linux-kernel - Re: "general protection fault in sctp_ulpevent_notify_peer_addr_change" and "general protection fault in sctp_ulpevent_nofity_peer_addr

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 12 Jan 2021 00:27:13 -0300
From:   Marcelo Ricardo Leitner <marcelo.leitner@...il.com>
To:     慕冬亮 <mudongliangabcd@...il.com>
Cc:     davem@...emloft.net, kuba@...nel.org,
        linux-kernel <linux-kernel@...r.kernel.org>,
        linux-sctp@...r.kernel.org, netdev@...r.kernel.org,
        nhorman@...driver.com, vyasevich@...il.com, rkovhaev@...il.com,
        syzkaller-bugs <syzkaller-bugs@...glegroups.com>
Subject: Re: "general protection fault in
 sctp_ulpevent_notify_peer_addr_change" and "general protection fault in
 sctp_ulpevent_nofity_peer_addr_change" should share the same root cause

On Tue, Jan 12, 2021 at 10:18:00AM +0800, 慕冬亮 wrote:
> Dear developers,
> 
> I find that "general protection fault in l2cap_sock_getsockopt" and
> "general protection fault in sco_sock_getsockopt" may be duplicated
> bugs from the same root cause.
> 
> First, by comparing the PoC similarity after own minimization, we find
> they share the same PoC. Second, the stack traces for both bug reports
> are the same except for the last function. And the different last
> functions are due to a function name change (typo fix) from
> "sctp_ulpevent_nofity_peer_addr_change" to
> "sctp_ulpevent_notify_peer_addr_change"

Not sure where you saw stack traces with this sctp function in it, but
the syzkaller reports from 17 Feb 2020 are not related to SCTP.

The one on sco_sock_getsockopt() seems to be lack of parameter
validation: it doesn't check if optval is big enough when handling
BT_PHY (which has the same value as SCTP_STATUS). It seems also miss a
check on if level != SOL_BLUETOOTH, but I may be wrong here.

l2cap_sock_getsockopt also lacks checking optlen.

  Marcelo