From: "Shannon Sterz" <s.sterz@proxmox.com>
To: "Dominik Csapak" <d.csapak@proxmox.com>,
<pve-devel@lists.proxmox.com>, <pbs-devel@lists.proxmox.com>
Subject: Re: [PATCH proxmox v3 2/6] http: tls: use legacy behavior when PROXMOX_NEW_TLS_CHECK is not set
Date: Thu, 25 Jun 2026 13:19:40 +0200 [thread overview]
Message-ID: <DJI38NI2NUXK.3FOV231WMXSZF@proxmox.com> (raw)
In-Reply-To: <20260617085949.1528300-3-d.csapak@proxmox.com>
On Wed Jun 17, 2026 at 10:59 AM CEST, Dominik Csapak wrote:
> if that environment variable is not set to "1", give the openssl result
> priority, and potentially ignore a given fingerprint that is not
> matching. If that's the case, print a warning.
>
> Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
-->8 snip 8<--
> /// Intended as an openssl verification callback.
> ///
> -/// The following things are checked:
> +/// If the 'PROXMOX_NEW_TLS_CHECK' environment variable is set to "1",
> +/// the following things are checked:
> ///
> /// * If no fingerprint is given, return the openssl verification result
> -/// * If a fingerprint is given, do:
> -/// * Ignore all non-leaf certificates/
> +/// * If a fingerprint is given, ignore all non-leaf certificates
> +///
> +/// Otherwise, we trust the openssl result if the whole chain was trusted
> pub fn openssl_verify_callback(
> openssl_valid: bool,
> ctx: &mut X509StoreContextRef,
> expected_fp: Option<&str>,
> ) -> Result<(), SslVerifyError> {
> let trust_openssl = ctx.error() != X509VerifyResult::APPLICATION_VERIFICATION;
> - if expected_fp.is_none() && openssl_valid && trust_openssl {
> - return Ok(());
> +
> + let new_check = matches!(std::env::var("PROXMOX_NEW_TLS_CHECK").as_deref(), Ok("1"));
> +
> + if openssl_valid && trust_openssl {
> + if new_check && expected_fp.is_none() {
> + return Ok(());
> + }
> +
> + // legacy mode: skip all valid certs except the leaf, so we can warn if fingerprint does not match
> + if !new_check && ctx.error_depth() > 0 {
this check is probably wrong from what i can tell. assume a certificate
chain with the certificate of the root ca being in the systems trust
store and no fingerprint was provided:
1. the root ca's certificate is provided. since it is in the systems
trust store `openssl_valid` and `trust_openssl` are true. we are in
the old verification logic, so `new_check` is `false`.
`ctx.error_depth()` is `> 0` as we aren't at the leaf yet.
2. the facts outlined hold true for all certificates in the chain until
we reach the leaf certificate. `trust_openssl` is never set to false,
as that would happen further down the callback (not that setting an
error here would *not* make the behavior correct).
3. the leaf certificate will not enter this block, as
`ctx.error_depth() > 0` no longer holds true.
> + return Ok(());
> + }
> }
>
> let cert = match ctx.current_cert() {
> @@ -50,7 +62,7 @@ pub fn openssl_verify_callback(
> };
>
> if ctx.error_depth() > 0 {
3. (cont.) same here
> - // openssl was not valid, but we want to continue, so save that we don't trust openssl
> + // if openssl is not valid, and we want to continue, save that we don't trust openssl
> ctx.set_error(X509VerifyResult::APPLICATION_VERIFICATION);
> return Ok(());
> }
> @@ -65,6 +77,13 @@ pub fn openssl_verify_callback(
> ctx.set_error(X509VerifyResult::OK);
> Ok(())
> } else {
> + if !new_check && openssl_valid && trust_openssl {
4. since no fingerprint was provided, we won't get here
> + log::warn!(
> + "Certificate chain valid, but fingerprint does not match, ignoring fingerprint! To prioritize the fingerprint, set `PROXMOX_NEW_TLS_CHECK=1` in your environment."
> + );
> + return Ok(());
> + }
> +
> Err(SslVerifyError::FingerprintMismatch {
> fingerprint,
> expected: expected_fp.to_string(),
5. instead the `else` branch of the outer-most `if` statement here will
be taken, reporting the certificate as
`SslVerifyError::UntrustedCertificate`.
however, that is wrong. the certificate should be trusted as openssl
validated the certificate and no fingerprint was provided.
next prev parent reply other threads:[~2026-06-25 11:19 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-17 8:59 [PATCH proxmox{,-backup,-websocket-tunnel} v3 0/6] unify openssl callback logic Dominik Csapak
2026-06-17 8:59 ` [PATCH proxmox v3 1/6] http: factor out openssl verification callback Dominik Csapak
2026-06-25 11:19 ` Shannon Sterz
2026-06-17 8:59 ` [PATCH proxmox v3 2/6] http: tls: use legacy behavior when PROXMOX_NEW_TLS_CHECK is not set Dominik Csapak
2026-06-25 11:19 ` Shannon Sterz [this message]
2026-06-17 8:59 ` [PATCH proxmox v3 3/6] client: use proxmox-http's openssl verification callback Dominik Csapak
2026-06-25 11:19 ` Shannon Sterz
2026-06-17 8:59 ` [PATCH proxmox-backup v3 4/6] pbs-client: use proxmox-https openssl callback Dominik Csapak
2026-06-25 11:19 ` Shannon Sterz
2026-06-17 8:59 ` [PATCH proxmox-backup v3 5/6] pbs-client: honor already verified fingerprint Dominik Csapak
2026-06-17 8:59 ` [PATCH proxmox-websocket-tunnel v3 6/6] use proxmox-http's openssl callback Dominik Csapak
2026-06-25 11:19 ` Shannon Sterz
2026-06-25 11:19 ` [PATCH proxmox{,-backup,-websocket-tunnel} v3 0/6] unify openssl callback logic Shannon Sterz
2026-06-25 11:22 ` [PATCH proxmox 1/3] http: tls: move PROXMOX_NEW_TLS_CHECK env var name into constant Shannon Sterz
2026-06-25 11:22 ` [PATCH proxmox 2/3] http: tls: implement `PartialEq` for `SslVerifyError` Shannon Sterz
2026-06-25 11:22 ` [PATCH proxmox 3/3] http: tls: add integration tests for openssl verify callbacks Shannon Sterz
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=DJI38NI2NUXK.3FOV231WMXSZF@proxmox.com \
--to=s.sterz@proxmox.com \
--cc=d.csapak@proxmox.com \
--cc=pbs-devel@lists.proxmox.com \
--cc=pve-devel@lists.proxmox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox