Coverity reported this as a leak, but the test queue is actually
getting freed later and does not need to be freed locally in add_path
This basically reverts c0863e5bc6
Relax the pre-check for local user certificate. Before we used to check
that the CA provided (if any) was used to verify both the peer identity
and the local certificate chain. However, there seem to be networks
that use different CAs to sign AP/Radius certificates and certificates
issued to users.
Drop the ca_certs argument from l_certchain_verify, but keep the call
there to make sure the certificate chain is indeed a chain as a sanity
check.
If the config file passed in is not found we would continue and
eventually something else would fail. Instead immediately raise an
exception to be more clear on what is actually failing.
The commit/confirm processing was incorrectly subtracting 2 from
the length when they should be subtracting 6. As with the other
similar change, the length is validated with mpdu_validate so
subtracting 6 will not cause an overflow.
This function was returning a boolean and the expected return was
a signed integer. Since this function actually returned false in
all cases the check for a success (0) return always worked.
The comment about the 'standard code path' was removed as this is
no longer valid.
If an authentication frame of length <= 5 is sent sae will overflow an
integer. The original cause of this was due to incorrectly using the
sizeof(struct mmpdu_header). The header can be either 24 or 28 bytes
depending on fc.order. sizeof does not account for this so 28 is always
the calculated length.
This, in addition to hostapd not including a group number when rejecting,
cause this erroneous length calculation to be worked around as seen in
the removed comment. The comment is still valid (and described again
in another location) but the actual check for len == 4 is not correct.
To fix this we now rely on mpdu_validate to check that the authentication
frame is valid, and then subtract the actual header length using
mmpdu_header_len rather than sizeof. Doing this lets us also remove the
length check since it was validated previously.
A recent change checked the return value of ie_parse_rsne_from_data
inside the ptk 1/4 handler. This seemed safe, but actually caused
the eapol unit test to fail.
The reason was because eapol was parsing the IEs assuming they were
an RSN, when they could be a WPA IE (WPA1 not WPA2). The WPA case
does not end up using the rsn_info at all, so having rsn_info
uninitialized did not pose a problem. After adding the return value
check it was found this fails every time for WPA1.
Since the rsn_info is not needed for WPA1 we can only do the RSN
parse for WPA2 and leave rsn_info uninitialized.
The intent here was to validate that the frequency is a multiple of 5
and lies in a certain range. Somehow the channel was checked for being
a multiple of 5 instead.
The logic here intended to check whether all required attributes were
available. However, it set the parse_error to true instead of
have_required to false as intended.
This dumper probably intended to update pos after invoking strncpy.
However, strncpy returns the number of bytes that *would* have been
copied and so the logic gets a bit complex to get completely right.
Instead, switch to using l_string since this is inside the monitor and
not particularly performance critical.
Replace uses of strcpy by the safer l_strlcpy. Note that both of these
functions can only be called with a buffer of max 253 bytes (the
identity string), so this is purely a precautionary measure.
Technically there's no problem here as l_queue_remove does not
dereference the pointer. Still, it confuses certain static analysis
tools in the current form. Reordering this will not change the behavior
at all.
