Re: diag_dump_*

[Gary Thomas <gary@mlbassoc.com>]

On 2011-08-09 09:33, Michael Bergandi wrote:
> Hi all,
>
> I have been poking around looking at the diag_dump_* routines in
> infra/current/src/diag.cxx and I see the following test in
> diag_vdump_buf_with_offset() that makes me wonder what the intent was:
>
> diag_vdump_buf_with_offset(__printf_fun *pf,
>                             cyg_uint8     *p,
>                             CYG_ADDRWORD   s,
>                             cyg_uint8     *base)
> {
>      int i, c;
>      if ((CYG_ADDRWORD)s>  (CYG_ADDRWORD)p) {
>          s = (CYG_ADDRWORD)s - (CYG_ADDRWORD)p;
>      }
>
> My understanding of the parameters are:
>
> pf = pointer to function to use to print
> p = pointer to buffer that you want to print
> s = size (number of bytes to print)
> base = base address to use for printing the addresses
>
> I hope this is right. I can't find any documentation that actually
> says what these parameters are. There is no documentation for the
> function in diag.cxx, diag.hxx, or the reference manual.
>
> So, the test above doesn't make much sense based on what I understand
> the parameters to be.
>
> If I stretch my brain a bit, it looks like 's' is being treated as a
> pointer in the test and not a size or length value. Is this right or
> what is intended?

's' can either be a size (length of buffer) or an upper address bound.
These calls would be equivalent:
   diag_dump_buf(p, 256);
      -or-
   diag_dump_buf(p, (char *)p+256);

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------