90e504dc25
In commit r9968 "wctdm24xxp: Allow more than one outstanding read at a time" I introduced a regression where the transmit FIFO on the data channel of a B400M could get locked up. The result would be constant HDLC overflows when writing to the data channel. This regression did not make it into any releases and did not exhibit itself when crossing spans on a single B400M module. This is a partial revert of commit r9968. Signed-off-by: Shaun Ruffell <sruffell@digium.com> git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9977 a0bf4364-ded3-4de4-8d8a-66a801d63aff