This fixes clients being able to connect with a certicate that was not
signed by the configured CA when SSL_VERIFY_FAIL_IF_NO_PEER_CERT is set.
The issue only appeared when a client connects via IP address directly and
not use a hostname.
When the hostname was used to connect, the SNI 'callback lws_mbedtls_sni_cb'
overwrote the invalid verfiy mode of
MBEDTLS_SSL_VERIFY_OPTIONAL with MBEDTLS_SSL_VERIFY_REQUIRED by
calling SSL_set_SSL_CTX.
Signed-off-by: Daniel Danzberger <daniel@dd-wrt.com>
mbedtls seemed to realize that they went overboard with the privacy stuff
on v3.0 and removed some of it. Introduce support for those members that
are only private on exactly v3.0 and unprotected before and after.
Ensure we still work with mbedtls_ssl_conf_verify() as well as
mbedtls_ssl_set_verify() if that's what we have got.
Make sure mbedtls tls validation is noisy and fast.
Disable Xenial + mbedtls in sai, it fails but not when the same
tests are run from the commandline. Very few people will be
using Xenial (2016 Ubuntu release) with mbedtls.
mbedtls validation was broken by an earlier patch on main... fix it and add
a CI test also using the wrong CA cert so this can be caught straight away
from now on.
Adapt mbedtls support for compatibility with v3, while maintaining
compatibility with v2.
Notice v3 has removed the ability to encrypt with pubkey and
decrypt with privkey. Openssl still has it, atm with v3 these
fall back to encrypt with privkey and decrypt with pubkey.
> The RSA module no longer supports private-key operations with the
> public key or vice versa. As a consequence, RSA operation functions
> no longer have a mode parameter. If you were calling RSA operations
> with the normal mode (public key for verification or encryption,
> private key for signature or decryption), remove the
> MBEDTLS_MODE_PUBLIC or MBEDTLS_MODE_PRIVATE argument. If you were
> calling RSA operations with the wrong mode, which rarely makes sense
>from a security perspective, this is no longer supported.
This is a huge patch that should be a global NOP.
For unix type platforms it enables -Wconversion to issue warnings (-> error)
for all automatic casts that seem less than ideal but are normally concealed
by the toolchain.
This is things like passing an int to a size_t argument. Once enabled, I
went through all args on my default build (which build most things) and
tried to make the removed default cast explicit.
With that approach it neither change nor bloat the code, since it compiles
to whatever it was doing before, just with the casts made explicit... in a
few cases I changed some length args from int to size_t but largely left
the causes alone.
From now on, new code that is relying on less than ideal casting
will complain and nudge me to improve it by warnings.
Nobody uses this from the wrapper... it brings in mbedtls_x509_crt_info()
even if you're using -ffunctionsections for a loss of about 1.7KB on Arm.
Let's chop it out...
The mbedtls openssl wrapper wants to use exports from mbedtls' net_sockets.c,
but this is only supposed to work on *nix and windows. Typically people
are using mbedtls on RTOS type platforms and to use it, net_sockets.c
needs some hacking.
Try to avoid that situation by porting the two exports we need into the
lws plat code and call from the wrapper.
This just supports "http/1.1" alpn and lets the client know it
can use keepalive earlier if it affirmitively negotiated
"http/1.1" on alpn.
mbedTLS wrapper needed a small adaptation to also allow
per-client-ssl control of the alpn negotiation list.
AG: unlike openssl, mbedtls does not load the system trust store.
So this change will make client tls operations that work OK on openssl fail on
mbedtls unless you provide the correct CA cert.
This allows lws to distinguish between untrusted CAs, hostname
mismatches, expired certificates.
NOTE: LCCSCF_ALLOW_SELFSIGNED actually allows for untrusted CAs, and
will also skip hostname verification. This is somewhat a limitiation of
the current lws verification process.
AG: improve error reporting up to the CLIENT_CONNECTION_ERROR argument
and add a note specific to mbedtls in the test client. Adapt the test
client to note the CA requirement if built with mbedTLS. Adapt the
minimal test clients to have the CAs available and use them if mbedTLS.
Under the condition the associated client went away (turn off WLAN at the
client), we can spin forever mistaking NET_SEND_FAILED for WANT_WRITE,
which was also true. This makes sure we understand that was fatal
immediately.
This adds a single api on lws that allows querying elements from the
peer certificate on a connection.
The api works the same regardless of the TLS backend.
1) The original wrapper logic for accounting for partial
sends on mbedtls is broken... adapt it
2) mbedtls has an additional restriction you MUST come
back to retry with EXACTLY the unsent part that you
originally asked for, or it loses coherency in the TLS
tunnel.
3) the wrapper is inconsistent between read WANT state
dynamically reported from mbedtls apis and the
SSL_want() api results. Check both on read.
- introduce lib/tls/mbedtls lib/tls/openssl
- move wrapper into lib/tls/mbedtls/wrapper
- introduce private helpers to hide backend
This patch doesn't replace or remove the wrapper, it moves it
to lib/tls/mbedtls/wrapper.
But it should be now that the ONLY functions directly consuming
wrapper apis are isolated in
- lib/tls/mbedtls/client.c (180 lines)
- lib/tls/mbedtls/server.c (317 lines)
- lib/tls/mbedtls/ssl.c (325 lines)
In particular there are no uses of openssl or mbedtls-related
constants outside of ./lib/tls any more.
