Re: [pdm-devel] [PATCH proxmox 2/4] access-control: add acl api feature

["Shannon Sterz" <s.sterz@proxmox.com>]

On Fri Apr 11, 2025 at 12:53 PM CEST, Dominik Csapak wrote:
> On 4/11/25 12:29, Shannon Sterz wrote:
>> On Wed Apr 9, 2025 at 2:58 PM CEST, Shannon Sterz wrote:
>>> On Wed Apr 9, 2025 at 1:39 PM CEST, Dominik Csapak wrote:
>>>> On 4/9/25 13:01, Dietmar Maurer wrote:
>>>> maybe something like this for the update case (untested, please verify before using this!):
>>>> (the diff is for pbs, where the code was copied from)
>>>>
>>>> this also includes a reformatted check for the token,non-token, same user checks
>>>> that are IMHO more readable than what we currently have
>>>> with the match, i think it's much more obvious that all cases are handled
>>>>
>>>> ---
>>>>        let user_info = CachedUserInfo::new()?;
>>>>
>>>> -    let top_level_privs = user_info.lookup_privs(&current_auth_id, &["access", "acl"]);
>>>> -    if top_level_privs & PRIV_PERMISSIONS_MODIFY == 0 {
>>>> +    let has_modify_permission = user_info
>>>> +        .check_privs(
>>>> +            &current_auth_id,
>>>> +            &["access", "acl"],
>>>> +            PRIV_PERMISSIONS_MODIFY,
>>>> +            false,
>>>> +        )
>>>> +        .is_ok();
>>>> +
>>>> +    if !has_modify_permission {
>>>>            if group.is_some() {
>>>>                bail!("Unprivileged users are not allowed to create group ACL item.");
>>>>            }
>>>>
>>>>            match &auth_id {
>>>>                Some(auth_id) => {
>>>> -                if current_auth_id.is_token() {
>>>> -                    bail!("Unprivileged API tokens can't set ACL items.");
>>>> -                } else if !auth_id.is_token() {
>>>> -                    bail!("Unprivileged users can only set ACL items for API tokens.");
>>>> -                } else if auth_id.user() != current_auth_id.user() {
>>>> -                    bail!("Unprivileged users can only set ACL items for their own API tokens.");
>>>> +                let same_user = auth_id.user() == current_auth_id.user();
>>>> +                match (current_auth_id.is_token(), auth_id.is_token(), same_user) {
>>>> +                    (true, _, _) => bail!("Unprivileged API tokens can't set ACL items."),
>>>> +                    (false, false, _) => {
>>>> +                        bail!("Unprivileged users can only set ACL items for API tokens.")
>>>> +                    }
>>>> +                    (false, true, true) => {
>>>> +                        // users are always allowed to modify ACLs for their own tokens
>>>> +                    }
>>>> +                    (false, true, false) => {
>>>> +                        bail!("Unprivileged users can only set ACL items for their own API tokens.")
>>>> +                    }
>>>>                    }
>>>>                }
>>>>                None => {
>>>> ---
>>
>> had another think about this, i'd tend towards something like the below.
>> the match statement is a nice idea, but it couples together things that
>> aren't really related. for example, why pull in the
>> `current_auth_id.is_token()` check, but not the `group.is_some()` check?
>> having match statements with tuples like this is making the code more
>> complex. imo, this is simpler:
>>
>> ```rs
>>      let unprivileged_user = CachedUserInfo::new()?
>>          .check_privs(
>>              &current_auth_id,
>>              &["access", "acl"],
>>              access_conf.acl_modify_privileges(),
>>              access_conf.allow_partial_permission_match(),
>>          )
>>          .is_err();
>>
>>      if unprivileged_user {
>>          if group.is_some() {
>>              bail!("Unprivileged users are not allowed to create group ACL item.");
>>          }
>>
>>          let auth_id = auth_id.as_ref().ok_or_else(|| {
>>              format_err!("Unprivileged user needs to provide auth_id to update ACL item.")
>>          })?;
>>
>>          if current_auth_id.is_token() {
>>              bail!("Unprivileged API tokens can't set ACL items.");
>>          }
>>
>>          if !auth_id.is_token() {
>>              bail!("Unprivileged users can only set ACL items for API tokens.");
>>          }
>>
>>          if current_auth_id != *auth_id {
>>              bail!("Unprivileged users can only set ACL items for their own API tokens.");
>>          }
>>      }
>> ```
>>
>> what do you think?
>
> I see what you mean, and yes i think it's more readable, but what I really wanted to convey with my
> approach was to clarify which condition is ok
>
> we currently try to filter out all invalid states, and it is not really obvious what
> condition makes the code continue at first glance
>
> Maybe we have to approach it completely different, for example check only the valid cases first
> and let that pass through, and then fail with the specific errors and have a fallback error
> for all other cases. that way we can't come into a situation where we forget/overlook some edge
> case.

oh, i mean we could just add a comment to the first if statement there
something like:

```rs
// check that if a user with insufficient permissions is changing acl
// entries, that they only modify their own api tokens' entries.
// unprivileged api tokens are not allowed to modify anything.
if unprivileged_user {
...
```

alternatively, we could do this which is closer to your suggestion in
the last comment

```rs
    if unprivileged_user {
        if group.is_none()
            && !current_auth_id.is_token()
            // check that an entry for an auth_id is being edited and
            // that it is a token for the user that is making the edit
            && auth_id
                .as_ref()
                .map(|id| id.is_token() && current_auth_id.user() == id.user())
                .unwrap_or_default()
        {
            // a user is directly editing the privileges of their own token, this is always
            // allowed
        } else {
            if group.is_some() {
                bail!("Unprivileged users are not allowed to create group ACL item.");
            }

            let auth_id = auth_id.as_ref().ok_or_else(|| {
                format_err!("Unprivileged user needs to provide auth_id to update ACL item.")
            })?;

            if current_auth_id.is_token() {
                bail!("Unprivileged API tokens can't set ACL items.");
            }

            if !auth_id.is_token() {
                bail!("Unprivileged users can only set ACL items for API tokens.");
            }

            if current_auth_id.user() != auth_id.user() {
                bail!("Unprivileged users can only set ACL items for their own API tokens.");
            }

            // this should not be reachable, but just in case, bail here
            bail!("Unprivileged user is trying to set an invalid ACL item.")
        }
    }
```

i think that respects your initial intend and also has a fail-safe just
in case something got overlooked or is changed later on.


_______________________________________________
pdm-devel mailing list
pdm-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pdm-devel