Wireshark-dev: Re: [Wireshark-dev] Thinking of some changes to packet-scsi.c ...
From: Richard Sharpe <[email protected]>
Date: Sat, 4 Dec 2010 16:24:35 -0800
On Sat, Dec 4, 2010 at 3:46 PM, ronnie sahlberg
<[email protected]> wrote:
> Yeah
>
> The scsi dissector does need some tlc
> to clean up and to harmonize the dissection according to the latest
> versions of the standard.
>
> For handling flags,
> you may want to look at some examples such as
>
> ...
>   static const int *aca_fields[] = {
>        &hf_scsi_inq_aerc,      /* obsolete in spc3 and forward */
>            &hf_scsi_inq_trmtsk,        /* obsolete in spc2 and forward */
>            &hf_scsi_inq_normaca,
>            &hf_scsi_inq_hisup,
>            &hf_scsi_inq_rdf,
>            NULL
>    };
> ...
>        /* aca flags */
>        proto_tree_add_bitmask(tree, tvb_v, offset_v,
> hf_scsi_inq_acaflags, ett_scsi_inq_acaflags, aca_fields, FALSE);
>        offset_v+=1
>
> which can simplify creating/dissecting bitfields.
>
>
>
> this would be a reasonable summer of code project for a student,
> go through all of spc/sbc/ssc/smc/mmc and update wireshark accordingly.
> at the same time, and to have something to generate data to test with,
> they could at the same time
> enhance dbench.samba.org and its iscsi backend to be able to generate such i/o.

Well, I have an immediate need for this so I will have to steal some
of that work from a student :-)

In addition, I have an immediate need for dbench to generate more
iSCSI IOs so I will probably be doing something in that line in the
next couple of weeks.

My need relates to SCST and PR and clusters ...

-- 
Regards,
Richard Sharpe