Thierry FOURNIER [Mon, 17 Aug 2015 16:38:24 +0000 (18:38 +0200)]
BUG/MEDIUM: vars: segfault during the configuration parsing
This bug is introduced by the patch 48a9cd104d07919fdd941b36e9f927a20d55edd3
"MINOR: vars: reduce the code size of some wrappers". A dereferencement was
removed.
Baptiste Assmann [Mon, 17 Aug 2015 12:25:02 +0000 (14:25 +0200)]
MINOR: server SRV_ADMF_CMAINT flag doesn't imply SRV_ADMF_FMAINT
The newly created server flag SRV_ADMF_CMAINT means that the server is
in 'disabled' mode because of configuration statement 'disabled'.
The flag SRV_ADMF_FMAINT should not be set anymore in such case and is
reserved only when the server is Forced in maintenance mode from the
stats socket.
Thierry FOURNIER [Sun, 16 Aug 2015 10:03:39 +0000 (12:03 +0200)]
BUG/MEDIUM: stream: The stream doen't inherit SC from the session
During the processing of tcp-request connection, the stream doesn't exists, so the
stick counters are stored in the session. When the stream is created it must
inherit from the session sc.
MEDIUM: cli: rely on the map's output type instead of the sample type
Next patch will remove sample_storage->type, and the only user is the
"show map" feature on the CLI which can use the map's output type instead.
Let's do that first.
MINOR: sample: Add ipv6 to ipv4 and sint to ipv6 casts
The RFC4291 says that when the IPv6 adress have the followin form:
0000::ffff:a.b.c.d, if can be converted to an IPv4 adress. This patch
enable this conversion in casts.
As the sint can be casted as ipv4, and ipv4 can be casted as ipv6, we
can directly cast sint as ipv6 using the RFC4291.
MINOR: Move http method enum from proto_http to sample
This is useful to prevent cross includes. The header file sample.h
needs to include proto_http, stick_tables.h will need to include
sample.h and proto_http includes stick_tables.h.
I choose to move the known http method define because this enum is
mainly used in sample.h. This enum is used for the sample type method.
MINOR: vars: reduce the code size of some wrappers
Some function are just a wrappers. This patch reduce the size of
this wrapper for improving the readability. One check is moved
from the wrapper to the main function, and some middle vars are
removed.
Willy Tarreau [Tue, 11 Aug 2015 09:36:45 +0000 (11:36 +0200)]
MEDIUM: config: emit a warning on a frontend without listener
Commit c6678e2 ("MEDIUM: config: authorize frontend and listen without bind")
completely removed the test for bind lines in frontends in order to make it
easier for automated tools to generate configs (eg: replacing a bind with
another one passing via a temporary config without any bind line). The
problem is that some common mistakes are totally hidden now. For example,
this apparently valid entry is silently ignored :
listen 1.2.3.4:8000
server s1 127.0.0.1:8000
Hint: 1.2.3.4:8000 is mistakenly the proxy name here.
Thus instead we now emit a warning to indicate that a frontend was found
with no listener. This should be backported to 1.5 to help spot abnormal
configurations.
Willy Tarreau [Mon, 10 Aug 2015 16:59:40 +0000 (18:59 +0200)]
MAJOR: http: remove references to appsession
appsessions started to be deprecated with the introduction of stick
tables, and the latter are much more powerful and flexible, and in
addition they are replicated between nodes and maintained across
reloads. Let's now remove appsession completely.
core.register_task(function()
while true do
core.Alert("LOLOLOLOLOL")
end
end)
I'd always get a timeout error starting the registered function.
The problem lies as far as I can tell in the fact that is possible for
now_ms to not change (is this maybe a problem on my config/system?)
until the expiration check happens, in the resume function that
actually kickstarts the lua task, making HAProxy think that expiration
time for the task is up, if I understand correctly tasks are meant to
never really timeout.
BUG/MEDIUM: DNS resolution response parsing broken
In some cases, parsing of the DNS response is broken and the response is
considered as invalid, despite being valid.
The current patch fixes this issue. It's a temporary solution until I
rework the response parsing to store the response buffer into a real DNS
packet structure.
Dragan Dosen [Fri, 7 Aug 2015 14:41:23 +0000 (16:41 +0200)]
MINOR: 51d: unable to start haproxy without "51degrees-data-file"
This patch adds a few checks on "global._51degrees.data_file_path" and allows
haproxy to start even when the pattern or trie data file is not specified.
If the "51d" converter is used, a new function "_51d_conv_check" will check
"global._51degrees.data_file_path" and displays a warning if necessary.
In src/haproxy.c, the global 51Degrees "cache_size" has moved outside of the
FIFTYONEDEGREES_H_PATTERN_INCLUDED ifdef block.
Willy Tarreau [Wed, 5 Aug 2015 15:16:33 +0000 (17:16 +0200)]
MEDIUM: backend: add the "http-reuse aggressive" strategy
This strategy is less extreme than "always", it only dispatches first
requests to validated reused connections, and moves a connection from
the idle list to the safe list once it has seen a second request, thus
proving that it could be reused.
Willy Tarreau [Wed, 5 Aug 2015 14:35:23 +0000 (16:35 +0200)]
MINOR: server: add a list of safe, already reused idle connections
These ones are considered safe as they have already been reused.
They will be useful in "aggressive" and "always" http-reuse modes
in order to place the first request of a connection with the least
risk.
Willy Tarreau [Wed, 5 Aug 2015 14:02:46 +0000 (16:02 +0200)]
MEDIUM: backend: implement "http-reuse safe"
The "safe" mode consists in picking existing connections only when
processing a request that's not the first one from a connection. This
ensures that in case where the server finally times out and closes, the
client can decide to replay idempotent requests.
Willy Tarreau [Thu, 6 Aug 2015 09:37:10 +0000 (11:37 +0200)]
MAJOR: backend: improve the connection reuse mechanism
Now instead of closing the existing connection attached to the
stream interface, we first check if the one we pick was attached to
another stream interface, in which case the connections are swapped
if possible (eg: if the current connection is not private). That way
the previous connection remains attached to an existing session and
significantly increases the chances of being reused.
Willy Tarreau [Tue, 4 Aug 2015 18:45:52 +0000 (20:45 +0200)]
MAJOR: backend: initial work towards connection reuse
In connect_server(), if we don't have a connection attached to the
stream-int, we first look into the server's idle_conns list and we
pick the first one there, we detach it from its owner if it had one.
If we used to have a connection, we close it.
This mechanism works well but doesn't scale : as servers increase,
the likeliness that the connection attached to the stream interface
doesn't match the server and gets closed increases.
Willy Tarreau [Wed, 5 Aug 2015 12:12:31 +0000 (14:12 +0200)]
MINOR: config: add new setting "http-reuse"
For now it only supports "never", meaning that we never want to reuse a
shared connection, and "always", meaning that we can use any connection
that was not marked private. When "never" is set, this also implies that
no idle connection may become a shared one.
Willy Tarreau [Tue, 4 Aug 2015 17:24:13 +0000 (19:24 +0200)]
MINOR: connection: add a new flag CO_FL_PRIVATE
This flag is set on an outgoing connection when this connection gets
some properties that must not be shared with other connections, such
as dynamic transparent source binding, SNI or a proxy protocol header,
or an authentication challenge from the server. This will be needed
later to implement connection reuse.
Willy Tarreau [Wed, 5 Aug 2015 09:08:30 +0000 (11:08 +0200)]
MINOR: stream-int: make si_idle_conn() only accept valid connections
This function is now dedicated to idle connections only, which means
that it must not be used without any endpoint nor anything not a
connection. The connection remains attached to the stream interface.
Willy Tarreau [Tue, 4 Aug 2015 15:25:58 +0000 (17:25 +0200)]
MINOR: connection: add a new list member in the connection struct
This list member will be used to attach a connection to a list of
idle, reusable or queued connections. It's unused for now. Given
that it's not expected to be used more than a few times per session,
the member was put after the target, in the area starting at the
second cache line of the structure.
Willy Tarreau [Tue, 4 Aug 2015 15:19:06 +0000 (17:19 +0200)]
MINOR: server: add a list of private idle connections
For now it's not populated but we have the list entry. It will carry
all idle connections that sessions don't want to share. They may be
used later to reclaim connections upon socket shortage for example.
Willy Tarreau [Tue, 4 Aug 2015 18:44:05 +0000 (20:44 +0200)]
MINOR: stream-int: add new function si_detach_endpoint()
This function only detaches the endpoint from the stream-int and
optionally returns the original pointer. This will be needed to
steal idle connections from other connections.
Willy Tarreau [Wed, 5 Aug 2015 19:47:23 +0000 (21:47 +0200)]
MEDIUM: stream-int: simplify si_alloc_conn()
Since we now always call this function with the reuse parameter cleared,
let's simplify the function's logic as it cannot return the existing
connection anymore. The savings on this inline function are appreciable
(240 bytes) :
$ size haproxy.old haproxy.new
text data bss dec hex filename 1020383 40816 36928 1098127 10c18f haproxy.old 1020143 40816 36928 1097887 10c09f haproxy.new
Willy Tarreau [Tue, 4 Aug 2015 17:45:36 +0000 (19:45 +0200)]
MEDIUM: backend: don't call si_alloc_conn() when we reuse a valid connection
connect_server() already does most of the check that is done again in
si_alloc_conn(), so let's simply reuse the existing connection instead
of calling the function again. It will also simplify the connection
reuse.
Indeed, for reuse to be set, it also requires srv_conn to be valid. In the
end, the only situation where we have to release the existing connection
and allocate a new one is when reuse == 0.
Willy Tarreau [Tue, 4 Aug 2015 17:34:21 +0000 (19:34 +0200)]
CLEANUP: backend: factor out objt_server() in connect_server()
objt_server() is called multiple times at various places while some
places already make use of srv for this. Let's move the call at the
top of the function and use it all over the place.
DOC: resolve-prefer default value and default-server update
By the code, resolve-prefer defaults to IPv6 and is also available in
the default-server directive.
These information were missing or wrong in the documentation.
MINOR: stream: initialize the current_rule field to NULL on stream init
Currently it is possible for the current_rule field to be evaluated before
being set, leading to valgrind complaining:
==16783== Conditional jump or move depends on uninitialised value(s)
==16783== at 0x44E662: http_res_get_intercept_rule (proto_http.c:3730)
==16783== by 0x44E662: http_process_res_common (proto_http.c:6528)
==16783== by 0x4797B7: process_stream (stream.c:1851)
==16783== by 0x414634: process_runnable_tasks (task.c:238)
==16783== by 0x40B02F: run_poll_loop (haproxy.c:1528)
==16783== by 0x407F25: main (haproxy.c:1887)
Vincent Bernat [Thu, 16 Jul 2015 18:52:51 +0000 (20:52 +0200)]
BUILD: link with libdl if needed for Lua support
On platforms where the dl*() functions are not part of the libc, a
program linking Lua also needs to link to libdl.
Moreover, on platforms using a gold linker with the --as-needed flag,
the libdl library needs to be linked after linking Lua, otherwise, it
won't be marked as needed and will be discarded and its symbols won't be
present at the end of the linking phase.
Ubuntu enables the --as-needed flag by default. Other distributions may
advertise its use, like Gentoo.
Released version 1.6-dev3 with the following main changes :
- CLEANUP: sample: generalize sample_fetch_string() as sample_fetch_as_type()
- MEDIUM: http: Add new 'set-src' option to http-request
- DOC usesrc root privileges requirments
- BUG/MINOR: dns: wrong time unit for some DNS default parameters
- MINOR: proxy: bit field for proxy_find_best_match diff status
- MINOR: server: new server flag: SRV_F_FORCED_ID
- MINOR: server: server_find functions: id, name, best_match
- DOC: dns: fix chapters syntax
- BUILD/MINOR: tools: rename popcount to my_popcountl
- BUILD: add netbsd TARGET
- MEDIUM: 51Degrees code refactoring and cleanup
- MEDIUM: 51d: add LRU-based cache on User-Agent string detection
- DOC: add notes about the "51degrees-cache-size" parameter
- BUG/MEDIUM: 51d: possible incorrect operations on smp->data.str.str
- BUG/MAJOR: connection: fix TLV offset calculation for proxy protocol v2 parsing
- MINOR: Add sample fetch to detect Supported Elliptic Curves Extension
- BUG/MINOR: payload: Add volatile flag to smp_fetch_req_ssl_ec_ext
- BUG/MINOR: lua: type error in the arguments wrapper
- CLEANUP: vars: remove unused struct
- BUG/MINOR: http/sample: gmtime/localtime can fail
- MINOR: standard: add 64 bits conversion functions
- MAJOR: sample: converts uint and sint in 64 bits signed integer
- MAJOR: arg: converts uint and sint in sint
- MEDIUM: sample: switch to saturated arithmetic
- MINOR: vars: returns variable content
- MEDIUM: vars/sample: operators can use variables as parameter
- BUG/MINOR: ssl: fix smp_fetch_ssl_fc_session_id
- BUILD/MINOR: lua: fix a harmless build warning
- BUILD/MINOR: stats: fix build warning due to condition always true
- BUG/MAJOR: lru: fix unconditional call to free due to unexpected semi-colon
- BUG/MEDIUM: logs: fix improper systematic use of quotes with a few tags
- BUILD/MINOR: lua: ensure that hlua_ctx_destroy is properly defined
- BUG/MEDIUM: lru: fix possible memory leak when ->free() is used
- MINOR: vars: make the accounting not depend on the stream
- MEDIUM: vars: move the session variables to the session, not the stream
- BUG/MEDIUM: vars: do not freeze the connection when the expression cannot be fetched
- BUG/MAJOR: buffers: make the buffer_slow_realign() function respect output data
- BUG/MAJOR: tcp: tcp rulesets were still broken
- MINOR: stats: improve compression stats reporting
- MINOR: ssl: make self-generated certs also work with raw IPv6 addresses
- CLEANUP: ssl: make ssl_sock_generated_cert_serial() take a const
- CLEANUP: ssl: make ssl_sock_generate_certificate() use ssl_sock_generated_cert_serial()
- BUG/MINOR: log: missing some ARGC_* entries in fmt_directives()
- MINOR: args: add new context for servers
- MINOR: stream: maintain consistence between channel_forward and HTTP forward
- MINOR: ssl: provide ia function to set the SNI extension on a connection
- MEDIUM: ssl: add sni support on the server lines
- CLEANUP: stream: remove a useless call to si_detach()
- CLEANUP: stream-int: fix a few outdated comments about stream_int_register_handler()
- CLEANUP: stream-int: remove stream_int_unregister_handler() and si_detach()
- MINOR: stream-int: only use si_release_endpoint() to release a connection
- MINOR: standard: provide htonll() and ntohll()
- CLEANUP/MINOR: dns: dns_str_to_dn_label() only needs a const char
- BUG/MAJOR: dns: fix the length of the string to be copied
BUG/MAJOR: dns: fix the length of the string to be copied
Jan A. Bruder reported that some very specific hostnames on server
lines were causing haproxy to crash on startup. Given that hist
backtrace showed some heap corruption, it was obvious there was an
overflow somewhere. The bug in fact is a typo in dns_str_to_dn_label()
which mistakenly copies one extra byte from the host name into the
output value, thus effectively corrupting the structure.
The bug triggers while parsing the next server of similar length
after the corruption, which generally triggers at config time but
could theorically crash at any moment during runtime depending on
what malloc sizes are needed next. This is why it's tagged major.
No backport is needed, this bug was introduced in 1.6-dev2.
MEDIUM: vars/sample: operators can use variables as parameter
This patch allow the existing operators to take a variable as parameter.
This is useful to add the content of two variables. This patch modify
the behavior of operators.
This patch check calculus for overflow and returns capped values.
This permits to protect against integer overflow in certain operations
involving ratios, percentages, limits or anything. That can sometimes
be critically important with some operations (eg: content-length < X).
MAJOR: sample: converts uint and sint in 64 bits signed integer
This patch removes the 32 bits unsigned integer and the 32 bit signed
integer. It replaces these types by a unique type 64 bit signed.
This makes easy the usage of integer and clarify signed and unsigned use.
With the previous version, signed and unsigned are used ones in place of
others, and sometimes the converter loose the sign. For example, divisions
are processed with "unsigned", if one entry is negative, the result is
wrong.
Note that the integer pattern matching and dotted version pattern matching
are already working with signed 64 bits integer values.
There is one user-visible change : the "uint()" and "sint()" sample fetch
functions which used to return a constant integer have been replaced with
a new more natural, unified "int()" function. These functions were only
introduced in the latest 1.6-dev2 so there's no impact on regular
deployments.
These are the 64-bit equivalent of htonl() and ntohl(). They're a bit
tricky in order to avoid expensive operations.
The principle consists in letting the compiler detect we're playing
with a union and simplify most or all operations. The asm-optimized
htonl() version involving bswap (x86) / rev (arm) / other is a single
operation on little endian, or a NOP on big-endian. In both cases,
this lets the compiler "see" that we're rebuilding a 64-bit word from
two 32-bit quantities that fit into a 32-bit register. In big endian,
the whole code is optimized out. In little endian, with a decent compiler,
a few bswap and 2 shifts are left, which is the minimum acceptable.
This patch adds 3 functions for 64 bit integer conversion.
* lltoa_r : converts signed 64 bit integer to string
* read_uint64 : converts from string to signed 64 bits integer with capping
* read_int64 : converts from string to unsigned 64 bits integer with capping
This patch introduces three new functions which can be used to find a
server in a farm using different server information:
- server unique id (srv->puid)
- server name
- find best match using either name or unique id
When performing best matching, the following applies:
- use the server name first (if provided)
- use the server id if provided
in any case, the function can update the caller about mismatches
encountered.
This flag aims at reporting whether the server unique id (srv->puid) has
been forced by the administrator in HAProxy's configuration.
If not set, it means HAProxy has generated automatically the server's
unique id.
MINOR: proxy: bit field for proxy_find_best_match diff status
function proxy_find_best_match can update the caller by updating an int
provided in argument.
For now, proxy_find_best_match hardcode bit values 0x01, 0x02 and 0x04,
which is not understandable when reading a code exploiting them.
This patch defines 3 macros with a more explicit wording, so further
reading of a code exploiting the magic bit values will be understandable
more easily.
The man said that gmtime() and localtime() can return a NULL value.
This is not tested. It appears that all the values of a 32 bit integer
are valid, but it is better to check the return of these functions.
However, if the integer move from 32 bits to 64 bits, some 64 values
can be unsupported.
MINOR: stream-int: only use si_release_endpoint() to release a connection
Change si_alloc_conn() to call si_release_endpoint() instead of
open-coding the connection releasing code when reuse is disabled.
This fuses the code with the one already dealing with applets, makes
it shorter and helps centralizing the connection freeing logic at a
single place.
BUG/MINOR: dns: wrong time unit for some DNS default parameters
Madison May reported that the timeout applied by the default
configuration is inproperly set up.
This patch fix this:
- hold valid default to 10s
- timeout retry default to 1s
Thierry FOURNIER [Tue, 23 Jun 2015 20:11:04 +0000 (22:11 +0200)]
CLEANUP: vars: remove unused struct
The commit "MEDIUM: vars: move the session variables to the session, not the stream" (ebcd4844e82a4198ea5d98fe491a46267da1d1ec")
moves the variables from the stream to the session. It forgot to remove
the stream definition of the "vars_sess".
The new "sni" server directive takes a sample fetch expression and
uses its return value as a hostname sent as the TLS SNI extension.
A typical use case consists in forwarding the front connection's SNI
value to the server in a bridged HTTPS forwarder :
MINOR: ssl: provide ia function to set the SNI extension on a connection
ssl_sock_set_servername() is used to set the SNI hostname on an
outgoing connection. This function comes from code originally
provided by Christopher Faulet of Qualys.
MINOR: stream: maintain consistence between channel_forward and HTTP forward
When the HTTP forwarder is used, it resets msg->sov so that we know that
the parsing pointer has advanced by exactly (msg->eoh + msg->eol - msg->sov)
bytes which may have to be rewound in case we want to perform an HTTP fetch
after forwarding has started (eg: upon connect).
But when the backend is in TCP mode, there may be no HTTP forwarding
analyser installed, still we may want to perform these HTTP fetches in
case we have already ensured at the TCP layer that we have a properly
parsed HTTP transaction.
In order to solve this, we reset msg->sov before doing a channel_forward()
so that we can still compute http_rewind() on the pending data. That ensures
the buffer is always rewindable even in mixed TCP+HTTP mode.
BUG/MINOR: log: missing some ARGC_* entries in fmt_directives()
ARGC_CAP was not added to fmt_directives() which is used to format
error messages when failing to parse log format expressions. The
whole switch/case has been reorganized to match the declaration
order making it easier to spot missing values. The default is not
the "log" directive anymore but "undefined" asking to report the
bug.
Backport to 1.5 is not strictly needed but is desirable at least
for code sanity.
MINOR: Add sample fetch to detect Supported Elliptic Curves Extension
Clients that support ECC cipher suites SHOULD send the specified extension
within the SSL ClientHello message according to RFC4492, section 5.1. We
can use this extension to chain-proxy requests so that, on the same IP
address, a ECC compatible clients gets an EC certificate and a non-ECC
compatible client gets a regular RSA certificate. The main advantage of this
approach compared to the one presented by Dave Zhu on the mailing list
is that we can make it work with OpenSSL versions before 1.0.2.
MINOR: ssl: make self-generated certs also work with raw IPv6 addresses
The current method of retrieving the incoming connection's destination
address to hash it is not compatible with IPv6 nor the proxy protocol
because it directly tries to get an IPv4 address from the socket. Instead
we must ask the connection. This is only used when no SNI is provided.
BUG/MEDIUM: 51d: possible incorrect operations on smp->data.str.str
In src/51d.c, the function _51d_conv(), a final '\0' is added into
smp->data.str.str, which can cause a problem if the SMP_F_CONST flag is
set in smp->flags or if smp->data.str.size is not available.
This patch adds a check on smp->flags and smp->data.str.size, and copies
the smp->data.str.str to another buffer by using smp_dup(). If necessary,
the "const" flag is set after device detection. Also, this patch removes
the unnecessary call to chunk_reset() on temp argument.
MEDIUM: http: Add new 'set-src' option to http-request
This option enables overriding source IP address in a HTTP request. It is
useful when we want to set custom source IP (e.g. front proxy rewrites address,
but provides the correct one in headers) or we wan't to mask source IP address
for privacy or compliance.
It acts on any expression which produces correct IP address.
CLEANUP: sample: generalize sample_fetch_string() as sample_fetch_as_type()
This modification makes possible to use sample_fetch_string() in more places,
where we might need to fetch sample values which are not plain strings. This
way we don't need to fetch string, and convert it into another type afterwards.
When using aliased types, the caller should explicitly check which exact type
was returned (e.g. SMP_T_IPV4 or SMP_T_IPV6 for SMP_T_ADDR).
All usages of sample_fetch_string() are converted to use new function.
Compression stats were not easy to read and could be confusing because
the saving ratio could be taken for global savings while it was only
relative to compressible input. Let's make that a bit clearer using
the new tooltips with a bit more details and also report the effective
ratio over all output bytes.