Ahh crud.  I see the problem now.  The fiid_obj_get logic loops based on
the number of bytes covered by a field.  In this case, the 64bit field
covers 9 bytes (5 bits in the first byte, then 7 full bytes, then 3 bits
in the last byte).  The last iteration causes the bit-shifting issue.

I'll have to think about this more.  It may have to involve doing
changes to the bit-ops code as well.  

We'll ignore this issue in the head branch.  I'll just deal with it in
the fiid interface re-implementation branch.

Al

--
Albert Chu
[EMAIL PROTECTED]
925-422-5311
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory


----- Original Message -----
From: "Bala.A" <[EMAIL PROTECTED]>
Date: Monday, January 16, 2006 7:42 pm
Subject: Re: [Freeipmi-devel] fiid_obj_get bugs

> Hi Al,
> 
> >> I applied your patch and tested it.  It fixes for some cases and
> >> not for other cases.
> >
> > What did you find that didn't work?  I did some sanity/test 
> checking and
> > didn't find anything.  Of course I may have missed stuff.  Just 
> point me
> > in the right direction and I'll fix the issues.
> 
> Here are few cases;
> 
> 1. This works
>  fiid_template_t tmpl =
>    {
>      {64, "64bits"},
>      {3, "3bits"},
>      {5, "5bits"},
>      {0, ""}
>    };
> 
> 2. This works
>  fiid_template_t tmpl =
>    {
>      {3, "3bits"},
>      {5, "5bits"},
>      {64, "64bits"},
>      {0, ""}
>    };
> 
> 3. This fails , here we loose 8th byte
>  fiid_template_t tmpl =
>    {
>      {3, "3bits"},
>      {64, "64bits"},
>      {5, "7bits"},
>      {0, ""}
>    };
> 
> 
> Thanks,
> 
> Bala
> ---
> Free as in freedom
> http://www.gnu.org/
> 
> 
> 



_______________________________________________
Freeipmi-devel mailing list
Freeipmi-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/freeipmi-devel

Reply via email to