Commit graph

702 commits

Author SHA1 Message Date
Daniel Stenberg
a8bef39036
openssl: remove code handling default version
Since it is no longer actually kept as default internally, that's just
dead code.

Follow-up to 9d8998c994
Closes #19354
2025-11-05 14:14:58 +01:00
Joshua Rogers
0d5e24281d
vtls: check final cfilter node in find_ssl_filter
find_ssl_filter used while(cf && cf->next) and skipped the last node.
If the SSL filter was last, channel binding lookup failed and we returned
CURLE_BAD_FUNCTION_ARGUMENT. Switch to while(cf) so the tail is examined.

This bug was found with ZeroPath.

Closes #19229
2025-11-03 18:21:57 +01:00
Stefan Eissing
cccc65f051
openssl: check CURL_SSLVERSION_MAX_DEFAULT properly
The definition of these constants does not give a numeric ordering
and MAX_DEFAULT needs to be checked in addition of ciphers and QUIC
checks to apply correctly.

Fixes #19340
Reported-by: Peter Piekarski
Closes #19341
2025-11-03 16:31:22 +01:00
Daniel Stenberg
d4d7139e70
openssl: combine all the x509-store flags
... intead of overwriting the previous ones in ossl_populate_x509_store()

Pointed out by ZeroPath

Closes #19306
2025-10-31 23:24:38 +01:00
Stefan Eissing
b4630ed8fa
sectrust: fix verifystatus via sectrust
When openssl does not verify the certificate, but apple sectrust
does, we also pass it the ocsp stapled response when configured and
available.

When openssl does not verify the cert chain, it will also not be able
to verify the ocsp stapling. Do not call it if sectrust is the
verifier of the cert chain.

Fixes #19307
Reported-by: Harry Sintonen
Closes #19308
2025-10-31 23:10:35 +01:00
Daniel Stenberg
d646d5a130
openssl: fix the ocsp len arg to Curl_vtls_apple_verify
If it has no data, pass in a zero.

Fixes #19303
Reported-by: Harry Sintonen
Closes #19305
2025-10-31 23:09:05 +01:00
Stefan Eissing
c82a70628d ssl-session-cache: check use on config and availability
Replace the check if a ssl session cache is configured with
a function checking if it is configured *and* if an ssl session
cache is available.

During normal operations, a session cache is always there, however
for "connect-only" transfers this might not be the case. When such
transfers receive new sessions/tickets, they need to silently
discard those and not fail.

Reported-by: Marc Aldorasi

Fixes https://github.com/curl/curl/issues/18983
Closes https://github.com/curl/curl/pull/19251
2025-10-29 03:07:59 -04:00
Theo Buehler
2edce44065
vtls: remove call to PKCS12_PBE_add()
Curl is one of the last callers of PKCS12_PBE_add(). It has been a noop
since OpenSSL 0.9.8k (2006) stubbed it out when moving the built-in PBE
algorithms to a static table:
b8f702a0af

Closes #19201
2025-10-23 14:13:26 +02:00
Daniel Stenberg
9e15b07638
openssl: only try engine/provider if a certificate file/name is provided
Bug: https://issues.oss-fuzz.com/issues/435278402

Closes #19197
2025-10-23 07:54:34 +02:00
Daniel Stenberg
00cb679c04
openssl: remove dead code
A condition in infof_certstack() would always equal true after a
previous change.

Follow-up to e2a4de8a60

Pointed out by Coverity
Closes #19142
2025-10-19 23:48:28 +02:00
Yedaya Katsman
55e0526566
openssl: fix unable do typo in failf() calls
Closes #19149
2025-10-19 23:42:54 +02:00
Daniel Stenberg
8de898414c
openssl: free UI_METHOD on exit path
In providercheck(), when failing to open the "store", the exit path
would not previously free the created UI_METHOD and instead leak this
resource.

Pointed out by ZeroPath

Closes #19114
2025-10-18 12:54:19 +02:00
Daniel Stenberg
b9b8a7a5df
openssl: fix resource leak in provider error path
Pointed out by ZeroPath

Closes #19111
2025-10-18 00:40:13 +02:00
Daniel Stenberg
fbff1d5b90
openssl: avoid overwriting 'result' after error
Follow-up to eefd03c572

Pointed out by ZeroPath https://zeropath.com/
Closes #19099
2025-10-17 15:48:47 +02:00
Daniel Stenberg
3df71e6dc2
openssl: fail if more than MAX_ALLOWED_CERT_AMOUNT certs
Detect and prevent abuse or mistakes. Limit set to 100.

Closes #19091
2025-10-17 13:36:57 +02:00
Daniel Stenberg
e2a4de8a60
openssl: better return code checks when logging cert data
Pointed out by ZeroPath

Closes #19094
2025-10-17 12:41:04 +02:00
Stefan Eissing
d1d5855689
openssl: add comments regarding OCSP verification
To allow future reviewers of "security" reports to more easily find out
why code is this way.

Closes #18962
2025-10-09 14:44:28 +02:00
Daniel Stenberg
1ce6dff01a
openssl: fix peer certificate leak in channel binding
Reported-by: Stanislav Fort
Bug: https://hackerone.com/reports/3373640
Closes #18917
2025-10-09 09:10:37 +02:00
Joshua Rogers
4bfd7a9615
openssl: skip session resumption when verifystatus is set
Resumed TLS sessions skip OCSP stapled-response verification.
Force a full handshake so verifystatus() runs.

Closes #18902
2025-10-07 23:56:25 +02:00
Viktor Szakats
6f0e212f6e
tidy-up: miscellaneous (cont.)
- examples: replace magic numbers with `sizeof()`.
- typos: drop rules no longer needed after excluding tests/data.
- typos: move an exception inline.
- alpha-sort lists.
- fix indentation, whitespace.

Closes #18898
2025-10-06 22:33:38 +02:00
Viktor Szakats
b12da22db1
lib: stop overriding system printf symbols
After this patch, the codebase no longer overrides system printf
functions. Instead it explicitly calls either the curl printf functions
`curl_m*printf()` or the system ones using their original names.

Also:
- drop unused `curl_printf.h` includes.
- checksrc: ban system printf functions, allow where necessary.

Follow-up to db98daab05 #18844
Follow-up to 4deea9396b #18814

Closes #18866
2025-10-06 20:57:59 +02:00
Daniel Stenberg
e7a5184fa1
openssl: call SSL_get_error() with proper error
The error function should be called with the return code from the
previous call to SSL_shutdown() as argument.

Closes #18872
2025-10-06 11:31:54 +02:00
Viktor Szakats
34ad78da89
curlx: move Curl_strerror, use in src and tests, ban strerror globally
Also:
- tests/server: replace local `sstrerror()` with `curlx_strerror()`.
- tests/server: show the error code next to the string, where missing.
- curlx: use `curl_msnprintf()` when building for src and tests.
  (units was already using it.)
- lib: drop unused includes found along the way.
- curlx_strerror(): avoid compiler warning (and another similar one):
  ```
  In file included from servers.c:14:
  ../../lib/../../lib/curlx/strerr.c: In function ‘curlx_strerror’:
  ../../lib/../../lib/curlx/strerr.c:328:32: error: ‘snprintf’ output may be truncated before the last format character [-Werror=format-truncation=]
    328 |       SNPRINTF(buf, buflen, "%s", msg);
        |                                ^
  ../../lib/../../lib/curlx/strerr.c:47:18: note: ‘snprintf’ output 1 or more bytes (assuming 2) into a destination of size 1
     47 | #define SNPRINTF snprintf
        |                  ^
  ../../lib/../../lib/curlx/strerr.c:328:7: note: in expansion of macro ‘SNPRINTF’
    328 |       SNPRINTF(buf, buflen, "%s", msg);
        |       ^~~~~~~~
  ```

Follow-up to 45438c8d6f #18823

Closes #18840
2025-10-06 09:44:23 +02:00
Viktor Szakats
56026dae02
openssl: fix build for v1.0.2
```
lib/vtls/openssl.c: In function 'asn1_object_dump':
lib/vtls/openssl.c:299:42: error: passing argument 3 of 'i2t_ASN1_OBJECT' discards 'const' qualifier from pointer target type [-Werror=discarded-qualifiers]
  299 |   int i = i2t_ASN1_OBJECT(buf, (int)len, a);
      |                                          ^
In file included from /home/runner/djgpp/include/openssl/objects.h:965,
                 from /home/runner/djgpp/include/openssl/evp.h:94,
                 from /home/runner/djgpp/include/openssl/x509.h:73,
                 from /home/runner/djgpp/include/openssl/ssl.h:156,
                 from lib/curl_ntlm_core.c:71,
                 from bld/lib/CMakeFiles/libcurl_static.dir/Unity/unity_0_c.c:88:
/home/runner/djgpp/include/openssl/asn1.h:921:58: note: expected 'ASN1_OBJECT *' {aka 'struct asn1_object_st *'} but argument is of type 'const ASN1_OBJECT *' {aka 'const struct asn1_object_st *'}
  921 | int i2t_ASN1_OBJECT(char *buf, int buf_len, ASN1_OBJECT *a);
      |                                             ~~~~~~~~~~~~~^
```
Ref: https://github.com/curl/curl/actions/runs/18236773678/job/51931937131?pr=18039

Follow-up to bb46d42407 #18647

Closes #18841
2025-10-04 11:34:44 +02:00
Stefan Eissing
eefd03c572
ssl: support Apple SecTrust configurations
- configure/cmake support for enabling the option
- supported in OpenSSL and GnuTLS backends
- when configured, Apple SecTrust is the default trust store
  for peer verification. When one of the CURLOPT_* for adding
  certificates is used, that default does not apply.
- add documentation of build options and SSL use

Closes #18703
2025-10-03 12:02:23 +02:00
Daniel Stenberg
ea4ba6d9ef
lib: remove personal names from comments
- it's just too random who got mentioned
- we can't mention all, so better consistently mention none
- make sure they all are mentioned in THANKS
- also remove some unnecessary comment ramblings

Closes #18803
2025-10-02 14:22:57 +02:00
Viktor Szakats
95e50ad694
tidy-up: miscellaneous
- GHA/checkdocs: rename `spellcheck` job to `pyspelling` to say
  the exact tool used.
- GHA/checkdocs: restore a comment.
- GHA/linux: add `-B .` to a cmake configure to avoid warning, and
  future breakage.
- autotools: use correct casing for `Schannel`.
- doh: update RFC URL.
- drop redundant parenthesis.
- fix indentation, whitespace.

Closes #18756
2025-09-27 12:59:07 +02:00
Daniel Stenberg
16e0a2098d
openssl: fail the transfer if ossl_certchain() fails
Since it would indicate errors to the degree that continuing would just
risk hiding the earlier errors or make things weird.

Inspired by a report in Joshua's sarif data

Closes #18646
2025-09-25 22:25:20 +02:00
Stefan Eissing
442943fb8e
openssl: set io_need always
When OpenSSL reports SSL_ERROR_WANT_READ, set the io_need explicitly.
It should have already been set by the BIO, but be safe.

Reported in Joshua's sarif data

Closes #18733
2025-09-25 14:19:50 +02:00
Stefan Eissing
887b863b00
openssl: clear retry flag on x509 error
When loading the trust anchors and encountering an error, clear
a possibly set retry flag.

Reported in Joshua's sarif data

Closes #18724
2025-09-25 14:10:00 +02:00
Daniel Stenberg
bb46d42407
openssl: make the asn1_object_dump name null terminated
In case the buffer is too small.

Reported in Joshua's sarif data

Closes #18647
2025-09-20 23:12:15 +02:00
Viktor Szakats
1429858bce
tidy-up: update MS links, allow long URLs via checksrc
- update Microsoft documentation links.
  (also drop language designator where present.)

- checksrc: allow longer than 78 character lines if they
  contain a https URL. To make these links easier to use and parse.

- merge links that were split into two lines.

Closes #18626
2025-09-20 11:49:23 +02:00
Viktor Szakats
92f215fea1
build: address some -Weverything warnings, update picky warnings
`-Weverything` is not enabled by curl, and not recommended by LLVM,
because it may enable experimental options, and will result in new
fallouts after toolchain upgrades. This patch aims to fix/silence as much
as possible as found with llvm/clang 21.1.0. It also permanently enables
warnings that were fixed in source and deemed manageable in the future.
`-Wformat` warnings are addressed separately via #18343.

Fix/silence warnings in the source:
- typecheck-gcc.h: fix `-Wreserved-identifier`.
- lib: silence `-Wcast-function-type-strict`.
  For llvm 16+ or Apple clang 16+.
- asyn-ares: limit `HAPPY_EYEBALLS_DNS_TIMEOUT` to old c-ares versions.
- curl_trc: fix `-Wc++-hidden-decl`.
- doh: fix `-Wc++-keyword`.
- ftp: fix `-Wreserved-identifier`.
- ldap: fix `-Wreserved-identifier`.
- mqtt: comment unused macro to avoid warning.
- multi_ev: drop unused macros to avoid warnings.
- setopt: fix useless `break;` after `return;`.
- gtls, mbedtls, rustls: silence `-Wconditional-uninitialized`.
- socks_sspi, schannel, x509asn1: fix `-Wimplicit-int-enum-cast`.
- x509asn1: fix `-Wc++-keyword`.
- openssl: scope `OSSL_UI_METHOD_CAST` to avoid unused macro warning.
- libssh2, wolfssl: drop unused macros.
- curl_ngtcp2, curl_quiche, httpsrr, urlapi: drop/limit unused macros.
- tool_getparam: fix useless `break;` after `return;` or `break;`.
  Not normally enabled because it doesn't work with unity.
  https://github.com/llvm/llvm-project/issues/71046
- tool_operate: fix `-Wc++-keyword`.
- curlinfo: fix a `-Wunsafe-buffer-usage`.
- tests: silence `-Wformat-non-iso`.
- lib557: fix `-Wreserved-identifier`.
- lib1565: silence `-Wconditional-uninitialized`.

Enable the above clang warnings permanently in picky mode:
- `-Wc++-hidden-decl`
- `-Wc++-keyword` (except for Windows, where it collides with `wchar_t`)
- `-Wcast-function-type-strict`
- `-Wcast-function-type`
- `-Wconditional-uninitialized`
- `-Wformat-non-iso` (except for clang-cl)
- `-Wreserved-identifier`
- `-Wtentative-definition-compat`

Silence problematic `-Weverything` warnings globally (in picky mode):
- `-Wused-but-marked-unused` (88000+ hits) and
  `-Wdisabled-macro-expansion` (2600+ hits).
  Triggered by `typecheck-gcc.h` when building with clang 14+.
  Maybe there exists a way to fix within that header?
  Ref: https://discourse.llvm.org/t/removing-wused-but-marked-unused/55310
- `-Wunsafe-buffer-usage`. clang 16+. 7000+ hits.
  May be useful in theory, but such high volume of hits makes it
  impractical to review and possibly address. Meant for C++.
  Ref: https://clang.llvm.org/docs/SafeBuffers.html
  Ref: https://stackoverflow.com/questions/77017567/how-to-fix-code-to-avoid-warning-wunsafe-buffer-usage
  Ref: https://discourse.llvm.org/t/rfc-c-buffer-hardening/65734
  Ref: https://github.com/llvm/llvm-project/pull/111624
- `-Wimplicit-void-ptr-cast`. clang 21+. 1700+ hits.
  C++ warning, deemed pure noise.
  Ref: https://github.com/curl/curl/issues/18470#issuecomment-3253506266
- `-Wswitch-default` (180+ hits), `-Wswitch-enum` (190+ hits),
  `-Wcovered-switch-default` (20+ hits).
  Next to impossible to fix cleanly, esp. when the covered `case`
  branches depend on compile-time options.
- `-Wdocumentation-unknown-command` (8+ hits).
  Triggered in a few sources. Seems arbitrary and bogus.
- `-Wpadded` (550+ hits).
- `-Wc++-keyword` on Windows, where it collides with `wchar_t`.
  (100+ hits)
  Ref: https://github.com/llvm/llvm-project/issues/155988
- `-Wreserved-macro-identifier`. clang 13+. 5+ hits.
  Sometimes it's necessary to set external macros that use
  the reserved namespace. E.g. `_CRT_NONSTDC_NO_DEPRECATE`,
  `__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__`, `__NO_NET_API`,
  possibly `_REENTRANT`, and more.
  It's not worth trying to silence them individually.
- `-Wnonportable-system-include-path` with `clang-cl`.
  It'd be broken by doing what the warning suggests.
- `-Wformat-non-iso` for clang-cl.

CMake `PICKY_COMPILER=ON` (the default) or `./configure`
`--enable-warnings` (not the default) is required to enable these
silencing rules.

Also:
- autotools, cmake: fix Apple clang and mainline llvm version translations.
  Ref: https://en.wikipedia.org/wiki/Xcode#Toolchain_versions
- autotools, cmake: enable `-Warray-compare` for clang 20+.
  Follow-up to 4b7accda5a #17196
- cmake: fix to enable `-Wmissing-variable-declarations` at an earlier
  clang version.
- cmake: update internal logic to handle warning options with `+` in
  them.
- cmake: fix internal logic to match the whole option when looking
  into `CMAKE_C_FLAGS` for custom-disabled warnings.

Follow-up to b85cb8cb4e #18485

Closes #18477
2025-09-20 10:16:15 +02:00
Stefan Eissing
ff8dfd315c
aws-lc: re-enable large read-ahead with v1.61.0 again
AWS-LC fixed a bug with large read ahead buffers in v1.61.0. Check a
define introduced in that version to enable the large read ahead again.

AWS-LC issue: https://github.com/aws/aws-lc/issues/2650

Closes #18568
2025-09-16 16:55:11 +02:00
Stefan Eissing
b178d58266
quic: fix min TLS version handling
When switching to TSLv1.2 as default in
9d8998c994, this led to an explicit
setting of 1.2 on QUIC connections when using quictls, overriding the
already set min version of 1.3.

This leads to a ClientHello with TLS 1.2+1.3 offered on a QUIC connect
which is rejected by the Caddy server. Using ngtcp2 with OpenSSL 3.5+,
GnuTLS or AWS-LC is not affected.

Fixes #18518
Reported-by: fds242 on github
Closes #18520
2025-09-11 15:59:00 +02:00
Viktor Szakats
49145249be
tidy-up: drop stray "unused" comments
Closes #18453
2025-09-03 16:31:16 +02:00
Stefan Eissing
e65dc7fa23
aws-lc: do not use large buffer
test_10_08, uploading larger files for a h2 proxy, sporadically fails
with a decrpytion error on received data in AWS-LC. The frequency can
be increased by simulated network receive blocks.

Not setting a 4 * TLS record sized buffer, leaving AWS-LC at its
default buffer size seems to mitigate this problem.

Closes #18434
2025-09-01 23:27:11 +02:00
Viktor Szakats
f67090679f
openssl: sync an AWS-LC guard with BoringSSL
BoringSSL always used the same type:
https://boringssl.googlesource.com/boringssl/+/103ed08549a74af9f03363c633028faf9a475066
103ed08549

But, this codepath isn't built with BoringSSL, because it defines
`OPENSSL_NO_OCSP` via `opensslconf.h`.

Also drop an out-of-place `#endif` comment.

Ref: 20f4e94eeb #11568

Closes #18384
2025-08-23 22:57:30 +02:00
Viktor Szakats
1fc622db8e
openssl: drop single-use interim macro USE_OPENSSL_SRP
Closes #18383
2025-08-23 21:32:24 +02:00
Viktor Szakats
4f0e530c77
tidy-up: formatting
Closes #18373
2025-08-23 01:12:27 +02:00
Viktor Szakats
806d97d172
openssl: merge two #if blocks
Cherry-picked from #18330
Closes #18370
2025-08-22 22:52:51 +02:00
Viktor Szakats
b54039de2c
openssl: use RSA_flags() again with BoringSSL
Supported since 2017-12-18:
a0c87adbf0

Follow-up to cd276c3cca #2117

Closes #18369
2025-08-22 20:36:11 +02:00
Viktor Szakats
63bf7a0898
openssl: drop redundant HAVE_OPENSSL_VERSION macro
It served as an extra guard over `OPENSSL_VERSION_STRING`.

Also, document that `OPENSSL_VERSION_STRING` is supported by OpenSSL 3+.

Closes #18367
2025-08-22 18:55:26 +02:00
Viktor Szakats
0be7f382dc
openssl: add and use HAVE_BORINGSSL_LIKE internal macro
To cover the common case of guarding for both BoringSSL and AWS-LC.

Cherry-picked from #18330
Closes #18358
2025-08-22 17:49:21 +02:00
Viktor Szakats
8c29a29add
openssl: add and use HAVE_OPENSSL3 internal macro
Cherry-picked from #18330
Closes #18360
2025-08-22 17:21:29 +02:00
Viktor Szakats
db08e86cfb
openssl: drop more legacy cruft
- drop `ALLOW_RENEG` undocumented (insecure) build-time option.
- drop unnecessary check for `OPENSSL_VERSION_NUMBER`.
  It's present in all supported OpenSSL versions and forks.

Follow-up to 80c10c5d5d #18351
Follow-up to 59311bd3df #3293 #3283

Closes #18359
2025-08-22 17:21:28 +02:00
Viktor Szakats
80c10c5d5d
openssl: remove legacy cruft, document macro guards
- assume:
  - `BIO_CTRL_EOF`
  - `SSL_CTRL_SET_MSG_CALLBACK`
  - `SSL_CTRL_SET_MSG_CALLBACK`
  - `SSL_CTRL_SET_TLSEXT_HOSTNAME`
  - `SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER`
  - `SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS`
  - `SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG`
  - `SSL_OP_NO_COMPRESSION`
  - `SSL_OP_NO_TICKET`
  - `X509_V_FLAG_PARTIAL_CHAIN`
  - `X509_V_FLAG_TRUSTED_FIRST`
  They are present in all supported OpenSSL (and fork) versions.

- replace `SSL_ERROR_WANT_EARLY` with `SSL_ERROR_WANT_CLIENT_HELLO_CB`.
  The former appeared in OpenSSL 1.1.1-dev, but renamed before
  the stable release.

- document support for macros:
  - `ENGINE_CTRL_GET_CMD_FROM_NAME`
  - `SSL_ERROR_WANT_ASYNC_JOB`
  - `SSL_ERROR_WANT_ASYNC`
  - `SSL2_VERSION_MAJOR`
  - `TLS1_3_VERSION`

- drop legacy fallback for `CONF_MFLAGS_DEFAULT_SECTION`.
  It was there for OpenSSL 0.9.8 support.

- fix `SSL_CTRL_SET_MSG_CALLBACK` accidentally serving as a guard for
  OpenSSL (and forks) as a whole.

Tested OK with OpenSSL 1.0.2 and 1.1.0 in CI.

Closes #18351
2025-08-22 15:53:45 +02:00
Stefan Eissing
a5f0ab7995
openssl: auto-pause on verify callback retry
When an application install its own OpenSSL verify callback and that
callback invokes `SSL_set_retry_verify()`, the transfer is automatically
paused and does not progress the connect attempt any further until
unpaused via `curl_easy_pause().

Added test758 to verify.

Ref: #18284
Original PR by @Natris
Bug: https://curl.se/mail/lib-2025-08/0012.html
Closes #18288
2025-08-15 13:50:28 +02:00
Viktor Szakats
3eb00fa795
openssl: save and restore OpenSSL error queue in two functions
After merging #18228, I reviewed whether the clearing of the error queue
may interfere with preceding code. Turns out there may be a preceding
`SSL_Connect()` call.

This patch replaces the previous fix of clearing the error queue with
saving and restoring it in two functions which may be called between
the connect call and the `SSL_get_error()` call following it:
- `ossl_log_tls12_secret()`
- `Curl_ssl_setup_x509_store()`

The `ERR_set_mark()`, `ERR_pop_to_mark()` functions are present in all
supported OpenSSL and LibreSSL versions. Also in BoringSSL since its
initial commit.

OpenSSL may modify its error queue in all API calls that can fail.

Thanks-to: Viktor Dukhovni
Ref: https://github.com/curl/curl/issues/18190#issuecomment-3167702142
Ref: https://github.com/curl/curl/issues/18190#issuecomment-3169211739
Ref: https://github.com/curl/curl/issues/18190#issuecomment-3169988050

Follow-up to 8ec241bc99 #18228 #18190
Ref: e8b00fcd6a #10432 #10389
Fixes #18190
Closes #18234
2025-08-13 18:54:08 +02:00
Viktor Szakats
8ec241bc99
openssl: clear errors after a failed d2i_X509()
Without it, subsequent OpenSSL API calls may fail with an error caught
within the OpenSSL `d2i_X509()` (decode) call.

It was seen to happen when importing from the Windows certificate store
(e.g. with `--ca-native`), and any one of the certificates failed while
decoding, then skipped.

Behind the scene (and undocumented), the failed decode call is adding
an error to an internal OpenSSL error queue. This error is picked up
later, at the connect phase, by another OpenSSL API call, which happens
to check the error queue, without clearing it first. It made the connect
fail with the error collected earlier, while decoding the malformed and
discarded certificate.

Fix by explicitly clearing the error queue if the decode call fails.

Ref: https://docs.openssl.org/3.5/man3/d2i_X509/

`-vvvv` output before this patch:
```
[0-0] == Info: successfully imported Windows ROOT store
[0-0] == Info: successfully imported Windows CA store
[0-0] == Info: [SSL] SSL_connect() -> err=-1, detail=1
[0-0] == Info: TLS connect error: error:068000DD:asn1 encoding routines::illegal padding
[0-0] == Info: [SSL] cf_connect() -> 35, done=0
```

Mainline OpenSSL (as of 3.5.2) and quictls (as of 3.3.0) are affected.

LibreSSL is not affected. (I did not test BoringSSL and AWS-LC)

Assisted-by: Stefan Eissing
Reported-by: Michał Petryka
Fixes #18190

Closes #18228
2025-08-08 20:08:31 +02:00