To make sure that the HTTP/2 state is initialized correctly for
duplicated handles. It would otherwise easily generate "spurious"
PRIORITY frames to get sent over HTTP/2 connections when duplicated easy
handles were used.
Reported-by: Daniel Silverstone
Fixes#4303Closes#4442
This fix removes a use after free which can be triggered by
the internal cookie fuzzer, but otherwise is probably
impossible to trigger from an ordinary application.
The following program reproduces it:
curl_global_init(CURL_GLOBAL_DEFAULT);
CURL* handle=curl_easy_init();
CookieInfo* info=Curl_cookie_init(handle,NULL,NULL,false);
curl_easy_setopt(handle, CURLOPT_COOKIEJAR, "/dev/null");
Curl_flush_cookies(handle, true);
Curl_cookie_cleanup(info);
curl_easy_cleanup(handle);
curl_global_cleanup();
This was found through fuzzing.
Closes#4454
The 'share object' only sets the storage area for cookies. The "cookie
engine" still needs to be enabled or activated using the normal cookie
options.
This caused the curl command line tool to accidentally use cookies
without having been told to, since curl switched to using shared cookies
in 7.66.0.
Test 1166 verifies
Updated test 506
Fixes#4429Closes#4434
Prior to this change non-ssl/non-ssh connections that were reused set
TIMER_APPCONNECT [1]. Arguably that was incorrect since no SSL/SSH
handshake took place.
[1]: TIMER_APPCONNECT is publicly known as CURLINFO_APPCONNECT_TIME in
libcurl and %{time_appconnect} in the curl tool. It is documented as
"the time until the SSL/SSH handshake is completed".
Reported-by: Marcel Hernandez
Ref: https://github.com/curl/curl/issues/3760
Closes https://github.com/curl/curl/pull/3773
- convert some of them to H3BUF() calls to infof()
- remove some of them completely
- made DEBUG_HTTP3 defined only if CURLDEBUG is set for now
Closes#4421
The parser would check for a query part before fragment, which caused it
to do wrong when the fragment contains a question mark.
Extended test 1560 to verify.
Reported-by: Alex Konev
Fixes#4412Closes#4413
As libcurl now uses these 2 system functions, wrappers are needed on os400
to convert returned AF_UNIX sockaddrs to ascii.
This is a follow-up to commit 7fb54ef.
See also #4037.
Closes#4214
Otherwise curl may be told to use for instance pop3 to
communicate with the doh server, which most likely
is not what you want.
Found through fuzzing.
Closes#4406
It was already fixed for BoringSSL in commit a0f8fccb1e.
LibreSSL has had the second argument to SSL_CTX_set_min_proto_version
as uint16_t ever since the function was added in [0].
[0] 56f107201b
Closes https://github.com/curl/curl/pull/4397
Prior to this change when a server returned a socks5 connect error then
curl would parse the destination address:port from that data and show it
to the user as the destination:
curld -v --socks5 10.0.3.1:1080 http://google.com:99
* SOCKS5 communication to google.com:99
* SOCKS5 connect to IPv4 172.217.12.206 (locally resolved)
* Can't complete SOCKS5 connection to 253.127.0.0:26673. (1)
curl: (7) Can't complete SOCKS5 connection to 253.127.0.0:26673. (1)
That's incorrect because the address:port included in the connect error
is actually a bind address:port (typically unused) and not the
destination address:port. This fix changes curl to show the destination
information that curl sent to the server instead:
curld -v --socks5 10.0.3.1:1080 http://google.com:99
* SOCKS5 communication to google.com:99
* SOCKS5 connect to IPv4 172.217.7.14:99 (locally resolved)
* Can't complete SOCKS5 connection to 172.217.7.14:99. (1)
curl: (7) Can't complete SOCKS5 connection to 172.217.7.14:99. (1)
curld -v --socks5-hostname 10.0.3.1:1080 http://google.com:99
* SOCKS5 communication to google.com:99
* SOCKS5 connect to google.com:99 (remotely resolved)
* Can't complete SOCKS5 connection to google.com:99. (1)
curl: (7) Can't complete SOCKS5 connection to google.com:99. (1)
Ref: https://tools.ietf.org/html/rfc1928#section-6
Closes https://github.com/curl/curl/pull/4394
As the loop discards cookies without domain set. This bug would lead to
qsort() trying to sort uninitialized pointers. We have however not found
it a security problem.
Reported-by: Paul Dreik
Closes#4386
If the input hostname is "[", hlen will underflow to max of size_t when
it is subtracted with 2.
hostname[hlen] will then cause a warning by ubsanitizer:
runtime error: addition of unsigned offset to 0x<snip> overflowed to
0x<snip>
I think that in practice, the generated code will work, and the output
of hostname[hlen] will be the first character "[".
This can be demonstrated by the following program (tested in both clang
and gcc, with -O3)
int main() {
char* hostname=strdup("[");
size_t hlen = strlen(hostname);
hlen-=2;
hostname++;
printf("character is %d\n",+hostname[hlen]);
free(hostname-1);
}
I found this through fuzzing, and even if it seems harmless, the proper
thing is to return early with an error.
Closes#4389
CURLU_NO_AUTHORITY is intended for use with unknown schemes (i.e. not
"file:///") to override cURL's default demand that an authority exists.
Closes#4349
If the :authority pseudo header field doesn't contain an explicit port,
we assume it is valid for the default port, instead of rejecting the
request for all ports.
Ref: https://curl.haxx.se/mail/lib-2019-09/0041.htmlCloses#4365
If you set the same URL for target as for DoH (and it isn't a DoH
server), like "https://example.com" in both, the easy handles used for
the DoH requests could be left "dangling" and end up not getting freed.
Reported-by: Paul Dreik
Closes#4366
The undefined behaviour is annoying when running fuzzing with
sanitizers. The codegen is the same, but the meaning is now not up for
dispute. See https://cppinsights.io/s/516a2ff4
By incrementing the pointer first, both gcc and clang recognize this as
a bswap and optimizes it to a single instruction. See
https://godbolt.org/z/994ZpxCloses#4350
Added unit test case 1655 to verify.
Close#4352
the code correctly finds the flaws in the old code,
if one temporarily restores doh.c to the old version.
This is a protocol violation but apparently there are legacy proprietary
servers doing this.
Added test 336 and 337 to verify.
Reported-by: Philippe Marguinaud
Closes#4339
For FTPS transfers, curl gets close_notify on the data connection
without that being a signal to close the control connection!
Regression since 3f5da4e59a (7.65.0)
Reported-by: Zenju on github
Reviewed-by: Jay Satiro
Fixes#4329Closes#4340
Despite ldapp_err2string being documented by MS as returning a
PCHAR (char *), when UNICODE it is mapped to ldap_err2stringW and
returns PWCHAR (wchar_t *).
We have lots of code that expects ldap_err2string to return char *,
most of it failf used like this:
failf(data, "LDAP local: Some error: %s", ldap_err2string(rc));
Closes https://github.com/curl/curl/pull/4272
It needs to parse correctly. Otherwise it could be tricked into letting
through a-f using host names that libcurl would then resolve. Like
'[ab.be]'.
Reported-by: Thomas Vegas
Closes#4315
OpenSSL 1.1.0 adds SSL_CTX_set_<min|max>_proto_version() that we now use
when available. Existing code is preserved for older versions of
OpenSSL.
Closes#4304
Otherwise, a three byte response would make the smtp_state_ehlo_resp()
function misbehave.
Credit to OSS-Fuzz
Bug: https://crbug.com/oss-fuzz/16918
Assisted-by: Max Dymond
Closes#4287
This allows the function to figure out if a unix domain socket has a
file name or not associated with it! When a socket is created with
socketpair(), as done in the fuzzer testing, the path struct member is
uninitialized and must not be accessed.
Bug: https://crbug.com/oss-fuzz/16699Closes#4283
It could otherwise return an error even when closed correctly if GOAWAY
had been received previously.
Reported-by: Tom van der Woerdt
Fixes#4267Closes#4268
For a long time (since 7.28.1) we've returned error when setting the
value to 1 to make applications notice that we stopped supported the old
behavior for 1. Starting now, we treat 1 and 2 exactly the same.
Closes#4241
The quiche debug callback is global and can only be initialized once, so
make sure we don't do it multiple times (e.g. if multiple requests are
executed).
In addition this initializes the callback before the connection is
created, so we get logs for the handshake as well.
Closes#4236
When a username and password are provided in the URL, they were wrongly
removed from the stored URL so that subsequent uses of the same URL
wouldn't find the crendentials. This made doing HTTP auth with multiple
connections (like Digest) mishave.
Regression from 46e164069d (7.62.0)
Test case 335 added to verify.
Reported-by: Mike Crowe
Fixes#4228Closes#4229
SSL_VersionRangeGetDefault returns (TLSv1.0, TLSv1.2) as supported
range in NSS 3.45. It looks like the intention is to raise the minimum
version rather than lowering the maximum, so adjust accordingly. Note
that the caller (nss_setup_connect) initializes the version range to
(TLSv1.0, TLSv1.3), so there is no need to check for >= TLSv1.0 again.
Closes#4187
Reviewed-by: Daniel Stenberg
Reviewed-by: Kamil Dudka
This commit makes sending HTTP/3 request with nghttp3 work. It
minimally receives HTTP response and calls nghttp3 callbacks, but no
processing is made at the moment.
Closes#4215
This avoids EBADF errors from EPOLL_CTL_DEL operations in the
ephiperfifo.c example. EBADF is dangerous in multi-threaded
applications where I rely on epoll_ctl to operate on the same
epoll description from different threads.
Follow-up to eb9a604f8d
Bug: https://curl.haxx.se/mail/lib-2019-08/0026.htmlCloses#4211
So that users can mask in/out specific HTTP versions when Alt-Svc is
used.
- Removed "h2c" and updated test case accordingly
- Changed how the altsvc struct is laid out
- Added ifdefs to make the unittest run even in a quiche-tree
Closes#4201
RFC 7838 section 5:
When using an alternative service, clients SHOULD include an Alt-Used
header field in all requests.
Removed CURLALTSVC_ALTUSED again (feature is still EXPERIMENTAL thus
this is deemed ok).
You can disable sending this header just like you disable any other HTTP
header in libcurl.
Closes#4199
Even though it cannot fall-back to a lower HTTP version automatically. The
safer way to upgrade remains via CURLOPT_ALTSVC.
CURLOPT_H3 no longer has any bits that do anything and might be removed
before we remove the experimental label.
Updated the curl tool accordingly to use "--http3".
Closes#4197
This is only the libcurl part that provides the information. There's no
user of the parsed value. This change includes three new tests for the
parser.
Ref: #3794
Added the ability for the calling program to specify the authorisation
identity (authzid), the identity to act as, in addition to the
authentication identity (authcid) and password when using SASL PLAIN
authentication.
Fixes#3653Closes#3790
NOTE: This commit was cherry-picked and is part of a series of commits
that added the authzid feature for upcoming 7.66.0. The series was
temporarily reverted in db8ec1f so that it would not ship in a 7.65.x
patch release.
Closes https://github.com/curl/curl/pull/4186
Repeatedly we see problems where using curl_multi_wait() is difficult or
just awkward because if it has no file descriptor to wait for
internally, it returns immediately and leaves it to the caller to wait
for a small amount of time in order to avoid occasional busy-looping.
This is often missed or misunderstood, leading to underperforming
applications.
This change introduces curl_multi_poll() as a replacement drop-in
function that accepts the exact same set of arguments. This function
works identically to curl_multi_wait() - EXCEPT - for the case when
there's nothing to wait for internally, as then this function will by
itself wait for a "suitable" short time before it returns. This
effectiely avoids all risks of busy-looping and should also make it less
likely that apps "over-wait".
This also changes the curl tool to use this funtion internally when
doing parallel transfers and changes curl_easy_perform() to use it
internally.
Closes#4163
As the plan has been laid out in DEPRECATED. Update docs accordingly and
verify in test 1174. Now requires the option to be set to allow HTTP/0.9
responses.
Closes#4191
Allow pretty much anything to be part of the ALPN identifier. In
particular minus, which is used for "h3-20" (in-progress HTTP/3
versions) etc.
Updated test 356.
Closes#4182
If HTTPAUTH_GSSNEGOTIATE was used for a POST request and
gss_init_sec_context() failed, the POST request was sent
with empty body. This commit also restores the original
behavior of `curl --fail --negotiate`, which was changed
by commit 6c60355323.
Add regression tests 2077 and 2078 to cover this.
Fixes#3992Closes#4171
It was used (intended) to pass in the size of the 'socks' array that is
also passed to these functions, but was rarely actually checked/used and
the array is defined to a fixed size of MAX_SOCKSPEREASYHANDLE entries
that should be used instead.
Closes#4169
... to make CURLOPT_MAX_RECV_SPEED_LARGE and
CURLOPT_MAX_SEND_SPEED_LARGE work correctly on subsequent transfers that
reuse the same handle.
Fixed-by: Ironbars13 on github
Fixes#4084Closes#4161
If using the read callback for HTTP_POST, and POSTFIELDSIZE is not set,
automatically add a Transfer-Encoding: chunked header, same as it is
already done for HTTP_PUT, HTTP_POST_FORM and HTTP_POST_MIME. Update
test 1514 according to the new behaviour.
Closes#4138
- In curl_easy_reset attempt to resize the receive buffer to its default
size. If realloc fails then continue using the previous size.
Prior to this change curl_easy_reset did not properly handle resetting
the receive buffer (data->state.buffer). It reset the variable holding
its size (data->set.buffer_size) to the default size (READBUFFER_SIZE)
but then did not actually resize the buffer. If a user resized the
buffer by using CURLOPT_BUFFERSIZE to set the size smaller than the
default, later called curl_easy_reset and attempted to reuse the handle
then a heap overflow would very likely occur during that handle's next
transfer.
Reported-by: Felix Hädicke
Fixes https://github.com/curl/curl/issues/4143
Closes https://github.com/curl/curl/pull/4145
Specifying O_APPEND in conjunction with O_TRUNC and O_CREAT does not
make much sense. And this combination of flags is not accepted by all
SFTP servers (at least not Apache SSHD).
Fixes#4147Closes#4148
Curl_disconnect bails out if conn->easyq is not empty, detach_connection
needs to be called first to remove the current easy from the queue.
Fixes#4144Closes#4151
USe configure --with-ngtcp2 or --with-quiche
Using either option will enable a HTTP3 build.
Co-authored-by: Alessandro Ghedini <alessandro@ghedini.me>
Closes#3500
Several reasons:
- we can't add everyone who's helping out so its unfair to just a few
selected ones.
- we already list all helpers in THANKS and in RELEASE-NOTES for each
release
- we don't want to give the impression that some parts of the code is
"owned" or "controlled" by specific persons
Assisted-by: Daniel Gustafsson
Closes#4129
PK11_IsPresent() checks for the token for the given slot is available,
and sets needlogin flags for the PK11_Authenticate() call. Should it
return false, we should however treat it as an error and bail out.
Closes https://github.com/curl/curl/pull/4110
All protocols except for CURLPROTO_FILE/CURLPROTO_SMB and their TLS
counterpart were allowed for redirect. This vastly broadens the
exploitation surface in case of a vulnerability such as SSRF [1], where
libcurl-based clients are forced to make requests to arbitrary hosts.
For instance, CURLPROTO_GOPHER can be used to smuggle any TCP-based
protocol by URL-encoding a payload in the URI. Gopher will open a TCP
connection and send the payload.
Only HTTP/HTTPS and FTP are allowed. All other protocols have to be
explicitly enabled for redirects through CURLOPT_REDIR_PROTOCOLS.
[1]: https://www.acunetix.com/blog/articles/server-side-request-forgery-vulnerability/
Signed-off-by: Linos Giannopoulos <lgian@skroutz.gr>
Closes#4094
With CURLOPT_TIMECONDITION set, a header is automatically added (e.g.
If-Modified-Since). Allow this to be replaced or suppressed with
CURLOPT_HTTPHEADER.
Fixes#4103Closes#4109
- Return CURLE_REMOTE_ACCESS_DENIED for SMB access denied on file open.
Prior to this change CURLE_REMOTE_FILE_NOT_FOUND was returned instead.
Closes https://github.com/curl/curl/pull/4095
There were a leftover few prototypes of Curl_ functions that we used to
export but no longer do, this removes those prototypes and cleans up any
comments still referring to them.
Curl_write32_le(), Curl_strcpy_url(), Curl_strlen_url(), Curl_up_free()
Curl_concat_url(), Curl_detach_connnection(), Curl_http_setup_conn()
were made static in 05b100aee2.
Curl_http_perhapsrewind() made static in 574aecee20.
For the remainder, I didn't trawl the Git logs hard enough to capture
their exact time of deletion, but they were all gone: Curl_splayprint(),
Curl_http2_send_request(), Curl_global_host_cache_dtor(),
Curl_scan_cache_used(), Curl_hostcache_destroy(), Curl_second_connect(),
Curl_http_auth_stage() and Curl_close_connections().
Closes#4096
Reviewed-by: Daniel Stenberg <daniel@haxx.se>
The file suffix for dynamically loadable objects on macOS is .dylib,
which need to be added for the module definitions in order to get the
NSS TLS backend to work properly on macOS.
Closes https://github.com/curl/curl/pull/4046
The value of the maxPTDs parameter to PR_Init() has since at least
NSPR 2.1, which was released sometime in 1998, been marked ignored
as is accordingly not used in the initialization code. Setting it
to a value when calling PR_Init() is thus benign, but indicates an
intent which may be misleading. Reset the value to zero to improve
clarity.
Closes https://github.com/curl/curl/pull/4054
Change the logic around such that we only keep CRLs that NSS actually
ended up caching around for later deletion. If CERT_CacheCRL() fails
then there is little point in delaying the freeing of the CRL as it
is not used.
Closes https://github.com/curl/curl/pull/4053
Some editors and IDEs assume that source files use UTF-8 file encodings.
It also fixes the build with MSVC when /utf-8 command line option is
used (this option is mandatory for some other open-source projects, this
is useful when using the same options is desired for building all
libraries of a project).
Closes https://github.com/curl/curl/pull/4087
... since that needs UI_OpenSSL() which isn't provided when OpenSSL is
built with OPENSSL_NO_UI_CONSOLE which happens when OpenSSL is built for
UWP (with "VC-WIN32-UWP").
Reported-by: Vasily Lobaskin
Fixes#4073Closes#4077
The header buffer size calculation can from static analysis seem to
overlow as it performs an addition between two size_t variables and
stores the result in a size_t variable. Overflow is however guarded
against elsewhere since the input to the addition is regulated by
the maximum read buffer size. Clarify this with a comment since the
question was asked.
Reviewed-by: Daniel Stenberg <daniel@haxx.se>
To make sure a HTTP/2 stream registers the end of stream.
Bug #4043 made me find this problem but this fix doesn't correct the
reported issue.
Closes#4068
Certinfo gives the same result for all OpenSSL versions.
Also made printing RSA pubkeys consistent with older versions.
Reported-by: Michael Wallner
Fixes#3706Closes#4030
OpenSSL used to call exit(1) on syntax errors in OPENSSL_config(),
which is why we switched to CONF_modules_load_file() and introduced
a comment stating why. This behavior was however changed in OpenSSL
commit abdd677125f3a9e3082f8c5692203590fdb9b860, so remove the now
outdated and incorrect comment. The mentioned commit also declares
OPENSSL_config() deprecated so keep the current coding.
Closes#4033
Reviewed-by: Daniel Stenberg <daniel@haxx.se>
