fix:shellcraft mips dupsh wrong redirect

[leommxj]

No description provided.

[Arusekk]

This is actually caused by shellcraft.mips.linux.dup() missing, which is present on some other architectures, but still under different names (is this intended? i386.dupio(), but amd64.dup() and thumb.dup()).

pwntools/pwnlib/shellcraft/templates/amd64/linux/dup.asm

Line 13 in 6ea5722

/* dup() file descriptor ${sock} into stdin/stdout/stderr */

pwntools/pwnlib/shellcraft/templates/thumb/linux/dup.asm

Line 12 in 6ea5722

/* dup() file descriptor ${sock} into stdin/stdout/stderr */

pwntools/pwnlib/shellcraft/templates/i386/linux/dupio.asm

Line 14 in 6ea5722

/* dup() file descriptor ${sock} into stdin/stdout/stderr */

[zachriggle]

Sorry for the long delay. Would you mind making this a decrementing loop to save space? Similar to how i386.linux.dupio works:

<%
  dup       = common.label("dup")
  looplabel = common.label("loop")
%>

    /* dup() file descriptor ${sock} into stdin/stdout/stderr */
${dup}:
    ${mov('ebx', sock)}
    ${mov('ecx', 3)}
${looplabel}:
    dec ecx

    ${dup2('ebx', 'ecx')}
    jnz ${looplabel}

Separately, please rename this to dupio and update dupsh accordingly? The reason for the rename was to avoid conflicts with the syscall name, it looks like it wasn't applied everywhere.

[zachriggle]

Also, can you add a test for this so that we can be sure that it does not regress? See the "Example" block in mov.asm for how tests work.