/mandos/trunk

To get this branch, use:
bzr branch http://bzr.recompile.se/loggerhead/mandos/trunk

« back to all changes in this revision

Viewing changes to plugins.d/splashy.xml

  • Committer: teddy at recompile
  • Date: 2020-12-03 20:30:45 UTC
  • Revision ID: teddy@recompile.se-20201203203045-iqd6nq9y5nwalh1x
Minor fix of a test function

In dracut-module/password-agent, the test function
test_send_password_to_socket_EMSGSIZE() (which tests that the
send_password_to_socket() task function aborts properly when getting
EMSGSIZE when writing to the password socket), part of the test code
is supposed to find a message size which definitely does trigger
EMSGSIZE when send()ing to a socket.  Without a "break" in the proper
place, however, the size given is always exactly 1024 bytes too large.

This is very probably not a problem, since a too large message will
still be too large if it is increased by 1024 bytes, and send(2) in
practice checks the size before reading the buffer.  The biggest issue
would be if some version of send(2) would try to look at the last 1024
bytes of the message buffer before checking the message size; this
would then lead to a buffer over-read when running this test function.
(But even then there would be no security implications since the tests
are not run in the normal operation of the program.)

* dracut-module/password-agent.c
  (test_send_password_to_socket_EMSGSIZE): Break out early when ssret
  < 0 and errno == EMSGSIZE; don't allow loop to increase message_size
  again.

Show diffs side-by-side

added added

removed removed

Lines of Context:
135
135
        <para>
136
136
          These variables will normally be inherited from
137
137
          <citerefentry><refentrytitle>plugin-runner</refentrytitle>
138
 
          <manvolnum>8mandos</manvolnum></citerefentry>, which will
139
 
          normally have inherited them from
140
 
          <filename>/scripts/local-top/cryptroot</filename> in the
141
 
          initial <acronym>RAM</acronym> disk environment, which will
142
 
          have set them from parsing kernel arguments and
143
 
          <filename>/conf/conf.d/cryptroot</filename> (also in the
144
 
          initial RAM disk environment), which in turn will have been
145
 
          created when the initial RAM disk image was created by
146
 
          <filename
147
 
          >/usr/share/initramfs-tools/hooks/cryptroot</filename>, by
148
 
          extracting the information of the root file system from
149
 
          <filename >/etc/crypttab</filename>.
 
138
          <manvolnum>8mandos</manvolnum></citerefentry>, which might
 
139
          have in turn inherited them from its calling process.
150
140
        </para>
151
141
        <para>
152
142
          This behavior is meant to exactly mirror the behavior of
276
266
    <para>
277
267
      <citerefentry><refentrytitle>intro</refentrytitle>
278
268
      <manvolnum>8mandos</manvolnum></citerefentry>,
279
 
      <citerefentry><refentrytitle>crypttab</refentrytitle>
280
 
      <manvolnum>5</manvolnum></citerefentry>,
281
269
      <citerefentry><refentrytitle>plugin-runner</refentrytitle>
282
270
      <manvolnum>8mandos</manvolnum></citerefentry>,
283
271
      <citerefentry><refentrytitle>proc</refentrytitle>