in check mode. These calls
are used to allocate resource and perform certain slow operations.
Let's have a way to register some functions which need to be called
there instead of having this multitude of #ifdef in the init path.
Willy Tarreau [Wed, 21 Dec 2016 18:30:30 +0000 (19:30 +0100)]
CLEANUP: compression: use the build options list to report the algos
This removes 2 #ifdef, an include, an ugly construct and a wild "extern"
declaration from haproxy.c. The message indicating that compression is
*not* enabled is not there anymore.
Willy Tarreau [Wed, 21 Dec 2016 17:43:10 +0000 (18:43 +0100)]
MINOR: haproxy: add a registration for build options
Many extensions now report some build options to ease debugging, but
this is now being done at the expense of code maintainability. Let's
provide a registration function to do this so that we can start to
remove most of the #ifdefs from haproxy.c (18 currently just for a
single function).
Thierry FOURNIER [Sat, 17 Dec 2016 11:09:51 +0000 (12:09 +0100)]
BUG/MINOR: lua: memleak when Lua/cli fails
If the memory allocator fails, it return a bad code, and the execution
continue. If the Lua/cli initializer fails, the allocated struct is not
released.
When data are sent from a cosocket, the action is done in the context of the
applet running a lua stack and not in the context of the applet owning the
cosocket. So we must take care, explicitly, that this last applet have a buffer
(the req buffer of the cosocket).
Thierry FOURNIER [Fri, 16 Dec 2016 07:48:32 +0000 (08:48 +0100)]
MINOR/DOC: lua: just precise one thing
In the case of applet, the Lua context is taken from session
when we get the private values. This patch just update comments
associated to this action because it is not obvious.
Willy Tarreau [Fri, 16 Dec 2016 07:02:21 +0000 (08:02 +0100)]
CONTRIB: tcploop: add limits.h to fix build issue with some compilers
Just got this while cross-compiling :
tcploop.c: In function 'tcp_recv':
tcploop.c:444:48: error: 'INT_MAX' undeclared (first use in this function)
tcploop.c:444:48: note: each undeclared identifier is reported only once for each function it appears in
Willy Tarreau [Fri, 16 Dec 2016 17:47:27 +0000 (18:47 +0100)]
MINOR: appctx/cli: remove the "tlskeys" entry from the appctx union
This one now migrates to the general purpose cli.p0 for the ref pointer,
cli.i0 for the dump_all flag and cli.i1 for the dump_keys_index. A few
comments were added.
The applet.h file doesn't depend on openssl anymore. It's worth noting
that the previous dependency was accidental and only used to work because
all files including this one used to have openssl included prior to
loading this file.
Willy Tarreau [Fri, 16 Dec 2016 17:23:39 +0000 (18:23 +0100)]
MINOR: appctx/cli: remove the "server_state" entry from the appctx union
This one now migrates to the general purpose cli.p0 for the proxy pointer,
cli.p1 for the server pointer, and cli.i0 for the proxy's instance if only
one has to be dumped.
Willy Tarreau [Fri, 16 Dec 2016 11:37:03 +0000 (12:37 +0100)]
MINOR: cli: add two general purpose pointers and integers in the CLI struct
Most of the keywords don't need to have their own entry in the appctx
union, they just need to reuse some generic pointers like we've been
used to do in the appctx with st{0,1,2}. This patch adds p0, p1, i0, i1
and initializes them to zero before calling the parser. This way some
of the simplest existing keywords will be able to disappear from the
union.
It's worth noting that this is an extension to what was initially
attempted via the "private" member that I removed a few patches ago by
not understanding how it was supposed to be used. Here the fact that
we share the same union will force us to be stricter: the code either
uses the general purpose variables or it uses its own fields but not
both.
Willy Tarreau [Fri, 16 Dec 2016 11:33:47 +0000 (12:33 +0100)]
CLEANUP: stats: move a misplaced stats context initialization
This is a leftover from the cleanup campaign, the stats scope was still
initialized by the CLI instead of being initialized by the stats keyword
parsers. This should probably be backported to 1.7 to make the code more
consistent.
Willy Tarreau [Fri, 16 Dec 2016 11:14:12 +0000 (12:14 +0100)]
CLEANUP: applet: group all CLI contexts together
The appctx storage became a real mess along the years. It now contains
mostly CLI-specific parts that share the same storage as the "cli" part
which in fact only contains the fields needed to pass an error message
to the caller, and it also has room a few other regular applets which
may become more and more common.
This first patch moves the parts around in the union so that all
standard applet parts are grouped together and the CLI-specific ones
are grouped together. It also adds a few comments to indicate what
certain parts are used for since it's sometimes a bit confusing.
Willy Tarreau [Fri, 16 Dec 2016 16:59:25 +0000 (17:59 +0100)]
MINOR: cli: automatically enable a CLI I/O handler when there's no parser
Sometimes a registered keyword will not need any specific parsing nor
initialization, so it's annoying to have to write an empty parsing
function returning zero just for this.
This patch makes it possible to automatically call a keyword's I/O
handler of when the parsing function is not defined, while still allowing
a parser to set the I/O handler itself.
Thierry FOURNIER [Fri, 16 Dec 2016 12:07:22 +0000 (13:07 +0100)]
MINOR: lua/signals: Remove Lua part from signals.
The signals system embedded in Lua can be tranformed in general purpose
signals code. To reach this goal, this path removes the Lua part of the
signals.
This is an easy job, because Lua is useles with signal. I change just two
prototypes.
Thierry FOURNIER [Fri, 16 Dec 2016 10:54:07 +0000 (11:54 +0100)]
MEDIUM: lua: use memory pool for hlua struct in applets
The struct hlua size is 128 bytes. The size is the biggest of all the elements
of the union embedded in the appctx struct. With HTTP2, it is possible that this
appctx struct will be use many times for each connection, so the 128 bytes are
a little bit heavy for the global memory consomation.
This patch replace the embbeded hlua struct by a pointer and an associated memory
pool. Now, the memory for lua is allocated only if it is required.
Willy Tarreau [Fri, 16 Dec 2016 11:56:31 +0000 (12:56 +0100)]
BUG/MINOR: cli: "show cli sockets" would always report process 64
Another small bug in "show cli sockets" made the last fix always report
process 64 due to a signedness issue in the shift operation when building
the mask.