From: Jukka Marin <jmarin@teeri.jmp.fi>
   Date: Fri, 4 Aug 1995 10:15:58 +0300 (EET DST)
   X-Mailer: ELM [version 2.4 PL23]
   Content-Type: text
   Content-Length: 830       

   > Well, its ok for me to avoid chains if possible. But the tuning
   > decision about how big MCLBYTES is (which is kernel global) has
   > nothing to do with the sanity check for the chip ring buffer pointers.
   > In this case, the driver is in error, independent of the size of
   > MCLBYTES. I submitted a patch for this the day before yesterday.

   Oh, good.  The driver in -current has been pretty unusable with that bug.

   Hmm, in the code there's a place where it copies packets off the
   Ethernet card RAM.  It uses a word copy and divides the number of
   bytes to copy by 2.  I had to add 1 to the number of words to copy
   to make it work properly for packets with an odd number of bytes.
   Has this been fixed as well?  The symptoms were that, for example,
   during an rlogin session, the last character of a shell prompt was
   garbage.

     -jm