Small patch to md.texi for the movstr and clrstr patterns

[Hans-Peter Nilsson]

Hi.

The documentation for the "movstr" patterns lack a paragraph, IMHO:

Sun Jan 11 19:49:05 1998  Hans-Peter Nilsson  <hp@axis.se>

        * md.texi: Explain use of different modes in movstrM, clrstrM

*** md.texi.original    Sun Jan 11 14:40:21 1998
--- md.texi     Sun Jan 11 20:34:54 1998
*************** it may provide the value 4 for this oper
*** 1959,1964 ****
--- 1959,1970 ----
  These patterns need not give special consideration to the possibility
  that the source and destination strings might overlap.
  
+ The maximum size in bytes of the block to move is the maximum number in mode
+ @var{m}.  If you only have one @code{movstr@var{m}} insn which works for any
+ size, then mode @var{m} would be @code{Pmode}.  Patterns for smaller modes are
+ beneficial if their operands have otherwise less restrictions than the larger
+ modes.
+ 
  @cindex @code{clrstr@var{m}} instruction pattern
  @item @samp{clrstr@var{m}}
  Block clear instruction.  The addresses of the destination string is the
*************** The third operand is the known alignment
*** 1969,1974 ****
--- 1975,1982 ----
  of a @code{const_int} rtx.  Thus, if the compiler knows that the
  destination is word-aligned, it may provide the value 4 for this
  operand.
+ 
+ The mode @var{m} works as for @code{movstr@var{m}}.
  
  @cindex @code{cmpstr@var{m}} instruction pattern
  @item @samp{cmpstr@var{m}}

Or words to that effect.  This is visible in expr.c, but I believe it should
be documented as well.  There's a plausible misinterpretation: that the "move"
must be in multiples of sizeof (M), which means movstrqi may look less
restricted than movstrsi.  The patch is against the gcc snapshot of 971225,
but should work for egcs as well.

Best regards,
-- 
Hans-Peter Nilsson, Axis Communications AB, S - 223 70 LUND, SWEDEN
Hans-Peter.Nilsson@axis.se | http://www.axis.se/ | Tel +46 46191867,191800
Fax +46 46136130 | Implementing RFC 1855 compliancy; report loss of brain.