S.M.A.R.T. Monitoring Tools

Hi Marc,

> Is anybody planning on making smartmontools endian-safe ? I am using
> smartctl on a ppc (big-endian) board, it is usable but a lot of values
> are byte swapped including flags, LBA sectors, it also causes smartctl
> to think SMART is not supported forcing the use of -T permissive.

I'm really delighted to get your email.  I have been wondering for a long
time if smartmontools would work on a big-endian machine.

It should be possible to easily (and cleanly) fix all the endian-order
problems, by adding a switch that byte swaps all the two- and four-byte
quantities in the different SMART data structures, as soon as these are
read from the disk.

I can't do this myself since I have no way to test it -- but I'd be
interested in working with you to do this -- what do you think.

To get a flavor for what is involved, we'd need to write six functions
that byte-swap the requisite data in each of the six different SMART data
structures. To see what such a function would look like, see the function
fixsamsungerrorlog() in atacmds.c [this corrects some byte order
problems in a disk with bugs in the firmware].

Let me know if you are interested in doing this with me.

Cheers,
	Bruce

Marc,

> Is anybody planning on making smartmontools endian-safe ? I am using
> smartctl on a ppc (big-endian) board, it is usable but a lot of values
> are byte swapped including flags, LBA sectors, it also causes smartctl
> to think SMART is not supported forcing the use of -T permissive.

Just for fun, I added a few lines of code to smartmontools that:

(1) detect endian order at runtime
(2) fix endian order in self-test log.  This should correct the LBA sector
    numbering that you point out above, if you are refering to the
    LBA of first detected self-test errors.
and committed it to CVS. grep bigendian *.[hc] will find the new code.

Could you please see if this works OK by checking out the code from CVS,
and then seeing if the big-endian issues with the self-test error log are
gone?  If so, one of us can continue to fix these structures -- that would
be one of six finished.

Cheers,
	Bruce